mirror of https://github.com/vladmandic/human
4443 lines
1.1 MiB
4443 lines
1.1 MiB
|
|
/*
|
|
Human library
|
|
homepage: <https://github.com/vladmandic/human>
|
|
author: <https://github.com/vladmandic>'
|
|
*/
|
|
|
|
var kG=Object.create;var Jh=Object.defineProperty;var vG=Object.getOwnPropertyDescriptor;var CG=Object.getOwnPropertyNames;var IG=Object.getPrototypeOf,NG=Object.prototype.hasOwnProperty;var SG=r=>Jh(r,"__esModule",{value:!0});var dc=r=>{if(typeof require!="undefined")return require(r);throw new Error('Dynamic require of "'+r+'" is not supported')};var qt=(r,e)=>()=>(e||r((e={exports:{}}).exports,e),e.exports),We=(r,e)=>{for(var t in e)Jh(r,t,{get:e[t],enumerable:!0})},TG=(r,e,t)=>{if(e&&typeof e=="object"||typeof e=="function")for(let n of CG(e))!NG.call(r,n)&&n!=="default"&&Jh(r,n,{get:()=>e[n],enumerable:!(t=vG(e,n))||t.enumerable});return r},hc=r=>TG(SG(Jh(r!=null?kG(IG(r)):{},"default",r&&r.__esModule&&"default"in r?{get:()=>r.default,enumerable:!0}:{value:r,enumerable:!0})),r);var uT=qt(()=>{});var jT=qt((WT,d_)=>{(function(r,e,t){function n(i){var l=this,u=a();l.next=function(){var c=2091639*l.s0+l.c*23283064365386963e-26;return l.s0=l.s1,l.s1=l.s2,l.s2=c-(l.c=c|0)},l.c=1,l.s0=u(" "),l.s1=u(" "),l.s2=u(" "),l.s0-=u(i),l.s0<0&&(l.s0+=1),l.s1-=u(i),l.s1<0&&(l.s1+=1),l.s2-=u(i),l.s2<0&&(l.s2+=1),u=null}function o(i,l){return l.c=i.c,l.s0=i.s0,l.s1=i.s1,l.s2=i.s2,l}function s(i,l){var u=new n(i),c=l&&l.state,p=u.next;return p.int32=function(){return u.next()*4294967296|0},p.double=function(){return p()+(p()*2097152|0)*11102230246251565e-32},p.quick=p,c&&(typeof c=="object"&&o(c,u),p.state=function(){return o(u,{})}),p}function a(){var i=4022871197,l=function(u){u=u.toString();for(var c=0;c<u.length;c++){i+=u.charCodeAt(c);var p=.02519603282416938*i;i=p>>>0,p-=i,p*=i,i=p>>>0,p-=i,i+=p*4294967296}return(i>>>0)*23283064365386963e-26};return l}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.alea=s})(WT,typeof d_=="object"&&d_,typeof define=="function"&&define)});var HT=qt((UT,h_)=>{(function(r,e,t){function n(a){var i=this,l="";i.x=0,i.y=0,i.z=0,i.w=0,i.next=function(){var c=i.x^i.x<<11;return i.x=i.y,i.y=i.z,i.z=i.w,i.w^=i.w>>>19^c^c>>>8},a===(a|0)?i.x=a:l+=a;for(var u=0;u<l.length+64;u++)i.x^=l.charCodeAt(u)|0,i.next()}function o(a,i){return i.x=a.x,i.y=a.y,i.z=a.z,i.w=a.w,i}function s(a,i){var l=new n(a),u=i&&i.state,c=function(){return(l.next()>>>0)/4294967296};return c.double=function(){do var p=l.next()>>>11,m=(l.next()>>>0)/4294967296,f=(p+m)/(1<<21);while(f===0);return f},c.int32=l.next,c.quick=c,u&&(typeof u=="object"&&o(u,l),c.state=function(){return o(l,{})}),c}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.xor128=s})(UT,typeof h_=="object"&&h_,typeof define=="function"&&define)});var KT=qt((qT,g_)=>{(function(r,e,t){function n(a){var i=this,l="";i.next=function(){var c=i.x^i.x>>>2;return i.x=i.y,i.y=i.z,i.z=i.w,i.w=i.v,(i.d=i.d+362437|0)+(i.v=i.v^i.v<<4^(c^c<<1))|0},i.x=0,i.y=0,i.z=0,i.w=0,i.v=0,a===(a|0)?i.x=a:l+=a;for(var u=0;u<l.length+64;u++)i.x^=l.charCodeAt(u)|0,u==l.length&&(i.d=i.x<<10^i.x>>>4),i.next()}function o(a,i){return i.x=a.x,i.y=a.y,i.z=a.z,i.w=a.w,i.v=a.v,i.d=a.d,i}function s(a,i){var l=new n(a),u=i&&i.state,c=function(){return(l.next()>>>0)/4294967296};return c.double=function(){do var p=l.next()>>>11,m=(l.next()>>>0)/4294967296,f=(p+m)/(1<<21);while(f===0);return f},c.int32=l.next,c.quick=c,u&&(typeof u=="object"&&o(u,l),c.state=function(){return o(l,{})}),c}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.xorwow=s})(qT,typeof g_=="object"&&g_,typeof define=="function"&&define)});var YT=qt((XT,x_)=>{(function(r,e,t){function n(a){var i=this;i.next=function(){var u=i.x,c=i.i,p,m,f;return p=u[c],p^=p>>>7,m=p^p<<24,p=u[c+1&7],m^=p^p>>>10,p=u[c+3&7],m^=p^p>>>3,p=u[c+4&7],m^=p^p<<7,p=u[c+7&7],p=p^p<<13,m^=p^p<<9,u[c]=m,i.i=c+1&7,m};function l(u,c){var p,m,f=[];if(c===(c|0))m=f[0]=c;else for(c=""+c,p=0;p<c.length;++p)f[p&7]=f[p&7]<<15^c.charCodeAt(p)+f[p+1&7]<<13;for(;f.length<8;)f.push(0);for(p=0;p<8&&f[p]===0;++p);for(p==8?m=f[7]=-1:m=f[p],u.x=f,u.i=0,p=256;p>0;--p)u.next()}l(i,a)}function o(a,i){return i.x=a.x.slice(),i.i=a.i,i}function s(a,i){a==null&&(a=+new Date);var l=new n(a),u=i&&i.state,c=function(){return(l.next()>>>0)/4294967296};return c.double=function(){do var p=l.next()>>>11,m=(l.next()>>>0)/4294967296,f=(p+m)/(1<<21);while(f===0);return f},c.int32=l.next,c.quick=c,u&&(u.x&&o(u,l),c.state=function(){return o(l,{})}),c}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.xorshift7=s})(XT,typeof x_=="object"&&x_,typeof define=="function"&&define)});var JT=qt((ZT,y_)=>{(function(r,e,t){function n(a){var i=this;i.next=function(){var u=i.w,c=i.X,p=i.i,m,f;return i.w=u=u+1640531527|0,f=c[p+34&127],m=c[p=p+1&127],f^=f<<13,m^=m<<17,f^=f>>>15,m^=m>>>12,f=c[p]=f^m,i.i=p,f+(u^u>>>16)|0};function l(u,c){var p,m,f,d,h,g=[],x=128;for(c===(c|0)?(m=c,c=null):(c=c+"\0",m=0,x=Math.max(x,c.length)),f=0,d=-32;d<x;++d)c&&(m^=c.charCodeAt((d+32)%c.length)),d===0&&(h=m),m^=m<<10,m^=m>>>15,m^=m<<4,m^=m>>>13,d>=0&&(h=h+1640531527|0,p=g[d&127]^=m+h,f=p==0?f+1:0);for(f>=128&&(g[(c&&c.length||0)&127]=-1),f=127,d=4*128;d>0;--d)m=g[f+34&127],p=g[f=f+1&127],m^=m<<13,p^=p<<17,m^=m>>>15,p^=p>>>12,g[f]=m^p;u.w=h,u.X=g,u.i=f}l(i,a)}function o(a,i){return i.i=a.i,i.w=a.w,i.X=a.X.slice(),i}function s(a,i){a==null&&(a=+new Date);var l=new n(a),u=i&&i.state,c=function(){return(l.next()>>>0)/4294967296};return c.double=function(){do var p=l.next()>>>11,m=(l.next()>>>0)/4294967296,f=(p+m)/(1<<21);while(f===0);return f},c.int32=l.next,c.quick=c,u&&(u.X&&o(u,l),c.state=function(){return o(l,{})}),c}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.xor4096=s})(ZT,typeof y_=="object"&&y_,typeof define=="function"&&define)});var e1=qt((QT,b_)=>{(function(r,e,t){function n(a){var i=this,l="";i.next=function(){var c=i.b,p=i.c,m=i.d,f=i.a;return c=c<<25^c>>>7^p,p=p-m|0,m=m<<24^m>>>8^f,f=f-c|0,i.b=c=c<<20^c>>>12^p,i.c=p=p-m|0,i.d=m<<16^p>>>16^f,i.a=f-c|0},i.a=0,i.b=0,i.c=2654435769|0,i.d=1367130551,a===Math.floor(a)?(i.a=a/4294967296|0,i.b=a|0):l+=a;for(var u=0;u<l.length+20;u++)i.b^=l.charCodeAt(u)|0,i.next()}function o(a,i){return i.a=a.a,i.b=a.b,i.c=a.c,i.d=a.d,i}function s(a,i){var l=new n(a),u=i&&i.state,c=function(){return(l.next()>>>0)/4294967296};return c.double=function(){do var p=l.next()>>>11,m=(l.next()>>>0)/4294967296,f=(p+m)/(1<<21);while(f===0);return f},c.int32=l.next,c.quick=c,u&&(typeof u=="object"&&o(u,l),c.state=function(){return o(l,{})}),c}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.tychei=s})(QT,typeof b_=="object"&&b_,typeof define=="function"&&define)});var w_=qt(()=>{});var t1=qt((I_e,wg)=>{(function(r,e){var t=this,n=256,o=6,s=52,a="random",i=e.pow(n,o),l=e.pow(2,s),u=l*2,c=n-1,p;function m(w,_,I){var E=[];_=_==!0?{entropy:!0}:_||{};var D=g(h(_.entropy?[w,b(r)]:w==null?x():w,3),E),$=new f(E),R=function(){for(var M=$.g(o),G=i,j=0;M<l;)M=(M+j)*n,G*=n,j=$.g(1);for(;M>=u;)M/=2,G/=2,j>>>=1;return(M+j)/G};return R.int32=function(){return $.g(4)|0},R.quick=function(){return $.g(4)/4294967296},R.double=R,g(b($.S),r),(_.pass||I||function(M,G,j,U){return U&&(U.S&&d(U,$),M.state=function(){return d($,{})}),j?(e[a]=M,G):M})(R,D,"global"in _?_.global:this==e,_.state)}e["seed"+a]=m;function f(w){var _,I=w.length,E=this,D=0,$=E.i=E.j=0,R=E.S=[];for(I||(w=[I++]);D<n;)R[D]=D++;for(D=0;D<n;D++)R[D]=R[$=c&$+w[D%I]+(_=R[D])],R[$]=_;(E.g=function(M){for(var G,j=0,U=E.i,H=E.j,q=E.S;M--;)G=q[U=c&U+1],j=j*n+q[c&(q[U]=q[H=c&H+G])+(q[H]=G)];return E.i=U,E.j=H,j})(n)}function d(w,_){return _.i=w.i,_.j=w.j,_.S=w.S.slice(),_}function h(w,_){var I=[],E=typeof w,D;if(_&&E=="object")for(D in w)try{I.push(h(w[D],_-1))}catch($){}return I.length?I:E=="string"?w:w+"\0"}function g(w,_){for(var I=w+"",E,D=0;D<I.length;)_[c&D]=c&(E^=_[c&D]*19)+I.charCodeAt(D++);return b(_)}function x(){try{var w;return p&&(w=p.randomBytes)?w=w(n):(w=new Uint8Array(n),(t.crypto||t.msCrypto).getRandomValues(w)),b(w)}catch(E){var _=t.navigator,I=_&&_.plugins;return[+new Date,t,I,t.screen,b(r)]}}function b(w){return String.fromCharCode.apply(0,w)}if(g(e.random(),r),typeof wg=="object"&&wg.exports){wg.exports=m;try{p=w_()}catch(w){}}else typeof define=="function"&&define.amd&&define(function(){return m})})([],Math)});var __=qt((N_e,r1)=>{var qU=jT(),KU=HT(),XU=KT(),YU=YT(),ZU=JT(),JU=e1(),du=t1();du.alea=qU;du.xor128=KU;du.xorwow=XU;du.xorshift7=YU;du.xor4096=ZU;du.tychei=JU;r1.exports=du});var Pp=qt(()=>{});var UM=qt(()=>{});var HM=qt(()=>{});var qM=qt((Ax,MC)=>{var PC=function(){var r=typeof document!="undefined"&&document.currentScript?document.currentScript.src:void 0;return typeof __filename!="undefined"&&(r=r||__filename),function(e){e=e||{};function t(){return ee.buffer!=He&&ir(ee.buffer),ct}function n(){return ee.buffer!=He&&ir(ee.buffer),mt}function o(){return ee.buffer!=He&&ir(ee.buffer),Zt}function s(){return ee.buffer!=He&&ir(ee.buffer),cn}function a(){return ee.buffer!=He&&ir(ee.buffer),Un}var i=typeof e!="undefined"?e:{},l,u;i.ready=new Promise(function(N,F){l=N,u=F});var c={},p;for(p in i)i.hasOwnProperty(p)&&(c[p]=i[p]);var m=[],f="./this.program",d=function(N,F){throw F},h=!1,g=!1,x=!1,b=!1;h=typeof window=="object",g=typeof importScripts=="function",x=typeof process=="object"&&typeof process.versions=="object"&&typeof process.versions.node=="string",b=!h&&!x&&!g;var w=i.ENVIRONMENT_IS_PTHREAD||!1;w&&(He=i.buffer);var _="";function I(N){return i.locateFile?i.locateFile(N,_):_+N}var E,D,$,R,M,G;if(x){g?_=Pp().dirname(_)+"/":_=__dirname+"/",E=function(F,B){return M||(M=dc("fs")),G||(G=Pp()),F=G.normalize(F),M.readFileSync(F,B?null:"utf8")},$=function(F){var B=E(F,!0);return B.buffer||(B=new Uint8Array(B)),de(B.buffer),B},process.argv.length>1&&(f=process.argv[1].replace(/\\/g,"/")),m=process.argv.slice(2),process.on("uncaughtException",function(N){if(!(N instanceof _m))throw N}),process.on("unhandledRejection",bi),d=function(N){process.exit(N)},i.inspect=function(){return"[Emscripten Module object]"};var j;try{j=UM()}catch(N){throw console.error('The "worker_threads" module is not supported in this node.js build - perhaps a newer version is needed?'),N}global.Worker=j.Worker}else b?(typeof read!="undefined"&&(E=function(F){return read(F)}),$=function(F){var B;return typeof readbuffer=="function"?new Uint8Array(readbuffer(F)):(B=read(F,"binary"),de(typeof B=="object"),B)},typeof scriptArgs!="undefined"?m=scriptArgs:typeof arguments!="undefined"&&(m=arguments),typeof quit=="function"&&(d=function(N){quit(N)}),typeof print!="undefined"&&(typeof console=="undefined"&&(console={}),console.log=print,console.warn=console.error=typeof printErr!="undefined"?printErr:print)):(h||g)&&(g?_=self.location.href:typeof document!="undefined"&&document.currentScript&&(_=document.currentScript.src),typeof r!="undefined"&&r&&(_=r),_.indexOf("blob:")!==0?_=_.substr(0,_.lastIndexOf("/")+1):_="",x?(E=function(F,B){return M||(M=dc("fs")),G||(G=Pp()),F=G.normalize(F),M.readFileSync(F,B?null:"utf8")},$=function(F){var B=E(F,!0);return B.buffer||(B=new Uint8Array(B)),de(B.buffer),B}):(E=function(N){var F=new XMLHttpRequest;return F.open("GET",N,!1),F.send(null),F.responseText},g&&($=function(N){var F=new XMLHttpRequest;return F.open("GET",N,!1),F.responseType="arraybuffer",F.send(null),new Uint8Array(F.response)}),D=function(N,F,B){var K=new XMLHttpRequest;K.open("GET",N,!0),K.responseType="arraybuffer",K.onload=function(){if(K.status==200||K.status==0&&K.response){F(K.response);return}B()},K.onerror=B,K.send(null)}),R=function(N){document.title=N});x&&typeof performance=="undefined"&&(global.performance=HM().performance);var U=i.print||console.log.bind(console),H=i.printErr||console.warn.bind(console);for(p in c)c.hasOwnProperty(p)&&(i[p]=c[p]);c=null,i.arguments&&(m=i.arguments),i.thisProgram&&(f=i.thisProgram),i.quit&&(d=i.quit);var q=Atomics.load,X=Atomics.store,ne=Atomics.compareExchange,Y;i.wasmBinary&&(Y=i.wasmBinary);var re=i.noExitRuntime||!0;typeof WebAssembly!="object"&&bi("no native wasm support detected");var ee,ie,me=!1,ae;function de(N,F){N||bi("Assertion failed: "+F)}function he(N){var F=i["_"+N];return de(F,"Cannot call unknown function "+N+", make sure it is exported"),F}function xe(N,F,B,K,be){var ge={string:function(Hr){var fc=0;if(Hr!=null&&Hr!==0){var HI=(Hr.length<<2)+1;fc=cc(HI),Ke(Hr,fc,HI)}return fc},array:function(Hr){var fc=cc(Hr.length);return At(Hr,fc),fc}};function ye(Hr){return F==="string"?ve(Hr):F==="boolean"?Boolean(Hr):Hr}var Te=he(N),_t=[],kr=0;if(K)for(var gr=0;gr<K.length;gr++){var Ka=ge[B[gr]];Ka?(kr===0&&(kr=wm()),_t[gr]=Ka(K[gr])):_t[gr]=K[gr]}var mc=Te.apply(null,_t);return mc=ye(mc),kr!==0&&uc(kr),mc}function we(N,F,B,K){B=B||[];var be=B.every(function(ye){return ye==="number"}),ge=F!=="string";return ge&&be&&!K?he(N):function(){return xe(N,F,B,arguments,K)}}function De(N,F,B){for(var K=F+B,be="";!(F>=K);){var ge=N[F++];if(!ge)return be;if(!(ge&128)){be+=String.fromCharCode(ge);continue}var ye=N[F++]&63;if((ge&224)==192){be+=String.fromCharCode((ge&31)<<6|ye);continue}var Te=N[F++]&63;if((ge&240)==224?ge=(ge&15)<<12|ye<<6|Te:ge=(ge&7)<<18|ye<<12|Te<<6|N[F++]&63,ge<65536)be+=String.fromCharCode(ge);else{var _t=ge-65536;be+=String.fromCharCode(55296|_t>>10,56320|_t&1023)}}return be}function ve(N,F){return N?De(n(),N,F):""}function Ge(N,F,B,K){if(!(K>0))return 0;for(var be=B,ge=B+K-1,ye=0;ye<N.length;++ye){var Te=N.charCodeAt(ye);if(Te>=55296&&Te<=57343){var _t=N.charCodeAt(++ye);Te=65536+((Te&1023)<<10)|_t&1023}if(Te<=127){if(B>=ge)break;F[B++]=Te}else if(Te<=2047){if(B+1>=ge)break;F[B++]=192|Te>>6,F[B++]=128|Te&63}else if(Te<=65535){if(B+2>=ge)break;F[B++]=224|Te>>12,F[B++]=128|Te>>6&63,F[B++]=128|Te&63}else{if(B+3>=ge)break;F[B++]=240|Te>>18,F[B++]=128|Te>>12&63,F[B++]=128|Te>>6&63,F[B++]=128|Te&63}}return F[B]=0,B-be}function Ke(N,F,B){return Ge(N,n(),F,B)}function at(N){for(var F=0,B=0;B<N.length;++B){var K=N.charCodeAt(B);K>=55296&&K<=57343&&(K=65536+((K&1023)<<10)|N.charCodeAt(++B)&1023),K<=127?++F:K<=2047?F+=2:K<=65535?F+=3:F+=4}return F}function At(N,F){t().set(N,F)}function Et(N,F){return N%F>0&&(N+=F-N%F),N}var He,ct,mt,zt,vn,Zt,cn,Mr,Un;function ir(N){He=N,i.HEAP8=ct=new Int8Array(N),i.HEAP16=zt=new Int16Array(N),i.HEAP32=Zt=new Int32Array(N),i.HEAPU8=mt=new Uint8Array(N),i.HEAPU16=vn=new Uint16Array(N),i.HEAPU32=cn=new Uint32Array(N),i.HEAPF32=Mr=new Float32Array(N),i.HEAPF64=Un=new Float64Array(N)}var Cn=i.INITIAL_MEMORY||16777216;if(w)ee=i.wasmMemory,He=i.buffer;else if(i.wasmMemory)ee=i.wasmMemory;else if(ee=new WebAssembly.Memory({initial:Cn/65536,maximum:2147483648/65536,shared:!0}),!(ee.buffer instanceof SharedArrayBuffer))throw H("requested a shared WebAssembly.Memory but the returned buffer is not a SharedArrayBuffer, indicating that while the browser has SharedArrayBuffer it does not have WebAssembly threads support - you may need to set a flag"),x&&console.log("(on node you may need: --experimental-wasm-threads --experimental-wasm-bulk-memory and also use a recent version)"),Error("bad memory");ee&&(He=ee.buffer),Cn=He.byteLength,ir(He);var Ur,_r=[],pn=[],Fn=[],zl=[],Wa=[],Ps=!1,xi=!1;w||pn.push({func:function(){Uh()}});function pm(){if(!w){if(i.preRun)for(typeof i.preRun=="function"&&(i.preRun=[i.preRun]);i.preRun.length;)$h(i.preRun.shift());sc(_r)}}function Bl(){Ps=!0,!w&&sc(pn)}function Dy(){w||sc(Fn)}function Dh(){w||(xi=!0)}function mn(){if(!w){if(i.postRun)for(typeof i.postRun=="function"&&(i.postRun=[i.postRun]);i.postRun.length;)$y(i.postRun.shift());sc(Wa)}}function $h(N){_r.unshift(N)}function $y(N){Wa.unshift(N)}var yi=0,ja=null,Vl=null;function Ry(N){de(!w,"addRunDependency cannot be used in a pthread worker"),yi++,i.monitorRunDependencies&&i.monitorRunDependencies(yi)}function Fy(N){if(yi--,i.monitorRunDependencies&&i.monitorRunDependencies(yi),yi==0&&(ja!==null&&(clearInterval(ja),ja=null),Vl)){var F=Vl;Vl=null,F()}}i.preloadedImages={},i.preloadedAudios={};function bi(N){i.onAbort&&i.onAbort(N),w&&console.error("Pthread aborting at "+new Error().stack),N+="",H(N),me=!0,ae=1,N="abort("+N+"). Build with -s ASSERTIONS=1 for more info.";var F=new WebAssembly.RuntimeError(N);throw u(F),F}function Rh(N,F){return String.prototype.startsWith?N.startsWith(F):N.indexOf(F)===0}var oc="data:application/octet-stream;base64,";function Fh(N){return Rh(N,oc)}var Oy="file://";function Oh(N){return Rh(N,Oy)}var fn="tfjs-backend-wasm-threaded-simd.wasm";Fh(fn)||(fn=I(fn));function Ph(N){try{if(N==fn&&Y)return new Uint8Array(Y);if($)return $(N);throw"both async and sync fetching of the wasm failed"}catch(F){bi(F)}}function Py(){if(!Y&&(h||g)){if(typeof fetch=="function"&&!Oh(fn))return fetch(fn,{credentials:"same-origin"}).then(function(N){if(!N.ok)throw"failed to load wasm binary file at '"+fn+"'";return N.arrayBuffer()}).catch(function(){return Ph(fn)});if(D)return new Promise(function(N,F){D(fn,function(B){N(new Uint8Array(B))},F)})}return Promise.resolve().then(function(){return Ph(fn)})}function My(){var N={a:Sb};function F(ye,Te){var _t=ye.exports;if(i.asm=_t,Ur=i.asm.F,ie=Te,!w){var kr=Ae.unusedWorkers.length;Ae.unusedWorkers.forEach(function(gr){Ae.loadWasmModuleToWorker(gr,function(){--kr||Fy("wasm-instantiate")})})}}w||Ry("wasm-instantiate");function B(ye){F(ye.instance,ye.module)}function K(ye){return Py().then(function(Te){return WebAssembly.instantiate(Te,N)}).then(ye,function(Te){H("failed to asynchronously prepare wasm: "+Te),bi(Te)})}function be(){return!Y&&typeof WebAssembly.instantiateStreaming=="function"&&!Fh(fn)&&!Oh(fn)&&typeof fetch=="function"?fetch(fn,{credentials:"same-origin"}).then(function(ye){var Te=WebAssembly.instantiateStreaming(ye,N);return Te.then(B,function(_t){return H("wasm streaming compile failed: "+_t),H("falling back to ArrayBuffer instantiation"),K(B)})}):K(B)}if(i.instantiateWasm)try{var ge=i.instantiateWasm(N,F);return ge}catch(ye){return H("Module.instantiateWasm callback failed with error: "+ye),!1}return be().catch(u),{}}var Ly={9816:function(){throw"Canceled!"},9834:function(N,F){setTimeout(function(){BI(N,F)},0)}};function Mh(){Ae.initRuntime()}function sc(N){for(;N.length>0;){var F=N.shift();if(typeof F=="function"){F(i);continue}var B=F.func;typeof B=="number"?F.arg===void 0?Ur.get(B)():Ur.get(B)(F.arg):B(F.arg===void 0?null:F.arg)}}function mm(N,F){if(N<=0||N>t().length||N&!0||F<0)return-28;if(F==0)return 0;F>=2147483647&&(F=Infinity);var B=Atomics.load(o(),pc>>2),K=0;if(B==N){var be=Atomics.compareExchange(o(),pc>>2,B,0);if(be==B&&(--F,K=1,F<=0))return 1}var ge=Atomics.notify(o(),N>>2,F);if(ge>=0)return ge+K;throw"Atomics.notify returned an unexpected value "+ge}i._emscripten_futex_wake=mm;function zy(N){if(w)throw"Internal Error! killThread() can only ever be called from main application thread!";if(!N)throw"Internal Error! Null pthread_ptr in killThread!";o()[N+12>>2]=0;var F=Ae.pthreads[N];F.worker.terminate(),Ae.freeThreadData(F),Ae.runningWorkers.splice(Ae.runningWorkers.indexOf(F.worker),1),F.worker.pthread=void 0}function By(N){if(w)throw"Internal Error! cancelThread() can only ever be called from main application thread!";if(!N)throw"Internal Error! Null pthread_ptr in cancelThread!";var F=Ae.pthreads[N];F.worker.postMessage({cmd:"cancel"})}function Vy(N){if(w)throw"Internal Error! cleanupThread() can only ever be called from main application thread!";if(!N)throw"Internal Error! Null pthread_ptr in cleanupThread!";var F=Ae.pthreads[N];if(F){o()[N+12>>2]=0;var B=F.worker;Ae.returnWorkerToPool(B)}}var Ae={unusedWorkers:[],runningWorkers:[],initMainThreadBlock:function(){for(var N=Math.min(4,Math.max(1,(navigator.hardwareConcurrency||1)/2)),F=0;F<N;++F)Ae.allocateUnusedWorker()},initRuntime:function(){for(var N=Wl(228),F=0;F<228/4;++F)s()[N/4+F]=0;o()[N+12>>2]=N;var B=N+152;o()[B>>2]=B;for(var K=Wl(512),F=0;F<128;++F)s()[K/4+F]=0;Atomics.store(s(),N+100>>2,K),Atomics.store(s(),N+40>>2,N),Jb(N,!g,1),zI(N)},initWorker:function(){},pthreads:{},threadExitHandlers:[],setThreadStatus:function(){},runExitHandlers:function(){for(;Ae.threadExitHandlers.length>0;)Ae.threadExitHandlers.pop()();w&&lc()&&LI()},runExitHandlersAndDeinitThread:function(N,F){Atomics.store(s(),N+56>>2,1),Atomics.store(s(),N+60>>2,0),Ae.runExitHandlers(),Atomics.store(s(),N+4>>2,F),Atomics.store(s(),N+0>>2,1),mm(N+0,2147483647),Jb(0,0,0)},threadExit:function(N){var F=lc();F&&(Ae.runExitHandlersAndDeinitThread(F,N),w&&postMessage({cmd:"exit"}))},threadCancel:function(){Ae.runExitHandlersAndDeinitThread(lc(),-1),postMessage({cmd:"cancelDone"})},terminateAllThreads:function(){for(var N in Ae.pthreads){var F=Ae.pthreads[N];F&&F.worker&&Ae.returnWorkerToPool(F.worker)}Ae.pthreads={};for(var B=0;B<Ae.unusedWorkers.length;++B){var K=Ae.unusedWorkers[B];K.terminate()}Ae.unusedWorkers=[];for(var B=0;B<Ae.runningWorkers.length;++B){var K=Ae.runningWorkers[B],F=K.pthread;Ae.freeThreadData(F),K.terminate()}Ae.runningWorkers=[]},freeThreadData:function(N){if(!!N){if(N.threadInfoStruct){var F=o()[N.threadInfoStruct+100>>2];o()[N.threadInfoStruct+100>>2]=0,bm(F),bm(N.threadInfoStruct)}N.threadInfoStruct=0,N.allocatedOwnStack&&N.stackBase&&bm(N.stackBase),N.stackBase=0,N.worker&&(N.worker.pthread=null)}},returnWorkerToPool:function(N){Ae.runWithoutMainThreadQueuedCalls(function(){delete Ae.pthreads[N.pthread.threadInfoStruct],Ae.unusedWorkers.push(N),Ae.runningWorkers.splice(Ae.runningWorkers.indexOf(N),1),Ae.freeThreadData(N.pthread),N.pthread=void 0})},runWithoutMainThreadQueuedCalls:function(N){o()[UI>>2]=0;try{N()}finally{o()[UI>>2]=1}},receiveObjectTransfer:function(N){},loadWasmModuleToWorker:function(N,F){N.onmessage=function(B){var K=B.data,be=K.cmd;if(N.pthread&&(Ae.currentProxiedOperationCallerThread=N.pthread.threadInfoStruct),K.targetThread&&K.targetThread!=lc()){var ge=Ae.pthreads[K.targetThread];ge?ge.worker.postMessage(B.data,K.transferList):console.error('Internal error! Worker sent a message "'+be+'" to target pthread '+K.targetThread+", but that thread no longer exists!"),Ae.currentProxiedOperationCallerThread=void 0;return}if(be==="processQueuedMainThreadWork")Yb();else if(be==="spawnThread")Wh(B.data);else if(be==="cleanupThread")Vy(K.thread);else if(be==="killThread")zy(K.thread);else if(be==="cancelThread")By(K.thread);else if(be==="loaded")N.loaded=!0,F&&F(N),N.runPthread&&(N.runPthread(),delete N.runPthread);else if(be==="print")U("Thread "+K.threadId+": "+K.text);else if(be==="printErr")H("Thread "+K.threadId+": "+K.text);else if(be==="alert")alert("Thread "+K.threadId+": "+K.text);else if(be==="exit"){var ye=N.pthread&&Atomics.load(s(),N.pthread.threadInfoStruct+64>>2);ye&&Ae.returnWorkerToPool(N)}else if(be==="exitProcess")try{_G(K.returnCode)}catch(Te){if(Te instanceof _m)return;throw Te}else be==="cancelDone"?Ae.returnWorkerToPool(N):be==="objectTransfer"?Ae.receiveObjectTransfer(B.data):B.data.target==="setimmediate"?N.postMessage(B.data):H("worker sent an unknown command "+be);Ae.currentProxiedOperationCallerThread=void 0},N.onerror=function(B){H("pthread sent an error! "+B.filename+":"+B.lineno+": "+B.message)},x&&(N.on("message",function(B){N.onmessage({data:B})}),N.on("error",function(B){N.onerror(B)}),N.on("exit",function(B){})),N.postMessage({cmd:"load",urlOrBlob:i.mainScriptUrlOrBlob||r,wasmMemory:ee,wasmModule:ie})},allocateUnusedWorker:function(){var N=I("tfjs-backend-wasm-threaded-simd.worker.js");Ae.unusedWorkers.push(new Worker(N))},getNewWorker:function(){return Ae.unusedWorkers.length==0&&(Ae.allocateUnusedWorker(),Ae.loadWasmModuleToWorker(Ae.unusedWorkers[0])),Ae.unusedWorkers.length>0?Ae.unusedWorkers.pop():null},busySpinWait:function(N){for(var F=performance.now()+N;performance.now()<F;);}};function Gy(N,F){WI(N,F),uc(N)}i.establishStackSpace=Gy;function Wy(){return re}i.getNoExitRuntime=Wy;function jy(N,F){return Ur.get(N)(F)}i.invokeEntryPoint=jy;function Uy(N,F,B,K){bi("Assertion failed: "+ve(N)+", at: "+[F?ve(F):"unknown filename",B,K?ve(K):"unknown function"])}function Hy(N,F){var B=_main(N,F)}var Gl;x?Gl=function(){var N=process.hrtime();return N[0]*1e3+N[1]/1e6}:w?Gl=function(){return performance.now()-i.__performance_now_clock_drift}:typeof dateNow!="undefined"?Gl=dateNow:Gl=function(){return performance.now()};function qy(N){return o()[PI()>>2]=N,N}function Ky(N,F){if(w)return Ua(1,1,N,F)}function Xy(N,F){if(N==F)postMessage({cmd:"processQueuedMainThreadWork"});else if(w)postMessage({targetThread:N,cmd:"processThreadQueue"});else{var B=Ae.pthreads[N],K=B&&B.worker;if(!K)return;K.postMessage({cmd:"processThreadQueue"})}return 1}function Yy(){bi()}function Zy(N,F,B){var K=rb(F,B);return Ly[N].apply(null,K)}function Jy(N,F){}function Qy(N,F,B){if(N<=0||N>t().length||N&!0)return-28;if(h){if(Atomics.load(o(),N>>2)!=F)return-6;for(var be=performance.now(),ge=be+B,ye=Atomics.exchange(o(),pc>>2,N);;){if(be=performance.now(),be>ge)return ye=Atomics.exchange(o(),pc>>2,0),-73;if(ye=Atomics.exchange(o(),pc>>2,0),ye==0)break;if(Yb(),Atomics.load(o(),N>>2)!=F)return-6;ye=Atomics.exchange(o(),pc>>2,N)}return 0}else{var K=Atomics.wait(o(),N>>2,F,B);if(K==="timed-out")return-73;if(K==="not-equal")return-6;if(K==="ok")return 0;throw"Atomics.wait returned an unexpected value "+K}}function eb(N,F,B){n().copyWithin(N,F,F+B)}function tb(){return x?dc("os").cpus().length:navigator.hardwareConcurrency}function Ua(N,F){for(var B=arguments.length-2,K=wm(),be=B,ge=cc(be*8),ye=ge>>3,Te=0;Te<B;Te++){var _t=arguments[2+Te];a()[ye+Te]=_t}var kr=GI(N,be,ge,F);return uc(K),kr}var fm=[],dm=[];function rb(N,F){dm.length=0;var B;for(F>>=2;B=n()[N++];){var K=B<105;K&&F&1&&F++,dm.push(K?a()[F++>>1]:o()[F]),++F}return dm}function nb(N,F,B){fm.length=F;for(var K=B>>3,be=0;be<F;be++)fm[be]=a()[K+be];var ge=N<0,ye=ge?Ly[-N-1]:Nb[N];return ye.apply(null,fm)}function ob(){return n().length}function sb(N){try{return ee.grow(N-He.byteLength+65535>>>16),ir(ee.buffer),1}catch(F){}}function ib(N){var F=ob();if(N<=F)return!1;var B=2147483648;if(N>B)return!1;for(var K=1;K<=4;K*=2){var be=F*(1+.2/K);be=Math.min(be,N+100663296);var ge=Math.min(B,Et(Math.max(N,be),65536)),ye=sb(ge);if(ye)return!0}return!1}var Je={inEventHandler:0,removeAllEventListeners:function(){for(var N=Je.eventHandlers.length-1;N>=0;--N)Je._removeHandler(N);Je.eventHandlers=[],Je.deferredCalls=[]},registerRemoveEventListeners:function(){Je.removeEventListenersRegistered||(zl.push(Je.removeAllEventListeners),Je.removeEventListenersRegistered=!0)},deferredCalls:[],deferCall:function(N,F,B){function K(ye,Te){if(ye.length!=Te.length)return!1;for(var _t in ye)if(ye[_t]!=Te[_t])return!1;return!0}for(var be in Je.deferredCalls){var ge=Je.deferredCalls[be];if(ge.targetFunction==N&&K(ge.argsList,B))return}Je.deferredCalls.push({targetFunction:N,precedence:F,argsList:B}),Je.deferredCalls.sort(function(ye,Te){return ye.precedence<Te.precedence})},removeDeferredCalls:function(N){for(var F=0;F<Je.deferredCalls.length;++F)Je.deferredCalls[F].targetFunction==N&&(Je.deferredCalls.splice(F,1),--F)},canPerformEventHandlerRequests:function(){return Je.inEventHandler&&Je.currentEventHandler.allowsDeferredCalls},runDeferredCalls:function(){if(!!Je.canPerformEventHandlerRequests())for(var N=0;N<Je.deferredCalls.length;++N){var F=Je.deferredCalls[N];Je.deferredCalls.splice(N,1),--N,F.targetFunction.apply(null,F.argsList)}},eventHandlers:[],removeAllHandlersOnTarget:function(N,F){for(var B=0;B<Je.eventHandlers.length;++B)Je.eventHandlers[B].target==N&&(!F||F==Je.eventHandlers[B].eventTypeString)&&Je._removeHandler(B--)},_removeHandler:function(N){var F=Je.eventHandlers[N];F.target.removeEventListener(F.eventTypeString,F.eventListenerFunc,F.useCapture),Je.eventHandlers.splice(N,1)},registerOrRemoveHandler:function(N){var F=function(be){++Je.inEventHandler,Je.currentEventHandler=N,Je.runDeferredCalls(),N.handlerFunc(be),Je.runDeferredCalls(),--Je.inEventHandler};if(N.callbackfunc)N.eventListenerFunc=F,N.target.addEventListener(N.eventTypeString,F,N.useCapture),Je.eventHandlers.push(N),Je.registerRemoveEventListeners();else for(var B=0;B<Je.eventHandlers.length;++B)Je.eventHandlers[B].target==N.target&&Je.eventHandlers[B].eventTypeString==N.eventTypeString&&Je._removeHandler(B--)},queueEventHandlerOnThread_iiii:function(N,F,B,K,be){var ge=wm(),ye=cc(12);o()[ye>>2]=B,o()[ye+4>>2]=K,o()[ye+8>>2]=be,Zb(0,N,637534208,F,K,ye),uc(ge)},getTargetThreadForEventCallback:function(N){switch(N){case 1:return 0;case 2:return Ae.currentProxiedOperationCallerThread;default:return N}},getNodeNameForTarget:function(N){return N?N==window?"#window":N==screen?"#screen":N&&N.nodeName?N.nodeName:"":""},fullscreenEnabled:function(){return document.fullscreenEnabled||document.webkitFullscreenEnabled}};function ab(N){var F=at(N)+1,B=Wl(F);return Ke(N,B,F),B}function lb(N,F,B,K){var be=wm(),ge=cc(12),ye=0;F&&(ye=ab(F)),o()[ge>>2]=ye,o()[ge+4>>2]=B,o()[ge+8>>2]=K,Zb(0,N,657457152,0,ye,ge),uc(be)}function ub(N,F,B,K){F=F?ve(F):"",lb(N,F,B,K)}function cb(N){return N>2?ve(N):N}var pb=[0,typeof document!="undefined"?document:0,typeof window!="undefined"?window:0];function mb(N){N=cb(N);var F=pb[N]||(typeof document!="undefined"?document.querySelector(N):void 0);return F}function hm(N){return mb(N)}function Lh(N,F,B){var K=hm(N);if(!K)return-4;if(K.canvasSharedPtr&&(o()[K.canvasSharedPtr>>2]=F,o()[K.canvasSharedPtr+4>>2]=B),K.offscreenCanvas||!K.controlTransferredOffscreen){K.offscreenCanvas&&(K=K.offscreenCanvas);var be=!1;if(K.GLctxObject&&K.GLctxObject.GLctx){var ge=K.GLctxObject.GLctx.getParameter(2978);be=ge[0]===0&&ge[1]===0&&ge[2]===K.width&&ge[3]===K.height}K.width=F,K.height=B,be&&K.GLctxObject.GLctx.viewport(0,0,F,B)}else if(K.canvasSharedPtr){var ye=o()[K.canvasSharedPtr+8>>2];return ub(ye,N,F,B),1}else return-4;return 0}function zh(N,F,B){return w?Ua(2,1,N,F,B):Lh(N,F,B)}function fb(N,F,B){var K=hm(N);return K?Lh(N,F,B):zh(N,F,B)}function db(N){}function hb(N,F){}function gb(N){var F=N.getExtension("ANGLE_instanced_arrays");if(F)return N.vertexAttribDivisor=function(B,K){F.vertexAttribDivisorANGLE(B,K)},N.drawArraysInstanced=function(B,K,be,ge){F.drawArraysInstancedANGLE(B,K,be,ge)},N.drawElementsInstanced=function(B,K,be,ge,ye){F.drawElementsInstancedANGLE(B,K,be,ge,ye)},1}function xb(N){var F=N.getExtension("OES_vertex_array_object");if(F)return N.createVertexArray=function(){return F.createVertexArrayOES()},N.deleteVertexArray=function(B){F.deleteVertexArrayOES(B)},N.bindVertexArray=function(B){F.bindVertexArrayOES(B)},N.isVertexArray=function(B){return F.isVertexArrayOES(B)},1}function yb(N){var F=N.getExtension("WEBGL_draw_buffers");if(F)return N.drawBuffers=function(B,K){F.drawBuffersWEBGL(B,K)},1}function bb(N){return!!(N.multiDrawWebgl=N.getExtension("WEBGL_multi_draw"))}var ft={counter:1,buffers:[],programs:[],framebuffers:[],renderbuffers:[],textures:[],uniforms:[],shaders:[],vaos:[],contexts:{},offscreenCanvases:{},timerQueriesEXT:[],programInfos:{},stringCache:{},unpackAlignment:4,recordError:function(F){ft.lastError||(ft.lastError=F)},getNewId:function(N){for(var F=ft.counter++,B=N.length;B<F;B++)N[B]=null;return F},getSource:function(N,F,B,K){for(var be="",ge=0;ge<F;++ge){var ye=K?o()[K+ge*4>>2]:-1;be+=ve(o()[B+ge*4>>2],ye<0?void 0:ye)}return be},createContext:function(N,F){var B=N.getContext("webgl",F);if(!B)return 0;var K=ft.registerContext(B,F);return K},registerContext:function(N,F){var B=Wl(8);o()[B+4>>2]=lc();var K={handle:B,attributes:F,version:F.majorVersion,GLctx:N};return N.canvas&&(N.canvas.GLctxObject=K),ft.contexts[B]=K,(typeof F.enableExtensionsByDefault=="undefined"||F.enableExtensionsByDefault)&&ft.initExtensions(K),B},makeContextCurrent:function(N){return ft.currentContext=ft.contexts[N],i.ctx=Ha=ft.currentContext&&ft.currentContext.GLctx,!(N&&!Ha)},getContext:function(N){return ft.contexts[N]},deleteContext:function(N){ft.currentContext===ft.contexts[N]&&(ft.currentContext=null),typeof Je=="object"&&Je.removeAllHandlersOnTarget(ft.contexts[N].GLctx.canvas),ft.contexts[N]&&ft.contexts[N].GLctx.canvas&&(ft.contexts[N].GLctx.canvas.GLctxObject=void 0),bm(ft.contexts[N].handle),ft.contexts[N]=null},initExtensions:function(N){if(N||(N=ft.currentContext),!N.initExtensionsDone){N.initExtensionsDone=!0;var F=N.GLctx;gb(F),xb(F),yb(F),F.disjointTimerQueryExt=F.getExtension("EXT_disjoint_timer_query"),bb(F);var B=F.getSupportedExtensions()||[];B.forEach(function(K){K.indexOf("lose_context")<0&&K.indexOf("debug")<0&&F.getExtension(K)})}},populateUniformTable:function(N){for(var F=ft.programs[N],B=ft.programInfos[N]={uniforms:{},maxUniformLength:0,maxAttributeLength:-1,maxUniformBlockNameLength:-1},K=B.uniforms,be=Ha.getProgramParameter(F,35718),ge=0;ge<be;++ge){var ye=Ha.getActiveUniform(F,ge),Te=ye.name;B.maxUniformLength=Math.max(B.maxUniformLength,Te.length+1),Te.slice(-1)=="]"&&(Te=Te.slice(0,Te.lastIndexOf("[")));var _t=Ha.getUniformLocation(F,Te);if(_t){var kr=ft.getNewId(ft.uniforms);K[Te]=[ye.size,kr],ft.uniforms[kr]=_t;for(var gr=1;gr<ye.size;++gr){var Ka=Te+"["+gr+"]";_t=Ha.getUniformLocation(F,Ka),kr=ft.getNewId(ft.uniforms),ft.uniforms[kr]=_t}}}}},wb=["default","low-power","high-performance"];function _b(N,F){var B=F>>2,K=o()[B+(24>>2)],be={alpha:!!o()[B+(0>>2)],depth:!!o()[B+(4>>2)],stencil:!!o()[B+(8>>2)],antialias:!!o()[B+(12>>2)],premultipliedAlpha:!!o()[B+(16>>2)],preserveDrawingBuffer:!!o()[B+(20>>2)],powerPreference:wb[K],failIfMajorPerformanceCaveat:!!o()[B+(28>>2)],majorVersion:o()[B+(32>>2)],minorVersion:o()[B+(36>>2)],enableExtensionsByDefault:o()[B+(40>>2)],explicitSwapControl:o()[B+(44>>2)],proxyContextToMainThread:o()[B+(48>>2)],renderViaOffscreenBackBuffer:o()[B+(52>>2)]},ge=hm(N);if(!ge||be.explicitSwapControl)return 0;var ye=ft.createContext(ge,be);return ye}function kb(N,F){return _b(N,F)}var ic={mappings:{},buffers:[null,[],[]],printChar:function(N,F){var B=ic.buffers[N];F===0||F===10?((N===1?U:H)(De(B,0)),B.length=0):B.push(F)},varargs:void 0,get:function(){ic.varargs+=4;var N=o()[ic.varargs-4>>2];return N},getStr:function(N){var F=ve(N);return F},get64:function(N,F){return N}};function Bh(N){return w?Ua(3,1,N):0}function Vh(N,F,B,K,be){if(w)return Ua(4,1,N,F,B,K,be)}function Gh(N,F,B,K){if(w)return Ua(5,1,N,F,B,K);for(var be=0,ge=0;ge<B;ge++){for(var ye=o()[F+ge*8>>2],Te=o()[F+(ge*8+4)>>2],_t=0;_t<Te;_t++)ic.printChar(N,n()[ye+_t]);be+=Te}return o()[K>>2]=be,0}function vb(N){var F=Ae.threadExitHandlers.pop();N&&F()}function Cb(N,F){Ae.threadExitHandlers.push(function(){Ur.get(N)(F)})}function Wh(N){if(w)throw"Internal Error! spawnThread() can only ever be called from main application thread!";var F=Ae.getNewWorker();if(F.pthread!==void 0)throw"Internal error!";if(!N.pthread_ptr)throw"Internal error, no pthread ptr!";Ae.runningWorkers.push(F);for(var B=Wl(128*4),K=0;K<128;++K)o()[B+K*4>>2]=0;var be=N.stackBase+N.stackSize,ge=Ae.pthreads[N.pthread_ptr]={worker:F,stackBase:N.stackBase,stackSize:N.stackSize,allocatedOwnStack:N.allocatedOwnStack,threadInfoStruct:N.pthread_ptr},ye=ge.threadInfoStruct>>2;Atomics.store(s(),ye+(64>>2),N.detached),Atomics.store(s(),ye+(100>>2),B),Atomics.store(s(),ye+(40>>2),ge.threadInfoStruct),Atomics.store(s(),ye+(80>>2),N.stackSize),Atomics.store(s(),ye+(76>>2),be),Atomics.store(s(),ye+(104>>2),N.stackSize),Atomics.store(s(),ye+(104+8>>2),be),Atomics.store(s(),ye+(104+12>>2),N.detached);var Te=MI(),_t=Te+40;Atomics.store(s(),ye+(172>>2),_t),F.pthread=ge;var kr={cmd:"run",start_routine:N.startRoutine,arg:N.arg,threadInfoStruct:N.pthread_ptr,stackBase:N.stackBase,stackSize:N.stackSize};F.runPthread=function(){kr.time=performance.now(),F.postMessage(kr,N.transferList)},F.loaded&&(F.runPthread(),delete F.runPthread)}function Ib(N,F,B,K){if(typeof SharedArrayBuffer=="undefined")return H("Current environment does not support SharedArrayBuffer, pthreads are not available!"),6;if(!N)return H("pthread_create called with a null thread pointer!"),28;var be=[],ge=0;if(w&&(be.length===0||ge))return VI(687865856,N,F,B,K);if(ge)return ge;var ye=0,Te=0,_t=0;F&&F!=-1?(ye=o()[F>>2],ye+=81920,Te=o()[F+8>>2],_t=o()[F+12>>2]!==0):ye=2097152;var kr=Te==0;kr?Te=jI(16,ye):(Te-=ye,de(Te>0));for(var gr=Wl(228),Ka=0;Ka<228>>2;++Ka)s()[(gr>>2)+Ka]=0;o()[N>>2]=gr,o()[gr+12>>2]=gr;var mc=gr+152;o()[mc>>2]=mc;var Hr={stackBase:Te,stackSize:ye,allocatedOwnStack:kr,detached:_t,startRoutine:B,pthread_ptr:gr,arg:K,transferList:be};return w?(Hr.cmd="spawnThread",postMessage(Hr,be)):Wh(Hr),0}function jh(N){if(w)return Ua(6,1,N);switch(N){case 30:return 16384;case 85:var F=2147483648;return F/16384;case 132:case 133:case 12:case 137:case 138:case 15:case 235:case 16:case 17:case 18:case 19:case 20:case 149:case 13:case 10:case 236:case 153:case 9:case 21:case 22:case 159:case 154:case 14:case 77:case 78:case 139:case 82:case 68:case 67:case 164:case 11:case 29:case 47:case 48:case 95:case 52:case 51:case 46:return 200809;case 27:case 246:case 127:case 128:case 23:case 24:case 160:case 161:case 181:case 182:case 242:case 183:case 184:case 243:case 244:case 245:case 165:case 178:case 179:case 49:case 50:case 168:case 169:case 175:case 170:case 171:case 172:case 97:case 76:case 32:case 173:case 35:case 80:case 81:case 79:return-1;case 176:case 177:case 7:case 155:case 8:case 157:case 125:case 126:case 92:case 93:case 129:case 130:case 131:case 94:case 91:return 1;case 74:case 60:case 69:case 70:case 4:return 1024;case 31:case 42:case 72:return 32;case 87:case 26:case 33:return 2147483647;case 34:case 1:return 47839;case 38:case 36:return 99;case 43:case 37:return 2048;case 0:return 2097152;case 3:return 65536;case 28:return 32768;case 44:return 32767;case 75:return 16384;case 39:return 1e3;case 89:return 700;case 71:return 256;case 40:return 255;case 2:return 100;case 180:return 64;case 25:return 20;case 5:return 16;case 6:return 6;case 73:return 4;case 84:return typeof navigator=="object"&&navigator.hardwareConcurrency||1}return qy(28),-1}w||Ae.initMainThreadBlock();var Ha,Nb=[null,Ky,zh,Bh,Vh,Gh,jh],Sb={e:Uy,r:Hy,x:Xy,b:Yy,y:Zy,j:Jy,c:Qy,d:mm,f:Gl,p:eb,z:tb,u:nb,q:ib,v:fb,i:db,t:hb,w:kb,m:Bh,n:Vh,g:Gh,o:Mh,a:ee||i.wasmMemory,k:vb,l:Cb,h:Ib,s:jh},OI=My(),Uh=i.___wasm_call_ctors=function(){return(Uh=i.___wasm_call_ctors=i.asm.A).apply(null,arguments)},Tb=i._init=function(){return(Tb=i._init=i.asm.B).apply(null,arguments)},Ab=i._register_tensor=function(){return(Ab=i._register_tensor=i.asm.C).apply(null,arguments)},Eb=i._dispose_data=function(){return(Eb=i._dispose_data=i.asm.D).apply(null,arguments)},Db=i._dispose=function(){return(Db=i._dispose=i.asm.E).apply(null,arguments)},$b=i._Abs=function(){return($b=i._Abs=i.asm.G).apply(null,arguments)},Rb=i._Add=function(){return(Rb=i._Add=i.asm.H).apply(null,arguments)},Fb=i._AddN=function(){return(Fb=i._AddN=i.asm.I).apply(null,arguments)},Ob=i._All=function(){return(Ob=i._All=i.asm.J).apply(null,arguments)},Pb=i._Any=function(){return(Pb=i._Any=i.asm.K).apply(null,arguments)},Mb=i._ArgMax=function(){return(Mb=i._ArgMax=i.asm.L).apply(null,arguments)},Lb=i._AvgPool=function(){return(Lb=i._AvgPool=i.asm.M).apply(null,arguments)},zb=i._BatchMatMul=function(){return(zb=i._BatchMatMul=i.asm.N).apply(null,arguments)},Bb=i._Ceil=function(){return(Bb=i._Ceil=i.asm.O).apply(null,arguments)},Vb=i._ClipByValue=function(){return(Vb=i._ClipByValue=i.asm.P).apply(null,arguments)},Gb=i._Conv2D=function(){return(Gb=i._Conv2D=i.asm.Q).apply(null,arguments)},Wb=i._Conv2DBackpropInput=function(){return(Wb=i._Conv2DBackpropInput=i.asm.R).apply(null,arguments)},jb=i._Cos=function(){return(jb=i._Cos=i.asm.S).apply(null,arguments)},Ub=i._CropAndResize=function(){return(Ub=i._CropAndResize=i.asm.T).apply(null,arguments)},Hb=i._Cumsum=function(){return(Hb=i._Cumsum=i.asm.U).apply(null,arguments)},qb=i._DepthToSpace=function(){return(qb=i._DepthToSpace=i.asm.V).apply(null,arguments)},Kb=i._DepthwiseConv2dNative=function(){return(Kb=i._DepthwiseConv2dNative=i.asm.W).apply(null,arguments)},Hh=i._Equal=function(){return(Hh=i._Equal=i.asm.X).apply(null,arguments)},qh=i._Exp=function(){return(qh=i._Exp=i.asm.Y).apply(null,arguments)},Kh=i._FlipLeftRight=function(){return(Kh=i._FlipLeftRight=i.asm.Z).apply(null,arguments)},gm=i._Floor=function(){return(gm=i._Floor=i.asm._).apply(null,arguments)},ac=i._FloorDiv=function(){return(ac=i._FloorDiv=i.asm.$).apply(null,arguments)},Xb=i._FusedBatchNorm=function(){return(Xb=i._FusedBatchNorm=i.asm.aa).apply(null,arguments)},xm=i._FusedConv2D=function(){return(xm=i._FusedConv2D=i.asm.ba).apply(null,arguments)},Z=i._FusedDepthwiseConv2D=function(){return(Z=i._FusedDepthwiseConv2D=i.asm.ca).apply(null,arguments)},se=i._Gather=function(){return(se=i._Gather=i.asm.da).apply(null,arguments)},ke=i._GatherNd=function(){return(ke=i._GatherNd=i.asm.ea).apply(null,arguments)},lt=i._Greater=function(){return(lt=i._Greater=i.asm.fa).apply(null,arguments)},Jt=i._GreaterEqual=function(){return(Jt=i._GreaterEqual=i.asm.ga).apply(null,arguments)},Bt=i._LeakyRelu=function(){return(Bt=i._LeakyRelu=i.asm.ha).apply(null,arguments)},tt=i._Less=function(){return(tt=i._Less=i.asm.ia).apply(null,arguments)},rt=i._LessEqual=function(){return(rt=i._LessEqual=i.asm.ja).apply(null,arguments)},Nr=i._Log=function(){return(Nr=i._Log=i.asm.ka).apply(null,arguments)},wi=i._LogicalAnd=function(){return(wi=i._LogicalAnd=i.asm.la).apply(null,arguments)},_i=i._Max=function(){return(_i=i._Max=i.asm.ma).apply(null,arguments)},Xh=i._MaxPool=function(){return(Xh=i._MaxPool=i.asm.na).apply(null,arguments)},ym=i._Maximum=function(){return(ym=i._Maximum=i.asm.oa).apply(null,arguments)},In=i._Mean=function(){return(In=i._Mean=i.asm.pa).apply(null,arguments)},qa=i._Min=function(){return(qa=i._Min=i.asm.qa).apply(null,arguments)},Yh=i._Minimum=function(){return(Yh=i._Minimum=i.asm.ra).apply(null,arguments)},FV=i._MirrorPad=function(){return(FV=i._MirrorPad=i.asm.sa).apply(null,arguments)},OV=i._Multiply=function(){return(OV=i._Multiply=i.asm.ta).apply(null,arguments)},PV=i._Neg=function(){return(PV=i._Neg=i.asm.ua).apply(null,arguments)},MV=i._NonMaxSuppressionV3=function(){return(MV=i._NonMaxSuppressionV3=i.asm.va).apply(null,arguments)},LV=i._NonMaxSuppressionV4=function(){return(LV=i._NonMaxSuppressionV4=i.asm.wa).apply(null,arguments)},zV=i._NonMaxSuppressionV5=function(){return(zV=i._NonMaxSuppressionV5=i.asm.xa).apply(null,arguments)},BV=i._NotEqual=function(){return(BV=i._NotEqual=i.asm.ya).apply(null,arguments)},VV=i._OneHot=function(){return(VV=i._OneHot=i.asm.za).apply(null,arguments)},GV=i._PadV2=function(){return(GV=i._PadV2=i.asm.Aa).apply(null,arguments)},WV=i._Pow=function(){return(WV=i._Pow=i.asm.Ba).apply(null,arguments)},jV=i._Prelu=function(){return(jV=i._Prelu=i.asm.Ca).apply(null,arguments)},UV=i._Prod=function(){return(UV=i._Prod=i.asm.Da).apply(null,arguments)},HV=i._RealDiv=function(){return(HV=i._RealDiv=i.asm.Ea).apply(null,arguments)},qV=i._Relu=function(){return(qV=i._Relu=i.asm.Fa).apply(null,arguments)},KV=i._Relu6=function(){return(KV=i._Relu6=i.asm.Ga).apply(null,arguments)},XV=i._ResizeBilinear=function(){return(XV=i._ResizeBilinear=i.asm.Ha).apply(null,arguments)},YV=i._Reverse=function(){return(YV=i._Reverse=i.asm.Ia).apply(null,arguments)},ZV=i._RotateWithOffset=function(){return(ZV=i._RotateWithOffset=i.asm.Ja).apply(null,arguments)},JV=i._Round=function(){return(JV=i._Round=i.asm.Ka).apply(null,arguments)},QV=i._Rsqrt=function(){return(QV=i._Rsqrt=i.asm.La).apply(null,arguments)},eG=i._ScatterNd=function(){return(eG=i._ScatterNd=i.asm.Ma).apply(null,arguments)},tG=i._SelectV2=function(){return(tG=i._SelectV2=i.asm.Na).apply(null,arguments)},rG=i._Sigmoid=function(){return(rG=i._Sigmoid=i.asm.Oa).apply(null,arguments)},nG=i._Sin=function(){return(nG=i._Sin=i.asm.Pa).apply(null,arguments)},oG=i._Softmax=function(){return(oG=i._Softmax=i.asm.Qa).apply(null,arguments)},sG=i._Sqrt=function(){return(sG=i._Sqrt=i.asm.Ra).apply(null,arguments)},iG=i._Square=function(){return(iG=i._Square=i.asm.Sa).apply(null,arguments)},aG=i._SquaredDifference=function(){return(aG=i._SquaredDifference=i.asm.Ta).apply(null,arguments)},lG=i._Step=function(){return(lG=i._Step=i.asm.Ua).apply(null,arguments)},uG=i._StridedSlice=function(){return(uG=i._StridedSlice=i.asm.Va).apply(null,arguments)},cG=i._Sub=function(){return(cG=i._Sub=i.asm.Wa).apply(null,arguments)},pG=i._Sum=function(){return(pG=i._Sum=i.asm.Xa).apply(null,arguments)},mG=i._Tan=function(){return(mG=i._Tan=i.asm.Ya).apply(null,arguments)},fG=i._Tanh=function(){return(fG=i._Tanh=i.asm.Za).apply(null,arguments)},dG=i._Tile=function(){return(dG=i._Tile=i.asm._a).apply(null,arguments)},hG=i._TopK=function(){return(hG=i._TopK=i.asm.$a).apply(null,arguments)},gG=i._Transform=function(){return(gG=i._Transform=i.asm.ab).apply(null,arguments)},xG=i._Transpose=function(){return(xG=i._Transpose=i.asm.bb).apply(null,arguments)},yG=i.__FusedMatMul=function(){return(yG=i.__FusedMatMul=i.asm.cb).apply(null,arguments)},Wl=i._malloc=function(){return(Wl=i._malloc=i.asm.db).apply(null,arguments)},bm=i._free=function(){return(bm=i._free=i.asm.eb).apply(null,arguments)},PI=i.___errno_location=function(){return(PI=i.___errno_location=i.asm.fb).apply(null,arguments)},MI=i._emscripten_get_global_libc=function(){return(MI=i._emscripten_get_global_libc=i.asm.gb).apply(null,arguments)},lc=i._pthread_self=function(){return(lc=i._pthread_self=i.asm.hb).apply(null,arguments)},LI=i.___pthread_tsd_run_dtors=function(){return(LI=i.___pthread_tsd_run_dtors=i.asm.ib).apply(null,arguments)},Yb=i._emscripten_main_thread_process_queued_calls=function(){return(Yb=i._emscripten_main_thread_process_queued_calls=i.asm.jb).apply(null,arguments)},bG=i._emscripten_current_thread_process_queued_calls=function(){return(bG=i._emscripten_current_thread_process_queued_calls=i.asm.kb).apply(null,arguments)},zI=i._emscripten_register_main_browser_thread_id=function(){return(zI=i._emscripten_register_main_browser_thread_id=i.asm.lb).apply(null,arguments)},BI=i.__emscripten_do_dispatch_to_thread=function(){return(BI=i.__emscripten_do_dispatch_to_thread=i.asm.mb).apply(null,arguments)},VI=i._emscripten_sync_run_in_main_thread_4=function(){return(VI=i._emscripten_sync_run_in_main_thread_4=i.asm.nb).apply(null,arguments)},GI=i._emscripten_run_in_main_runtime_thread_js=function(){return(GI=i._emscripten_run_in_main_runtime_thread_js=i.asm.ob).apply(null,arguments)},Zb=i.__emscripten_call_on_thread=function(){return(Zb=i.__emscripten_call_on_thread=i.asm.pb).apply(null,arguments)},wG=i._emscripten_tls_init=function(){return(wG=i._emscripten_tls_init=i.asm.qb).apply(null,arguments)},Jb=i.__emscripten_thread_init=function(){return(Jb=i.__emscripten_thread_init=i.asm.rb).apply(null,arguments)},wm=i.stackSave=function(){return(wm=i.stackSave=i.asm.sb).apply(null,arguments)},uc=i.stackRestore=function(){return(uc=i.stackRestore=i.asm.tb).apply(null,arguments)},cc=i.stackAlloc=function(){return(cc=i.stackAlloc=i.asm.ub).apply(null,arguments)},WI=i._emscripten_stack_set_limits=function(){return(WI=i._emscripten_stack_set_limits=i.asm.vb).apply(null,arguments)},jI=i._memalign=function(){return(jI=i._memalign=i.asm.wb).apply(null,arguments)},UI=i.__emscripten_allow_main_runtime_queued_calls=9808,pc=i.__emscripten_main_thread_futex=11432;i.cwrap=we,i.PThread=Ae,i.PThread=Ae,i.wasmMemory=ee,i.ExitStatus=_m;var Zh;function _m(N){this.name="ExitStatus",this.message="Program terminated with exit("+N+")",this.status=N}Vl=function N(){Zh||Qb(),Zh||(Vl=N)};function Qb(N){if(N=N||m,yi>0)return;if(w){l(i),Bl(),postMessage({cmd:"loaded"});return}if(pm(),yi>0)return;function F(){Zh||(Zh=!0,i.calledRun=!0,!me&&(Bl(),Dy(),l(i),i.onRuntimeInitialized&&i.onRuntimeInitialized(),mn()))}i.setStatus?(i.setStatus("Running..."),setTimeout(function(){setTimeout(function(){i.setStatus("")},1),F()},1)):F()}i.run=Qb;function _G(N,F){if(!(F&&re&&N===0)){if(!F&&w)throw postMessage({cmd:"exitProcess",returnCode:N}),new _m(N);re||(Ae.terminateAllThreads(),ae=N,Dh(),i.onExit&&i.onExit(N),me=!0),d(N,new _m(N))}}if(i.preInit)for(typeof i.preInit=="function"&&(i.preInit=[i.preInit]);i.preInit.length>0;)i.preInit.pop()();return w&&(re=!1,Ae.initWorker()),Qb(),e.ready}}();typeof Ax=="object"&&typeof MC=="object"?MC.exports=PC:typeof define=="function"&&define.amd?define([],function(){return PC}):typeof Ax=="object"&&(Ax.WasmBackendModuleThreadedSimd=PC)});var XM=qt((Ex,zC)=>{var LC=function(){var r=typeof document!="undefined"&&document.currentScript?document.currentScript.src:void 0;return typeof __filename!="undefined"&&(r=r||__filename),function(e){e=e||{};var t=typeof e!="undefined"?e:{},n,o;t.ready=new Promise(function(Z,se){n=Z,o=se});var s={},a;for(a in t)t.hasOwnProperty(a)&&(s[a]=t[a]);var i=[],l="./this.program",u=function(Z,se){throw se},c=!1,p=!1,m=!1,f=!1;c=typeof window=="object",p=typeof importScripts=="function",m=typeof process=="object"&&typeof process.versions=="object"&&typeof process.versions.node=="string",f=!c&&!m&&!p;var d="";function h(Z){return t.locateFile?t.locateFile(Z,d):d+Z}var g,x,b,w,_,I;m?(p?d=Pp().dirname(d)+"/":d=__dirname+"/",g=function(se,ke){return _||(_=dc("fs")),I||(I=Pp()),se=I.normalize(se),_.readFileSync(se,ke?null:"utf8")},b=function(se){var ke=g(se,!0);return ke.buffer||(ke=new Uint8Array(ke)),U(ke.buffer),ke},process.argv.length>1&&(l=process.argv[1].replace(/\\/g,"/")),i=process.argv.slice(2),process.on("uncaughtException",function(Z){if(!(Z instanceof Xb))throw Z}),process.on("unhandledRejection",Ps),u=function(Z){process.exit(Z)},t.inspect=function(){return"[Emscripten Module object]"}):f?(typeof read!="undefined"&&(g=function(se){return read(se)}),b=function(se){var ke;return typeof readbuffer=="function"?new Uint8Array(readbuffer(se)):(ke=read(se,"binary"),U(typeof ke=="object"),ke)},typeof scriptArgs!="undefined"?i=scriptArgs:typeof arguments!="undefined"&&(i=arguments),typeof quit=="function"&&(u=function(Z){quit(Z)}),typeof print!="undefined"&&(typeof console=="undefined"&&(console={}),console.log=print,console.warn=console.error=typeof printErr!="undefined"?printErr:print)):(c||p)&&(p?d=self.location.href:typeof document!="undefined"&&document.currentScript&&(d=document.currentScript.src),r&&(d=r),d.indexOf("blob:")!==0?d=d.substr(0,d.lastIndexOf("/")+1):d="",g=function(Z){var se=new XMLHttpRequest;return se.open("GET",Z,!1),se.send(null),se.responseText},p&&(b=function(Z){var se=new XMLHttpRequest;return se.open("GET",Z,!1),se.responseType="arraybuffer",se.send(null),new Uint8Array(se.response)}),x=function(Z,se,ke){var lt=new XMLHttpRequest;lt.open("GET",Z,!0),lt.responseType="arraybuffer",lt.onload=function(){if(lt.status==200||lt.status==0&<.response){se(lt.response);return}ke()},lt.onerror=ke,lt.send(null)},w=function(Z){document.title=Z});var E=t.print||console.log.bind(console),D=t.printErr||console.warn.bind(console);for(a in s)s.hasOwnProperty(a)&&(t[a]=s[a]);s=null,t.arguments&&(i=t.arguments),t.thisProgram&&(l=t.thisProgram),t.quit&&(u=t.quit);var $;t.wasmBinary&&($=t.wasmBinary);var R=t.noExitRuntime||!0;typeof WebAssembly!="object"&&Ps("no native wasm support detected");var M,G=!1,j;function U(Z,se){Z||Ps("Assertion failed: "+se)}function H(Z){var se=t["_"+Z];return U(se,"Cannot call unknown function "+Z+", make sure it is exported"),se}function q(Z,se,ke,lt,Jt){var Bt={string:function(In){var qa=0;if(In!=null&&In!==0){var Yh=(In.length<<2)+1;qa=gm(Yh),ie(In,qa,Yh)}return qa},array:function(In){var qa=gm(In.length);return me(In,qa),qa}};function tt(In){return se==="string"?re(In):se==="boolean"?Boolean(In):In}var rt=H(Z),Nr=[],wi=0;if(lt)for(var _i=0;_i<lt.length;_i++){var Xh=Bt[ke[_i]];Xh?(wi===0&&(wi=qh()),Nr[_i]=Xh(lt[_i])):Nr[_i]=lt[_i]}var ym=rt.apply(null,Nr);return ym=tt(ym),wi!==0&&Kh(wi),ym}function X(Z,se,ke,lt){ke=ke||[];var Jt=ke.every(function(tt){return tt==="number"}),Bt=se!=="string";return Bt&&Jt&&!lt?H(Z):function(){return q(Z,se,ke,arguments,lt)}}var ne=typeof TextDecoder!="undefined"?new TextDecoder("utf8"):void 0;function Y(Z,se,ke){for(var lt=se+ke,Jt=se;Z[Jt]&&!(Jt>=lt);)++Jt;if(Jt-se>16&&Z.subarray&&ne)return ne.decode(Z.subarray(se,Jt));for(var Bt="";se<Jt;){var tt=Z[se++];if(!(tt&128)){Bt+=String.fromCharCode(tt);continue}var rt=Z[se++]&63;if((tt&224)==192){Bt+=String.fromCharCode((tt&31)<<6|rt);continue}var Nr=Z[se++]&63;if((tt&240)==224?tt=(tt&15)<<12|rt<<6|Nr:tt=(tt&7)<<18|rt<<12|Nr<<6|Z[se++]&63,tt<65536)Bt+=String.fromCharCode(tt);else{var wi=tt-65536;Bt+=String.fromCharCode(55296|wi>>10,56320|wi&1023)}}return Bt}function re(Z,se){return Z?Y(xe,Z,se):""}function ee(Z,se,ke,lt){if(!(lt>0))return 0;for(var Jt=ke,Bt=ke+lt-1,tt=0;tt<Z.length;++tt){var rt=Z.charCodeAt(tt);if(rt>=55296&&rt<=57343){var Nr=Z.charCodeAt(++tt);rt=65536+((rt&1023)<<10)|Nr&1023}if(rt<=127){if(ke>=Bt)break;se[ke++]=rt}else if(rt<=2047){if(ke+1>=Bt)break;se[ke++]=192|rt>>6,se[ke++]=128|rt&63}else if(rt<=65535){if(ke+2>=Bt)break;se[ke++]=224|rt>>12,se[ke++]=128|rt>>6&63,se[ke++]=128|rt&63}else{if(ke+3>=Bt)break;se[ke++]=240|rt>>18,se[ke++]=128|rt>>12&63,se[ke++]=128|rt>>6&63,se[ke++]=128|rt&63}}return se[ke]=0,ke-Jt}function ie(Z,se,ke){return ee(Z,xe,se,ke)}function me(Z,se){he.set(Z,se)}function ae(Z,se){return Z%se>0&&(Z+=se-Z%se),Z}var de,he,xe,we,De,ve,Ge,Ke,at;function At(Z){de=Z,t.HEAP8=he=new Int8Array(Z),t.HEAP16=we=new Int16Array(Z),t.HEAP32=ve=new Int32Array(Z),t.HEAPU8=xe=new Uint8Array(Z),t.HEAPU16=De=new Uint16Array(Z),t.HEAPU32=Ge=new Uint32Array(Z),t.HEAPF32=Ke=new Float32Array(Z),t.HEAPF64=at=new Float64Array(Z)}var Et=t.INITIAL_MEMORY||16777216,He,ct=[],mt=[],zt=[],vn=[],Zt=!1;mt.push({func:function(){Mh()}});function cn(){if(t.preRun)for(typeof t.preRun=="function"&&(t.preRun=[t.preRun]);t.preRun.length;)Cn(t.preRun.shift());ja(ct)}function Mr(){Zt=!0,ja(mt)}function Un(){ja(zt)}function ir(){if(t.postRun)for(typeof t.postRun=="function"&&(t.postRun=[t.postRun]);t.postRun.length;)Ur(t.postRun.shift());ja(vn)}function Cn(Z){ct.unshift(Z)}function Ur(Z){vn.unshift(Z)}var _r=0,pn=null,Fn=null;function zl(Z){_r++,t.monitorRunDependencies&&t.monitorRunDependencies(_r)}function Wa(Z){if(_r--,t.monitorRunDependencies&&t.monitorRunDependencies(_r),_r==0&&(pn!==null&&(clearInterval(pn),pn=null),Fn)){var se=Fn;Fn=null,se()}}t.preloadedImages={},t.preloadedAudios={};function Ps(Z){t.onAbort&&t.onAbort(Z),Z+="",D(Z),G=!0,j=1,Z="abort("+Z+"). Build with -s ASSERTIONS=1 for more info.";var se=new WebAssembly.RuntimeError(Z);throw o(se),se}function xi(Z,se){return String.prototype.startsWith?Z.startsWith(se):Z.indexOf(se)===0}var pm="data:application/octet-stream;base64,";function Bl(Z){return xi(Z,pm)}var Dy="file://";function Dh(Z){return xi(Z,Dy)}var mn="tfjs-backend-wasm.wasm";Bl(mn)||(mn=h(mn));function $h(Z){try{if(Z==mn&&$)return new Uint8Array($);if(b)return b(Z);throw"both async and sync fetching of the wasm failed"}catch(se){Ps(se)}}function $y(){if(!$&&(c||p)){if(typeof fetch=="function"&&!Dh(mn))return fetch(mn,{credentials:"same-origin"}).then(function(Z){if(!Z.ok)throw"failed to load wasm binary file at '"+mn+"'";return Z.arrayBuffer()}).catch(function(){return $h(mn)});if(x)return new Promise(function(Z,se){x(mn,function(ke){Z(new Uint8Array(ke))},se)})}return Promise.resolve().then(function(){return $h(mn)})}function yi(){var Z={a:My};function se(tt,rt){var Nr=tt.exports;t.asm=Nr,M=t.asm.i,At(M.buffer),He=t.asm.o,Wa("wasm-instantiate")}zl("wasm-instantiate");function ke(tt){se(tt.instance)}function lt(tt){return $y().then(function(rt){return WebAssembly.instantiate(rt,Z)}).then(tt,function(rt){D("failed to asynchronously prepare wasm: "+rt),Ps(rt)})}function Jt(){return!$&&typeof WebAssembly.instantiateStreaming=="function"&&!Bl(mn)&&!Dh(mn)&&typeof fetch=="function"?fetch(mn,{credentials:"same-origin"}).then(function(tt){var rt=WebAssembly.instantiateStreaming(tt,Z);return rt.then(ke,function(Nr){return D("wasm streaming compile failed: "+Nr),D("falling back to ArrayBuffer instantiation"),lt(ke)})}):lt(ke)}if(t.instantiateWasm)try{var Bt=t.instantiateWasm(Z,se);return Bt}catch(tt){return D("Module.instantiateWasm callback failed with error: "+tt),!1}return Jt().catch(o),{}}function ja(Z){for(;Z.length>0;){var se=Z.shift();if(typeof se=="function"){se(t);continue}var ke=se.func;typeof ke=="number"?se.arg===void 0?He.get(ke)():He.get(ke)(se.arg):ke(se.arg===void 0?null:se.arg)}}function Vl(){Ps()}function Ry(Z,se,ke){xe.copyWithin(Z,se,se+ke)}function Fy(){return xe.length}function bi(Z){try{return M.grow(Z-de.byteLength+65535>>>16),At(M.buffer),1}catch(se){}}function Rh(Z){var se=Fy(),ke=2147483648;if(Z>ke)return!1;for(var lt=1;lt<=4;lt*=2){var Jt=se*(1+.2/lt);Jt=Math.min(Jt,Z+100663296);var Bt=Math.min(ke,ae(Math.max(Z,Jt),65536)),tt=bi(Bt);if(tt)return!0}return!1}var oc={mappings:{},buffers:[null,[],[]],printChar:function(Z,se){var ke=oc.buffers[Z];se===0||se===10?((Z===1?E:D)(Y(ke,0)),ke.length=0):ke.push(se)},varargs:void 0,get:function(){oc.varargs+=4;var Z=ve[oc.varargs-4>>2];return Z},getStr:function(Z){var se=re(Z);return se},get64:function(Z,se){return Z}};function Fh(Z){return 0}function Oy(Z,se,ke,lt,Jt){}function Oh(Z,se,ke,lt){for(var Jt=0,Bt=0;Bt<ke;Bt++){for(var tt=ve[se+Bt*8>>2],rt=ve[se+(Bt*8+4)>>2],Nr=0;Nr<rt;Nr++)oc.printChar(Z,xe[tt+Nr]);Jt+=rt}return ve[lt>>2]=Jt,0}function fn(){return 6}function Ph(Z){return ve[Hh()>>2]=Z,Z}function Py(Z){switch(Z){case 30:return 16384;case 85:var se=2147483648;return se/16384;case 132:case 133:case 12:case 137:case 138:case 15:case 235:case 16:case 17:case 18:case 19:case 20:case 149:case 13:case 10:case 236:case 153:case 9:case 21:case 22:case 159:case 154:case 14:case 77:case 78:case 139:case 82:case 68:case 67:case 164:case 11:case 29:case 47:case 48:case 95:case 52:case 51:case 46:return 200809;case 27:case 246:case 127:case 128:case 23:case 24:case 160:case 161:case 181:case 182:case 242:case 183:case 184:case 243:case 244:case 245:case 165:case 178:case 179:case 49:case 50:case 168:case 169:case 175:case 170:case 171:case 172:case 97:case 76:case 32:case 173:case 35:case 80:case 81:case 79:return-1;case 176:case 177:case 7:case 155:case 8:case 157:case 125:case 126:case 92:case 93:case 129:case 130:case 131:case 94:case 91:return 1;case 74:case 60:case 69:case 70:case 4:return 1024;case 31:case 42:case 72:return 32;case 87:case 26:case 33:return 2147483647;case 34:case 1:return 47839;case 38:case 36:return 99;case 43:case 37:return 2048;case 0:return 2097152;case 3:return 65536;case 28:return 32768;case 44:return 32767;case 75:return 16384;case 39:return 1e3;case 89:return 700;case 71:return 256;case 40:return 255;case 2:return 100;case 180:return 64;case 25:return 20;case 5:return 16;case 6:return 6;case 73:return 4;case 84:return typeof navigator=="object"&&navigator.hardwareConcurrency||1}return Ph(28),-1}var My={a:Vl,d:Ry,e:Rh,f:Fh,c:Oy,b:Oh,g:fn,h:Py},Ly=yi(),Mh=t.___wasm_call_ctors=function(){return(Mh=t.___wasm_call_ctors=t.asm.j).apply(null,arguments)},sc=t._init=function(){return(sc=t._init=t.asm.k).apply(null,arguments)},mm=t._register_tensor=function(){return(mm=t._register_tensor=t.asm.l).apply(null,arguments)},zy=t._dispose_data=function(){return(zy=t._dispose_data=t.asm.m).apply(null,arguments)},By=t._dispose=function(){return(By=t._dispose=t.asm.n).apply(null,arguments)},Vy=t._Abs=function(){return(Vy=t._Abs=t.asm.p).apply(null,arguments)},Ae=t._Add=function(){return(Ae=t._Add=t.asm.q).apply(null,arguments)},Gy=t._AddN=function(){return(Gy=t._AddN=t.asm.r).apply(null,arguments)},Wy=t._All=function(){return(Wy=t._All=t.asm.s).apply(null,arguments)},jy=t._Any=function(){return(jy=t._Any=t.asm.t).apply(null,arguments)},Uy=t._ArgMax=function(){return(Uy=t._ArgMax=t.asm.u).apply(null,arguments)},Hy=t._AvgPool=function(){return(Hy=t._AvgPool=t.asm.v).apply(null,arguments)},Gl=t._BatchMatMul=function(){return(Gl=t._BatchMatMul=t.asm.w).apply(null,arguments)},qy=t._Ceil=function(){return(qy=t._Ceil=t.asm.x).apply(null,arguments)},Ky=t._ClipByValue=function(){return(Ky=t._ClipByValue=t.asm.y).apply(null,arguments)},Xy=t._Conv2D=function(){return(Xy=t._Conv2D=t.asm.z).apply(null,arguments)},Yy=t._Conv2DBackpropInput=function(){return(Yy=t._Conv2DBackpropInput=t.asm.A).apply(null,arguments)},Zy=t._Cos=function(){return(Zy=t._Cos=t.asm.B).apply(null,arguments)},Jy=t._CropAndResize=function(){return(Jy=t._CropAndResize=t.asm.C).apply(null,arguments)},Qy=t._Cumsum=function(){return(Qy=t._Cumsum=t.asm.D).apply(null,arguments)},eb=t._DepthToSpace=function(){return(eb=t._DepthToSpace=t.asm.E).apply(null,arguments)},tb=t._DepthwiseConv2dNative=function(){return(tb=t._DepthwiseConv2dNative=t.asm.F).apply(null,arguments)},Ua=t._Equal=function(){return(Ua=t._Equal=t.asm.G).apply(null,arguments)},fm=t._Exp=function(){return(fm=t._Exp=t.asm.H).apply(null,arguments)},dm=t._FlipLeftRight=function(){return(dm=t._FlipLeftRight=t.asm.I).apply(null,arguments)},rb=t._Floor=function(){return(rb=t._Floor=t.asm.J).apply(null,arguments)},nb=t._FloorDiv=function(){return(nb=t._FloorDiv=t.asm.K).apply(null,arguments)},ob=t._FusedBatchNorm=function(){return(ob=t._FusedBatchNorm=t.asm.L).apply(null,arguments)},sb=t._FusedConv2D=function(){return(sb=t._FusedConv2D=t.asm.M).apply(null,arguments)},ib=t._FusedDepthwiseConv2D=function(){return(ib=t._FusedDepthwiseConv2D=t.asm.N).apply(null,arguments)},Je=t._Gather=function(){return(Je=t._Gather=t.asm.O).apply(null,arguments)},ab=t._GatherNd=function(){return(ab=t._GatherNd=t.asm.P).apply(null,arguments)},lb=t._Greater=function(){return(lb=t._Greater=t.asm.Q).apply(null,arguments)},ub=t._GreaterEqual=function(){return(ub=t._GreaterEqual=t.asm.R).apply(null,arguments)},cb=t._LeakyRelu=function(){return(cb=t._LeakyRelu=t.asm.S).apply(null,arguments)},pb=t._Less=function(){return(pb=t._Less=t.asm.T).apply(null,arguments)},mb=t._LessEqual=function(){return(mb=t._LessEqual=t.asm.U).apply(null,arguments)},hm=t._Log=function(){return(hm=t._Log=t.asm.V).apply(null,arguments)},Lh=t._LogicalAnd=function(){return(Lh=t._LogicalAnd=t.asm.W).apply(null,arguments)},zh=t._Max=function(){return(zh=t._Max=t.asm.X).apply(null,arguments)},fb=t._MaxPool=function(){return(fb=t._MaxPool=t.asm.Y).apply(null,arguments)},db=t._Maximum=function(){return(db=t._Maximum=t.asm.Z).apply(null,arguments)},hb=t._Mean=function(){return(hb=t._Mean=t.asm._).apply(null,arguments)},gb=t._Min=function(){return(gb=t._Min=t.asm.$).apply(null,arguments)},xb=t._Minimum=function(){return(xb=t._Minimum=t.asm.aa).apply(null,arguments)},yb=t._MirrorPad=function(){return(yb=t._MirrorPad=t.asm.ba).apply(null,arguments)},bb=t._Multiply=function(){return(bb=t._Multiply=t.asm.ca).apply(null,arguments)},ft=t._Neg=function(){return(ft=t._Neg=t.asm.da).apply(null,arguments)},wb=t._NonMaxSuppressionV3=function(){return(wb=t._NonMaxSuppressionV3=t.asm.ea).apply(null,arguments)},_b=t._NonMaxSuppressionV4=function(){return(_b=t._NonMaxSuppressionV4=t.asm.fa).apply(null,arguments)},kb=t._NonMaxSuppressionV5=function(){return(kb=t._NonMaxSuppressionV5=t.asm.ga).apply(null,arguments)},ic=t._NotEqual=function(){return(ic=t._NotEqual=t.asm.ha).apply(null,arguments)},Bh=t._OneHot=function(){return(Bh=t._OneHot=t.asm.ia).apply(null,arguments)},Vh=t._PadV2=function(){return(Vh=t._PadV2=t.asm.ja).apply(null,arguments)},Gh=t._Pow=function(){return(Gh=t._Pow=t.asm.ka).apply(null,arguments)},vb=t._Prelu=function(){return(vb=t._Prelu=t.asm.la).apply(null,arguments)},Cb=t._Prod=function(){return(Cb=t._Prod=t.asm.ma).apply(null,arguments)},Wh=t._RealDiv=function(){return(Wh=t._RealDiv=t.asm.na).apply(null,arguments)},Ib=t._Relu=function(){return(Ib=t._Relu=t.asm.oa).apply(null,arguments)},jh=t._Relu6=function(){return(jh=t._Relu6=t.asm.pa).apply(null,arguments)},Ha=t._ResizeBilinear=function(){return(Ha=t._ResizeBilinear=t.asm.qa).apply(null,arguments)},Nb=t._Reverse=function(){return(Nb=t._Reverse=t.asm.ra).apply(null,arguments)},Sb=t._RotateWithOffset=function(){return(Sb=t._RotateWithOffset=t.asm.sa).apply(null,arguments)},OI=t._Round=function(){return(OI=t._Round=t.asm.ta).apply(null,arguments)},Uh=t._Rsqrt=function(){return(Uh=t._Rsqrt=t.asm.ua).apply(null,arguments)},Tb=t._ScatterNd=function(){return(Tb=t._ScatterNd=t.asm.va).apply(null,arguments)},Ab=t._SelectV2=function(){return(Ab=t._SelectV2=t.asm.wa).apply(null,arguments)},Eb=t._Sigmoid=function(){return(Eb=t._Sigmoid=t.asm.xa).apply(null,arguments)},Db=t._Sin=function(){return(Db=t._Sin=t.asm.ya).apply(null,arguments)},$b=t._Softmax=function(){return($b=t._Softmax=t.asm.za).apply(null,arguments)},Rb=t._Sqrt=function(){return(Rb=t._Sqrt=t.asm.Aa).apply(null,arguments)},Fb=t._Square=function(){return(Fb=t._Square=t.asm.Ba).apply(null,arguments)},Ob=t._SquaredDifference=function(){return(Ob=t._SquaredDifference=t.asm.Ca).apply(null,arguments)},Pb=t._Step=function(){return(Pb=t._Step=t.asm.Da).apply(null,arguments)},Mb=t._StridedSlice=function(){return(Mb=t._StridedSlice=t.asm.Ea).apply(null,arguments)},Lb=t._Sub=function(){return(Lb=t._Sub=t.asm.Fa).apply(null,arguments)},zb=t._Sum=function(){return(zb=t._Sum=t.asm.Ga).apply(null,arguments)},Bb=t._Tan=function(){return(Bb=t._Tan=t.asm.Ha).apply(null,arguments)},Vb=t._Tanh=function(){return(Vb=t._Tanh=t.asm.Ia).apply(null,arguments)},Gb=t._Tile=function(){return(Gb=t._Tile=t.asm.Ja).apply(null,arguments)},Wb=t._TopK=function(){return(Wb=t._TopK=t.asm.Ka).apply(null,arguments)},jb=t._Transform=function(){return(jb=t._Transform=t.asm.La).apply(null,arguments)},Ub=t._Transpose=function(){return(Ub=t._Transpose=t.asm.Ma).apply(null,arguments)},Hb=t.__FusedMatMul=function(){return(Hb=t.__FusedMatMul=t.asm.Na).apply(null,arguments)},qb=t._malloc=function(){return(qb=t._malloc=t.asm.Oa).apply(null,arguments)},Kb=t._free=function(){return(Kb=t._free=t.asm.Pa).apply(null,arguments)},Hh=t.___errno_location=function(){return(Hh=t.___errno_location=t.asm.Qa).apply(null,arguments)},qh=t.stackSave=function(){return(qh=t.stackSave=t.asm.Ra).apply(null,arguments)},Kh=t.stackRestore=function(){return(Kh=t.stackRestore=t.asm.Sa).apply(null,arguments)},gm=t.stackAlloc=function(){return(gm=t.stackAlloc=t.asm.Ta).apply(null,arguments)};t.cwrap=X;var ac;function Xb(Z){this.name="ExitStatus",this.message="Program terminated with exit("+Z+")",this.status=Z}Fn=function Z(){ac||xm(),ac||(Fn=Z)};function xm(Z){if(Z=Z||i,_r>0||(cn(),_r>0))return;function se(){ac||(ac=!0,t.calledRun=!0,!G&&(Mr(),Un(),n(t),t.onRuntimeInitialized&&t.onRuntimeInitialized(),ir()))}t.setStatus?(t.setStatus("Running..."),setTimeout(function(){setTimeout(function(){t.setStatus("")},1),se()},1)):se()}if(t.run=xm,t.preInit)for(typeof t.preInit=="function"&&(t.preInit=[t.preInit]);t.preInit.length>0;)t.preInit.pop()();return xm(),e.ready}}();typeof Ex=="object"&&typeof zC=="object"?zC.exports=LC:typeof define=="function"&&define.amd?define([],function(){return LC}):typeof Ex=="object"&&(Ex.WasmBackendModule=LC)});var MB=qt((PB,wI)=>{(function(r,e,t){function n(i){var l=this,u=a();l.next=function(){var c=2091639*l.s0+l.c*23283064365386963e-26;return l.s0=l.s1,l.s1=l.s2,l.s2=c-(l.c=c|0)},l.c=1,l.s0=u(" "),l.s1=u(" "),l.s2=u(" "),l.s0-=u(i),l.s0<0&&(l.s0+=1),l.s1-=u(i),l.s1<0&&(l.s1+=1),l.s2-=u(i),l.s2<0&&(l.s2+=1),u=null}function o(i,l){return l.c=i.c,l.s0=i.s0,l.s1=i.s1,l.s2=i.s2,l}function s(i,l){var u=new n(i),c=l&&l.state,p=u.next;return p.int32=function(){return u.next()*4294967296|0},p.double=function(){return p()+(p()*2097152|0)*11102230246251565e-32},p.quick=p,c&&(typeof c=="object"&&o(c,u),p.state=function(){return o(u,{})}),p}function a(){var i=4022871197,l=function(u){u=String(u);for(var c=0;c<u.length;c++){i+=u.charCodeAt(c);var p=.02519603282416938*i;i=p>>>0,p-=i,p*=i,i=p>>>0,p-=i,i+=p*4294967296}return(i>>>0)*23283064365386963e-26};return l}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.alea=s})(PB,typeof wI=="object"&&wI,typeof define=="function"&&define)});var zB=qt((LB,_I)=>{(function(r,e,t){function n(a){var i=this,l="";i.x=0,i.y=0,i.z=0,i.w=0,i.next=function(){var c=i.x^i.x<<11;return i.x=i.y,i.y=i.z,i.z=i.w,i.w^=i.w>>>19^c^c>>>8},a===(a|0)?i.x=a:l+=a;for(var u=0;u<l.length+64;u++)i.x^=l.charCodeAt(u)|0,i.next()}function o(a,i){return i.x=a.x,i.y=a.y,i.z=a.z,i.w=a.w,i}function s(a,i){var l=new n(a),u=i&&i.state,c=function(){return(l.next()>>>0)/4294967296};return c.double=function(){do var p=l.next()>>>11,m=(l.next()>>>0)/4294967296,f=(p+m)/(1<<21);while(f===0);return f},c.int32=l.next,c.quick=c,u&&(typeof u=="object"&&o(u,l),c.state=function(){return o(l,{})}),c}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.xor128=s})(LB,typeof _I=="object"&&_I,typeof define=="function"&&define)});var VB=qt((BB,kI)=>{(function(r,e,t){function n(a){var i=this,l="";i.next=function(){var c=i.x^i.x>>>2;return i.x=i.y,i.y=i.z,i.z=i.w,i.w=i.v,(i.d=i.d+362437|0)+(i.v=i.v^i.v<<4^(c^c<<1))|0},i.x=0,i.y=0,i.z=0,i.w=0,i.v=0,a===(a|0)?i.x=a:l+=a;for(var u=0;u<l.length+64;u++)i.x^=l.charCodeAt(u)|0,u==l.length&&(i.d=i.x<<10^i.x>>>4),i.next()}function o(a,i){return i.x=a.x,i.y=a.y,i.z=a.z,i.w=a.w,i.v=a.v,i.d=a.d,i}function s(a,i){var l=new n(a),u=i&&i.state,c=function(){return(l.next()>>>0)/4294967296};return c.double=function(){do var p=l.next()>>>11,m=(l.next()>>>0)/4294967296,f=(p+m)/(1<<21);while(f===0);return f},c.int32=l.next,c.quick=c,u&&(typeof u=="object"&&o(u,l),c.state=function(){return o(l,{})}),c}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.xorwow=s})(BB,typeof kI=="object"&&kI,typeof define=="function"&&define)});var WB=qt((GB,vI)=>{(function(r,e,t){function n(a){var i=this;i.next=function(){var u=i.x,c=i.i,p,m,f;return p=u[c],p^=p>>>7,m=p^p<<24,p=u[c+1&7],m^=p^p>>>10,p=u[c+3&7],m^=p^p>>>3,p=u[c+4&7],m^=p^p<<7,p=u[c+7&7],p=p^p<<13,m^=p^p<<9,u[c]=m,i.i=c+1&7,m};function l(u,c){var p,m,f=[];if(c===(c|0))m=f[0]=c;else for(c=""+c,p=0;p<c.length;++p)f[p&7]=f[p&7]<<15^c.charCodeAt(p)+f[p+1&7]<<13;for(;f.length<8;)f.push(0);for(p=0;p<8&&f[p]===0;++p);for(p==8?m=f[7]=-1:m=f[p],u.x=f,u.i=0,p=256;p>0;--p)u.next()}l(i,a)}function o(a,i){return i.x=a.x.slice(),i.i=a.i,i}function s(a,i){a==null&&(a=+new Date);var l=new n(a),u=i&&i.state,c=function(){return(l.next()>>>0)/4294967296};return c.double=function(){do var p=l.next()>>>11,m=(l.next()>>>0)/4294967296,f=(p+m)/(1<<21);while(f===0);return f},c.int32=l.next,c.quick=c,u&&(u.x&&o(u,l),c.state=function(){return o(l,{})}),c}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.xorshift7=s})(GB,typeof vI=="object"&&vI,typeof define=="function"&&define)});var UB=qt((jB,CI)=>{(function(r,e,t){function n(a){var i=this;i.next=function(){var u=i.w,c=i.X,p=i.i,m,f;return i.w=u=u+1640531527|0,f=c[p+34&127],m=c[p=p+1&127],f^=f<<13,m^=m<<17,f^=f>>>15,m^=m>>>12,f=c[p]=f^m,i.i=p,f+(u^u>>>16)|0};function l(u,c){var p,m,f,d,h,g=[],x=128;for(c===(c|0)?(m=c,c=null):(c=c+"\0",m=0,x=Math.max(x,c.length)),f=0,d=-32;d<x;++d)c&&(m^=c.charCodeAt((d+32)%c.length)),d===0&&(h=m),m^=m<<10,m^=m>>>15,m^=m<<4,m^=m>>>13,d>=0&&(h=h+1640531527|0,p=g[d&127]^=m+h,f=p==0?f+1:0);for(f>=128&&(g[(c&&c.length||0)&127]=-1),f=127,d=4*128;d>0;--d)m=g[f+34&127],p=g[f=f+1&127],m^=m<<13,p^=p<<17,m^=m>>>15,p^=p>>>12,g[f]=m^p;u.w=h,u.X=g,u.i=f}l(i,a)}function o(a,i){return i.i=a.i,i.w=a.w,i.X=a.X.slice(),i}function s(a,i){a==null&&(a=+new Date);var l=new n(a),u=i&&i.state,c=function(){return(l.next()>>>0)/4294967296};return c.double=function(){do var p=l.next()>>>11,m=(l.next()>>>0)/4294967296,f=(p+m)/(1<<21);while(f===0);return f},c.int32=l.next,c.quick=c,u&&(u.X&&o(u,l),c.state=function(){return o(l,{})}),c}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.xor4096=s})(jB,typeof CI=="object"&&CI,typeof define=="function"&&define)});var qB=qt((HB,II)=>{(function(r,e,t){function n(a){var i=this,l="";i.next=function(){var c=i.b,p=i.c,m=i.d,f=i.a;return c=c<<25^c>>>7^p,p=p-m|0,m=m<<24^m>>>8^f,f=f-c|0,i.b=c=c<<20^c>>>12^p,i.c=p=p-m|0,i.d=m<<16^p>>>16^f,i.a=f-c|0},i.a=0,i.b=0,i.c=2654435769|0,i.d=1367130551,a===Math.floor(a)?(i.a=a/4294967296|0,i.b=a|0):l+=a;for(var u=0;u<l.length+20;u++)i.b^=l.charCodeAt(u)|0,i.next()}function o(a,i){return i.a=a.a,i.b=a.b,i.c=a.c,i.d=a.d,i}function s(a,i){var l=new n(a),u=i&&i.state,c=function(){return(l.next()>>>0)/4294967296};return c.double=function(){do var p=l.next()>>>11,m=(l.next()>>>0)/4294967296,f=(p+m)/(1<<21);while(f===0);return f},c.int32=l.next,c.quick=c,u&&(typeof u=="object"&&o(u,l),c.state=function(){return o(l,{})}),c}e&&e.exports?e.exports=s:t&&t.amd?t(function(){return s}):this.tychei=s})(HB,typeof II=="object"&&II,typeof define=="function"&&define)});var XB=qt((KB,Cy)=>{(function(r,e,t){var n=256,o=6,s=52,a="random",i=t.pow(n,o),l=t.pow(2,s),u=l*2,c=n-1,p;function m(w,_,I){var E=[];_=_==!0?{entropy:!0}:_||{};var D=g(h(_.entropy?[w,b(e)]:w==null?x():w,3),E),$=new f(E),R=function(){for(var M=$.g(o),G=i,j=0;M<l;)M=(M+j)*n,G*=n,j=$.g(1);for(;M>=u;)M/=2,G/=2,j>>>=1;return(M+j)/G};return R.int32=function(){return $.g(4)|0},R.quick=function(){return $.g(4)/4294967296},R.double=R,g(b($.S),e),(_.pass||I||function(M,G,j,U){return U&&(U.S&&d(U,$),M.state=function(){return d($,{})}),j?(t[a]=M,G):M})(R,D,"global"in _?_.global:this==t,_.state)}function f(w){var _,I=w.length,E=this,D=0,$=E.i=E.j=0,R=E.S=[];for(I||(w=[I++]);D<n;)R[D]=D++;for(D=0;D<n;D++)R[D]=R[$=c&$+w[D%I]+(_=R[D])],R[$]=_;(E.g=function(M){for(var G,j=0,U=E.i,H=E.j,q=E.S;M--;)G=q[U=c&U+1],j=j*n+q[c&(q[U]=q[H=c&H+G])+(q[H]=G)];return E.i=U,E.j=H,j})(n)}function d(w,_){return _.i=w.i,_.j=w.j,_.S=w.S.slice(),_}function h(w,_){var I=[],E=typeof w,D;if(_&&E=="object")for(D in w)try{I.push(h(w[D],_-1))}catch($){}return I.length?I:E=="string"?w:w+"\0"}function g(w,_){for(var I=w+"",E,D=0;D<I.length;)_[c&D]=c&(E^=_[c&D]*19)+I.charCodeAt(D++);return b(_)}function x(){try{var w;return p&&(w=p.randomBytes)?w=w(n):(w=new Uint8Array(n),(r.crypto||r.msCrypto).getRandomValues(w)),b(w)}catch(E){var _=r.navigator,I=_&&_.plugins;return[+new Date,r,I,r.screen,b(e)]}}function b(w){return String.fromCharCode.apply(0,w)}if(g(t.random(),e),typeof Cy=="object"&&Cy.exports){Cy.exports=m;try{p=w_()}catch(w){}}else typeof define=="function"&&define.amd?define(function(){return m}):t["seed"+a]=m})(typeof self!="undefined"?self:KB,[],Math)});var NI=qt((bjt,YB)=>{var Lne=MB(),zne=zB(),Bne=VB(),Vne=WB(),Gne=UB(),Wne=qB(),nc=XB();nc.alea=Lne;nc.xor128=zne;nc.xorwow=Bne;nc.xorshift7=Vne;nc.xor4096=Gne;nc.tychei=Wne;YB.exports=nc});var CV=qt(()=>{});var km={};We(km,{bin:()=>nN,browser:()=>uN,default:()=>AG,dependencies:()=>lN,description:()=>XI,devDependencies:()=>iN,jsdelivr:()=>QI,license:()=>sN,main:()=>ZI,miniprogram:()=>rN,module:()=>JI,name:()=>qI,private:()=>YI,repository:()=>oN,scripts:()=>aN,types:()=>tN,unpkg:()=>eN,version:()=>KI});var qI="@tensorflow/tfjs",KI="3.6.0",XI="An open-source machine learning framework.",YI=!1,ZI="dist/tf.node.js",JI="dist/index.js",QI="dist/tf.min.js",eN="dist/tf.min.js",tN="dist/index.d.ts",rN="dist/miniprogram",nN={"tfjs-custom-module":"dist/tools/custom_module/cli.js"},oN={type:"git",url:"https://github.com/tensorflow/tfjs.git"},sN="Apache-2.0",iN={"@babel/core":"^7.9.0","@babel/polyfill":"^7.10.4","@babel/preset-env":"^7.9.5","@rollup/plugin-commonjs":"^11.0.2","@rollup/plugin-node-resolve":"^7.1.1","@rollup/plugin-typescript":"^3.0.0","@types/argparse":"^1.0.38","@types/jasmine":"2.8.7","@types/node":"~10.17.50","@types/shelljs":"^0.8.4","@types/yargs":"^15.0.7","clang-format":"~1.2.2",commander:"~2.14.1",jasmine:"3.1.0","jasmine-core":"~3.1.0",karma:"~6.3.2","karma-browserstack-launcher":"~1.6.0","karma-chrome-launcher":"~2.2.0","karma-firefox-launcher":"~1.1.0","karma-jasmine":"~1.1.1","karma-typescript":"~5.5.1","karma-typescript-es6-transform":"^5.5.1","npm-run-all":"~4.1.3",rimraf:"~2.6.2",rollup:"~2.3.2","rollup-plugin-babel":"^4.4.0","rollup-plugin-terser":"~7.0.2","rollup-plugin-visualizer":"~4.2.2",shelljs:"~0.8.1","ts-node":"~8.8.2",tslint:"~5.11.0","tslint-no-circular-imports":"~0.5.0",typescript:"3.5.3",yalc:"1.0.0-pre.50"},aN={build:"tsc && yarn build-cli && yarn bundle","build-ci":"tsc && yarn build-cli && yarn bundle-ci",bundle:"rollup -c","bundle-ci":"rollup -c --ci","build-core":"cd ../tfjs-core && yarn && yarn build","build-core-ci":"cd ../tfjs-core && yarn && yarn build-ci","build-layers":"cd ../tfjs-layers && yarn && yarn build","build-layers-ci":"cd ../tfjs-layers && yarn && yarn build-ci","build-converter":"cd ../tfjs-converter && yarn && yarn build","build-converter-ci":"cd ../tfjs-converter && yarn && yarn build-ci","build-data":"cd ../tfjs-data && yarn && yarn build","build-data-ci":"cd ../tfjs-data && yarn && yarn build-ci","build-backend-cpu":"cd ../tfjs-backend-cpu && yarn && yarn build","build-backend-cpu-ci":"cd ../tfjs-backend-cpu && yarn && yarn build-ci","build-backend-webgl":"cd ../tfjs-backend-webgl && yarn && yarn build","build-backend-webgl-ci":"cd ../tfjs-backend-webgl && yarn && yarn build-ci","build-deps":"yarn build-core && yarn build-layers && yarn build-converter && yarn build-data && yarn build-backend-cpu && yarn build-backend-webgl","build-deps-ci":"yarn build-core-ci && yarn build-layers-ci && yarn build-converter-ci && yarn build-data-ci && yarn build-backend-cpu-ci && yarn build-backend-webgl-ci","build-cli":"tsc --project ./tools/custom_module/tsconfig.json && chmod +x ./dist/tools/custom_module/cli.js","run-custom-build":"ts-node -s ./tools/custom_module/cli.ts","build-npm":"./scripts/build-npm.sh","link-local":"yalc link","publish-local":"yarn build-npm && yalc push","publish-npm":"npm publish",lint:"tslint -p . -t verbose",coverage:"KARMA_COVERAGE=1 karma start --browsers='Chrome' --singleRun",test:"yarn && yarn build-deps && yarn build && karma start","test-dev":"karma start","test-tools":"ts-node --project ./tools/custom_module/tsconfig.json run_tools_tests.ts","test-ci":"./scripts/test-ci.sh"},lN={"@tensorflow/tfjs-backend-cpu":"3.6.0","@tensorflow/tfjs-backend-webgl":"3.6.0","@tensorflow/tfjs-converter":"3.6.0","@tensorflow/tfjs-core":"3.6.0","@tensorflow/tfjs-data":"3.6.0","@tensorflow/tfjs-layers":"3.6.0",argparse:"^1.0.10",chalk:"^4.1.0","core-js":"3","regenerator-runtime":"^0.13.5",yargs:"^16.0.3"},uN={"node-fetch":!1,util:!1,crypto:!1},AG={name:qI,version:KI,description:XI,private:YI,main:ZI,module:JI,jsdelivr:QI,unpkg:eN,types:tN,miniprogram:rN,bin:nN,repository:oN,license:sN,devDependencies:iN,scripts:aN,dependencies:lN,browser:uN};var vm={};We(vm,{browser:()=>SN,default:()=>EG,dependencies:()=>NN,description:()=>mN,devDependencies:()=>CN,engines:()=>_N,jsdelivr:()=>hN,"jsnext:main":()=>yN,license:()=>vN,main:()=>dN,miniprogram:()=>wN,module:()=>bN,name:()=>cN,private:()=>fN,repository:()=>kN,scripts:()=>IN,sideEffects:()=>TN,types:()=>xN,unpkg:()=>gN,version:()=>pN});var cN="@tensorflow/tfjs-core",pN="3.6.0",mN="Hardware-accelerated JavaScript library for machine intelligence",fN=!1,dN="dist/tf-core.node.js",hN="dist/tf-core.min.js",gN="dist/tf-core.min.js",xN="dist/index.d.ts",yN="dist/index.js",bN="dist/index.js",wN="dist/miniprogram",_N={yarn:">= 1.3.2"},kN={type:"git",url:"https://github.com/tensorflow/tfjs-core.git"},vN="Apache-2.0",CN={"@bazel/bazelisk":"^1.3.0","@bazel/typescript":"^0.27.8","@rollup/plugin-commonjs":"^11.0.2","@rollup/plugin-node-resolve":"^7.1.1","@rollup/plugin-typescript":"^3.0.0","@tensorflow/tfjs-backend-cpu":"link:../tfjs-backend-cpu","@types/jasmine":"~3.0.0","@types/node":"~9.6.0","@types/node-fetch":"~2.1.2","clang-format":"~1.2.4",jasmine:"~3.1.0","jasmine-core":"~3.1.0",karma:"6.3.2","karma-browserstack-launcher":"~1.6.0","karma-chrome-launcher":"~3.1.0","karma-jasmine":"~4.0.1","karma-typescript":"~5.5.1","npm-run-all":"~4.1.3",rimraf:"~2.6.2",rollup:"~2.3.2","rollup-plugin-terser":"~5.3.0","rollup-plugin-visualizer":"~3.3.2",shelljs:"~0.8.3","ts-node":"~8.8.2",tslint:"~5.11.0","tslint-no-circular-imports":"~0.5.0",typescript:"3.5.3",yalc:"~1.0.0-pre.21",yargs:"~13.2.2"},IN={"build-ci":"./scripts/enumerate-tests.js --ci && tsc && yarn bundle-ci && yarn build-test-snippets",build:"node ./scripts/enumerate-tests.js && tsc && yarn bundle",bundle:"rollup -c","bundle-ci":"rollup -c --ci","build-npm":"./scripts/build-npm.sh","build-deps":"yarn build && yarn build-cpu-backend","build-cpu-backend":"cd ../tfjs-backend-cpu && yarn && yarn build","build-cpu-backend-ci":"cd ../tfjs-backend-cpu && yarn && yarn build-ci","build:bazel":"bazelisk build //...","build-test-snippets":"yarn tsc --project ./scripts/test_snippets/tsconfig.json","format-all":"clang-format -i -style=Google --glob=src/**/*.ts","link-local":"yalc link","publish-local":"rimraf dist/ && yarn build && rollup -c && yalc push","publish-npm":"npm publish",lint:"tslint -p . -t verbose",coverage:"KARMA_COVERAGE=1 karma start --browsers='Chrome' --singleRun",test:"yarn && yarn build-deps && karma start","test-dev":"karma start","test-ci":"./scripts/test-ci.sh","test-webworker":"karma start --worker","run-browserstack":"karma start --browserstack","test-bundle-size":"./scripts/test-bundle-size.js","test-node":"rimraf dist/ && yarn build-deps && yarn build && ts-node --transpile-only --skip-ignore -P tsconfig.test.json dist/test_node.js","test-node-dev":"tsc && ts-node --transpile-only --skip-ignore -P tsconfig.test.json dist/test_node.js","test-node-ci":"ts-node --transpile-only -P tsconfig.test.json dist/test_node.js","test-async-backends":"rimraf dist/ && yarn build && ts-node --transpile-only -P tsconfig.test.json dist/test_async_backends.js","test-async-backends-ci":"ts-node --transpile-only -P tsconfig.test.json dist/test_async_backends.js","test-snippets":"yarn build && yarn build-cpu-backend && ts-node -P tsconfig.test.json ./scripts/test_snippets/test_snippets.ts","test-snippets-ci":"ts-node -P tsconfig.test.json ./scripts/test_snippets/test_snippets.ts"},NN={"@types/offscreencanvas":"~2019.3.0","@types/seedrandom":"2.4.27","@types/webgl-ext":"0.0.30","node-fetch":"~2.6.1",seedrandom:"2.4.3"},SN={"node-fetch":!1,util:!1,crypto:!1,worker_threads:!1},TN=["./dist/index.js","./dist/engine.js","./dist/tensor.js","./dist/base_side_effects.js","./dist/flags.js","./dist/platforms/*.js","./dist/register_all_gradients.js","./dist/public/chained_ops/*.js","./dist/io/*.js"],EG={name:cN,version:pN,description:mN,private:fN,main:dN,jsdelivr:hN,unpkg:gN,types:xN,"jsnext:main":yN,module:bN,miniprogram:wN,engines:_N,repository:kN,license:vN,devDependencies:CN,scripts:IN,dependencies:NN,browser:SN,sideEffects:TN};var Cm={};We(Cm,{browser:()=>UN,default:()=>DG,dependencies:()=>jN,description:()=>DN,devDependencies:()=>VN,jsdelivr:()=>FN,"jsnext:main":()=>MN,license:()=>BN,main:()=>RN,miniprogram:()=>zN,module:()=>LN,name:()=>AN,peerDependencies:()=>WN,private:()=>$N,scripts:()=>GN,types:()=>PN,unpkg:()=>ON,version:()=>EN});var AN="@tensorflow/tfjs-data",EN="3.6.0",DN="TensorFlow Data API in JavaScript",$N=!1,RN="dist/tf-data.node.js",FN="dist/tf-data.min.js",ON="dist/tf-data.min.js",PN="dist/index.d.ts",MN="dist/index.js",LN="dist/index.js",zN="dist/miniprogram",BN="Apache-2.0",VN={"@rollup/plugin-commonjs":"^11.0.2","@rollup/plugin-node-resolve":"^7.1.1","@rollup/plugin-typescript":"^3.0.0","@tensorflow/tfjs-backend-cpu":"3.6.0","@tensorflow/tfjs-core":"3.6.0","@tensorflow/tfjs-layers":"3.6.0","@types/jasmine":"~2.5.53","@types/seedrandom":"^2.4.27","@types/utf8":"~2.1.6","clang-format":"~1.2.2","http-server":"~0.12.3",jasmine:"3.1.0","jasmine-core":"~3.1.0",karma:"~6.3.1","karma-chrome-launcher":"~2.2.0","karma-firefox-launcher":"~1.1.0","karma-jasmine":"~1.1.1","karma-typescript":"~5.5.1","karma-typescript-es6-transform":"^5.0.2",nyc:"^15.1.0",rimraf:"~2.6.2",rollup:"~2.3.2","rollup-plugin-terser":"~7.0.2","rollup-plugin-visualizer":"~3.3.2","ts-node":"~7.0.0",tslint:"~6.1.3","tslint-no-circular-imports":"^0.7.0",typescript:"3.5.3",yalc:"^1.0.0-pre.50"},GN={build:"tsc && yarn bundle","build-ci":"tsc && yarn bundle-ci",bundle:"rollup -c","bundle-ci":"rollup -c --ci","build-core":"cd ../tfjs-core && yarn && yarn build","build-core-ci":"cd ../tfjs-core && yarn && yarn build-ci","build-layers":"cd ../tfjs-layers && yarn && yarn build","build-backend-cpu":"cd ../tfjs-backend-cpu && yarn && yarn build","build-backend-cpu-ci":"cd ../tfjs-backend-cpu && yarn && yarn build-ci","build-layers-ci":"cd ../tfjs-layers && yarn && yarn build-ci","build-deps":"yarn build-core && yarn build-layers && yarn build-backend-cpu","build-deps-ci":"yarn build-core-ci && yarn build-layers-ci && yarn build-backend-cpu-ci","build-npm":"./scripts/build-npm.sh","link-local":"yalc link","publish-local":"rimraf dist/ && yarn build-npm && yalc push","publish-npm":"npm publish",test:"yarn && yarn build-deps && yarn build && ts-node --transpile-only --project tsconfig.test.json src/test_node.ts","test-dev":"tsc && ts-node --transpile-only --project tsconfig.test.json src/test_node.ts","test-browsers":"karma start --browsers='Chrome,Firefox'","test-ci":"ts-node --transpile-only --skip-ignore -P tsconfig.test.json src/test_node.ts","test-snippets":"yarn && yarn build-deps && yarn build && ts-node --skip-ignore --project tsconfig.test.json ./scripts/test_snippets.ts","test-snippets-ci":"ts-node --skip-ignore --project tsconfig.test.json ./scripts/test_snippets.ts",coverage:"yarn nyc yarn ts-node --transpile-only -P tsconfig.test.json src/test_node.ts",lint:"tslint -p . -t verbose"},WN={"@tensorflow/tfjs-core":"3.6.0",seedrandom:"~2.4.3"},jN={"@types/node-fetch":"^2.1.2","node-fetch":"~2.6.1"},UN={fs:!1,"node-fetch":!1,string_decoder:!1,crypto:!1},DG={name:AN,version:EN,description:DN,private:$N,main:RN,jsdelivr:FN,unpkg:ON,types:PN,"jsnext:main":MN,module:LN,miniprogram:zN,license:BN,devDependencies:VN,scripts:GN,peerDependencies:WN,dependencies:jN,browser:UN};var Im={};We(Im,{default:()=>$G,description:()=>KN,devDependencies:()=>oS,jsdelivr:()=>tS,"jsnext:main":()=>QN,license:()=>XN,main:()=>ZN,miniprogram:()=>nS,module:()=>eS,name:()=>HN,peerDependencies:()=>iS,private:()=>YN,scripts:()=>sS,types:()=>JN,unpkg:()=>rS,version:()=>qN});var HN="@tensorflow/tfjs-layers",qN="3.6.0",KN="TensorFlow layers API in JavaScript",XN="Apache-2.0 AND MIT",YN=!1,ZN="dist/tf-layers.node.js",JN="dist/index.d.ts",QN="dist/index.js",eS="dist/index.js",tS="dist/tf-layers.min.js",rS="dist/tf-layers.min.js",nS="dist/miniprogram",oS={"@babel/polyfill":"^7.8.7","@rollup/plugin-commonjs":"^11.0.2","@rollup/plugin-node-resolve":"^7.1.1","@rollup/plugin-typescript":"^3.0.0","@tensorflow/tfjs-backend-cpu":"3.6.0","@tensorflow/tfjs-backend-webgl":"3.6.0","@tensorflow/tfjs-core":"3.6.0","@types/jasmine":"~2.5.53","clang-format":"~1.2.2","http-server":"~0.12.3",jasmine:"~3.1.0","jasmine-core":"~3.1.0",karma:"~6.3.1","karma-browserstack-launcher":"~1.6.0","karma-chrome-launcher":"~2.2.0","karma-firefox-launcher":"~1.1.0","karma-jasmine":"~1.1.1","karma-typescript":"~5.5.1","karma-typescript-es6-transform":"^5.0.2",rimraf:"~2.6.2",rollup:"~2.3.2","rollup-plugin-terser":"~7.0.2","rollup-plugin-visualizer":"~3.3.2","ts-node":"~8.8.2",tslint:"~6.1.3","tslint-no-circular-imports":"^0.7.0",typescript:"3.5.3",yalc:"~1.0.0-pre.50"},sS={prep:"yarn install && yarn build-ci",build:"tsc && yarn bundle","build-ci":"tsc && yarn bundle-ci",bundle:"rollup -c","bundle-ci":"rollup -c --ci","build-core":"cd ../tfjs-core && yarn && yarn build","build-backend-cpu":"cd ../tfjs-backend-cpu && yarn && yarn build","build-backend-cpu-ci":"cd ../tfjs-backend-cpu && yarn && yarn build-ci","build-backend-webgl":"cd ../tfjs-backend-webgl && yarn && yarn build","build-backend-webgl-ci":"cd ../tfjs-backend-webgl && yarn && yarn build-ci","build-core-ci":"cd ../tfjs-core && yarn && yarn build-ci","build-deps":"yarn build-core && yarn build-backend-cpu && yarn build-backend-webgl","build-deps-ci":"yarn build-core-ci && yarn build-backend-cpu-ci && yarn build-backend-webgl-ci","build-npm":"./scripts/build-npm.sh",format:"./tools/clang_format_ts.sh","link-local":"yalc link","publish-local":"yarn build-npm && yalc push","publish-npm":"npm publish",coverage:"KARMA_COVERAGE=1 karma start --browsers='Chrome' --singleRun",test:"yarn && yarn build-deps && karma start","test-dev":"karma start","test-ci":"./scripts/test-ci.sh","test-snippets":"yarn && yarn build-deps && yarn build && ts-node --skip-ignore -s ./scripts/test_snippets.ts","test-snippets-ci":"ts-node --skip-ignore -s ./scripts/test_snippets.ts","run-browserstack":"karma start --browsers='bs_chrome_mac' --singleRun --reporters='dots,karma-typescript'",lint:"tslint -p . -t verbose"},iS={"@tensorflow/tfjs-core":"3.6.0"},$G={name:HN,version:qN,description:KN,license:XN,private:YN,main:ZN,types:JN,"jsnext:main":QN,module:eS,jsdelivr:tS,unpkg:rS,miniprogram:nS,devDependencies:oS,scripts:sS,peerDependencies:iS};var Nm={};We(Nm,{default:()=>RG,description:()=>uS,devDependencies:()=>wS,jsdelivr:()=>hS,"jsnext:main":()=>pS,license:()=>yS,main:()=>cS,miniprogram:()=>gS,module:()=>mS,name:()=>aS,peerDependencies:()=>bS,repository:()=>xS,scripts:()=>_S,types:()=>fS,unpkg:()=>dS,version:()=>lS});var aS="@tensorflow/tfjs-converter",lS="3.6.0",uS="Tensorflow model converter for javascript",cS="dist/tf-converter.node.js",pS="dist/index.js",mS="dist/index.js",fS="dist/index.d.ts",dS="dist/tf-converter.min.js",hS="dist/tf-converter.min.js",gS="dist/miniprogram",xS={type:"git",url:"https://github.com/tensorflow/tfjs-converter.git"},yS="Apache-2.0",bS={"@tensorflow/tfjs-core":"3.6.0"},wS={"@rollup/plugin-commonjs":"^11.0.2","@rollup/plugin-node-resolve":"^7.1.1","@rollup/plugin-replace":"^2.3.3","@rollup/plugin-typescript":"^3.0.0","@tensorflow/tfjs-backend-cpu":"3.6.0","@tensorflow/tfjs-core":"3.6.0","@types/argparse":"^1.0.38","@types/deep-equal":"^1.0.1","@types/jasmine":"~2.8.6","@types/long":"~3.0.32","@types/node-fetch":"1.6.9",ajv:"~6.3.0",argparse:"^1.0.10","babel-core":"~6.26.3","babel-plugin-external-helpers":"~6.22.0","babel-preset-env":"~1.7.0","clang-format":"~1.2.2",copyfiles:"~1.2.0","deep-equal":"^1.0.1","jasmine-core":"~3.5.0","node-fetch":"~2.6.1",opn:"~5.1.0",protobufjs:"~6.8.6",rimraf:"~2.6.2",rollup:"~2.3.2","rollup-plugin-terser":"~7.0.2","rollup-plugin-visualizer":"~3.3.2","ts-morph":"^7.1.3","ts-node":"~8.8.2",tslint:"~6.1.3","tslint-no-circular-imports":"~0.7.0",typescript:"3.5.3",yalc:"~1.0.0-pre.50"},_S={build:"yarn gen-json --test && yarn gen-kernel2ops && tsc && yarn bundle","build-ci":"yarn gen-json --test && yarn gen-kernel2ops && tsc && yarn bundle-ci",bundle:"rollup -c","bundle-ci":"rollup -c --ci","build-core":"cd ../tfjs-core && yarn && yarn build","build-backend-cpu":"cd ../tfjs-backend-cpu && yarn && yarn build","build-backend-cpu-ci":"cd ../tfjs-backend-cpu && yarn && yarn build-ci","build-core-ci":"cd ../tfjs-core && yarn && yarn build-ci","build-deps":"yarn build-core && yarn build-backend-cpu","build-deps-ci":"yarn build-core-ci && yarn build-backend-cpu","build-npm":"./scripts/build-npm.sh","link-local":"yalc link","publish-local":"yarn build-npm && yalc push","publish-npm":"npm publish",test:"yarn && yarn build-deps && yarn build && yarn gen-json --test && yarn gen-kernel2ops && ts-node --transpile-only -P tsconfig.test.json src/run_tests.ts","test-ci":"ts-node --transpile-only --skip-ignore -P tsconfig.test.json src/run_tests.ts","test-dev":"tsc && ts-node --transpile-only -P tsconfig.test.json src/run_tests.ts","test-snippets":"yarn && yarn build-deps && yarn build && ts-node --skip-ignore -s ./scripts/test_snippets.ts","test-snippets-ci":"ts-node --skip-ignore -s ./scripts/test_snippets.ts",lint:"tslint -p . -t verbose","make-version":"sh -c ./scripts/make-version","gen-doc":"ts-node -s ./scripts/gen_doc.ts","gen-json":"ts-node -s ./scripts/gen_json.ts","model-summary":"ts-node -s ./tools/model_summary.ts",pb2json:"ts-node -s ./tools/pb2json_converter.ts","build-pip-package":"yarn gen-json --test && cd python && ./build-pip-package.sh --test /tmp/tfjs-pips","run-python-tests":"yarn gen-json --test && cd python && ./run-python-tests.sh","gen-kernel2ops":"ts-node -s scripts/kernels_to_ops.ts --out metadata/kernel2op.json"},RG={name:aS,version:lS,description:uS,main:cS,"jsnext:main":pS,module:mS,types:fS,unpkg:dS,jsdelivr:hS,miniprogram:gS,repository:xS,license:yS,peerDependencies:bS,devDependencies:wS,scripts:_S};var FG=1e-7,OG=1e-4,Xa=class{constructor(e,t){this.backend=e,this.dataMover=t,this.data=new WeakMap,this.dataIdsCount=0}get(e){return this.data.has(e)||this.dataMover.moveData(this.backend,e),this.data.get(e)}set(e,t){this.dataIdsCount++,this.data.set(e,t)}has(e){return this.data.has(e)}delete(e){return this.dataIdsCount--,this.data.delete(e)}numDataIds(){return this.dataIdsCount}},Ms=class{refCount(e){return Hn("refCount")}incRef(e){return Hn("incRef")}timerAvailable(){return!0}time(e){return Hn("time")}read(e){return Hn("read")}readSync(e){return Hn("readSync")}numDataIds(){return Hn("numDataIds")}disposeData(e,t){return Hn("disposeData")}write(e,t,n){return Hn("write")}move(e,t,n,o,s){return Hn("move")}memory(){return Hn("memory")}floatPrecision(){return Hn("floatPrecision")}epsilon(){return this.floatPrecision()===32?FG:OG}dispose(){return Hn("dispose")}};function Hn(r){throw new Error(`'${r}' not yet implemented or not found in the registry. This kernel may not be supported by the tfjs backend you have chosen`)}function kS(r){let e=r.length,t=0,n=0;for(;e>0;)n=Math.random()*e|0,e--,t=r[e],r[e]=r[n],r[n]=t}function PG(r,e){if(r.length!==e.length)throw new Error(`Array sizes must match to be shuffled together First array length was ${r.length}Second array length was ${e.length}`);let t=r.length,n,o,s=0;for(;t>0;)s=Math.random()*t|0,t--,n=r[t],o=e[t],r[t]=r[s],e[t]=e[s],r[s]=n,e[s]=o}function gc(r,e,t){return Math.max(r,Math.min(e,t))}function MG(r){return r%2==0?r:r+1}function LG(r){let e=0;for(let t=0;t<r.length;t++)e+=r[t];return e}function zG(r,e){let t=Math.random();return e*t+(1-t)*r}function BG(r,e){let t=0;for(let n=0;n<r.length;n++){let o=Number(r[n])-Number(e[n]);t+=o*o}return t}function T(r,e){if(!r)throw new Error(typeof e=="string"?e:e())}function Dt(r,e,t=""){T(Xr(r,e),()=>t+` Shapes ${r} and ${e} must match`)}function On(r){T(r!=null,()=>"The input to the tensor constructor must be a non-null value.")}function bo(r,e=[],t=!1){if(e==null&&(e=[]),Array.isArray(r)||ar(r)&&!t)for(let n=0;n<r.length;++n)bo(r[n],e,t);else e.push(r);return e}function nt(r){if(r.length===0)return 1;let e=r[0];for(let t=1;t<r.length;t++)e*=r[t];return e}function VG(r){return r.length===0}function Xr(r,e){if(r===e)return!0;if(r==null||e==null||r.length!==e.length)return!1;for(let t=0;t<r.length;t++)if(r[t]!==e[t])return!1;return!0}function ot(r){return r%1==0}function GG(r){if(Math.tanh!=null)return Math.tanh(r);if(r===Infinity)return 1;if(r===-Infinity)return-1;{let e=Math.exp(2*r);return(e-1)/(e+1)}}function WG(r){let e=Math.ceil(Math.sqrt(r));return[e,Math.ceil(r/e)]}function jG(r){let e=new Uint32Array(r);for(let t=0;t<r;++t)e[t]=t;return kS(e),e}function jl(r,e){return e<=r.length?r:r+" ".repeat(e-r.length)}function UG(r,e=n=>0,t){return new Promise((n,o)=>{let s=0,a=()=>{if(r()){n();return}s++;let i=e(s);if(t!=null&&s>=t){o();return}setTimeout(a,i)};a()})}function HG(r,e){let t=1,n=-1;for(let s=0;s<r.length;++s)if(r[s]>=0)t*=r[s];else if(r[s]===-1){if(n!==-1)throw Error(`Shapes can only have 1 implicit size. Found -1 at dim ${n} and dim ${s}`);n=s}else if(r[s]<0)throw Error(`Shapes can not be < 0. Found ${r[s]} at dim ${s}`);if(n===-1){if(e>0&&e!==t)throw Error(`Size(${e}) must match the product of shape ${r}`);return r}if(t===0)throw Error(`Cannot infer the missing size in [${r}] when there are 0 elements`);if(e%t!=0)throw Error(`The implicit shape can't be a fractional number. Got ${e} / ${t}`);let o=r.slice();return o[n]=e/t,o}function nr(r,e){let t=e.length;return r=r==null?e.map((n,o)=>o):[].concat(r),T(r.every(n=>n>=-t&&n<t),()=>`All values in axis param must be in range [-${t}, ${t}) but got axis ${r}`),T(r.every(n=>ot(n)),()=>`All values in axis param must be integers but got axis ${r}`),r.map(n=>n<0?t+n:n)}function ew(r,e){let t=[],n=[],o=e!=null&&Array.isArray(e)&&e.length===0,s=e==null||o?null:nr(e,r).sort(),a=0;for(let i=0;i<r.length;++i){if(s!=null){if(s[a]===i&&r[i]!==1)throw new Error(`Can't squeeze axis ${i} since its dim '${r[i]}' is not 1`);(s[a]==null||s[a]>i)&&r[i]===1&&(t.push(r[i]),n.push(i)),s[a]<=i&&a++}r[i]!==1&&(t.push(r[i]),n.push(i))}return{newShape:t,keptDims:n}}function tw(r,e){let t=null;if(r==null||r==="float32")t=new Float32Array(e);else if(r==="int32")t=new Int32Array(e);else if(r==="bool")t=new Uint8Array(e);else throw new Error(`Unknown data type ${r}`);return t}function rw(r,e){let t=null;if(r==null||r==="float32")t=new Float32Array(e);else if(r==="int32")t=new Int32Array(e);else if(r==="bool")t=new Uint8Array(e);else if(r==="string")t=new Array(e);else throw new Error(`Unknown data type ${r}`);return t}function nw(r,e){for(let t=0;t<r.length;t++){let n=r[t];if(isNaN(n)||!isFinite(n))throw Error(`A tensor of type ${e} being uploaded contains ${n}.`)}}function ow(r){return r==="bool"||r==="complex64"||r==="float32"||r==="int32"||r==="string"}function qG(r,e){return!(e==="complex64"||e==="float32"&&r!=="complex64"||e==="int32"&&r!=="float32"&&r!=="complex64"||e==="bool"&&r==="bool")}function ar(r){return r instanceof Float32Array||r instanceof Int32Array||r instanceof Uint8Array}function Qh(r){if(r==="float32"||r==="int32")return 4;if(r==="complex64")return 8;if(r==="bool")return 1;throw new Error(`Unknown dtype ${r}`)}function sw(r){if(r==null)return 0;let e=0;return r.forEach(t=>e+=t.length),e}function wo(r){return typeof r=="string"||r instanceof String}function vS(r){return typeof r=="boolean"}function CS(r){return typeof r=="number"}function xc(r){return Array.isArray(r)?xc(r[0]):r instanceof Float32Array?"float32":r instanceof Int32Array||r instanceof Uint8Array?"int32":CS(r)?"float32":wo(r)?"string":vS(r)?"bool":"float32"}function Ls(r){return!!(r&&r.constructor&&r.call&&r.apply)}function yc(r,e){for(let t=e;t<r;++t)if(r%t==0)return t;return r}function zs(r){let e=r.length;if(e<2)return[];let t=new Array(e-1);t[e-2]=r[e-1];for(let n=e-3;n>=0;--n)t[n]=t[n+1]*r[n+1];return t}function IS(r,e,t,n=!1){let o=new Array;if(e.length===1){let s=e[0]*(n?2:1);for(let a=0;a<s;a++)o[a]=t[r+a]}else{let s=e[0],a=e.slice(1),i=a.reduce((l,u)=>l*u)*(n?2:1);for(let l=0;l<s;l++)o[l]=IS(r+l*i,a,t,n)}return o}function Ul(r,e,t=!1){if(r.length===0)return e[0];let n=r.reduce((o,s)=>o*s)*(t?2:1);if(n===0)return[];if(n!==e.length)throw new Error(`[${r}] does not match the input size ${e.length}${t?" for a complex tensor":""}.`);return IS(0,r,e,t)}function Sm(r,e){let t=bc(r,e);for(let n=0;n<t.length;n++)t[n]=1;return t}function bc(r,e){if(e==null||e==="float32"||e==="complex64")return new Float32Array(r);if(e==="int32")return new Int32Array(r);if(e==="bool")return new Uint8Array(r);throw new Error(`Unknown data type ${e}`)}function KG(r,e){let t=r.reduce((n,o)=>n*o,1);if(e==null||e==="float32")return Ul(r,new Float32Array(t));if(e==="int32")return Ul(r,new Int32Array(t));if(e==="bool")return Ul(r,new Uint8Array(t));throw new Error(`Unknown data type ${e}`)}function Tm(r){r.forEach(e=>{T(Number.isInteger(e)&&e>=0,()=>`Tensor must have a shape comprised of positive integers but got shape [${r}].`)})}function XG(r,e,t){if(e===0)return 0;if(e===1)return r[0];let n=r[r.length-1];for(let o=0;o<r.length-1;++o)n+=t[o]*r[o];return n}function YG(r,e,t){if(e===0)return[];if(e===1)return[r];let n=new Array(e);for(let o=0;o<n.length-1;++o)n[o]=Math.floor(r/t[o]),r-=n[o]*t[o];return n[n.length-1]=r,n}function Am(r){return r&&r.then&&typeof r.then=="function"}var NS="tfjsflags",eg=class{constructor(e){this.global=e,this.flags={},this.flagRegistry={},this.urlFlags={},this.getQueryParams=ZG,this.populateURLFlags()}setPlatform(e,t){this.platform!=null&&console.warn(`Platform ${this.platformName} has already been set. Overwriting the platform with ${t}.`),this.platformName=e,this.platform=t}registerFlag(e,t,n){if(this.flagRegistry[e]={evaluationFn:t,setHook:n},this.urlFlags[e]!=null){let o=this.urlFlags[e];console.warn(`Setting feature override from URL ${e}: ${o}.`),this.set(e,o)}}async getAsync(e){return e in this.flags?this.flags[e]:(this.flags[e]=await this.evaluateFlag(e),this.flags[e])}get(e){if(e in this.flags)return this.flags[e];let t=this.evaluateFlag(e);if(Am(t))throw new Error(`Flag ${e} cannot be synchronously evaluated. Please use getAsync() instead.`);return this.flags[e]=t,this.flags[e]}getNumber(e){return this.get(e)}getBool(e){return this.get(e)}getFlags(){return this.flags}get features(){return this.flags}set(e,t){if(this.flagRegistry[e]==null)throw new Error(`Cannot set flag ${e} as it has not been registered.`);this.flags[e]=t,this.flagRegistry[e].setHook!=null&&this.flagRegistry[e].setHook(t)}evaluateFlag(e){if(this.flagRegistry[e]==null)throw new Error(`Cannot evaluate flag '${e}': no evaluation function found.`);return this.flagRegistry[e].evaluationFn()}setFlags(e){this.flags=Object.assign({},e)}reset(){this.flags={},this.urlFlags={},this.populateURLFlags()}populateURLFlags(){if(typeof this.global=="undefined"||typeof this.global.location=="undefined"||typeof this.global.location.search=="undefined")return;let e=this.getQueryParams(this.global.location.search);NS in e&&e[NS].split(",").forEach(n=>{let[o,s]=n.split(":");this.urlFlags[o]=QG(o,s)})}};function ZG(r){let e={};return r.replace(/[?&]([^=?&]+)(?:=([^&]*))?/g,(t,...n)=>(JG(e,n[0],n[1]),n.join("="))),e}function JG(r,e,t){r[decodeURIComponent(e)]=decodeURIComponent(t||"")}function QG(r,e){if(e=e.toLowerCase(),e==="true"||e==="false")return e==="true";if(`${+e}`===e)return+e;throw new Error(`Could not parse value flag value ${e} for flag ${r}.`)}function W(){return iw}var iw=null;function SS(r){iw=r}var aw;function lw(){if(aw==null){let r;if(typeof window!="undefined")r=window;else if(typeof global!="undefined")r=global;else if(typeof process!="undefined")r=process;else if(typeof self!="undefined")r=self;else throw new Error("Could not find a global object");aw=r}return aw}function eW(){let r=lw();return r._tfGlobals==null&&(r._tfGlobals=new Map),r._tfGlobals}function Em(r,e){let t=eW();if(t.has(r))return t.get(r);{let n=e();return t.set(r,n),t.get(r)}}var Bs="Abs",ki="Acos",vi="Acosh",Pn="Add",_o="AddN",Ci="All",Ii="Any",ko="ArgMax",Ya="ArgMin",Ni="Asin",Si="Asinh",Ti="Atan",Ai="Atanh",Ei="Atan2",vo="AvgPool",wc="AvgPoolGrad",Za="AvgPool3D",_c="AvgPool3DGrad",Co="BatchMatMul",Ja="BatchToSpaceND",kc="Bincount",TS="BroadcastTo",qn="Cast",Io="Ceil",Kn="ClipByValue",vc="Complex",Qa="ComplexAbs",Vs="Concat",No="Conv2D",Cc="Conv2DBackpropFilter",So="Conv2DBackpropInput",el="Conv3D",Ic="Conv3DBackpropFilterV2",Nc="Conv3DBackpropInputV2",To="Cos",Di="Cosh",Ao="Cumsum",$i="CropAndResize",Sc="DenseBincount",Ri="DepthToSpace",Eo="DepthwiseConv2dNative",Tc="DepthwiseConv2dNativeBackpropFilter",Ac="DepthwiseConv2dNativeBackpropInput",Ec="Diag",tl="Dilation2D",Dm="Dilation2DBackpropInput",$m="Dilation2DBackpropFilter",Do="RealDiv",Dc="Einsum",Fi="Elu",$c="EluGrad",Oi="Erf",Pi="Equal",$o="Exp",Gs="ExpandDims",Mi="Expm1",Rc="FFT",rl="Fill",Li="FlipLeftRight",Ro="Floor",Fo="FloorDiv",Oo="FusedBatchNorm",Ws="GatherV2",zi="GatherNd",Bi="Greater",Po="GreaterEqual",Xn="Identity",Fc="IFFT",Oc="Imag",Vi="IsFinite",Gi="IsInf",Wi="IsNan",Mo="LeakyRelu",ji="Less",Ui="LessEqual",Pc="LinSpace",Lo="Log",Hi="Log1p",qi="LogicalAnd",Hl="LogicalNot",ql="LogicalOr",AS="LogSoftmax",nl="LRN",Mc="LRNGrad",zo="Max",Bo="Maximum",Vo="MaxPool",Lc="MaxPoolGrad",ol="MaxPool3D",zc="MaxPool3DGrad",Bc="MaxPoolWithArgmax",Go="Mean",Wo="Min",jo="Minimum",Uo="MirrorPad",Ki="Mod",Vc="Multinomial",Ho="Multiply",js="Neg",Xi="NotEqual",Yi="NonMaxSuppressionV3",Zi="NonMaxSuppressionV4",Ji="NonMaxSuppressionV5",Us="OnesLike",qo="OneHot",Hs="Pack",Ko="PadV2",toe="Pool",Xo="Pow",Yo="Prelu",Qi="Prod",sl="Range",Gc="Real",ea="Reciprocal",Zo="Relu",qs="Reshape",il="ResizeNearestNeighbor",Wc="ResizeNearestNeighborGrad",Jo="ResizeBilinear",jc="ResizeBilinearGrad",Qo="Relu6",es="Reverse",ts="Round",rs="Rsqrt",ta="ScatterNd",Ks="Select",ra="Selu",Xs="Slice",ns="Sin",na="Sinh",oa="Sign",os="Sigmoid",sa="Softplus",ss="Sqrt",is="Sum",al="SpaceToBatchND",Ys="SplitV",as="Softmax",Uc="SparseFillEmptyRows",Hc="SparseReshape",qc="SparseToDense",ls="SquaredDifference",ll="Square",ia="StridedSlice",us="Sub",cs="Tan",ps="Tanh",Mn="Tile",aa="TopK",la="Transform",ms="Transpose",Kc="Unique",Zs="Unpack",ul="UnsortedSegmentSum",Js="ZerosLike",Yn="Step",Rm="FromPixels",ua="RotateWithOffset",Qs="_FusedMatMul",ei="FusedConv2D",ti="FusedDepthwiseConv2D";var Xc=Em("kernelRegistry",()=>new Map),Fm=Em("gradRegistry",()=>new Map);function Om(r,e){let t=cw(r,e);return Xc.get(t)}function uw(r){return Fm.get(r)}function tg(r){let e=Xc.entries(),t=[];for(;;){let{done:n,value:o}=e.next();if(n)break;let[s,a]=o,[i]=s.split("_");i===r&&t.push(a)}return t}function Kl(r){let{kernelName:e,backendName:t}=r,n=cw(e,t);Xc.has(n)&&console.warn(`The kernel '${e}' for backend '${t}' is already registered`),Xc.set(n,r)}function ES(r){let{kernelName:e}=r;Fm.has(e)&&W().getBool("DEBUG")&&console.warn(`Overriding the gradient for '${e}'`),Fm.set(e,r)}function soe(r,e){let t=cw(r,e);if(!Xc.has(t))throw new Error(`The kernel '${r}' for backend '${e}' is not registered`);Xc.delete(t)}function ioe(r){if(!Fm.has(r))throw new Error(`The gradient '${r}' for backend is not registered`);Fm.delete(r)}function aoe(r,e){tg(r).forEach(n=>{let o=Object.assign({},n,{backendName:e});Kl(o)})}function cw(r,e){return`${e}_${r}`}var y={};We(y,{arraysEqual:()=>Xr,assert:()=>T,assertNonNegativeIntegerDimensions:()=>Tm,assertNonNull:()=>On,assertShapesMatch:()=>Dt,bytesFromStringArray:()=>sw,bytesPerElement:()=>Qh,checkConversionForErrors:()=>nw,clamp:()=>gc,computeStrides:()=>zs,createScalarValue:()=>tW,createShuffledIndices:()=>jG,decodeString:()=>Zc,distSquared:()=>BG,encodeString:()=>cl,fetch:()=>nW,flatten:()=>bo,getArrayFromDType:()=>rw,getTypedArrayFromDType:()=>tw,hasEncodingLoss:()=>qG,indexToLoc:()=>YG,inferDtype:()=>xc,inferFromImplicitShape:()=>HG,isBoolean:()=>vS,isFunction:()=>Ls,isInt:()=>ot,isNumber:()=>CS,isPromise:()=>Am,isScalarShape:()=>VG,isString:()=>wo,isTypedArray:()=>ar,isValidDtype:()=>ow,locToIndex:()=>XG,makeOnesTypedArray:()=>Sm,makeZerosNestedTypedArray:()=>KG,makeZerosTypedArray:()=>bc,nearestDivisor:()=>yc,nearestLargerEven:()=>MG,now:()=>Xl,parseAxisParam:()=>nr,randUniform:()=>zG,repeatedTry:()=>UG,rightPad:()=>jl,shuffle:()=>kS,shuffleCombo:()=>PG,sizeFromShape:()=>nt,sizeToSquarishShape:()=>WG,squeezeShape:()=>ew,sum:()=>LG,tanh:()=>GG,toNestedArray:()=>Ul,toTypedArray:()=>Yc});function tW(r,e){return e==="string"?cl(r):Yc([r],e)}function rW(r,e){return r instanceof Float32Array&&e==="float32"||r instanceof Int32Array&&e==="int32"||r instanceof Uint8Array&&e==="bool"}function Yc(r,e){if(e==="string")throw new Error("Cannot convert a string[] to a TypedArray");if(Array.isArray(r)&&(r=bo(r)),W().getBool("DEBUG")&&nw(r,e),rW(r,e))return r;if(e==null||e==="float32"||e==="complex64")return new Float32Array(r);if(e==="int32")return new Int32Array(r);if(e==="bool"){let t=new Uint8Array(r.length);for(let n=0;n<t.length;++n)Math.round(r[n])!==0&&(t[n]=1);return t}else throw new Error(`Unknown data type ${e}`)}function Xl(){return W().platform.now()}function nW(r,e){return W().platform.fetch(r,e)}function cl(r,e="utf-8"){return e=e||"utf-8",W().platform.encode(r,e)}function Zc(r,e="utf-8"){return e=e||"utf-8",W().platform.decode(r,e)}var pw=class{constructor(e,t){this.backendTimer=e,this.logger=t,t==null&&(this.logger=new DS)}profileKernel(e,t,n){let o,s=()=>{o=n()},a,i=Xl();if(this.backendTimer.timerAvailable())a=this.backendTimer.time(s);else{s();for(let u of o)u.dataSync();a=Promise.resolve({kernelMs:Xl()-i})}if(W().getBool("CHECK_COMPUTATION_FOR_ERRORS"))for(let u=0;u<o.length;u++){let c=o[u];c.data().then(p=>{oW(p,c.dtype,e)})}return{kernelName:e,outputs:o,inputs:t,timeMs:a.then(u=>u.kernelMs),extraInfo:a.then(u=>u.getExtraProfileInfo!=null?u.getExtraProfileInfo():"")}}logKernelProfile(e){let{kernelName:t,outputs:n,timeMs:o,inputs:s,extraInfo:a}=e;n.forEach(i=>{Promise.all([i.data(),o,a]).then(l=>{this.logger.logKernelProfile(t,i,l[0],l[1],s,l[2])})})}};function oW(r,e,t){if(e!=="float32")return!1;for(let n=0;n<r.length;n++){let o=r[n];if(isNaN(o)||!isFinite(o))return console.warn(`Found ${o} in the result of '${t}'`),!0}return!1}var DS=class{logKernelProfile(e,t,n,o,s,a){let i=typeof o=="number"?jl(`${o}ms`,9):o.error,l=jl(e,25),u=t.rank,c=t.size,p=jl(t.shape.toString(),14),m="";for(let f in s){let d=s[f];if(d!=null){let h=d.shape||t.shape,g=h.length;m+=`${f}: ${g}D ${g>0?h:""} `}}console.log(`%c${l} %c${i} %c${u}D ${p} %c${c} %c${m} %c${a}`,"font-weight:bold","color:red","color:blue","color: orange","color: green","color: steelblue")}};function $S(r,e,t){let n={},o={};for(let l=0;l<e.length;l++)n[e[l].id]=!0;for(let l=0;l<r.length;l++){let u=r[l],c=u.inputs;for(let p in c){let m=c[p],f=!1;for(let d=0;d<e.length;d++)if(n[m.id]){u.outputs.forEach(h=>n[h.id]=!0),f=!0,o[u.id]=!0;break}if(f)break}}let s={};s[t.id]=!0;let a={};for(let l=r.length-1;l>=0;l--){let u=r[l],c=u.inputs;for(let p=0;p<u.outputs.length;p++)if(s[u.outputs[p].id]){for(let m in c)s[c[m].id]=!0,a[u.id]=!0;break}}let i=[];for(let l=0;l<r.length;l++){let u=r[l];if(o[u.id]&&a[u.id]){let c={};for(let m in u.inputs){let f=u.inputs[m];n[f.id]&&(c[m]=f)}let p=Object.assign({},u);p.inputs=c,p.outputs=u.outputs,i.push(p)}}return i}function RS(r,e,t,n){for(let o=e.length-1;o>=0;o--){let s=e[o],a=[];if(s.outputs.forEach(l=>{let u=r[l.id];u!=null?a.push(u):a.push(null)}),s.gradient==null)throw new Error(`Cannot compute gradient: gradient function not found for ${s.kernelName}.`);let i=s.gradient(a);for(let l in s.inputs){if(!(l in i))throw new Error(`Cannot backprop through input ${l}. Available gradients found: ${Object.keys(i)}.`);let u=t(()=>i[l]());if(u.dtype!=="float32")throw new Error(`Error in gradient for op ${s.kernelName}. The gradient of input ${l} must have 'float32' dtype, but has '${u.dtype}'`);let c=s.inputs[l];if(!Xr(u.shape,c.shape))throw new Error(`Error in gradient for op ${s.kernelName}. The gradient of input '${l}' has shape '${u.shape}', which does not match the shape of the input '${c.shape}'`);if(r[c.id]==null)r[c.id]=u;else{let p=r[c.id];r[c.id]=n(p,u),p.dispose()}}}}var FS=20,Pm=3,mw=7;function OS(r,e,t,n){let o=zs(e),s=sW(r,e,t,o),a=e.length,i=rg(r,e,t,o,s),l=["Tensor"];return n&&(l.push(` dtype: ${t}`),l.push(` rank: ${a}`),l.push(` shape: [${e}]`),l.push(" values:")),l.push(i.map(u=>" "+u).join(`
|
|
`)),l.join(`
|
|
`)}function sW(r,e,t,n){let o=nt(e),s=n[n.length-1],a=new Array(s).fill(0),i=e.length,l=t==="complex64"?Lm(r):r;if(i>1)for(let u=0;u<o/s;u++){let c=u*s;for(let p=0;p<s;p++)a[p]=Math.max(a[p],Mm(l[c+p],0,t).length)}return a}function Mm(r,e,t){let n;return Array.isArray(r)?n=`${parseFloat(r[0].toFixed(mw))} + ${parseFloat(r[1].toFixed(mw))}j`:wo(r)?n=`'${r}'`:t==="bool"?n=PS(r):n=parseFloat(r.toFixed(mw)).toString(),jl(n,e)}function PS(r){return r===0?"false":"true"}function rg(r,e,t,n,o,s=!0){let a=t==="complex64"?2:1,i=e[0],l=e.length;if(l===0){if(t==="complex64"){let h=Lm(r);return[Mm(h[0],0,t)]}return t==="bool"?[PS(r[0])]:[r[0].toString()]}if(l===1){if(i>FS){let g=Pm*a,x=Array.from(r.slice(0,g)),b=Array.from(r.slice((i-Pm)*a,i*a));return t==="complex64"&&(x=Lm(x),b=Lm(b)),["["+x.map((w,_)=>Mm(w,o[_],t)).join(", ")+", ..., "+b.map((w,_)=>Mm(w,o[i-Pm+_],t)).join(", ")+"]"]}let h=t==="complex64"?Lm(r):Array.from(r);return["["+h.map((g,x)=>Mm(g,o[x],t)).join(", ")+"]"]}let u=e.slice(1),c=n.slice(1),p=n[0]*a,m=[];if(i>FS){for(let h=0;h<Pm;h++){let g=h*p,x=g+p;m.push(...rg(r.slice(g,x),u,t,c,o,!1))}m.push("...");for(let h=i-Pm;h<i;h++){let g=h*p,x=g+p;m.push(...rg(r.slice(g,x),u,t,c,o,h===i-1))}}else for(let h=0;h<i;h++){let g=h*p,x=g+p;m.push(...rg(r.slice(g,x),u,t,c,o,h===i-1))}let f=l===2?",":"";m[0]="["+m[0]+f;for(let h=1;h<m.length-1;h++)m[h]=" "+m[h]+f;let d=`,
|
|
`;for(let h=2;h<l;h++)d+=`
|
|
`;return m[m.length-1]=" "+m[m.length-1]+"]"+(s?"":d),m}function Lm(r){let e=[];for(let t=0;t<r.length;t+=2)e.push([r[t],r[t+1]]);return e}var ut=class{constructor(e,t,n){if(this.dtype=t,this.shape=e.slice(),this.size=nt(e),n!=null){let o=n.length;T(o===this.size,()=>`Length of values '${o}' does not match the size inferred by the shape '${this.size}'.`)}if(t==="complex64")throw new Error("complex64 dtype TensorBuffers are not supported. Please create a TensorBuffer for the real and imaginary parts separately and call tf.complex(real, imag).");this.values=n||rw(t,this.size),this.strides=zs(e)}set(e,...t){t.length===0&&(t=[0]),T(t.length===this.rank,()=>`The number of provided coordinates (${t.length}) must match the rank (${this.rank})`);let n=this.locToIndex(t);this.values[n]=e}get(...e){e.length===0&&(e=[0]);let t=0;for(let o of e){if(o<0||o>=this.shape[t]){let s=`Requested out of range element at ${e}. Buffer shape=${this.shape}`;throw new Error(s)}t++}let n=e[e.length-1];for(let o=0;o<e.length-1;++o)n+=this.strides[o]*e[o];return this.values[n]}locToIndex(e){if(this.rank===0)return 0;if(this.rank===1)return e[0];let t=e[e.length-1];for(let n=0;n<e.length-1;++n)t+=this.strides[n]*e[n];return t}indexToLoc(e){if(this.rank===0)return[];if(this.rank===1)return[e];let t=new Array(this.shape.length);for(let n=0;n<t.length-1;++n)t[n]=Math.floor(e/this.strides[n]),e-=t[n]*this.strides[n];return t[t.length-1]=e,t}get rank(){return this.shape.length}toTensor(){return ri().makeTensor(this.values,this.shape,this.dtype)}},ri=null,Jc=null,iW=null;function MS(r){ri=r}function LS(r){Jc=r}function zS(r){iW=r}var Pe=class{constructor(e,t,n,o){this.kept=!1,this.isDisposedInternal=!1,this.shape=e.slice(),this.dtype=t||"float32",this.size=nt(e),this.strides=zs(e),this.dataId=n,this.id=o,this.rankType=this.rank<5?this.rank.toString():"higher"}get rank(){return this.shape.length}async buffer(){let e=await this.data();return Jc.buffer(this.shape,this.dtype,e)}bufferSync(){return Jc.buffer(this.shape,this.dtype,this.dataSync())}async array(){let e=await this.data();return Ul(this.shape,e,this.dtype==="complex64")}arraySync(){return Ul(this.shape,this.dataSync(),this.dtype==="complex64")}async data(){this.throwIfDisposed();let e=ri().read(this.dataId);if(this.dtype==="string"){let t=await e;try{return t.map(n=>Zc(n))}catch(n){throw new Error("Failed to decode the string bytes into utf-8. To get the original bytes, call tensor.bytes().")}}return e}dataSync(){this.throwIfDisposed();let e=ri().readSync(this.dataId);if(this.dtype==="string")try{return e.map(t=>Zc(t))}catch(t){throw new Error("Failed to decode the string bytes into utf-8. To get the original bytes, call tensor.bytes().")}return e}async bytes(){this.throwIfDisposed();let e=await ri().read(this.dataId);return this.dtype==="string"?e:new Uint8Array(e.buffer)}dispose(){this.isDisposed||(ri().disposeTensor(this),this.isDisposedInternal=!0)}get isDisposed(){return this.isDisposedInternal}throwIfDisposed(){if(this.isDisposed)throw new Error("Tensor is disposed.")}print(e=!1){return Jc.print(this,e)}clone(){return this.throwIfDisposed(),Jc.clone(this)}toString(e=!1){let t=this.dataSync();return OS(t,this.shape,this.dtype,e)}cast(e){return this.throwIfDisposed(),Jc.cast(this,e)}variable(e=!0,t,n){return this.throwIfDisposed(),ri().makeVariable(this,e,t,n)}};Object.defineProperty(Pe,Symbol.hasInstance,{value:r=>!!r&&r.data!=null&&r.dataSync!=null&&r.throwIfDisposed!=null});function P(){return Em("Tensor",()=>Pe)}P();var pl=class extends Pe{constructor(e,t,n,o){super(e.shape,e.dtype,e.dataId,o);this.trainable=t,this.name=n}assign(e){if(e.dtype!==this.dtype)throw new Error(`dtype of the new value (${e.dtype}) and previous value (${this.dtype}) must match`);if(!Xr(e.shape,this.shape))throw new Error(`shape of the new value (${e.shape}) and previous value (${this.shape}) must match`);ri().disposeTensor(this),this.dataId=e.dataId,ri().incRef(this,null)}dispose(){ri().disposeVariable(this),this.isDisposedInternal=!0}};Object.defineProperty(pl,Symbol.hasInstance,{value:r=>r instanceof Pe&&r.assign!=null&&r.assign instanceof Function});var Zn={};We(Zn,{assertTypesMatch:()=>yw,getTensorsInContainer:()=>zm,isTensorInList:()=>lW,makeTypesMatch:()=>je});var fw;(function(r){r.R0="R0",r.R1="R1",r.R2="R2",r.R3="R3",r.R4="R4",r.R5="R5",r.R6="R6"})(fw||(fw={}));var dw;(function(r){r.float32="float32",r.int32="int32",r.bool="int32",r.complex64="complex64"})(dw||(dw={}));var hw;(function(r){r.float32="float32",r.int32="int32",r.bool="bool",r.complex64="complex64"})(hw||(hw={}));var gw;(function(r){r.float32="float32",r.int32="float32",r.bool="float32",r.complex64="complex64"})(gw||(gw={}));var xw;(function(r){r.float32="complex64",r.int32="complex64",r.bool="complex64",r.complex64="complex64"})(xw||(xw={}));var aW={float32:gw,int32:dw,bool:hw,complex64:xw};function lr(r,e){if(r==="string"||e==="string"){if(r==="string"&&e==="string")return"string";throw new Error(`Can not upcast ${r} with ${e}`)}return aW[r][e]}function Yl(r){return lr(r,"int32")}function je(r,e){if(r.dtype===e.dtype)return[r,e];let t=lr(r.dtype,e.dtype);return[r.cast(t),e.cast(t)]}function yw(r,e){T(r.dtype===e.dtype,()=>`The dtypes of the first(${r.dtype}) and second(${e.dtype}) input must match`)}function lW(r,e){return e.some(t=>t.id===r.id)}function zm(r){let e=[],t=new Set;return BS(r,e,t),e}function BS(r,e,t){if(r==null)return;if(r instanceof Pe){e.push(r);return}if(!uW(r))return;let n=r;for(let o in n){let s=n[o];t.has(s)||(t.add(s),BS(s,e,t))}}function uW(r){return Array.isArray(r)||typeof r=="object"}function bw(r){return r.kernelName!=null}var ww=class{constructor(){this.registeredVariables={},this.nextTapeNodeId=0,this.numBytes=0,this.numTensors=0,this.numStringTensors=0,this.numDataBuffers=0,this.gradientDepth=0,this.kernelDepth=0,this.scopeStack=[],this.numDataMovesStack=[],this.nextScopeId=0,this.tensorInfo=new WeakMap,this.profiling=!1,this.activeProfile={newBytes:0,newTensors:0,peakBytes:0,kernels:[],result:null,get kernelNames(){return Array.from(new Set(this.kernels.map(e=>e.name)))}}}dispose(){for(let e in this.registeredVariables)this.registeredVariables[e].dispose()}},Zl=class{constructor(e){this.ENV=e,this.registry={},this.registryFactory={},this.pendingBackendInitId=0,this.state=new ww}async ready(){if(this.pendingBackendInit!=null)return this.pendingBackendInit.then(()=>{});if(this.backendInstance!=null)return;let e=this.getSortedBackends();for(let t=0;t<e.length;t++){let n=e[t];if(await this.initializeBackend(n).success){await this.setBackend(n);return}}throw new Error("Could not initialize any backends, all backend initializations failed.")}get backend(){if(this.pendingBackendInit!=null)throw new Error(`Backend '${this.backendName}' has not yet been initialized. Make sure to await tf.ready() or await tf.setBackend() before calling other methods`);if(this.backendInstance==null){let{name:e,asyncInit:t}=this.initializeBackendsAndReturnBest();if(t)throw new Error(`The highest priority backend '${e}' has not yet been initialized. Make sure to await tf.ready() or await tf.setBackend() before calling other methods`);this.setBackend(e)}return this.backendInstance}backendNames(){return Object.keys(this.registryFactory)}findBackend(e){if(!(e in this.registry))if(e in this.registryFactory){let{asyncInit:t}=this.initializeBackend(e);if(t)return null}else return null;return this.registry[e]}findBackendFactory(e){return e in this.registryFactory?this.registryFactory[e].factory:null}registerBackend(e,t,n=1){return e in this.registryFactory?(console.warn(`${e} backend was already registered. Reusing existing backend factory.`),!1):(this.registryFactory[e]={factory:t,priority:n},!0)}async setBackend(e){if(this.registryFactory[e]==null)throw new Error(`Backend name '${e}' not found in registry`);if(this.backendName=e,this.registry[e]==null){this.backendInstance=null;let{success:t,asyncInit:n}=this.initializeBackend(e);if(!(n?await t:t))return!1}return this.backendInstance=this.registry[e],this.setupRegisteredKernels(),this.profiler=new pw(this.backendInstance),!0}setupRegisteredKernels(){tg(this.backendName).forEach(t=>{t.setupFunc!=null&&t.setupFunc(this.backendInstance)})}disposeRegisteredKernels(e){tg(e).forEach(n=>{n.disposeFunc!=null&&n.disposeFunc(this.registry[e])})}initializeBackend(e){let t=this.registryFactory[e];if(t==null)throw new Error(`Cannot initialize backend ${e}, no registration found.`);try{let n=t.factory();if(n&&!(n instanceof Ms)&&typeof n.then=="function"){let o=++this.pendingBackendInitId,s=n.then(a=>o<this.pendingBackendInitId?!1:(this.registry[e]=a,this.pendingBackendInit=null,!0)).catch(a=>(o<this.pendingBackendInitId||(this.pendingBackendInit=null,console.warn(`Initialization of backend ${e} failed`),console.warn(a.stack||a.message)),!1));return this.pendingBackendInit=s,{success:s,asyncInit:!0}}else return this.registry[e]=n,{success:!0,asyncInit:!1}}catch(n){return console.warn(`Initialization of backend ${e} failed`),console.warn(n.stack||n.message),{success:!1,asyncInit:!1}}}removeBackend(e){if(!(e in this.registryFactory))throw new Error(`${e} backend not found in registry`);this.backendName===e&&this.pendingBackendInit!=null&&this.pendingBackendInitId++,e in this.registry&&(this.disposeRegisteredKernels(e),this.registry[e].dispose(),delete this.registry[e]),delete this.registryFactory[e],this.backendName===e&&(this.pendingBackendInit=null,this.backendName=null,this.backendInstance=null)}getSortedBackends(){if(Object.keys(this.registryFactory).length===0)throw new Error("No backend found in registry.");return Object.keys(this.registryFactory).sort((e,t)=>this.registryFactory[t].priority-this.registryFactory[e].priority)}initializeBackendsAndReturnBest(){let e=this.getSortedBackends();for(let t=0;t<e.length;t++){let n=e[t],{success:o,asyncInit:s}=this.initializeBackend(n);if(s||o)return{name:n,asyncInit:s}}throw new Error("Could not initialize any backends, all backend initializations failed.")}moveData(e,t){let n=this.state.tensorInfo.get(t),o=n.backend,s=this.readSync(t),a=o.refCount(t);o.disposeData(t,!0),n.backend=e,e.move(t,s,n.shape,n.dtype,a),this.shouldCheckForMemLeaks()&&this.state.numDataMovesStack[this.state.numDataMovesStack.length-1]++}tidy(e,t){let n=null;if(t==null){if(typeof e!="function")throw new Error("Please provide a function to tidy()");t=e}else{if(typeof e!="string"&&!(e instanceof String))throw new Error("When calling with two arguments, the first argument to tidy() must be a string");if(typeof t!="function")throw new Error("When calling with two arguments, the 2nd argument to tidy() must be a function");n=e}let o;return this.scopedRun(()=>this.startScope(n),()=>this.endScope(o),()=>(o=t(),o instanceof Promise&&console.error("Cannot return a Promise inside of tidy."),o))}scopedRun(e,t,n){e();try{let o=n();return t(),o}catch(o){throw t(),o}}nextTensorId(){return Zl.nextTensorId++}nextVariableId(){return Zl.nextVariableId++}clone(e){let t=A.runKernel(Xn,{x:e}),n={x:e},o=a=>({x:()=>{let i="float32",l={x:a},u={dtype:i};return A.runKernel(qn,l,u)}}),s=[];return this.addTapeNode(this.state.activeScope.name,n,[t],o,s,{}),t}runKernel(e,t,n){if(!(Om(e,this.backendName)!=null))throw new Error(`Kernel '${e}' not registered for backend '${this.backendName}'`);return this.runKernelFunc({kernelName:e,inputs:t,attrs:n})}shouldCheckForMemLeaks(){return this.ENV.getBool("IS_TEST")}checkKernelForMemLeak(e,t,n){let o=this.backend.numDataIds(),s=0;n.forEach(l=>{s+=l.dtype==="complex64"?3:1});let a=this.state.numDataMovesStack[this.state.numDataMovesStack.length-1],i=o-t-s-a;if(i>0)throw new Error(`Backend '${this.backendName}' has an internal memory leak (${i} data ids) after running '${e}'`)}runKernelFunc(e){let t,n=[],o=this.isTapeOn(),s=this.state.numBytes,a=this.state.numTensors;this.shouldCheckForMemLeaks()&&this.state.numDataMovesStack.push(0);let i;this.backendName==null&&this.backend;let l,u=bw(e)?e.kernelName:this.state.activeScope!=null?this.state.activeScope.name:"";if(bw(e)){let{kernelName:d,inputs:h,attrs:g}=e;this.backendName==null&&this.backend;let x=Om(d,this.backendName);T(x!=null,()=>`Cannot find registered kernel '${d}' for backend '${this.backendName}'`),i=()=>{let b=this.backend.numDataIds();l=x.kernelFunc({inputs:h,attrs:g,backend:this.backend});let w=Array.isArray(l)?l:[l];this.shouldCheckForMemLeaks()&&this.checkKernelForMemLeak(d,b,w);let _=w.map(I=>{if(I.rank!=null)return I;let{dataId:E,shape:D,dtype:$}=I;return this.makeTensorFromDataId(E,D,$)});if(o){let I=this.getTensorsForGradient(d,h,_);n=this.saveTensorsForBackwardMode(I)}return _}}else{let{forwardFunc:d}=e,h=g=>{!o||(n=g.map(x=>this.keep(this.clone(x))))};i=()=>{let g=this.backend.numDataIds();l=this.tidy(()=>d(this.backend,h));let x=Array.isArray(l)?l:[l];return this.shouldCheckForMemLeaks()&&this.checkKernelForMemLeak(u,g,x),x}}let{inputs:c,attrs:p}=e,m=bw(e)?null:e.backwardsFunc,f;return this.scopedRun(()=>this.state.kernelDepth++,()=>this.state.kernelDepth--,()=>{!this.ENV.getBool("DEBUG")&&!this.state.profiling?t=i():(f=this.profiler.profileKernel(u,c,()=>i()),this.ENV.getBool("DEBUG")&&this.profiler.logKernelProfile(f),t=f.outputs)}),o&&this.addTapeNode(u,c,t,m,n,p),this.state.profiling&&this.state.activeProfile.kernels.push({name:u,bytesAdded:this.state.numBytes-s,totalBytesSnapshot:this.state.numBytes,tensorsAdded:this.state.numTensors-a,totalTensorsSnapshot:this.state.numTensors,inputShapes:Object.keys(c).map(d=>c[d]!=null?c[d].shape:null),outputShapes:t.map(d=>d.shape),kernelTimeMs:f.timeMs,extraInfo:f.extraInfo}),Array.isArray(l)?t:t[0]}saveTensorsForBackwardMode(e){return e.map(n=>this.keep(this.clone(n)))}getTensorsForGradient(e,t,n){let o=uw(e);if(o!=null){let s=o.inputsToSave||[],a=o.outputsToSave||[],i;o.saveAllInputs?(T(Array.isArray(t),()=>"saveAllInputs is true, expected inputs to be an array."),i=Object.keys(t).map(u=>t[u])):i=s.map(u=>t[u]);let l=n.filter((u,c)=>a[c]);return i.concat(l)}return[]}makeTensor(e,t,n,o){if(e==null)throw new Error("Values passed to engine.makeTensor() are null");n=n||"float32",o=o||this.backend;let s=e;n==="string"&&wo(e[0])&&(s=e.map(l=>cl(l)));let a=o.write(s,t,n),i=new Pe(t,n,a,this.nextTensorId());if(this.trackTensor(i,o),n==="string"){let l=this.state.tensorInfo.get(a),u=sw(s);this.state.numBytes+=u-l.bytes,l.bytes=u}return i}makeTensorFromDataId(e,t,n,o){n=n||"float32";let s=new Pe(t,n,e,this.nextTensorId());return this.trackTensor(s,o),s}makeVariable(e,t=!0,n,o){n=n||this.nextVariableId().toString(),o!=null&&o!==e.dtype&&(e=e.cast(o));let s=new pl(e,t,n,this.nextTensorId());if(this.state.registeredVariables[s.name]!=null)throw new Error(`Variable with name ${s.name} was already registered`);return this.state.registeredVariables[s.name]=s,this.incRef(s,this.backend),s}trackTensor(e,t){this.state.numTensors++,e.dtype==="string"&&this.state.numStringTensors++;let n=0;e.dtype!=="complex64"&&e.dtype!=="string"&&(n=e.size*Qh(e.dtype)),this.state.numBytes+=n,this.state.tensorInfo.has(e.dataId)||(this.state.numDataBuffers++,this.state.tensorInfo.set(e.dataId,{backend:t||this.backend,dtype:e.dtype,shape:e.shape,bytes:n})),e instanceof pl||this.track(e)}incRef(e,t){this.trackTensor(e,t),this.backend.incRef(e.dataId)}removeDataId(e,t){this.state.tensorInfo.has(e)&&this.state.tensorInfo.get(e).backend===t&&(this.state.tensorInfo.delete(e),this.state.numDataBuffers--)}disposeTensor(e){if(!this.state.tensorInfo.has(e.dataId))return;let t=this.state.tensorInfo.get(e.dataId);if(this.state.numTensors--,e.dtype==="string"&&(this.state.numStringTensors--,this.state.numBytes-=t.bytes),e.dtype!=="complex64"&&e.dtype!=="string"){let n=e.size*Qh(e.dtype);this.state.numBytes-=n}t.backend.disposeData(e.dataId)&&this.removeDataId(e.dataId,t.backend)}disposeVariables(){for(let e in this.state.registeredVariables){let t=this.state.registeredVariables[e];this.disposeVariable(t)}}disposeVariable(e){this.disposeTensor(e),this.state.registeredVariables[e.name]!=null&&delete this.state.registeredVariables[e.name]}memory(){let e=this.backend.memory();return e.numTensors=this.state.numTensors,e.numDataBuffers=this.state.numDataBuffers,e.numBytes=this.state.numBytes,this.state.numStringTensors>0&&(e.unreliable=!0,e.reasons==null&&(e.reasons=[]),e.reasons.push("Memory usage by string tensors is approximate (2 bytes per character)")),e}async profile(e){this.state.profiling=!0;let t=this.state.numBytes,n=this.state.numTensors;this.state.activeProfile.kernels=[],this.state.activeProfile.result=await e(),this.state.profiling=!1,this.state.activeProfile.peakBytes=Math.max(...this.state.activeProfile.kernels.map(o=>o.totalBytesSnapshot)),this.state.activeProfile.newBytes=this.state.numBytes-t,this.state.activeProfile.newTensors=this.state.numTensors-n;for(let o of this.state.activeProfile.kernels)o.kernelTimeMs=await o.kernelTimeMs,o.extraInfo=await o.extraInfo;return this.state.activeProfile}isTapeOn(){return this.state.gradientDepth>0&&this.state.kernelDepth===0}addTapeNode(e,t,n,o,s,a){let i={id:this.state.nextTapeNodeId++,kernelName:e,inputs:t,outputs:n,saved:s},l=uw(e);l!=null&&(o=l.gradFunc),o!=null&&(i.gradient=u=>(u=u.map((c,p)=>{if(c==null){let m=n[p],f=bc(m.size,m.dtype);return this.makeTensor(f,m.shape,m.dtype)}return c}),o(u.length>1?u:u[0],s,a))),this.state.activeTape.push(i)}keep(e){return e.kept=!0,e}startTape(){this.state.gradientDepth===0&&(this.state.activeTape=[]),this.state.gradientDepth++}endTape(){this.state.gradientDepth--}startScope(e){let t={track:[],name:"unnamed scope",id:this.state.nextScopeId++};e&&(t.name=e),this.state.scopeStack.push(t),this.state.activeScope=t}endScope(e){let t=zm(e),n=new Set(t.map(s=>s.id));for(let s=0;s<this.state.activeScope.track.length;s++){let a=this.state.activeScope.track[s];!a.kept&&!n.has(a.id)&&a.dispose()}let o=this.state.scopeStack.pop();this.state.activeScope=this.state.scopeStack.length===0?null:this.state.scopeStack[this.state.scopeStack.length-1],t.forEach(s=>{!s.kept&&s.scopeId===o.id&&this.track(s)})}gradients(e,t,n,o=!1){if(T(t.length>0,()=>"gradients() received an empty list of xs."),n!=null&&n.dtype!=="float32")throw new Error(`dy must have 'float32' dtype, but has '${n.dtype}'`);let s=this.scopedRun(()=>this.startTape(),()=>this.endTape(),()=>this.tidy("forward",e));T(s instanceof Pe,()=>"The result y returned by f() must be a tensor.");let a=$S(this.state.activeTape,t,s);if(!o&&a.length===0&&t.length>0)throw new Error("Cannot compute gradient of y=f(x) with respect to x. Make sure that the f you passed encloses all operations that lead from x to y.");return this.tidy("backward",()=>{let i={};i[s.id]=n==null?cW(s.shape):n,RS(i,a,u=>this.tidy(u),pW);let l=t.map(u=>i[u.id]);return this.state.gradientDepth===0&&(this.state.activeTape.forEach(u=>{for(let c of u.saved)c.dispose()}),this.state.activeTape=null),{value:s,grads:l}})}customGrad(e){return T(Ls(e),()=>"The f passed in customGrad(f) must be a function."),(...t)=>{T(t.every(i=>i instanceof Pe),()=>"The args passed in customGrad(f)(x1, x2,...) must all be tensors");let n,o={};t.forEach((i,l)=>{o[l]=i});let s=(i,l)=>(n=e(...t,l),T(n.value instanceof Pe,()=>"The function f passed in customGrad(f) must return an object where `obj.value` is a tensor"),T(Ls(n.gradFunc),()=>"The function f passed in customGrad(f) must return an object where `obj.gradFunc` is a function."),n.value),a=(i,l)=>{let u=n.gradFunc(i,l),c=Array.isArray(u)?u:[u];T(c.length===t.length,()=>"The function f passed in customGrad(f) must return an object where `obj.gradFunc` is a function that returns the same number of tensors as inputs passed to f(...)."),T(c.every(m=>m instanceof Pe),()=>"The function f passed in customGrad(f) must return an object where `obj.gradFunc` is a function that returns a list of only tensors.");let p={};return c.forEach((m,f)=>{p[f]=()=>m}),p};return this.runKernelFunc({forwardFunc:s,backwardsFunc:a,inputs:o})}}readSync(e){return this.state.tensorInfo.get(e).backend.readSync(e)}read(e){return this.state.tensorInfo.get(e).backend.read(e)}async time(e){let t=Xl(),n=await this.backend.time(e);return n.wallMs=Xl()-t,n}track(e){return this.state.activeScope!=null&&(e.scopeId=this.state.activeScope.id,this.state.activeScope.track.push(e)),e}get registeredVariables(){return this.state.registeredVariables}reset(){this.pendingBackendInitId++,this.state.dispose(),this.ENV.reset(),this.state=new ww;for(let e in this.registry)this.disposeRegisteredKernels(e),this.registry[e].dispose(),delete this.registry[e];this.backendName=null,this.backendInstance=null,this.pendingBackendInit=null}};Zl.nextTensorId=0;Zl.nextVariableId=0;function cW(r){let e=Sm(nt(r),"float32");return A.makeTensor(e,r,"float32")}function _w(){let r=lw();if(r._tfengine==null){let e=new eg(r);r._tfengine=new Zl(e)}return SS(r._tfengine.ENV),MS(()=>r._tfengine),r._tfengine}var A=_w();function pW(r,e){let t={a:r,b:e};return A.runKernel(Pn,t)}var Jl={};We(Jl,{isBrowser:()=>kw,isMobile:()=>fW});function mW(){return typeof navigator!="undefined"&&navigator!=null}function fW(r){if(r||mW()){if(r||(r=navigator),r.product==="ReactNative")return!0;let e=r.userAgent||r.vendor||window.opera;return/(android|bb\d+|meego).+mobile|avantgo|bada\/|blackberry|blazer|compal|elaine|fennec|hiptop|iemobile|ip(hone|od)|iris|kindle|lge |maemo|midp|mmp|mobile.+firefox|netfront|opera m(ob|in)i|palm( os)?|phone|p(ixi|re)\/|plucker|pocket|psp|series(4|6)0|symbian|treo|up\.(browser|link)|vodafone|wap|windows ce|xda|xiino/i.test(e)||/1207|6310|6590|3gso|4thp|50[1-6]i|770s|802s|a wa|abac|ac(er|oo|s\-)|ai(ko|rn)|al(av|ca|co)|amoi|an(ex|ny|yw)|aptu|ar(ch|go)|as(te|us)|attw|au(di|\-m|r |s )|avan|be(ck|ll|nq)|bi(lb|rd)|bl(ac|az)|br(e|v)w|bumb|bw\-(n|u)|c55\/|capi|ccwa|cdm\-|cell|chtm|cldc|cmd\-|co(mp|nd)|craw|da(it|ll|ng)|dbte|dc\-s|devi|dica|dmob|do(c|p)o|ds(12|\-d)|el(49|ai)|em(l2|ul)|er(ic|k0)|esl8|ez([4-7]0|os|wa|ze)|fetc|fly(\-|_)|g1 u|g560|gene|gf\-5|g\-mo|go(\.w|od)|gr(ad|un)|haie|hcit|hd\-(m|p|t)|hei\-|hi(pt|ta)|hp( i|ip)|hs\-c|ht(c(\-| |_|a|g|p|s|t)|tp)|hu(aw|tc)|i\-(20|go|ma)|i230|iac( |\-|\/)|ibro|idea|ig01|ikom|im1k|inno|ipaq|iris|ja(t|v)a|jbro|jemu|jigs|kddi|keji|kgt( |\/)|klon|kpt |kwc\-|kyo(c|k)|le(no|xi)|lg( g|\/(k|l|u)|50|54|\-[a-w])|libw|lynx|m1\-w|m3ga|m50\/|ma(te|ui|xo)|mc(01|21|ca)|m\-cr|me(rc|ri)|mi(o8|oa|ts)|mmef|mo(01|02|bi|de|do|t(\-| |o|v)|zz)|mt(50|p1|v )|mwbp|mywa|n10[0-2]|n20[2-3]|n30(0|2)|n50(0|2|5)|n7(0(0|1)|10)|ne((c|m)\-|on|tf|wf|wg|wt)|nok(6|i)|nzph|o2im|op(ti|wv)|oran|owg1|p800|pan(a|d|t)|pdxg|pg(13|\-([1-8]|c))|phil|pire|pl(ay|uc)|pn\-2|po(ck|rt|se)|prox|psio|pt\-g|qa\-a|qc(07|12|21|32|60|\-[2-7]|i\-)|qtek|r380|r600|raks|rim9|ro(ve|zo)|s55\/|sa(ge|ma|mm|ms|ny|va)|sc(01|h\-|oo|p\-)|sdk\/|se(c(\-|0|1)|47|mc|nd|ri)|sgh\-|shar|sie(\-|m)|sk\-0|sl(45|id)|sm(al|ar|b3|it|t5)|so(ft|ny)|sp(01|h\-|v\-|v )|sy(01|mb)|t2(18|50)|t6(00|10|18)|ta(gt|lk)|tcl\-|tdg\-|tel(i|m)|tim\-|t\-mo|to(pl|sh)|ts(70|m\-|m3|m5)|tx\-9|up(\.b|g1|si)|utst|v400|v750|veri|vi(rg|te)|vk(40|5[0-3]|\-v)|vm40|voda|vulc|vx(52|53|60|61|70|80|81|83|85|98)|w3c(\-| )|webc|whit|wi(g |nc|nw)|wmlb|wonu|x700|yas\-|your|zeto|zte\-/i.test(e.substr(0,4))}return!1}function kw(){return typeof window!="undefined"&&window.document!=null||typeof WorkerGlobalScope!="undefined"}var fs=W();fs.registerFlag("DEBUG",()=>!1,r=>{r&&console.warn("Debugging mode is ON. The output of every math call will be downloaded to CPU and checked for NaNs. This significantly impacts performance.")});fs.registerFlag("IS_BROWSER",()=>kw());fs.registerFlag("IS_NODE",()=>typeof process!="undefined"&&typeof process.versions!="undefined"&&typeof process.versions.node!="undefined");fs.registerFlag("IS_CHROME",()=>typeof navigator!="undefined"&&navigator!=null&&navigator.userAgent!=null&&/Chrome/.test(navigator.userAgent)&&/Google Inc/.test(navigator.vendor));fs.registerFlag("PROD",()=>!1);fs.registerFlag("TENSORLIKE_CHECK_SHAPE_CONSISTENCY",()=>fs.getBool("DEBUG"));fs.registerFlag("DEPRECATION_WARNINGS_ENABLED",()=>!0);fs.registerFlag("IS_TEST",()=>!1);fs.registerFlag("CHECK_COMPUTATION_FOR_ERRORS",()=>!0);fs.registerFlag("WRAP_TO_IMAGEBITMAP",()=>!1);function Ar(r,e){let t=r;if(ar(r))return e==="string"?[]:[r.length];if(!Array.isArray(r))return[];let n=[];for(;Array.isArray(t)||ar(t)&&e!=="string";)n.push(t.length),t=t[0];return Array.isArray(r)&&W().getBool("TENSORLIKE_CHECK_SHAPE_CONSISTENCY")&&VS(r,n,[]),n}function VS(r,e,t){if(t=t||[],!Array.isArray(r)&&!ar(r)){T(e.length===0,()=>`Element arr[${t.join("][")}] is a primitive, but should be an array/TypedArray of ${e[0]} elements`);return}T(e.length>0,()=>`Element arr[${t.join("][")}] should be a primitive, but is an array of ${r.length} elements`),T(r.length===e[0],()=>`Element arr[${t.join("][")}] should have ${e[0]} elements, but has ${r.length} elements`);let n=e.slice(1);for(let o=0;o<r.length;++o)VS(r[o],n,t.concat(o))}function GS(r,e,t,n){if(r!=="string_or_numeric"){if(r==null)throw new Error("Expected dtype cannot be null.");if(r!=="numeric"&&r!==e||r==="numeric"&&e==="string")throw new Error(`Argument '${t}' passed to '${n}' must be ${r} tensor, but got ${e} tensor`)}}function k(r,e,t,n="numeric"){if(r instanceof Pe)return GS(n,r.dtype,e,t),r;let o=xc(r);if(o!=="string"&&["bool","int32","float32"].indexOf(n)>=0&&(o=n),GS(n,o,e,t),r==null||!ar(r)&&!Array.isArray(r)&&typeof r!="number"&&typeof r!="boolean"&&typeof r!="string"){let l=r==null?"null":r.constructor.name;throw new Error(`Argument '${e}' passed to '${t}' must be a Tensor or TensorLike, but got '${l}'`)}let s=Ar(r,o);!ar(r)&&!Array.isArray(r)&&(r=[r]);let i=o!=="string"?Yc(r,o):bo(r,[],!0);return A.makeTensor(i,s,o)}function ca(r,e,t,n="numeric"){if(!Array.isArray(r))throw new Error(`Argument ${e} passed to ${t} must be a \`Tensor[]\` or \`TensorLike[]\``);return r.map((s,a)=>k(s,`${e}[${a}]`,t,n))}var WS="__op";function S(r){let e=Object.keys(r);if(e.length!==1)throw new Error(`Please provide an object with a single key (operation name) mapping to a function. Got an object with ${e.length} keys.`);let t=e[0],n=r[t];t.endsWith("_")&&(t=t.substring(0,t.length-1)),t=t+WS;let o=(...s)=>{A.startScope(t);try{let a=n(...s);return Am(a)&&console.error("Cannot return a Promise inside of tidy."),A.endScope(a),a}catch(a){throw A.endScope(null),a}};return Object.defineProperty(o,"name",{value:t,configurable:!0}),o}function dW(r,e){let t=k(r,"real","complex"),n=k(e,"imag","complex");Dt(t.shape,n.shape,`real and imag shapes, ${t.shape} and ${n.shape}, must match in call to tf.complex().`);let o={real:t,imag:n};return A.runKernel(vc,o)}var Nn=S({complex_:dW});function Yr(r,e,t,n){if(n==null&&(n=xc(r)),n==="complex64")throw new Error("Cannot construct a complex64 tensor directly. Please use tf.complex(real, imag).");if(!ar(r)&&!Array.isArray(r)&&typeof r!="number"&&typeof r!="boolean"&&typeof r!="string")throw new Error("values passed to tensor(values) must be a number/boolean/string or an array of numbers/booleans/strings, or a TypedArray");if(e!=null){Tm(e);let o=nt(e),s=nt(t);T(o===s,()=>`Based on the provided shape, [${e}], the tensor should have ${o} values but has ${s}`);for(let a=0;a<t.length;++a){let i=t[a],l=a===t.length-1?i!==nt(e.slice(a)):!0;T(t[a]===e[a]||!l,()=>`Error creating a new Tensor. Inferred shape (${t}) does not match the provided shape (${e}). `)}}return!ar(r)&&!Array.isArray(r)&&(r=[r]),e=e||t,r=n!=="string"?Yc(r,n):bo(r,[],!0),A.makeTensor(r,e,n)}function vr(r,e,t){let n=Ar(r,t);return Yr(r,e,n,t)}var Bm={float32:4,float16:2,int32:4,uint16:2,uint8:1,bool:1,complex64:8};var ng=4;async function jS(r,e){let t=[],n=[],o=Array.isArray(r)?r.map(a=>a.name):Object.keys(r);for(let a=0;a<o.length;++a){let i=o[a],l=Array.isArray(r)?r[a].tensor:r[i];if(l.dtype!=="float32"&&l.dtype!=="int32"&&l.dtype!=="bool"&&l.dtype!=="string"&&l.dtype!=="complex64")throw new Error(`Unsupported dtype in weight '${i}': ${l.dtype}`);let u={name:i,shape:l.shape,dtype:l.dtype};if(l.dtype==="string"){let c=new Promise(async p=>{let m=await l.bytes(),f=m.reduce((g,x)=>g+x.length,0)+ng*m.length,d=new Uint8Array(f),h=0;for(let g=0;g<m.length;g++){let x=m[g],b=new Uint8Array(new Uint32Array([x.length]).buffer);d.set(b,h),h+=ng,d.set(x,h),h+=x.length}p(d)});n.push(c)}else n.push(l.data());e!=null&&(u.group=e),t.push(u)}let s=await Promise.all(n);return{data:hW(s),specs:t}}function og(r,e){let t={},n,o=0;for(let s of e){let a=s.name,i=s.dtype,l=s.shape,u=nt(l),c;if("quantization"in s){let p=s.quantization;if(p.dtype==="uint8"||p.dtype==="uint16"){if(!("min"in p&&"scale"in p))throw new Error(`Weight ${s.name} with quantization ${p.dtype} doesn't have corresponding metadata min and scale.`)}else if(p.dtype==="float16"){if(i!=="float32")throw new Error(`Weight ${s.name} is quantized with ${p.dtype} which only supports weights of type float32 not ${i}.`)}else throw new Error(`Weight ${s.name} has unknown quantization dtype ${p.dtype}. Supported quantization dtypes are: 'uint8', 'uint16', and 'float16'.`);let m=Bm[p.dtype],f=r.slice(o,o+u*m),d=p.dtype==="uint8"?new Uint8Array(f):new Uint16Array(f);if(i==="float32")if(p.dtype==="uint8"||p.dtype==="uint16"){c=new Float32Array(d.length);for(let h=0;h<d.length;h++){let g=d[h];c[h]=g*p.scale+p.min}}else if(p.dtype==="float16")n===void 0&&(n=bW()),c=n(d);else throw new Error(`Unsupported quantization type ${p.dtype} for weight type float32.`);else if(i==="int32"){if(p.dtype!=="uint8"&&p.dtype!=="uint16")throw new Error(`Unsupported quantization type ${p.dtype} for weight type int32.`);c=new Int32Array(d.length);for(let h=0;h<d.length;h++){let g=d[h];c[h]=Math.round(g*p.scale+p.min)}}else throw new Error(`Unsupported dtype in weight '${a}': ${i}`);o+=u*m}else if(i==="string"){let p=nt(s.shape);c=[];for(let m=0;m<p;m++){let f=new Uint32Array(r.slice(o,o+ng))[0];o+=ng;let d=new Uint8Array(r.slice(o,o+f));c.push(d),o+=f}}else{let p=Bm[i],m=r.slice(o,o+u*p);if(i==="float32")c=new Float32Array(m);else if(i==="int32")c=new Int32Array(m);else if(i==="bool")c=new Uint8Array(m);else if(i==="complex64"){c=new Float32Array(m);let f=new Float32Array(c.length/2),d=new Float32Array(c.length/2);for(let x=0;x<f.length;x++)f[x]=c[x*2],d[x]=c[x*2+1];let h=vr(f,l,"float32"),g=vr(d,l,"float32");t[a]=Nn(h,g),h.dispose(),g.dispose()}else throw new Error(`Unsupported dtype in weight '${a}': ${i}`);o+=u*p}i!=="complex64"&&(t[a]=vr(c,l,i))}return t}function hW(r){if(r===null)throw new Error(`Invalid input value: ${JSON.stringify(r)}`);let e=0,t=[];r.forEach(s=>{if(e+=s.byteLength,t.push(s.byteLength===s.buffer.byteLength?s:new s.constructor(s)),!(s instanceof Float32Array||s instanceof Int32Array||s instanceof Uint8Array))throw new Error(`Unsupported TypedArray subtype: ${s.constructor.name}`)});let n=new Uint8Array(e),o=0;return t.forEach(s=>{n.set(new Uint8Array(s.buffer),o),o+=s.byteLength}),n.buffer}var vw=typeof Buffer!="undefined"&&(typeof Blob=="undefined"||typeof atob=="undefined"||typeof btoa=="undefined");function US(r){return vw?Buffer.byteLength(r):new Blob([r]).size}function HS(r){if(vw)return Buffer.from(r).toString("base64");let e=new Uint8Array(r),t="";for(let n=0,o=e.length;n<o;n++)t+=String.fromCharCode(e[n]);return btoa(t)}function qS(r){if(vw){let n=Buffer.from(r,"base64");return n.buffer.slice(n.byteOffset,n.byteOffset+n.byteLength)}let e=atob(r),t=new Uint8Array(e.length);for(let n=0;n<e.length;++n)t.set([e.charCodeAt(n)],n);return t.buffer}function Qc(r){if(r.length===1)return r[0];let e=0;r.forEach(o=>{e+=o.byteLength});let t=new Uint8Array(e),n=0;return r.forEach(o=>{t.set(new Uint8Array(o),n),n+=o.byteLength}),t.buffer}function Cw(r){let e="/";for(r=r.trim();r.endsWith(e);)r=r.slice(0,r.length-1);let t=r.split(e);return t[t.length-1]}function ni(r){if(r.modelTopology instanceof ArrayBuffer)throw new Error("Expected JSON model topology, received ArrayBuffer.");return{dateSaved:new Date,modelTopologyType:"JSON",modelTopologyBytes:r.modelTopology==null?0:US(JSON.stringify(r.modelTopology)),weightSpecsBytes:r.weightSpecs==null?0:US(JSON.stringify(r.weightSpecs)),weightDataBytes:r.weightData==null?0:r.weightData.byteLength}}function gW(){let r=t=>{let n=t<<13,o=0;for(;(n&8388608)==0;)o-=8388608,n<<=1;return n&=~8388608,o+=947912704,n|o},e=new Uint32Array(2048);e[0]=0;for(let t=1;t<1024;t++)e[t]=r(t);for(let t=1024;t<2048;t++)e[t]=939524096+(t-1024<<13);return e}function xW(){let r=new Uint32Array(64);r[0]=0,r[31]=1199570944,r[32]=2147483648,r[63]=3347054592;for(let e=1;e<31;e++)r[e]=e<<23;for(let e=33;e<63;e++)r[e]=2147483648+(e-32<<23);return r}function yW(){let r=new Uint32Array(64);for(let e=0;e<64;e++)r[e]=1024;return r[0]=r[32]=0,r}function bW(){let r=gW(),e=xW(),t=yW();return n=>{let o=new ArrayBuffer(4*n.length),s=new Uint32Array(o);for(let a=0;a<n.length;a++){let i=n[a],l=r[t[i>>10]+(i&1023)]+e[i>>10];s[a]=l}return new Float32Array(o)}}var St=class{constructor(){this.saveRouters=[],this.loadRouters=[]}static getInstance(){return St.instance==null&&(St.instance=new St),St.instance}static registerSaveRouter(e){St.getInstance().saveRouters.push(e)}static registerLoadRouter(e){St.getInstance().loadRouters.push(e)}static getSaveHandlers(e){return St.getHandlers(e,"save")}static getLoadHandlers(e,t){return St.getHandlers(e,"load",t)}static getHandlers(e,t,n){let o=[];return(t==="load"?St.getInstance().loadRouters:St.getInstance().saveRouters).forEach(a=>{let i=a(e,n);i!==null&&o.push(i)}),o}},KS=r=>St.registerSaveRouter(r),XS=r=>St.registerLoadRouter(r),YS=r=>St.getSaveHandlers(r),ZS=(r,e)=>St.getLoadHandlers(r,e);var Iw="tensorflowjs",Nw=1,Ql="models_store",ml="model_info_store";function JS(){if(!W().getBool("IS_BROWSER"))throw new Error("Failed to obtain IndexedDB factory because the current environmentis not a web browser.");let r=typeof window=="undefined"?self:window,e=r.indexedDB||r.mozIndexedDB||r.webkitIndexedDB||r.msIndexedDB||r.shimIndexedDB;if(e==null)throw new Error("The current browser does not appear to support IndexedDB.");return e}function Sw(r){let e=r.result;e.createObjectStore(Ql,{keyPath:"modelPath"}),e.createObjectStore(ml,{keyPath:"modelPath"})}var pa=class{constructor(e){if(this.indexedDB=JS(),e==null||!e)throw new Error("For IndexedDB, modelPath must not be null, undefined or empty.");this.modelPath=e}async save(e){if(e.modelTopology instanceof ArrayBuffer)throw new Error("BrowserLocalStorage.save() does not support saving model topology in binary formats yet.");return this.databaseAction(this.modelPath,e)}async load(){return this.databaseAction(this.modelPath)}databaseAction(e,t){return new Promise((n,o)=>{let s=this.indexedDB.open(Iw,Nw);s.onupgradeneeded=()=>Sw(s),s.onsuccess=()=>{let a=s.result;if(t==null){let i=a.transaction(Ql,"readonly"),u=i.objectStore(Ql).get(this.modelPath);u.onsuccess=()=>{if(u.result==null)return a.close(),o(new Error(`Cannot find model with path '${this.modelPath}' in IndexedDB.`));n(u.result.modelArtifacts)},u.onerror=c=>(a.close(),o(u.error)),i.oncomplete=()=>a.close()}else{let i=ni(t),l=a.transaction(ml,"readwrite"),u=l.objectStore(ml),c=u.put({modelPath:this.modelPath,modelArtifactsInfo:i}),p;c.onsuccess=()=>{p=a.transaction(Ql,"readwrite");let f=p.objectStore(Ql).put({modelPath:this.modelPath,modelArtifacts:t,modelArtifactsInfo:i});f.onsuccess=()=>n({modelArtifactsInfo:i}),f.onerror=d=>{u=l.objectStore(ml);let h=u.delete(this.modelPath);h.onsuccess=()=>(a.close(),o(f.error)),h.onerror=g=>(a.close(),o(f.error))}},c.onerror=m=>(a.close(),o(c.error)),l.oncomplete=()=>{p==null?a.close():p.oncomplete=()=>a.close()}}},s.onerror=a=>o(s.error)})}};pa.URL_SCHEME="indexeddb://";var QS=r=>W().getBool("IS_BROWSER")&&!Array.isArray(r)&&r.startsWith(pa.URL_SCHEME)?wW(r.slice(pa.URL_SCHEME.length)):null;St.registerSaveRouter(QS);St.registerLoadRouter(QS);function wW(r){return new pa(r)}function _W(r){return r.startsWith(pa.URL_SCHEME)?r.slice(pa.URL_SCHEME.length):r}var Tw=class{constructor(){this.indexedDB=JS()}async listModels(){return new Promise((e,t)=>{let n=this.indexedDB.open(Iw,Nw);n.onupgradeneeded=()=>Sw(n),n.onsuccess=()=>{let o=n.result,s=o.transaction(ml,"readonly"),i=s.objectStore(ml).getAll();i.onsuccess=()=>{let l={};for(let u of i.result)l[u.modelPath]=u.modelArtifactsInfo;e(l)},i.onerror=l=>(o.close(),t(i.error)),s.oncomplete=()=>o.close()},n.onerror=o=>t(n.error)})}async removeModel(e){return e=_W(e),new Promise((t,n)=>{let o=this.indexedDB.open(Iw,Nw);o.onupgradeneeded=()=>Sw(o),o.onsuccess=()=>{let s=o.result,a=s.transaction(ml,"readwrite"),i=a.objectStore(ml),l=i.get(e),u;l.onsuccess=()=>{if(l.result==null)return s.close(),n(new Error(`Cannot find model with path '${e}' in IndexedDB.`));{let c=i.delete(e),p=()=>{u=s.transaction(Ql,"readwrite");let f=u.objectStore(Ql).delete(e);f.onsuccess=()=>t(l.result.modelArtifactsInfo),f.onerror=d=>n(l.error)};c.onsuccess=p,c.onerror=m=>(p(),s.close(),n(l.error))}},l.onerror=c=>(s.close(),n(l.error)),a.oncomplete=()=>{u==null?s.close():u.oncomplete=()=>s.close()}},o.onerror=s=>n(o.error)})}};var ma="/",ep="tensorflowjs_models",eT="info",kW="model_topology",vW="weight_specs",CW="weight_data",IW="model_metadata";function tT(r){return{info:[ep,r,eT].join(ma),topology:[ep,r,kW].join(ma),weightSpecs:[ep,r,vW].join(ma),weightData:[ep,r,CW].join(ma),modelMetadata:[ep,r,IW].join(ma)}}function NW(r){let e=r.split(ma);if(e.length<3)throw new Error(`Invalid key format: ${r}`);return e.slice(1,e.length-1).join(ma)}function SW(r){return r.startsWith(fa.URL_SCHEME)?r.slice(fa.URL_SCHEME.length):r}var fa=class{constructor(e){if(!W().getBool("IS_BROWSER")||typeof window=="undefined"||typeof window.localStorage=="undefined")throw new Error("The current environment does not support local storage.");if(this.LS=window.localStorage,e==null||!e)throw new Error("For local storage, modelPath must not be null, undefined or empty.");this.modelPath=e,this.keys=tT(this.modelPath)}async save(e){if(e.modelTopology instanceof ArrayBuffer)throw new Error("BrowserLocalStorage.save() does not support saving model topology in binary formats yet.");{let t=JSON.stringify(e.modelTopology),n=JSON.stringify(e.weightSpecs),o=ni(e);try{this.LS.setItem(this.keys.info,JSON.stringify(o)),this.LS.setItem(this.keys.topology,t),this.LS.setItem(this.keys.weightSpecs,n),this.LS.setItem(this.keys.weightData,HS(e.weightData));let s={format:e.format,generatedBy:e.generatedBy,convertedBy:e.convertedBy};return e.signature!=null&&(s.signature=e.signature),e.userDefinedMetadata!=null&&(s.userDefinedMetadata=e.userDefinedMetadata),e.modelInitializer!=null&&(s.modelInitializer=e.modelInitializer),this.LS.setItem(this.keys.modelMetadata,JSON.stringify(s)),{modelArtifactsInfo:o}}catch(s){throw this.LS.removeItem(this.keys.info),this.LS.removeItem(this.keys.topology),this.LS.removeItem(this.keys.weightSpecs),this.LS.removeItem(this.keys.weightData),this.LS.removeItem(this.keys.modelMetadata),new Error(`Failed to save model '${this.modelPath}' to local storage: size quota being exceeded is a possible cause of this failure: modelTopologyBytes=${o.modelTopologyBytes}, weightSpecsBytes=${o.weightSpecsBytes}, weightDataBytes=${o.weightDataBytes}.`)}}}async load(){let e=JSON.parse(this.LS.getItem(this.keys.info));if(e==null)throw new Error(`In local storage, there is no model with name '${this.modelPath}'`);if(e.modelTopologyType!=="JSON")throw new Error("BrowserLocalStorage does not support loading non-JSON model topology yet.");let t={},n=JSON.parse(this.LS.getItem(this.keys.topology));if(n==null)throw new Error(`In local storage, the topology of model '${this.modelPath}' is missing.`);t.modelTopology=n;let o=JSON.parse(this.LS.getItem(this.keys.weightSpecs));if(o==null)throw new Error(`In local storage, the weight specs of model '${this.modelPath}' are missing.`);t.weightSpecs=o;let s=this.LS.getItem(this.keys.modelMetadata);if(s!=null){let i=JSON.parse(s);t.format=i.format,t.generatedBy=i.generatedBy,t.convertedBy=i.convertedBy,i.signature!=null&&(t.signature=i.signature),i.userDefinedMetadata!=null&&(t.userDefinedMetadata=i.userDefinedMetadata),i.modelInitializer!=null&&(t.modelInitializer=i.modelInitializer)}let a=this.LS.getItem(this.keys.weightData);if(a==null)throw new Error(`In local storage, the binary weight values of model '${this.modelPath}' are missing.`);return t.weightData=qS(a),t}};fa.URL_SCHEME="localstorage://";var rT=r=>W().getBool("IS_BROWSER")&&!Array.isArray(r)&&r.startsWith(fa.URL_SCHEME)?TW(r.slice(fa.URL_SCHEME.length)):null;St.registerSaveRouter(rT);St.registerLoadRouter(rT);function TW(r){return new fa(r)}var Aw=class{constructor(){T(W().getBool("IS_BROWSER"),()=>"Current environment is not a web browser"),T(typeof window=="undefined"||typeof window.localStorage!="undefined",()=>"Current browser does not appear to support localStorage"),this.LS=window.localStorage}async listModels(){let e={},t=ep+ma,n=ma+eT;for(let o=0;o<this.LS.length;++o){let s=this.LS.key(o);if(s.startsWith(t)&&s.endsWith(n)){let a=NW(s);e[a]=JSON.parse(this.LS.getItem(s))}}return e}async removeModel(e){e=SW(e);let t=tT(e);if(this.LS.getItem(t.info)==null)throw new Error(`Cannot find model at path '${e}'`);let n=JSON.parse(this.LS.getItem(t.info));return this.LS.removeItem(t.info),this.LS.removeItem(t.topology),this.LS.removeItem(t.weightSpecs),this.LS.removeItem(t.weightData),n}};var tp="://",Zr=class{constructor(){this.managers={}}static getInstance(){return Zr.instance==null&&(Zr.instance=new Zr),Zr.instance}static registerManager(e,t){T(e!=null,()=>"scheme must not be undefined or null."),e.endsWith(tp)&&(e=e.slice(0,e.indexOf(tp))),T(e.length>0,()=>"scheme must not be an empty string.");let n=Zr.getInstance();T(n.managers[e]==null,()=>`A model store manager is already registered for scheme '${e}'.`),n.managers[e]=t}static getManager(e){let t=this.getInstance().managers[e];if(t==null)throw new Error(`Cannot find model manager for scheme '${e}'`);return t}static getSchemes(){return Object.keys(this.getInstance().managers)}};function sg(r){if(r.indexOf(tp)===-1)throw new Error(`The url string provided does not contain a scheme. Supported schemes are: ${Zr.getSchemes().join(",")}`);return{scheme:r.split(tp)[0],path:r.split(tp)[1]}}async function nT(r,e,t=!1){T(r!==e,()=>`Old path and new path are the same: '${r}'`);let n=St.getLoadHandlers(r);T(n.length>0,()=>`Copying failed because no load handler is found for source URL ${r}.`),T(n.length<2,()=>`Copying failed because more than one (${n.length}) load handlers for source URL ${r}.`);let o=n[0],s=St.getSaveHandlers(e);T(s.length>0,()=>`Copying failed because no save handler is found for destination URL ${e}.`),T(s.length<2,()=>`Copying failed because more than one (${n.length}) save handlers for destination URL ${e}.`);let a=s[0],i=sg(r).scheme,l=sg(r).path,u=i===sg(r).scheme,c=await o.load();t&&u&&await Zr.getManager(i).removeModel(l);let p=await a.save(c);return t&&!u&&await Zr.getManager(i).removeModel(l),p.modelArtifactsInfo}async function oT(){let r=Zr.getSchemes(),e={};for(let t of r){let n=await Zr.getManager(t).listModels();for(let o in n){let s=t+tp+o;e[s]=n[o]}}return e}async function sT(r){let e=sg(r);return Zr.getManager(e.scheme).removeModel(e.path)}async function iT(r,e){return nT(r,e,!1)}async function aT(r,e){return nT(r,e,!0)}var lT=class{fetch(e,t){return fetch(e,t)}now(){return performance.now()}encode(e,t){if(t!=="utf-8"&&t!=="utf8")throw new Error(`Browser's encoder only supports utf-8, but got ${t}`);return this.textEncoder==null&&(this.textEncoder=new TextEncoder),this.textEncoder.encode(e)}decode(e,t){return new TextDecoder(t).decode(e)}};if(W().get("IS_BROWSER")){W().setPlatform("browser",new lT);try{Zr.registerManager(fa.URL_SCHEME,new Aw)}catch(r){}try{Zr.registerManager(pa.URL_SCHEME,new Tw)}catch(r){}}var AW={importFetch:()=>uT()},Ew;var cT=class{constructor(){this.util=dc("util"),this.textEncoder=new this.util.TextEncoder}fetch(e,t){return W().global.fetch!=null?W().global.fetch(e,t):(Ew==null&&(Ew=AW.importFetch()),Ew(e,t))}now(){let e=process.hrtime();return e[0]*1e3+e[1]/1e6}encode(e,t){if(t!=="utf-8"&&t!=="utf8")throw new Error(`Node built-in encoder only supports utf-8, but got ${t}`);return this.textEncoder.encode(e)}decode(e,t){return e.length===0?"":new this.util.TextDecoder(t).decode(e)}};W().get("IS_NODE")&&W().setPlatform("node",new cT);function Ce(r,e="float32",t){return e=e||"float32",Tm(r),new ut(r,e,t)}function EW(r,e){let t=k(r,"x","cast");if(!ow(e))throw new Error(`Failed to cast to unknown dtype ${e}`);if(e==="string"&&t.dtype!=="string"||e!=="string"&&t.dtype==="string")throw new Error("Only strings can be casted to strings");let n={x:t},o={dtype:e};return A.runKernel(qn,n,o)}var oe=S({cast_:EW});function DW(r){let t={x:k(r,"x","clone","string_or_numeric")};return A.runKernel(Xn,t)}var Ln=S({clone_:DW});function Dw(r,e=!1){console.log(r.toString(e))}_w();var $W={buffer:Ce,cast:oe,clone:Ln,print:Dw};LS($W);var Er={};We(Er,{browserFiles:()=>fT,browserHTTPRequest:()=>gT,concatenateArrayBuffers:()=>Qc,copyModel:()=>iT,decodeWeights:()=>og,encodeWeights:()=>jS,fromMemory:()=>yT,getLoadHandlers:()=>ZS,getModelArtifactsInfoForJSON:()=>ni,getSaveHandlers:()=>YS,http:()=>lg,isHTTPScheme:()=>ag,listModels:()=>oT,loadWeights:()=>dT,moveModel:()=>aT,registerLoadRouter:()=>XS,registerSaveRouter:()=>KS,removeModel:()=>sT,weightsLoaderFactory:()=>Fw,withSaveHandler:()=>bT});var RW="model",FW=".json",OW=".weights.bin";function pT(r){return new Promise(e=>setTimeout(e)).then(r)}var fl=class{constructor(e){if(!W().getBool("IS_BROWSER"))throw new Error("browserDownloads() cannot proceed because the current environment is not a browser.");e.startsWith(fl.URL_SCHEME)&&(e=e.slice(fl.URL_SCHEME.length)),(e==null||e.length===0)&&(e=RW),this.modelTopologyFileName=e+FW,this.weightDataFileName=e+OW}async save(e){if(typeof document=="undefined")throw new Error("Browser downloads are not supported in this environment since `document` is not present");let t=window.URL.createObjectURL(new Blob([e.weightData],{type:"application/octet-stream"}));if(e.modelTopology instanceof ArrayBuffer)throw new Error("BrowserDownloads.save() does not support saving model topology in binary formats yet.");{let n=[{paths:["./"+this.weightDataFileName],weights:e.weightSpecs}],o={modelTopology:e.modelTopology,format:e.format,generatedBy:e.generatedBy,convertedBy:e.convertedBy,weightsManifest:n};e.signature!=null&&(o.signature=e.signature),e.userDefinedMetadata!=null&&(o.userDefinedMetadata=e.userDefinedMetadata),e.modelInitializer!=null&&(o.modelInitializer=e.modelInitializer);let s=window.URL.createObjectURL(new Blob([JSON.stringify(o)],{type:"application/json"})),a=this.jsonAnchor==null?document.createElement("a"):this.jsonAnchor;if(a.download=this.modelTopologyFileName,a.href=s,await pT(()=>a.dispatchEvent(new MouseEvent("click"))),e.weightData!=null){let i=this.weightDataAnchor==null?document.createElement("a"):this.weightDataAnchor;i.download=this.weightDataFileName,i.href=t,await pT(()=>i.dispatchEvent(new MouseEvent("click")))}return{modelArtifactsInfo:ni(e)}}}};fl.URL_SCHEME="downloads://";var mT=class{constructor(e){if(e==null||e.length<1)throw new Error(`When calling browserFiles, at least 1 file is required, but received ${e}`);this.files=e}async load(){let e=this.files[0],t=this.files.slice(1);return new Promise((n,o)=>{let s=new FileReader;s.onload=a=>{let i=JSON.parse(a.target.result),l=i.modelTopology;if(l==null){o(new Error(`modelTopology field is missing from file ${e.name}`));return}t.length===0&&n({modelTopology:l});let u=i.weightsManifest;if(u==null){o(new Error(`weightManifest field is missing from file ${e.name}`));return}let c;try{c=this.checkManifestAndWeightFiles(u,t)}catch(d){o(d);return}let p=[],m=[],f=[];u.forEach(d=>{d.paths.forEach(h=>{m.push(h),f.push(null)}),p.push(...d.weights)}),u.forEach(d=>{d.paths.forEach(h=>{let g=new FileReader;g.onload=x=>{let b=x.target.result,w=m.indexOf(h);if(f[w]=b,f.indexOf(null)===-1){let _={modelTopology:l,weightSpecs:p,weightData:Qc(f),format:i.format,generatedBy:i.generatedBy,convertedBy:i.convertedBy};i.signature!=null&&(_.signature=i.signature),i.userDefinedMetadata!=null&&(_.userDefinedMetadata=i.userDefinedMetadata),i.modelInitializer!=null&&(_.modelInitializer=i.modelInitializer),n(_)}},g.onerror=x=>o(`Failed to weights data from file of path '${h}'.`),g.readAsArrayBuffer(c[h])})})},s.onerror=a=>o(`Failed to read model topology and weights manifest JSON from file '${e.name}'. BrowserFiles supports loading Keras-style tf.Model artifacts only.`),s.readAsText(e)})}checkManifestAndWeightFiles(e,t){let n=[],o=t.map(a=>Cw(a.name)),s={};for(let a of e)a.paths.forEach(i=>{let l=Cw(i);if(n.indexOf(l)!==-1)throw new Error(`Duplicate file basename found in weights manifest: '${l}'`);if(n.push(l),o.indexOf(l)===-1)throw new Error(`Weight file with basename '${l}' is not provided.`);s[i]=t[o.indexOf(l)]});if(n.length!==t.length)throw new Error(`Mismatch in the number of files in weights manifest (${n.length}) and the number of weight files provided (${t.length}).`);return s}},PW=r=>W().getBool("IS_BROWSER")&&!Array.isArray(r)&&r.startsWith(fl.URL_SCHEME)?MW(r.slice(fl.URL_SCHEME.length)):null;St.registerSaveRouter(PW);function MW(r="model"){return new fl(r)}function fT(r){return new mT(r)}function $w(r,e,t,n){a(r),t=t==null?0:t,n=n==null?1:n,i(t,n);let o=0,s=l=>(l.then(u=>{let c=t+ ++o/r.length*(n-t);return e(c),u}),l);function a(l){T(l!=null&&Array.isArray(l)&&l.length>0,()=>"promises must be a none empty array")}function i(l,u){T(l>=0&&l<=1,()=>`Progress fraction must be in range [0, 1], but got startFraction ${l}`),T(u>=0&&u<=1,()=>`Progress fraction must be in range [0, 1], but got endFraction ${u}`),T(u>=l,()=>`startFraction must be no more than endFraction, but got startFraction ${l} and endFraction ${u}`)}return Promise.all(r.map(s))}async function Rw(r,e){e==null&&(e={});let t=e.fetchFunc==null?W().platform.fetch:e.fetchFunc,n=r.map(p=>t(p,e.requestInit,{isBinary:!0})),o=0,s=.5,i=(e.onProgress==null?await Promise.all(n):await $w(n,e.onProgress,o,s)).map(p=>p.arrayBuffer()),l=.5,u=1;return e.onProgress==null?await Promise.all(i):await $w(i,e.onProgress,l,u)}async function dT(r,e="",t,n){return Fw(a=>Rw(a,{requestInit:n}))(r,e,t)}function Fw(r){return async(e,t="",n)=>{let o=e.map(()=>!1),s={},a=n!=null?n.map(()=>!1):[],i=[];if(e.forEach((f,d)=>{let h=0;f.weights.forEach(g=>{let x="quantization"in g?g.quantization.dtype:g.dtype,b=Bm[x]*nt(g.shape),w=()=>{o[d]=!0,s[d]==null&&(s[d]=[]),s[d].push({manifestEntry:g,groupOffset:h,sizeBytes:b})};n!=null?n.forEach((_,I)=>{_===g.name&&(w(),a[I]=!0)}):w(),i.push(g.name),h+=b})}),!a.every(f=>f)){let f=n.filter((d,h)=>!a[h]);throw new Error(`Could not find weights in manifest with names: ${f.join(", ")}.
|
|
Manifest JSON has weights with names: ${i.join(", ")}.`)}let l=o.reduce((f,d,h)=>(d&&f.push(h),f),[]),u=[];l.forEach(f=>{e[f].paths.forEach(d=>{let h=t+(t.endsWith("/")?"":"/")+d;u.push(h)})});let c=await r(u),p={},m=0;return l.forEach(f=>{let d=e[f].paths.length,h=0;for(let _=0;_<d;_++)h+=c[m+_].byteLength;let g=new ArrayBuffer(h),x=new Uint8Array(g),b=0;for(let _=0;_<d;_++){let I=new Uint8Array(c[m+_]);x.set(I,b),b+=I.byteLength}s[f].forEach(_=>{let I=g.slice(_.groupOffset,_.groupOffset+_.sizeBytes),E=og(I,[_.manifestEntry]);for(let D in E)p[D]=E[D]}),m+=d}),p}}var LW="application/octet-stream",zW="application/json",ig=class{constructor(e,t){if(this.DEFAULT_METHOD="POST",t==null&&(t={}),this.weightPathPrefix=t.weightPathPrefix,this.onProgress=t.onProgress,this.weightUrlConverter=t.weightUrlConverter,t.fetchFunc!=null?(T(typeof t.fetchFunc=="function",()=>"Must pass a function that matches the signature of `fetch` (see https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API)"),this.fetch=t.fetchFunc):this.fetch=W().platform.fetch,T(e!=null&&e.length>0,()=>"URL path for http must not be null, undefined or empty."),Array.isArray(e)&&T(e.length===2,()=>`URL paths for http must have a length of 2, (actual length is ${e.length}).`),this.path=e,t.requestInit!=null&&t.requestInit.body!=null)throw new Error("requestInit is expected to have no pre-existing body, but has one.");this.requestInit=t.requestInit||{}}async save(e){if(e.modelTopology instanceof ArrayBuffer)throw new Error("BrowserHTTPRequest.save() does not support saving model topology in binary formats yet.");let t=Object.assign({method:this.DEFAULT_METHOD},this.requestInit);t.body=new FormData;let n=[{paths:["./model.weights.bin"],weights:e.weightSpecs}],o={modelTopology:e.modelTopology,format:e.format,generatedBy:e.generatedBy,convertedBy:e.convertedBy,weightsManifest:n};e.signature!=null&&(o.signature=e.signature),e.userDefinedMetadata!=null&&(o.userDefinedMetadata=e.userDefinedMetadata),e.modelInitializer!=null&&(o.modelInitializer=e.modelInitializer),t.body.append("model.json",new Blob([JSON.stringify(o)],{type:zW}),"model.json"),e.weightData!=null&&t.body.append("model.weights.bin",new Blob([e.weightData],{type:LW}),"model.weights.bin");let s=await this.fetch(this.path,t);if(s.ok)return{modelArtifactsInfo:ni(e),responses:[s]};throw new Error(`BrowserHTTPRequest.save() failed due to HTTP response status ${s.status}.`)}async load(){let e=await this.fetch(this.path,this.requestInit);if(!e.ok)throw new Error(`Request to ${this.path} failed with status code ${e.status}. Please verify this URL points to the model JSON of the model to load.`);let t;try{t=await e.json()}catch(d){let h=`Failed to parse model JSON of response from ${this.path}.`;throw this.path.endsWith(".pb")?h+=" Your path contains a .pb file extension. Support for .pb models have been removed in TensorFlow.js 1.0 in favor of .json models. You can re-convert your Python TensorFlow model using the TensorFlow.js 1.0 conversion scripts or you can convert your.pb models with the 'pb2json'NPM script in the tensorflow/tfjs-converter repository.":h+=" Please make sure the server is serving valid JSON for this request.",new Error(h)}let n=t.modelTopology,o=t.weightsManifest,s=t.generatedBy,a=t.convertedBy,i=t.format,l=t.signature,u=t.userDefinedMetadata;if(n==null&&o==null)throw new Error(`The JSON from HTTP path ${this.path} contains neither model topology or manifest for weights.`);let c,p;o!=null&&([c,p]=await this.loadWeights(o));let m={modelTopology:n,weightSpecs:c,weightData:p,generatedBy:s,convertedBy:a,format:i};l!=null&&(m.signature=l),u!=null&&(m.userDefinedMetadata=u);let f=t.modelInitializer;return f&&(m.modelInitializer=f),m}async loadWeights(e){let t=Array.isArray(this.path)?this.path[1]:this.path,[n,o]=BW(t),s=this.weightPathPrefix||n,a=[];for(let c of e)a.push(...c.weights);let i=[],l=[];for(let c of e)for(let p of c.paths)this.weightUrlConverter!=null?l.push(this.weightUrlConverter(p)):i.push(s+p+o);this.weightUrlConverter&&i.push(...await Promise.all(l));let u=await Rw(i,{requestInit:this.requestInit,fetchFunc:this.fetch,onProgress:this.onProgress});return[a,Qc(u)]}};ig.URL_SCHEME_REGEX=/^https?:\/\//;function BW(r){let e=r.lastIndexOf("/"),t=r.lastIndexOf("?"),n=r.substring(0,e),o=t>e?r.substring(t):"";return[n+"/",o]}function ag(r){return r.match(ig.URL_SCHEME_REGEX)!=null}var hT=(r,e)=>{if(typeof fetch=="undefined"&&(e==null||e.fetchFunc==null))return null;{let t=!0;if(Array.isArray(r)?t=r.every(n=>ag(n)):t=ag(r),t)return lg(r,e)}return null};St.registerSaveRouter(hT);St.registerLoadRouter(hT);function lg(r,e){return new ig(r,e)}function gT(r,e){return lg(r,e)}var ug=class{constructor(e){this.modelArtifacts=e}async load(){return this.modelArtifacts}},xT=class{constructor(e){this.saveHandler=e}async save(e){return this.saveHandler(e)}};function yT(r,e,t,n){return arguments.length===1?r.modelTopology!=null||r.weightSpecs!=null?new ug(r):(console.warn("Please call tf.io.fromMemory() with only one argument. The argument should be of type ModelArtifacts. The multi-argument signature of tf.io.fromMemory() has been deprecated and will be removed in a future release."),new ug({modelTopology:r})):(console.warn("Please call tf.io.fromMemory() with only one argument. The argument should be of type ModelArtifacts. The multi-argument signature of tf.io.fromMemory() has been deprecated and will be removed in a future release."),new ug({modelTopology:r,weightSpecs:e,weightData:t,trainingConfig:n}))}function bT(r){return new xT(r)}var _T={};We(_T,{confusionMatrix:()=>wT});function VW(r,e,t=!1,n=!1){let o=k(r,"a","matMul"),s=k(e,"b","matMul");[o,s]=je(o,s);let a={a:o,b:s},i={transposeA:t,transposeB:n};return A.runKernel(Co,a,i)}var ze=S({matMul_:VW});function GW(r,e,t=1,n=0){if(e<2)throw new Error(`Error in oneHot: depth must be >=2, but it is ${e}`);let s={indices:k(r,"indices","oneHot","int32")},a={depth:e,onValue:t,offValue:n};return A.runKernel(qo,s,a)}var ds=S({oneHot_:GW});function WW(r,e){let t=k(r,"x","transpose");if(e==null&&(e=t.shape.map((s,a)=>a).reverse()),T(t.rank===e.length,()=>`Error in transpose: rank of input ${t.rank} must match length of perm ${e}.`),e.forEach(s=>{T(s>=0&&s<t.rank,()=>`All entries in 'perm' must be between 0 and ${t.rank-1} but got ${e}`)}),t.rank<=1)return t.clone();let n={x:t},o={perm:e};return A.runKernel(ms,n,o)}var Ue=S({transpose_:WW});function jW(r,e,t){let n=k(r,"labels","confusionMatrix"),o=k(e,"predictions","confusionMatrix");T(t==null||t>0&&Number.isInteger(t),()=>`If provided, numClasses must be a positive integer, but got ${t}`),T(n.rank===1,()=>`Expected the rank of labels to be 1, but got ${n.rank}`),T(o.rank===1,()=>`Expected the rank of predictions to be 1, but got ${o.rank}`),T(n.shape[0]===o.shape[0],()=>`Mismatch in the number of examples: ${n.shape[0]} vs. ${o.shape[0]}. Labels and predictions should have the same number of elements.`),T(t>0&&Number.isInteger(t),()=>`numClasses is required to be a positive integer, but got ${t}`);let s=ds(oe(n,"int32"),t),a=ds(oe(o,"int32"),t),i=Ue(s),l=ze(i,a);return oe(l,"int32")}var wT=S({confusionMatrix_:jW});var cg={};We(cg,{fromPixels:()=>ZW,fromPixelsAsync:()=>XW,toPixels:()=>YW});function Ow(r,e,t){if(On(r),e!=null&&e.length!==3)throw new Error("tensor3d() requires shape to have three numbers");let n=Ar(r,t);if(n.length!==3&&n.length!==1)throw new Error("tensor3d() requires values to be number[][][] or flat/TypedArray");if(n.length===1&&e==null)throw new Error("tensor3d() requires shape to be provided when `values` are a flat array");return Yr(r,e,n,t)}var rp;function kT(r,e=3){if(e>4)throw new Error("Cannot construct Tensor with more than 4 channels from pixels.");if(r==null)throw new Error("pixels passed to tf.browser.fromPixels() can not be null");let t=!1,n=!1,o=!1,s=!1,a=!1,i=!1;if(r.data instanceof Uint8Array)t=!0;else if(typeof ImageData!="undefined"&&r instanceof ImageData)n=!0;else if(typeof HTMLVideoElement!="undefined"&&r instanceof HTMLVideoElement)o=!0;else if(typeof HTMLImageElement!="undefined"&&r instanceof HTMLImageElement)s=!0;else if(r.getContext!=null)a=!0;else if(typeof ImageBitmap!="undefined"&&r instanceof ImageBitmap)i=!0;else throw new Error(`pixels passed to tf.browser.fromPixels() must be either an HTMLVideoElement, HTMLImageElement, HTMLCanvasElement, ImageData in browser, or OffscreenCanvas, ImageData in webworker or {data: Uint32Array, width: number, height: number}, but was ${r.constructor.name}`);if(o){let d=2;if(o&&r.readyState<d)throw new Error("The video element has not loaded data yet. Please wait for `loadeddata` event on the <video> element.")}if(Om(Rm,A.backendName)!=null){let d={pixels:r},h={numChannels:e};return A.runKernel(Rm,d,h)}let[u,c]=o?[r.videoWidth,r.videoHeight]:[r.width,r.height],p;a?p=r.getContext("2d").getImageData(0,0,u,c).data:n||t?p=r.data:(s||o||i)&&(rp==null&&(rp=document.createElement("canvas").getContext("2d")),rp.canvas.width=u,rp.canvas.height=c,rp.drawImage(r,0,0,u,c),p=rp.getImageData(0,0,u,c).data);let m;if(e===4)m=new Int32Array(p);else{let d=u*c;m=new Int32Array(d*e);for(let h=0;h<d;h++)for(let g=0;g<e;++g)m[h*e+g]=p[h*4+g]}return Ow(m,[c,u,e],"int32")}function UW(r){return r!=null&&r.data instanceof Uint8Array}function HW(){return typeof window!="undefined"&&typeof ImageBitmap!="undefined"&&window.hasOwnProperty("createImageBitmap")}function qW(r){return r!=null&&r.width!==0&&r.height!==0}function KW(r){return HW()&&!(r instanceof ImageBitmap)&&qW(r)&&!UW(r)}async function XW(r,e=3){let t=null;if(W().getBool("WRAP_TO_IMAGEBITMAP")&&KW(r)){let n;try{n=await createImageBitmap(r,{premultiplyAlpha:"none"})}catch(o){n=null}n!=null&&n.width===r.width&&n.height===r.height?t=n:t=r}else t=r;return kT(t,e)}async function YW(r,e){let t=k(r,"img","toPixels");if(!(r instanceof Pe)){let u=t;t=oe(u,"int32"),u.dispose()}if(t.rank!==2&&t.rank!==3)throw new Error(`toPixels only supports rank 2 or 3 tensors, got rank ${t.rank}.`);let[n,o]=t.shape.slice(0,2),s=t.rank===2?1:t.shape[2];if(s>4||s===2)throw new Error(`toPixels only supports depth of size 1, 3 or 4 but got ${s}`);if(t.dtype!=="float32"&&t.dtype!=="int32")throw new Error(`Unsupported type for toPixels: ${t.dtype}. Please use float32 or int32 tensors.`);let a=await t.data(),i=t.dtype==="float32"?255:1,l=new Uint8ClampedArray(o*n*4);for(let u=0;u<n*o;++u){let c=[0,0,0,255];for(let m=0;m<s;m++){let f=a[u*s+m];if(t.dtype==="float32"){if(f<0||f>1)throw new Error(`Tensor values for a float32 Tensor must be in the range [0 - 1] but encountered ${f}.`)}else if(t.dtype==="int32"&&(f<0||f>255))throw new Error(`Tensor values for a int32 Tensor must be in the range [0 - 255] but encountered ${f}.`);s===1?(c[0]=f*i,c[1]=f*i,c[2]=f*i):c[m]=f*i}let p=u*4;l[p+0]=Math.round(c[0]),l[p+1]=Math.round(c[1]),l[p+2]=Math.round(c[2]),l[p+3]=Math.round(c[3])}if(e!=null){e.width=o,e.height=n;let u=e.getContext("2d"),c=new ImageData(l,o,n);u.putImageData(c,0,0)}return t!==r&&t.dispose(),l}var ZW=S({fromPixels_:kT});var pg={};We(pg,{prepareAndValidate:()=>vT});function vT(r,e){let t=r.shape.length,n=e.shape.length;if(t<1)throw new Error(`tf.gatherND() expects the input to be rank 1 or higher, but the rank was ${t}.`);if(n<1)throw new Error(`tf.gatherND() expects the indices to be rank 1 or higher, but the rank was ${n}.`);if(e.dtype!=="int32")throw new Error(`tf.gatherND() expects the indices to be int32 type, but the dtype was ${e.dtype}.`);if(e.shape[n-1]>t)throw new Error(`index innermost dimension length must be <= tensor rank; saw: ${e.shape[n-1]} vs. ${t}`);if(nt(r.shape)===0)throw new Error(`Requested more than 0 entries, but input is empty. Input shape: ${r.shape}.`);let o=e.shape,s=o[o.length-1],a=1;for(let p=0;p<o.length-1;++p)a*=o[p];let i=r.shape,l=o.slice();l.pop();let u=1;for(let p=s;p<t;++p)u*=i[p],l.push(i[p]);let c=[...zs(r.shape).map(p=>p/u),1].slice(0,s);return[l,a,u,c]}var fg={};We(fg,{calculateShapes:()=>CT,validateInput:()=>mg,validateUpdateShape:()=>Pw});function Pw(r,e,t){let n=e.rank>1?e.shape[e.rank-1]:1,o=e.rank>1?e.rank-1:1,s=`Must have updates.shape = indices.shape[:batchDim] + shape[sliceDim:], got updates.shape: ${t.shape}, indices.shape: ${e.shape}, shape: ${r}, sliceDim: ${n}, and batchDim: ${o}.`;if(t.rank<o)throw new Error(s+` update.rank < ${o}. `);if(r.length<n+(t.rank-o))throw new Error(s+` Output shape length < ${n+(t.rank-o)}`);if(t.rank!==o+r.length-n)throw new Error(s+` update.rank != ${o+r.length-n}`);for(let a=0;a<o;++a)if(t.shape[a]!==e.shape[a])throw new Error(s+` updates.shape[${a}] (${t.shape[a]}) != indices.shape[${a}] (${e.shape[a]}).`);for(let a=0;a<t.rank-o;++a)if(t.shape[a+o]!==r[a+n])throw new Error(s+` updates.shape[${a+o}] (${t.shape[a+o]}) != shape[${a+o}] (${r[a+o]})`)}function mg(r,e,t){if(e.rank<1)throw new Error(`tf.scatterND() expects the indices to be rank 1 or higher, but the rank was ${e.rank}.`);if(r.rank<1)throw new Error(`tf.scatterND() expects the updates to be rank 1 or higher, but the rank was ${r.rank}.`);if(e.dtype!=="int32")throw new Error(`The dtype of 'indices' should be int32, but got dtype: ${e.dtype}`);if(t.length<1)throw new Error(`Output rank must be greater or equal to 1, but got shape: ${t}`);if(t.length===0){if(e.size===0)throw new Error(`Indices specified for empty output. indices shape: ${e.shape}`);if(r.size===0)throw new Error(`Updates specified for empty output. updates shape: ${r.shape}`)}Pw(t,e,r)}function CT(r,e,t){let n=e.shape.length,o=n>1?e.shape[n-1]:1,s=t.length,a=1;for(let p=o;p<s;++p)a*=t[p];let i=o<1?1:o,l=nt(e.shape)/i,u=[...zs(t.slice(0,o)),1],c=nt(t);return{sliceRank:o,numUpdates:l,sliceSize:a,strides:u,outputSize:c}}var or={};We(or,{assertParamsValid:()=>JW,computeFlatOffset:()=>ej,computeOutShape:()=>IT,getNormalizedAxes:()=>AT,isSliceContinous:()=>QW,maskToAxes:()=>dg,parseSliceParams:()=>Mw,sliceInfo:()=>tj,startForAxis:()=>RT,startIndicesWithElidedDims:()=>ET,stopForAxis:()=>FT,stopIndicesWithElidedDims:()=>DT,stridesForAxis:()=>$T,stridesWithElidedDims:()=>NT});function JW(r,e,t){let n=r.shape.length;T(n===e.length,()=>`Error in slice${n}D: Length of begin ${e} must match the rank of the array (${n}).`),T(n===t.length,()=>`Error in slice${n}D: Length of size ${t} must match the rank of the array (${n}).`);for(let o=0;o<n;++o)T(e[o]+t[o]<=r.shape[o],()=>`Error in slice${n}D: begin[${o}] + size[${o}] (${e[o]+t[o]}) would overflow input.shape[${o}] (${r.shape[o]})`)}function dg(r){let e=[],t=0;for(;r>0;)r&1&&e.push(t),r/=2,t++;return e}function IT(r,e,t){let n=[];for(let o=0;o<r.length;o++)n[o]=Math.ceil((e[o]-r[o])/t[o]);return n}function NT(r,e,t,n){let o=[...r];for(let s=o.length;s<n.length;s++)o.push(1);for(let s=0;s<t;s++)s===0?o[e]=1:(o.splice(e,0,1),o.pop());return o}function ST(r,e,t){return t<=r?t:t-(e-1)}function TT(r,e){let t=[];for(let n=0;n<r;n++)t.push(e+n);return t}function AT(r,e,t,n,o,s,a,i,l){let u=r.length,c=new Array(u),p=new Array(u),m=new Array(u);if(e.length&&t>0){let f=e[0],d=t+1;c=ET(a,f,d,n,r),p=DT(i,f,d,o,r),m=NT(s,f,d,r)}else for(let f=0;f<u;f++)c[f]=RT(a,n,s,r,f,l),p[f]=FT(i,o,s,r,f,l),m[f]=$T(s,f,l);return{begin:c,end:p,strides:m}}function ET(r,e,t,n,o){let s=[...o],a=TT(t,e);for(let i=0;i<s.length;i++)if(a.indexOf(i)>-1)s[i]=0;else{let l=ST(e,t,i),u=n[l];r&1<<l&&(u=0),s[i]=u}return s}function DT(r,e,t,n,o){let s=[...o],a=TT(t,e);for(let i=0;i<s.length;i++)if(a.indexOf(i)>-1)s[i]=Number.MAX_SAFE_INTEGER;else{let l=ST(e,t,i),u=n[l];r&1<<l&&(u=Number.MAX_SAFE_INTEGER),s[i]=u}for(let i=0;i<s.length;i++){let l=o[i];s[i]<0&&(s[i]+=l),s[i]=gc(0,s[i],o[i])}return s}function $T(r,e,t){let n=r[e];return(t&1<<e||n==null)&&(n=1),n}function RT(r,e,t,n,o,s){let a=e[o],i=t[o]||1;(r&1<<o||s&1<<o||a==null)&&(i>0?a=Number.MIN_SAFE_INTEGER:a=Number.MAX_SAFE_INTEGER);let l=n[o];return a<0&&(a+=l),a=gc(0,a,l-1),a}function FT(r,e,t,n,o,s){let a=e[o],i=t[o]||1;(r&1<<o||s&1<<o||a==null)&&(i>0?a=Number.MAX_SAFE_INTEGER:a=Number.MIN_SAFE_INTEGER);let l=n[o];return a<0&&(a+=l),i>0?a=gc(0,a,l):a=gc(-1,a,l-1),a}function QW(r,e,t){let n=t.length;for(let o=0;o<t.length;o++)if(t[o]>1){n=o;break}for(let o=n+1;o<t.length;o++)if(e[o]>0||t[o]!==r[o])return!1;return!0}function ej(r,e){let t=r.length>0?r[r.length-1]:1;for(let n=0;n<r.length-1;n++)t+=r[n]*e[n];return t}function Mw(r,e,t){let n,o=r.shape.length;typeof e=="number"?n=[e,...new Array(o-1).fill(0)]:e.length<o?n=e.concat(new Array(o-e.length).fill(0)):n=e.slice(),n.forEach(a=>{T(a!==-1,()=>"slice() does not support negative begin indexing.")});let s;return t==null?s=new Array(o).fill(-1):typeof t=="number"?s=[t,...new Array(o-1).fill(-1)]:t.length<o?s=t.concat(new Array(o-t.length).fill(-1)):s=t,s=s.map((a,i)=>a>=0?a:(T(a===-1,()=>`Negative size values should be exactly -1 but got ${a} for the slice() size at index ${i}.`),r.shape[i]-n[i])),[n,s]}function tj(r,e,t,n,o,s,a,i,l){let u=e.slice(),c=t.slice(),p=n;n==null&&(p=new Array(u.length));let m=dg(a);if(m.length>1)throw new Error("Multiple ellipses in slice is not allowed.");if(a!==0&&i!==0)throw new Error("Using both ellipsisMask and newAxisMask is not yet supported.");if(a!==0&&l!==0)throw new Error("Using both ellipsisMask and shrinkAxisMask is not yet supported.");let f=r.length-u.length,d=dg(i),h=r.slice();d.forEach(D=>{u[D]=0,c[D]=1,h.splice(D,0,1)});let{begin:g,end:x,strides:b}=AT(h,m,f,u,c,p,o,s,a);u=g,c=x,p=b;let w=dg(l);w.forEach(D=>{c[D]=u[D]+1,p[D]=1});let _=IT(u,c,p),I=_.filter((D,$)=>w.indexOf($)===-1);return{nonStrided:p.every(D=>D===1),$begin:u,$end:c,$strides:p,size:_,newShape:h,outShape:I}}var Q={};We(Q,{Serializable:()=>hg,SerializationMap:()=>da,registerClass:()=>dn});var hg=class{getClassName(){return this.constructor.className}static fromConfig(e,t){return new e(t)}},da=class{constructor(){this.classNameMap={}}static getMap(){return da.instance==null&&(da.instance=new da),da.instance}static register(e){da.getMap().classNameMap[e.className]=[e,e.fromConfig]}};function dn(r){T(r.className!=null,()=>"Class being registered does not have the static className property defined."),T(typeof r.className=="string",()=>"className is required to be a string, but got type "+typeof r.className),T(r.className.length>0,()=>"Class being registered has an empty-string as its className, which is disallowed."),da.register(r)}var MT={};We(MT,{TEST_EPSILON_FLOAT16:()=>OT,encodeStrings:()=>PT,expectArrayBuffersEqual:()=>lj,expectArraysClose:()=>nj,expectArraysEqual:()=>sj,expectNumbersClose:()=>ij,expectPromiseToFail:()=>oj,expectValuesInRange:()=>aj,testEpsilon:()=>Lw});var rj=.001,OT=.1;function nj(r,e,t){return t==null&&(t=Lw()),zw(r,e,(n,o)=>Bw(n,o,t))}function Lw(){return A.backend.floatPrecision()===32?rj:OT}function zw(r,e,t){let n=!0;if((ar(r)||ar(e))&&(n=!1),ar(r)&&ar(e)&&(n=!0),n){let a=r.constructor.name,i=e.constructor.name;if(a!==i)throw new Error(`Arrays are of different type. Actual: ${a}. Expected: ${i}`)}if(Array.isArray(r)&&Array.isArray(e)){let a=Ar(r),i=Ar(e);if(!Xr(a,i))throw new Error(`Arrays have different shapes. Actual: [${a}]. Expected: [${i}]`)}let o=ar(r)?r:bo(r),s=ar(e)?e:bo(e);if(o.length!==s.length)throw new Error(`Arrays have different lengths actual: ${o.length} vs expected: ${s.length}.
|
|
Actual: ${o}.
|
|
Expected: ${s}.`);for(let a=0;a<s.length;++a){let i=o[a],l=s[a];if(!t(i,l))throw new Error(`Arrays differ: actual[${a}] = ${i}, expected[${a}] = ${l}.
|
|
Actual: ${o}.
|
|
Expected: ${s}.`)}}function oj(r,e){r().then(()=>e.fail(),()=>e())}function sj(r,e){let t=typeof e=="string"||typeof e=="number"||typeof e=="boolean"?[e]:e;return wo(r)||wo(r[0])||wo(e)||wo(e[0])?zw(r,t,(n,o)=>n==o):zw(r,e,(n,o)=>Bw(n,o,0))}function ij(r,e,t){if(t==null&&(t=Lw()),!Bw(r,e,t))throw new Error(`Numbers differ: actual === ${r}, expected === ${e}`)}function Bw(r,e,t){return!isFinite(r)&&!isFinite(e)?!0:!(isNaN(r)||isNaN(e)||Math.abs(r-e)>t)}function aj(r,e,t){for(let n=0;n<r.length;n++)if(r[n]<e||r[n]>t)throw new Error(`Value out of range:${r[n]} low: ${e}, high: ${t}`)}function lj(r,e){expect(new Float32Array(r)).toEqual(new Float32Array(e))}function PT(r){for(let e=0;e<r.length;e++){let t=r[e];Array.isArray(t)?PT(t):r[e]=cl(t)}return r}var uj="3.6.0";function Pae(){W().set("PROD",!0)}function Mae(){W().set("DEBUG",!0)}function Lae(){W().set("DEPRECATION_WARNINGS_ENABLED",!1),console.warn("TensorFlow.js deprecation warnings have been disabled.")}function Vw(r){W().getBool("DEPRECATION_WARNINGS_ENABLED")&&console.warn(r+" You can disable deprecation warnings with tf.disableDeprecationWarnings().")}zS(Vw);function zae(){A.disposeVariables()}function hs(){return A}function Vm(){return A.memory()}function Bae(r){return A.profile(r)}function V(r,e){return A.tidy(r,e)}function Ee(r){zm(r).forEach(t=>t.dispose())}function $t(r){return A.keep(r)}function Vae(r){return A.time(r)}function cj(r){return A.setBackend(r)}function Gae(){return A.ready()}function Wae(){return A.backendName}function jae(r){A.removeBackend(r)}function Uae(r){return A.findBackend(r)}function Hae(r){return A.findBackendFactory(r)}function np(r,e,t=1){return A.registerBackend(r,e,t)}function LT(){return A.backend}function qae(r,e){W().setPlatform(r,e)}function pj(r,e){let t=k(r,"a","add"),n=k(e,"b","add");[t,n]=je(t,n);let o={a:t,b:n};return A.runKernel(Pn,o)}var J=S({add_:pj});function mj(r,e){let t=k(r,"a","floorDiv"),n=k(e,"b","floorDiv");[t,n]=je(t,n);let o={a:t,b:n};return A.runKernel(Fo,o)}var eu=S({floorDiv_:mj});function fj(r,e){let t=k(r,"a","div"),n=k(e,"b","div");if([t,n]=je(t,n),t.dtype==="int32"&&n.dtype==="int32")return eu(t,n);let o={a:t,b:n},s={};return A.runKernel(Do,o,s)}var pe=S({div_:fj});function dj(r,e){let t=k(r,"a","mul"),n=k(e,"b","mul");[t,n]=je(t,n);let o={a:t,b:n};return A.runKernel(Ho,o)}var O=S({mul_:dj});function hj(r){let e=k(r,"x","abs");if(e.dtype==="complex64"){let t={x:e};return A.runKernel(Qa,t)}else{let t={x:e};return A.runKernel(Bs,t)}}var Tt=S({abs_:hj});function gj(r){let t={x:k(r,"x","acos")};return A.runKernel(ki,t)}var Gm=S({acos_:gj});function xj(r){let t={x:k(r,"x","acosh")};return A.runKernel(vi,t)}var Wm=S({acosh_:xj});function yj(r){T(Array.isArray(r),()=>"The argument passed to tf.addN() must be a list of tensors"),T(r.length>=1,()=>`Must pass at least one tensor to tf.addN(), but got ${r.length}`);let e=r.map((o,s)=>k(o,`tensors${s}`,"addN")),t=e[0];e.forEach(o=>{if(o.dtype!==t.dtype)throw new Error("All tensors passed to tf.addN() must have the same dtype")}),e.forEach(o=>{if(!Xr(o.shape,t.shape))throw new Error("All tensors passed to tf.addN() must have the same shape")});let n=e;return A.runKernel(_o,n)}var Gw=S({addN_:yj});function bj(r,e=null,t=!1){let o={x:k(r,"x","all","bool")},s={axis:e,keepDims:t};return A.runKernel(Ci,o,s)}var tu=S({all_:bj});function wj(r,e=null,t=!1){let o={x:k(r,"x","any","bool")},s={axis:e,keepDims:t};return A.runKernel(Ii,o,s)}var dl=S({any_:wj});function _j(r,e=0){let n={x:k(r,"x","argMax")},o={axis:e};return A.runKernel(ko,n,o)}var hl=S({argMax_:_j});function kj(r,e=0){let n={x:k(r,"x","argMin")},o={axis:e};return A.runKernel(Ya,n,o)}var jm=S({argMin_:kj});function vj(r){let t={x:k(r,"x","asin")};return A.runKernel(Ni,t)}var Um=S({asin_:vj});function Cj(r){let t={x:k(r,"x","asinh")};return A.runKernel(Si,t)}var Hm=S({asinh_:Cj});function Ij(r){let t={x:k(r,"x","atan")};return A.runKernel(Ti,t)}var qm=S({atan_:Ij});function Nj(r,e){let t=k(r,"a","atan2"),n=k(e,"b","atan2");[t,n]=je(t,n);let o={a:t,b:n};return A.runKernel(Ei,o)}var Km=S({atan2_:Nj});function Sj(r){let t={x:k(r,"x","atanh")};return A.runKernel(Ai,t)}var Xm=S({atanh_:Sj});function Tj(r,e,t,n,o="NHWC",s){let a=r[3],i=[...e,a],l=BT(o);return ru(r,i,t,s,n,null,null,l)}function Ww(r,e,t,n,o,s,a="channelsLast"){let[i,l]=gg(e),u;if(a==="channelsLast")u=[i,l,r[3],r[3]];else if(a==="channelsFirst")u=[i,l,r[1],r[1]];else throw new Error(`Unknown dataFormat ${a}`);return ru(r,u,t,n,o,s,!1,a)}function Aj(r,e,t,n,o,s,a="NDHWC"){let[i,l,u]=Uw(e),c,p;if(a==="NDHWC")p="channelsLast",c=[i,l,u,r[4],r[4]];else if(a==="NCDHW")p="channelsFirst",c=[i,l,u,r[1],r[1]];else throw new Error(`Unknown dataFormat ${a}`);return zT(r,c,t,n,o,!1,p,s)}function ru(r,e,t,n,o,s,a=!1,i="channelsLast"){let[l,u,c,p]=[-1,-1,-1,-1];if(i==="channelsLast")[l,u,c,p]=r;else if(i==="channelsFirst")[l,p,u,c]=r;else throw new Error(`Unknown dataFormat ${i}`);let[m,f,,d]=e,[h,g]=gg(t),[x,b]=gg(n),w=op(m,x),_=op(f,b),{padInfo:I,outHeight:E,outWidth:D}=$j(o,u,c,h,g,w,_,s,i),$=a?d*p:d,R;return i==="channelsFirst"?R=[l,$,E,D]:i==="channelsLast"&&(R=[l,E,D,$]),{batchSize:l,dataFormat:i,inHeight:u,inWidth:c,inChannels:p,outHeight:E,outWidth:D,outChannels:$,padInfo:I,strideHeight:h,strideWidth:g,filterHeight:m,filterWidth:f,effectiveFilterHeight:w,effectiveFilterWidth:_,dilationHeight:x,dilationWidth:b,inShape:r,outShape:R,filterShape:e}}function zT(r,e,t,n,o,s=!1,a="channelsLast",i){let[l,u,c,p,m]=[-1,-1,-1,-1,-1];if(a==="channelsLast")[l,u,c,p,m]=r;else if(a==="channelsFirst")[l,m,u,c,p]=r;else throw new Error(`Unknown dataFormat ${a}`);let[f,d,h,,g]=e,[x,b,w]=Uw(t),[_,I,E]=Uw(n),D=op(f,_),$=op(d,I),R=op(h,E),{padInfo:M,outDepth:G,outHeight:j,outWidth:U}=Rj(o,u,c,p,x,b,w,D,$,R,i),H=s?g*m:g,q;return a==="channelsFirst"?q=[l,H,G,j,U]:a==="channelsLast"&&(q=[l,G,j,U,H]),{batchSize:l,dataFormat:a,inDepth:u,inHeight:c,inWidth:p,inChannels:m,outDepth:G,outHeight:j,outWidth:U,outChannels:H,padInfo:M,strideDepth:x,strideHeight:b,strideWidth:w,filterDepth:f,filterHeight:d,filterWidth:h,effectiveFilterDepth:D,effectiveFilterHeight:$,effectiveFilterWidth:R,dilationDepth:_,dilationHeight:I,dilationWidth:E,inShape:r,outShape:q,filterShape:e}}function Ej(r,e,t,n,o){n==null&&(n=jw(r,e,t));let s=r[0],a=r[1],i=nu((s-e+2*n)/t+1,o),l=nu((a-e+2*n)/t+1,o);return[i,l]}function Dj(r,e,t,n,o,s){o==null&&(o=jw(r,e,n));let a=r[0],i=r[1],l=r[2],u=nu((a-e+2*o)/n+1,s),c=nu((i-e+2*o)/n+1,s),p=nu((l-e+2*o)/n+1,s);return[u,c,p,t]}function jw(r,e,t,n=1){let o=op(e,n);return Math.floor((r[0]*(t-1)-t+o)/2)}function gg(r){return typeof r=="number"?[r,r,r]:r.length===2?[r[0],r[1],1]:r}function Uw(r){return typeof r=="number"?[r,r,r]:r}function op(r,e){return e<=1?r:r+(r-1)*(e-1)}function $j(r,e,t,n,o,s,a,i,l){let u,c,p;if(typeof r=="number"){u={top:r,bottom:r,left:r,right:r,type:r===0?"VALID":"NUMBER"};let f=Ej([e,t],s,n,r,i);c=f[0],p=f[1]}else if(r==="same"){c=Math.ceil(e/n),p=Math.ceil(t/o);let m=Math.max(0,(c-1)*n+s-e),f=Math.max(0,(p-1)*o+a-t),d=Math.floor(m/2),h=m-d,g=Math.floor(f/2),x=f-g;u={top:d,bottom:h,left:g,right:x,type:"SAME"}}else if(r==="valid")u={top:0,bottom:0,left:0,right:0,type:"VALID"},c=Math.ceil((e-s+1)/n),p=Math.ceil((t-a+1)/o);else if(typeof r=="object"){let m=l==="channelsLast"?r[1][0]:r[2][0],f=l==="channelsLast"?r[1][1]:r[2][1],d=l==="channelsLast"?r[2][0]:r[3][0],h=l==="channelsLast"?r[2][1]:r[3][1];u={top:m,bottom:f,left:d,right:h,type:m===0&&f===0&&d===0&&h===0?"VALID":"EXPLICIT"},c=nu((e-s+m+f)/n+1,i),p=nu((t-a+d+h)/o+1,i)}else throw Error(`Unknown padding parameter: ${r}`);return{padInfo:u,outHeight:c,outWidth:p}}function Rj(r,e,t,n,o,s,a,i,l,u,c){let p,m,f,d;if(typeof r=="number"){p={top:r,bottom:r,left:r,right:r,front:r,back:r,type:r===0?"VALID":"NUMBER"};let g=Dj([e,t,n,1],i,1,o,r,c);m=g[0],f=g[1],d=g[2]}else if(r==="same"){m=Math.ceil(e/o),f=Math.ceil(t/s),d=Math.ceil(n/a);let h=(m-1)*o+i-e,g=(f-1)*s+l-t,x=(d-1)*a+u-n,b=Math.floor(h/2),w=h-b,_=Math.floor(g/2),I=g-_,E=Math.floor(x/2),D=x-E;p={top:_,bottom:I,left:E,right:D,front:b,back:w,type:"SAME"}}else if(r==="valid")p={top:0,bottom:0,left:0,right:0,front:0,back:0,type:"VALID"},m=Math.ceil((e-i+1)/o),f=Math.ceil((t-l+1)/s),d=Math.ceil((n-u+1)/a);else throw Error(`Unknown padding parameter: ${r}`);return{padInfo:p,outDepth:m,outHeight:f,outWidth:d}}function nu(r,e){if(!e)return Math.trunc(r);switch(e){case"round":return Math.round(r);case"ceil":return Math.ceil(r);case"floor":return Math.floor(r);default:throw new Error(`Unknown roundingMode ${e}`)}}function zn(r){let[e,t,n]=gg(r);return e===1&&t===1&&n===1}function Cr(r,e){return zn(r)||zn(e)}function BT(r){if(r==="NHWC")return"channelsLast";if(r==="NCHW")return"channelsFirst";throw new Error(`Unknown dataFormat ${r}`)}function Fj(r,e){let n={x:k(r,"x","reshape","string_or_numeric")},o={shape:e};return A.runKernel(qs,n,o)}var L=S({reshape_:Fj});function Oj(r,e,t,n,o){let s=k(r,"x","avgPool","float32"),a=1;T(Cr(t,a),()=>`Error in avgPool: Either strides or dilations must be 1. Got strides ${t} and dilations '${a}'`);let i=s,l=!1;s.rank===3&&(l=!0,i=L(s,[1,s.shape[0],s.shape[1],s.shape[2]])),T(i.rank===4,()=>`Error in avgPool: x must be rank 4 but got rank ${i.rank}.`),o!=null&&T(ot(n),()=>`Error in avgPool: pad must be an integer when using, dimRoundingMode ${o} but got pad ${n}.`);let u={x:i},c={filterSize:e,strides:t,pad:n,dimRoundingMode:o},p=A.runKernel(vo,u,c);return p=oe(p,s.dtype),l?L(p,[p.shape[1],p.shape[2],p.shape[3]]):p}var ha=S({avgPool_:Oj});function Pj(r,e,t,n,o,s="NDHWC"){let a=k(r,"x","avgPool3d","float32"),i=a,l=!1;a.rank===4&&(l=!0,i=L(a,[1,a.shape[0],a.shape[1],a.shape[2],a.shape[3]])),T(i.rank===5,()=>`Error in avgPool3d: x must be rank 5 but got rank ${i.rank}.`),T(s==="NDHWC",()=>`Error in avgPool3d: Only NDHWC is currently supported, but got dataFormat of ${s}`),o!=null&&T(ot(n),()=>`Error in avgPool3d: pad must be an integer when using, dimRoundingMode ${o} but got pad ${n}.`);let u={x:i},c={filterSize:e,strides:t,pad:n,dimRoundingMode:o,dataFormat:s},p=A.runKernel(Za,u,c);return p=oe(p,i.dtype),l?L(p,[p.shape[1],p.shape[2],p.shape[3],p.shape[4]]):p}var Ym=S({avgPool3d_:Pj});function Mj(r,e=0){T(r.length>=1,()=>"Pass at least one tensor to concat");let t=ca(r,"tensors","concat","string_or_numeric");if(t[0].dtype==="complex64"&&t.forEach(s=>{if(s.dtype!=="complex64")throw new Error(`Cannot concatenate complex64 tensors with a tensor
|
|
with dtype ${s.dtype}. `)}),t.length===1)return Ln(t[0]);let n=t,o={axis:e};return A.runKernel(Vs,n,o)}var Qe=S({concat_:Mj});function Lj(r){let t={x:k(r,"x","sigmoid")};return A.runKernel(os,t)}var Dr=S({sigmoid_:Lj});function zj(r,e,t){let n=k(r,"x","slice","string_or_numeric");if(n.rank===0)throw new Error("Slicing scalar is not possible");let o={x:n},s={begin:e,size:t};return A.runKernel(Xs,o,s)}var Re=S({slice_:zj});function Bj(r){let t={x:k(r,"x","tanh")};return A.runKernel(ps,t)}var gs=S({tanh_:Bj});function Vj(r,e,t,n,o,s){let a=k(r,"forgetBias","basicLSTMCell"),i=k(e,"lstmKernel","basicLSTMCell"),l=k(t,"lstmBias","basicLSTMCell"),u=k(n,"data","basicLSTMCell"),c=k(o,"c","basicLSTMCell"),p=k(s,"h","basicLSTMCell"),m=Qe([u,p],1),f=ze(m,i),d=J(f,l),h=d.shape[0],g=d.shape[1]/4,x=[h,g],b=Re(d,[0,0],x),w=Re(d,[0,g],x),_=Re(d,[0,g*2],x),I=Re(d,[0,g*3],x),E=J(O(Dr(b),gs(w)),O(c,Dr(J(a,_)))),D=O(gs(E),Dr(I));return[E,D]}var Gj=S({basicLSTMCell_:Vj});function Wj(r,e,t){let n=k(r,"x","batchToSpaceND"),o=e.reduce((i,l)=>i*l);T(n.rank>=1+e.length,()=>`input rank is ${n.rank} but should be > than blockShape.length ${e.length}`),T(t.length===e.length,()=>`crops.length is ${t.length} but should be equal to blockShape.length ${e.length}`),T(n.shape[0]%o==0,()=>`input tensor batch is ${n.shape[0]} but is not divisible by the product of the elements of blockShape ${e.join(" * ")} === ${o}`);let s={x:n},a={blockShape:e,crops:t};return A.runKernel(Ja,s,a)}var ga=S({batchToSpaceND_:Wj});function VT(r){let e;return r.rank===0||r.rank===1?e=L(r,[1,1,1,r.size]):r.rank===2?e=L(r,[1,1,r.shape[0],r.shape[1]]):r.rank===3?e=L(r,[1,r.shape[0],r.shape[1],r.shape[2]]):e=r,e}function jj(r,e,t,n,o,s){s==null&&(s=.001);let a=k(r,"x","batchNorm"),i=k(e,"mean","batchNorm"),l=k(t,"variance","batchNorm"),u;o!=null&&(u=k(o,"scale","batchNorm"));let c;n!=null&&(c=k(n,"offset","batchNorm")),T(i.rank===l.rank,()=>"Batch normalization gradient requires mean and variance to have equal ranks."),T(c==null||i.rank===c.rank,()=>"Batch normalization gradient requires mean and offset to have equal ranks."),T(u==null||i.rank===u.rank,()=>"Batch normalization gradient requires mean and scale to have equal ranks.");let m={x:VT(a),scale:u,offset:c,mean:i,variance:l},f={varianceEpsilon:s},d=A.runKernel(Oo,m,f);return L(d,a.shape)}var Jn=S({batchNorm_:jj});function Uj(r,e,t,n,o,s){let a=k(r,"x","batchNorm"),i=k(e,"mean","batchNorm"),l=k(t,"variance","batchNorm"),u;o!=null&&(u=k(o,"scale","batchNorm"));let c;return n!=null&&(c=k(n,"offset","batchNorm")),T(a.rank===2,()=>`Error in batchNorm2D: x must be rank 2 but got rank ${a.rank}.`),T(i.rank===2||i.rank===1,()=>`Error in batchNorm2D: mean must be rank 2 or rank 1 but got rank ${i.rank}.`),T(l.rank===2||l.rank===1,()=>`Error in batchNorm2D: variance must be rank 2 or rank 1 but got rank ${l.rank}.`),u!=null&&T(u.rank===2||u.rank===1,()=>`Error in batchNorm2D: scale must be rank 2 or rank 1 but got rank ${u.rank}.`),c!=null&&T(c.rank===2||c.rank===1,()=>`Error in batchNorm2D: offset must be rank 2 or rank 1 but got rank ${c.rank}.`),Jn(a,i,l,c,u,s)}var Hw=S({batchNorm2d_:Uj});function Hj(r,e,t,n,o,s){let a=k(r,"x","batchNorm"),i=k(e,"mean","batchNorm"),l=k(t,"variance","batchNorm"),u;o!=null&&(u=k(o,"scale","batchNorm"));let c;return n!=null&&(c=k(n,"offset","batchNorm")),T(a.rank===3,()=>`Error in batchNorm3D: x must be rank 3 but got rank ${a.rank}.`),T(i.rank===3||i.rank===1,()=>`Error in batchNorm3D: mean must be rank 3 or rank 1 but got rank ${i.rank}.`),T(l.rank===3||l.rank===1,()=>`Error in batchNorm3D: variance must be rank 3 or rank 1 but got rank ${l.rank}.`),u!=null&&T(u.rank===3||u.rank===1,()=>`Error in batchNorm3D: scale must be rank 3 or rank 1 but got rank ${u.rank}.`),c!=null&&T(c.rank===3||c.rank===1,()=>`Error in batchNorm3D: offset must be rank 3 or rank 1 but got rank ${c.rank}.`),Jn(a,i,l,c,u,s)}var qw=S({batchNorm3d_:Hj});function qj(r,e,t,n,o,s){let a=k(r,"x","batchNorm"),i=k(e,"mean","batchNorm"),l=k(t,"variance","batchNorm"),u;o!=null&&(u=k(o,"scale","batchNorm"));let c;return n!=null&&(c=k(n,"offset","batchNorm")),T(a.rank===4,()=>`Error in batchNorm4D: x must be rank 4 but got rank ${a.rank}.`),T(i.rank===4||i.rank===1,()=>`Error in batchNorm4D: mean must be rank 4 or rank 1 but got rank ${i.rank}.`),T(l.rank===4||l.rank===1,()=>`Error in batchNorm4D: variance must be rank 4 or rank 1 but got rank ${l.rank}.`),u!=null&&T(u.rank===4||u.rank===1,()=>`Error in batchNorm4D: scale must be rank 4 or rank 1 but got rank ${u.rank}.`),c!=null&&T(c.rank===4||c.rank===1,()=>`Error in batchNorm4D: offset must be rank 4 or rank 1 but got rank ${c.rank}.`),Jn(a,i,l,c,u,s)}var Kw=S({batchNorm4d_:qj});function Kj(r,e,t){let n=k(r,"x","bincount"),o=k(e,"weights","bincount");T(n.dtype==="int32",()=>`Error in bincount: input dtype must be int32, but got ${n.dtype}`),T(t>=0,()=>`size must be non-negative, but got ${t}.`),T(o.size===n.size||o.size===0,()=>`Error in bincount: weights must have the same size as input or0-length, but got input shape: ${n.shape}, weights shape: ${o.shape}.`);let s={x:n,weights:o},a={size:t};return A.runKernel(kc,s,a)}var Zm=S({bincount_:Kj});function Xj(r,e){let t=k(r,"broadcastTo","x"),n=t.shape;if(e.some(u=>!(u>0)||u%1!=0))throw new Error(`broadcastTo(): Invalid broadcast shape [${e}].`);if(e.length<t.rank)throw new Error(`broadcastTo(): shape.length=${e.length} < input.rank=${t.rank}.`);if(e.length>t.rank){let u=t.shape.slice();for(;u.length<e.length;)u.unshift(1);t=L(t,u)}let o=t.shape,s=Array.from(e);for(let u=e.length-1;u>=0;u--)if(o[u]===e[u])s[u]=1;else if(t.shape[u]!==1)throw new Error(`broadcastTo(): [${n}] cannot be broadcast to [${e}].`);if(s.map((u,c)=>u>1?c:-1).filter(u=>u>=0).length===0)return Ln(t);let i={x:t},l={reps:s};return A.runKernel(Mn,i,l)}var xa=S({broadcastTo_:Xj});function Yj(r){let t={x:k(r,"x","ceil")};return A.runKernel(Io,t)}var Jm=S({ceil_:Yj});function Zj(r,e,t){let n=k(r,"x","clipByValue");T(e<=t,()=>`Error in clip: min (${e}) must be less than or equal to max (${t}).`);let o={x:n},s={clipValueMin:e,clipValueMax:t};return A.runKernel(Kn,o,s)}var ur=S({clipByValue_:Zj});function Jj(r){return Qe(r,0)}var Xw=S({concat1d_:Jj});function Qj(r,e){return Qe(r,e)}var Yw=S({concat2d_:Qj});function e4(r,e){return Qe(r,e)}var Zw=S({concat3d_:e4});function t4(r,e){return Qe(r,e)}var Jw=S({concat4d_:t4});function r4(r,e,t,n,o="NHWC",s=[1,1],a){let i=k(r,"x","conv2d"),l=k(e,"filter","conv2d"),u=i,c=!1;i.rank===3&&(c=!0,u=L(i,[1,i.shape[0],i.shape[1],i.shape[2]])),T(u.rank===4,()=>`Error in conv2d: input must be rank 4, but got rank ${u.rank}.`),T(l.rank===4,()=>`Error in conv2d: filter must be rank 4, but got rank ${l.rank}.`),a!=null&&T(ot(n),()=>`Error in conv2d: pad must be an integer when using, dimRoundingMode ${a} but got pad ${n}.`);let p=o==="NHWC"?u.shape[3]:u.shape[1];T(p===l.shape[2],()=>`Error in conv2d: depth of input (${p}) must match input depth for filter ${l.shape[2]}.`),T(Cr(t,s),()=>`Error in conv2D: Either strides or dilations must be 1. Got strides ${t} and dilations '${s}'`);let m={x:u,filter:l},f={strides:t,pad:n,dataFormat:o,dilations:s,dimRoundingMode:a},d=A.runKernel(No,m,f);return c?L(d,[d.shape[1],d.shape[2],d.shape[3]]):d}var Jr=S({conv2d_:r4});function n4(r,e,t,n,o="NWC",s=1,a){let i=k(r,"x","conv1d"),l=k(e,"filter","conv1d"),u=i,c=!1;i.rank===2&&(c=!0,u=L(i,[1,i.shape[0],i.shape[1]])),T(u.rank===3,()=>`Error in conv1d: input must be rank 3, but got rank ${u.rank}.`),T(l.rank===3,()=>`Error in conv1d: filter must be rank 3, but got rank ${l.rank}.`),a!=null&&T(ot(n),()=>`Error in conv1d: pad must be an integer when using, dimRoundingMode ${a} but got pad ${n}.`),T(u.shape[2]===l.shape[1],()=>`Error in conv1d: depth of input (${u.shape[2]}) must match input depth for filter ${l.shape[1]}.`),T(Cr(t,s),()=>`Error in conv1D: Either stride or dilation must be 1. Got stride ${t} and dilation '${s}'`),T(o==="NWC",()=>`Error in conv1d: got dataFormat of ${o} but only NWC is currently supported.`);let p=L(l,[1,l.shape[0],l.shape[1],l.shape[2]]),m=L(u,[u.shape[0],1,u.shape[1],u.shape[2]]),g=Jr(m,p,[1,t],n,"NHWC",[1,s],a);return c?L(g,[g.shape[2],g.shape[3]]):L(g,[g.shape[0],g.shape[2],g.shape[3]])}var ou=S({conv1d_:n4});function o4(r,e,t,n,o,s="NHWC",a){T(r.length===e.rank,()=>`Length of inShape (${r.length}) and rank of dy (${e.rank}) must match`);let i=r,l=e,u=!1;e.rank===3&&(u=!0,l=L(e,[1,e.shape[0],e.shape[1],e.shape[2]]),i=[1,r[0],r[1],r[2]]),T(i.length===4,()=>`Error in conv2dDerInput: inShape must be length 4, but got length ${i.length}.`),T(l.rank===4,()=>`Error in conv2dDerInput: dy must be rank 4, but got rank ${l.rank}`),T(t.rank===4,()=>`Error in conv2dDerInput: filter must be rank 4, but got rank ${t.rank}`);let c=s==="NHWC"?i[3]:i[1],p=s==="NHWC"?l.shape[3]:l.shape[1];T(c===t.shape[2],()=>`Error in conv2dDerInput: depth of input (${c}) must match input depth for filter ${t.shape[2]}.`),T(p===t.shape[3],()=>`Error in conv2dDerInput: depth of output (${p}) must match output depth for filter ${t.shape[3]}.`),a!=null&&T(ot(o),()=>`Error in conv2dDerInput: pad must be an integer when using, dimRoundingMode ${a} but got pad ${o}.`);let m={dy:l,filter:t},f={strides:n,pad:o,dataFormat:s,dimRoundingMode:a,inputShape:i},d=A.runKernel(So,m,f);return u?L(d,[d.shape[1],d.shape[2],d.shape[3]]):d}var sp=S({conv2DBackpropInput_:o4});function s4(r,e,t,n,o,s){let a=k(r,"x","conv2dTranspose"),i=k(e,"filter","conv2dTranspose");return sp(t,a,i,n,o,"NHWC",s)}var su=S({conv2dTranspose_:s4});function i4(r,e,t,n,o="NDHWC",s=[1,1,1]){let a=k(r,"x","conv3d"),i=k(e,"filter","conv3d"),l=a,u=!1;a.rank===4&&(u=!0,l=L(a,[1,a.shape[0],a.shape[1],a.shape[2],a.shape[3]])),T(l.rank===5,()=>`Error in conv3d: input must be rank 5, but got rank ${l.rank}.`),T(i.rank===5,()=>`Error in conv3d: filter must be rank 5, but got rank ${i.rank}.`),T(l.shape[4]===i.shape[3],()=>`Error in conv3d: depth of input (${l.shape[4]}) must match input depth for filter ${i.shape[3]}.`),T(Cr(t,s),()=>`Error in conv3D: Either strides or dilations must be 1. Got strides ${t} and dilations '${s}'`),T(o==="NDHWC",()=>`Error in conv3d: got dataFormat of ${o} but only NDHWC is currently supported.`);let c={x:l,filter:i},p={strides:t,pad:n,dataFormat:o,dilations:s},m=A.runKernel(el,c,p);return u?L(m,[m.shape[1],m.shape[2],m.shape[3],m.shape[4]]):m}var Qm=S({conv3d_:i4});function a4(r,e,t,n,o){T(r.length===e.rank,()=>`Length of inShape (${r.length}) and rank of dy (${e.rank}) must match`);let s=r,a=e,i=!1;e.rank===4&&(i=!0,a=L(e,[1,e.shape[0],e.shape[1],e.shape[2],e.shape[3]]),s=[1,r[0],r[1],r[2],r[3]]);let l=s[4],u=a.shape[4];T(s.length===5,()=>`Error in conv3dDerInput: inShape must be length 5, but got length ${s.length}.`),T(a.rank===5,()=>`Error in conv3dDerInput: dy must be rank 5, but got rank ${a.rank}`),T(t.rank===5,()=>`Error in conv3dDerInput: filter must be rank 5, but got rank ${t.rank}`),T(l===t.shape[3],()=>`Error in conv3dDerInput: depth of input (${l}) must match input depth for filter ${t.shape[3]}.`),T(u===t.shape[4],()=>`Error in conv3dDerInput: depth of output (${u}) must match output depth for filter ${t.shape[4]}.`);let c={dy:a,filter:t},p={pad:o,strides:n,inputShape:s},m=A.runKernel(Nc,c,p);return i?L(m,[m.shape[1],m.shape[2],m.shape[3],m.shape[4]]):m}var xg=S({conv3DBackpropInput_:a4});function l4(r,e,t,n,o){let s=k(r,"x","conv3dTranspose"),a=k(e,"filter","conv3dTranspose");return xg(t,s,a,n,o)}var Qw=S({conv3dTranspose_:l4});function u4(r){let t={x:k(r,"x","cos")};return A.runKernel(To,t)}var ya=S({cos_:u4});function c4(r){let t={x:k(r,"x","cosh")};return A.runKernel(Di,t)}var iu=S({cosh_:c4});function p4(r,e=0,t=!1,n=!1){let s={x:k(r,"x","cumsum")},a={axis:e,exclusive:t,reverse:n};return A.runKernel(Ao,s,a)}var au=S({cumsum_:p4});function m4(r,e,t,n=!1){let o=k(r,"x","denseBincount"),s=k(e,"weights","denseBincount");T(o.dtype==="int32",()=>`Error in denseBincount: input dtype must be int32, but got ${o.dtype}`),T(o.rank<=2,()=>`Error in denseBincount: input must be at most rank 2, but got rank ${o.rank}.`),T(t>=0,()=>`size must be non-negative, but got ${t}.`),T(s.size===o.size||s.size===0,()=>`Error in denseBincount: weights must have the same shape as x or 0-length, but got x shape: ${o.shape}, weights shape: ${s.shape}.`);let a={x:o,weights:s},i={size:t,binaryOutput:n};return A.runKernel(Sc,a,i)}var e_=S({denseBincount_:m4});function f4(r,e,t="NHWC"){let n=k(r,"x","depthToSpace"),o=t==="NHWC"?n.shape[1]:n.shape[2],s=t==="NHWC"?n.shape[2]:n.shape[3],a=t==="NHWC"?n.shape[3]:n.shape[1];T(o*e>=0,()=>`Negative dimension size caused by overflow when multiplying
|
|
${o} and ${e} for depthToSpace with input shape
|
|
${n.shape}`),T(s*e>=0,()=>`Negative dimension size caused by overflow when multiplying
|
|
${s} and ${e} for depthToSpace with input shape
|
|
${n.shape}`),T(a%(e*e)==0,()=>`Dimension size must be evenly divisible by ${e*e} but is ${a} for depthToSpace with input shape ${n.shape}`);let i={x:n},l={blockSize:e,dataFormat:t};return A.runKernel(Ri,i,l)}var ef=S({depthToSpace_:f4});function d4(r,e,t,n,o="NHWC",s=[1,1],a){let i=k(r,"x","depthwiseConv2d"),l=k(e,"filter","depthwiseConv2d"),u=i,c=!1;i.rank===3&&(c=!0,u=L(i,[1,i.shape[0],i.shape[1],i.shape[2]])),T(u.rank===4,()=>`Error in depthwiseConv2d: input must be rank 4, but got rank ${u.rank}.`),T(l.rank===4,()=>`Error in depthwiseConv2d: filter must be rank 4, but got rank ${l.rank}.`),T(u.shape[3]===l.shape[2],()=>`Error in depthwiseConv2d: number of input channels (${u.shape[3]}) must match the inChannels dimension in filter ${l.shape[2]}.`),a!=null&&T(ot(n),()=>`Error in depthwiseConv2d: pad must be an integer when using, dimRoundingMode ${a} but got pad ${n}.`);let p={x:u,filter:l},m={strides:t,pad:n,dataFormat:o,dilations:s,dimRoundingMode:a},f=A.runKernel(Eo,p,m);return c?L(f,[f.shape[1],f.shape[2],f.shape[3]]):f}var xs=S({depthwiseConv2d_:d4});function h4(r){let t={x:k(r,"x","diag")};return A.runKernel(Ec,t)}var g4=S({diag_:h4});function x4(r,e,t,n,o=[1,1],s="NHWC"){let a=k(r,"x","dilation2d"),i=k(e,"filter","dilation2d");T(a.rank===3||a.rank===4,()=>`Error in dilation2d: input must be rank 3 or 4, but got rank ${a.rank}.`),T(i.rank===3,()=>`Error in dilation2d: filter must be rank 3, but got rank ${i.rank}.`),T(s==="NHWC",()=>`Error in dilation2d: Only NHWC is currently supported, but got dataFormat of ${s}`);let l=a,u=!1;a.rank===3&&(l=L(a,[1,a.shape[0],a.shape[1],a.shape[2]]),u=!0);let c={x:l,filter:i},p={strides:t,pad:n,dilations:o},m=A.runKernel(tl,c,p);return u?L(m,[m.shape[1],m.shape[2],m.shape[3]]):m}var tf=S({dilation2d_:x4});function y4(r,e){let t=r.length,n=[];for(let o=0;o<t;o++){let s=t-1-o,a=r[s]||1;(e[e.length-1-o]||1)>1&&a===1&&n.unshift(s)}return n}function kt(r,e){let t=[];for(let n=0;n<e.length;n++){let o=r[r.length-n-1],s=e.length-n-1,a=e[s];(o==null||o===1&&a>1)&&t.unshift(s)}return t}function Be(r,e){let t=[],n=Math.max(r.length,e.length);for(let o=0;o<n;o++){let s=r[r.length-o-1];s==null&&(s=1);let a=e[e.length-o-1];if(a==null&&(a=1),s===1)t.unshift(a);else if(a===1)t.unshift(s);else if(s!==a){let i=`Operands could not be broadcast together with shapes ${r} and ${e}.`;throw Error(i)}else t.unshift(s)}return t}function b4(r,e){let t=k(r,"a","equal"),n=k(e,"b","equal");[t,n]=je(t,n),Be(t.shape,n.shape);let o={a:t,b:n};return A.runKernel(Pi,o)}var Sn=S({equal_:b4});function w4(r,e,t){let n=k(e,"a","where"),o=k(t,"b","where"),s=k(r,"condition","where","bool"),a=Be(Be(s.shape,n.shape),o.shape),i=xa(s,a),l=xa(n,a),u=xa(o,a),c={condition:i,t:l,e:u};return A.runKernel(Ks,c)}var vt=S({where_:w4});function _4(r){let t={x:k(r,"x","zerosLike")};return A.runKernel(Js,t)}var Ie=S({zerosLike_:_4});function k4(r,e){let t=k(r,"a","div"),n=k(e,"b","div");[t,n]=je(t,n);let o=pe(t,n),s=Ie(o),a=Sn(n,s);return vt(a,s,o)}var rf=S({divNoNan_:k4});function v4(r,e){let t=k(r,"t1","dot"),n=k(e,"t2","dot");T((t.rank===1||t.rank===2)&&(n.rank===1||n.rank===2),()=>`Error in dot: inputs must all be rank 1 or 2, but got ranks ${t.rank} and ${n.rank}.`);let o=t.rank===1?t.size:t.shape[1],s=n.rank===1?n.size:n.shape[0];if(T(o===s,()=>`Error in dot: inner dimensions of inputs must match, but got ${o} and ${s}.`),t.rank===1&&n.rank===1){let a=L(t,[1,-1]),i=L(n,[-1,1]),l=ze(a,i);return L(l,[])}else if(t.rank===1&&n.rank===2){let a=L(t,[1,-1]),i=L(n,[n.shape[0],n.shape[1]]),l=ze(a,i);return L(l,[l.size])}else if(t.rank===2&&n.rank===1){let a=L(n,[-1,1]),i=ze(t,a);return L(i,[i.size])}else{let a=L(n,[n.shape[0],n.shape[1]]);return ze(t,a)}}var t_=S({dot_:v4});function C4(r,...e){let t=e.map((o,s)=>k(o,`tensors${s}`,"einsum")),n={equation:r};return A.runKernel(Dc,t,n)}var r_=S({einsum_:C4});function I4(r){let t={x:k(r,"x","elu")};return A.runKernel(Fi,t)}var ys=S({elu_:I4});function N4(r){let e=k(r,"x","erf");T(e.dtype==="int32"||e.dtype==="float32",()=>"Input dtype must be `int32` or `float32`."),e.dtype==="int32"&&(e=oe(e,"float32"));let t={x:e};return A.runKernel(Oi,t)}var nf=S({erf_:N4});function S4(r){let t={x:k(r,"x","exp")};return A.runKernel($o,t)}var Qt=S({exp_:S4});function T4(r,e=0){let t=k(r,"x","expandDims","string_or_numeric");T(e<=t.rank,()=>"Axis must be <= rank of the tensor");let n={input:t},o={dim:e};return A.runKernel(Gs,n,o)}var cr=S({expandDims_:T4});function A4(r){let t={x:k(r,"x","expm1")};return A.runKernel(Mi,t)}var of=S({expm1_:A4});function E4(r,e){let t=k(r,"x","tile","string_or_numeric");T(t.rank===e.length,()=>`Error in transpose: rank of input ${t.rank} must match length of reps ${e}.`);let n={x:t},o={reps:e};return A.runKernel(Mn,n,o)}var Bn=S({tile_:E4});function D4(r,e,t,n="float32"){e==null&&(e=r);let o=Ce([r,e],n),s=r<=e?r:e;for(let i=0;i<s;++i)o.set(1,i,i);let a=L(o.toTensor(),[r,e]);if(t==null)return a;if(t.length===1)return Bn(cr(a,0),[t[0],1,1]);if(t.length===2)return Bn(cr(cr(a,0),0),[t[0],t[1],1,1]);if(t.length===3)return Bn(cr(cr(cr(a,0),0),0),[t[0],t[1],t[2],1,1]);throw new Error(`eye() currently supports only 1D and 2D batchShapes, but received ${t.length}D.`)}var ip=S({eye_:D4});function bs(r,e,t){let n={shape:r,value:e,dtype:t};return A.runKernel(rl,{},n)}function $4(r){let t={x:k(r,"x","floor")};return A.runKernel(Ro,t)}var ws=S({floor_:$4});function R4(r,e,t=0,n=0){let o=k(r,"x","gather"),s=k(e,"indices","gather","int32"),a={x:o,indices:s},i={axis:t,batchDims:n};return A.runKernel(Ws,a,i)}var Qn=S({gather_:R4});function F4(r,e){let t=k(r,"a","greater"),n=k(e,"b","greater");[t,n]=je(t,n),Be(t.shape,n.shape);let o={a:t,b:n};return A.runKernel(Bi,o)}var Vt=S({greater_:F4});function O4(r,e){let t=k(r,"a","greaterEqual"),n=k(e,"b","greaterEqual");[t,n]=je(t,n),Be(t.shape,n.shape);let o={a:t,b:n};return A.runKernel(Po,o)}var hn=S({greaterEqual_:O4});function P4(r){let t={input:k(r,"input","imag")};return A.runKernel(Oc,t)}var lu=S({imag_:P4});function M4(r){let t={x:k(r,"x","isFinite")};return A.runKernel(Vi,t)}var n_=S({isFinite_:M4});function L4(r){let t={x:k(r,"x","isInf")};return A.runKernel(Gi,t)}var o_=S({isInf_:L4});function z4(r){let t={x:k(r,"x","isNaN")};return A.runKernel(Wi,t)}var sf=S({isNaN_:z4});function B4(r,e=.2){let n={x:k(r,"x","leakyRelu")},o={alpha:e};return A.runKernel(Mo,n,o)}var ba=S({leakyRelu_:B4});function V4(r,e){let t=k(r,"a","less"),n=k(e,"b","less");[t,n]=je(t,n),Be(t.shape,n.shape);let o={a:t,b:n};return A.runKernel(ji,o)}var uu=S({less_:V4});function G4(r,e){let t=k(r,"a","lessEqual"),n=k(e,"b","lessEqual");[t,n]=je(t,n),Be(t.shape,n.shape);let o={a:t,b:n};return A.runKernel(Ui,o)}var gn=S({lessEqual_:G4});function s_(r,e,t){if(t<=0)throw new Error("The number of values should be positive.");let n={start:r,stop:e,num:t};return A.runKernel(Pc,{},n)}function W4(r,e=5,t=1,n=1,o=.5){let s=k(r,"x","localResponseNormalization");T(s.rank===4||s.rank===3,()=>`Error in localResponseNormalization: x must be rank 3 or 4 but got
|
|
rank ${s.rank}.`),T(ot(e),()=>`Error in localResponseNormalization: depthRadius must be an integer but got depthRadius ${e}.`);let a=s,i=!1;s.rank===3&&(i=!0,a=L(s,[1,s.shape[0],s.shape[1],s.shape[2]]));let l={x:a},u={depthRadius:e,bias:t,alpha:n,beta:o},c=A.runKernel(nl,l,u);return i?L(c,[c.shape[1],c.shape[2],c.shape[3]]):c}var af=S({localResponseNormalization_:W4});function j4(r){let t={x:k(r,"x","log")};return A.runKernel(Lo,t)}var pr=S({log_:j4});function U4(r){let t={x:k(r,"x","log1p")};return A.runKernel(Hi,t)}var cu=S({log1p_:U4});function H4(r){return T(Ls(r),()=>"The f passed in grad(f) must be a function"),(e,t)=>{let n=k(e,"x","tf.grad","string_or_numeric"),o=t!=null?k(t,"dy","tf.grad"):null;return A.tidy(()=>{let{value:s,grads:a}=A.gradients(()=>r(n),[n],o);return o!=null&&Dt(s.shape,o.shape,"The shape of dy passed in grad(f)(x, dy) must match the shape returned by f(x)"),bg(a),a[0]})}}function q4(r){return T(Ls(r),()=>"The f passed in grads(f) must be a function"),(e,t)=>{T(Array.isArray(e),()=>"The args passed in grads(f)(args) must be an array of `Tensor`s or `TensorLike`s");let n=ca(e,"args","tf.grads","string_or_numeric"),o=t!=null?k(t,"dy","tf.grads"):null;return A.tidy(()=>{let{value:s,grads:a}=A.gradients(()=>r(...n),n,o);return o!=null&&Dt(s.shape,o.shape,"The shape of dy passed in grads(f)([x1,...], dy) must match the shape returned by f([x1,...])"),bg(a),a})}}function K4(r){return T(Ls(r),()=>"The f passed in valueAndGrad(f) must be a function"),(e,t)=>{T(e instanceof Pe,()=>"The x passed in valueAndGrad(f)(x) must be a tensor"),T(t==null||t instanceof Pe,()=>"The dy passed in valueAndGrad(f)(x, dy) must be a tensor");let{grads:n,value:o}=A.gradients(()=>r(e),[e],t);return bg(n),{grad:n[0],value:o}}}function X4(r){return T(Ls(r),()=>"The f passed in valueAndGrads(f) must be a function"),(e,t)=>{T(Array.isArray(e)&&e.every(o=>o instanceof Pe),()=>"The args passed in valueAndGrads(f)(args) must be array of tensors"),T(t==null||t instanceof Pe,()=>"The dy passed in valueAndGrads(f)(args, dy) must be a tensor");let n=A.gradients(()=>r(...e),e,t);return t!=null&&Dt(n.value.shape,t.shape,"The shape of dy passed in valueAndGrads(f)([x1,...], dy) must match the shape returned by f([x1,...])"),bg(n.grads),n}}function yg(r,e){T(Ls(r),()=>"The f passed in variableGrads(f) must be a function"),T(e==null||Array.isArray(e)&&e.every(u=>u instanceof pl),()=>"The varList passed in variableGrads(f, varList) must be an array of variables");let t=e!=null;if(!t){e=[];for(let u in A.registeredVariables)e.push(A.registeredVariables[u])}let n=t?e.filter(u=>!u.trainable):null,o=e.length;e=e.filter(u=>u.trainable),T(e.length>0,()=>`variableGrads() expects at least one of the input variables to be trainable, but none of the ${o} variables is trainable.`);let s=!0,{value:a,grads:i}=A.gradients(r,e,null,s);T(i.some(u=>u!=null),()=>"Cannot find a connection between any variable and the result of the loss function y=f(x). Please make sure the operations that use variables are inside the function f passed to minimize()."),T(a.rank===0,()=>`The f passed in variableGrads(f) must return a scalar, but it returned a rank-${a.rank} tensor`);let l={};return e.forEach((u,c)=>{i[c]!=null&&(l[u.name]=i[c])}),n!=null&&n.forEach(u=>l[u.name]=null),{value:a,grads:l}}function Qr(r){return A.customGrad(r)}function bg(r){if(r.filter(t=>t==null).length>0)throw new Error(`Cannot compute gradient of y=f(x) with respect to x. Make sure that
|
|
the f you passed encloses all operations that lead from x to y.`)}function Y4(r){let t={x:k(r,"x","neg")};return A.runKernel(js,t)}var qe=S({neg_:Y4});function Z4(r){let t={x:k(r,"x","softplus")};return A.runKernel(sa,t)}var eo=S({softplus_:Z4});function J4(r){let e=k(r,"x","logSigmoid");return Qr(n=>({value:qe(eo(qe(n))),gradFunc:a=>O(a,Dr(qe(n)))}))(e)}var i_=S({logSigmoid_:J4});function Q4(r,e=null,t=!1){let o={x:k(r,"x","max")},s={reductionIndices:e,keepDims:t};return A.runKernel(zo,o,s)}var mr=S({max_:Q4});function eU(r,e){let t=k(r,"a","sub"),n=k(e,"b","sub");[t,n]=je(t,n);let o={a:t,b:n};return A.runKernel(us,o)}var le=S({sub_:eU});function tU(r,e=null,t=!1){let n=k(r,"x","sum");n.dtype==="bool"&&(n=oe(n,"int32"));let o={x:n},s={axis:e,keepDims:t};return A.runKernel(is,o,s)}var fe=S({sum_:tU});function rU(r,e=-1){let t=k(r,"logits","logSoftmax");if(e===-1&&(e=t.rank-1),e!==t.rank-1)throw Error(`Log Softmax along a non-last dimension is not yet supported. Logits was rank ${t.rank} and axis was ${e}`);return Qr((o,s)=>{let a=!0,i=mr(o,e,!0),l=le(o,i),u=le(oe(l,"float32"),pr(fe(Qt(l),e,a)));return s([u]),{value:u,gradFunc:(p,m)=>{let[f]=m,d=!0,h=Qt(f);return le(p,O(fe(p,e,d),h))}}})(t)}var pu=S({logSoftmax_:rU});function a_(r,e){for(let t=0;t<r.length;++t)if(r[r.length-t-1]!==e-1-t)return!1;return!0}function GT(r,e,t){let n=r.length+e.length,o=[],s=0,a=0;for(let i=0;i<n;i++)t.indexOf(i)===-1?o.push(r[s++]):o.push(e[a++]);return o}function l_(r,e){let t=[],n=r.length;for(let s=0;s<n;s++)e.indexOf(s)===-1&&t.push(r[s]);let o=e.map(s=>r[s]);return[t,o]}function to(r,e){let t=e.map(n=>1);return GT(r,t,e)}function nU(r,e,t){T(a_(e,t),()=>`${r} supports only inner-most axes for now. Got axes ${e} and rank-${t} input.`)}function u_(r,e){if(a_(r,e))return null;let t=[];for(let n=0;n<e;++n)r.indexOf(n)===-1&&t.push(n);return r.forEach(n=>t.push(n)),t}function lf(r){return r.map((e,t)=>[t,e]).sort((e,t)=>e[1]-t[1]).map(e=>e[0])}function oU(r,e){let t=[];for(let n=e-r;n<e;++n)t.push(n);return t}function sU(r,e=null,t=!1){let n=k(r,"x","logSumExp"),o=nr(e,n.shape),s=mr(n,o,!0),a=le(n,s),i=Qt(a),l=fe(i,o),u=pr(l),c=J(L(s,u.shape),u);if(t){let p=to(c.shape,o);return L(c,p)}return c}var uf=S({logSumExp_:sU});function iU(r,e){let t=k(r,"a","logicalAnd","bool"),n=k(e,"b","logicalAnd","bool");Be(t.shape,n.shape);let o={a:t,b:n};return A.runKernel(qi,o)}var yr=S({logicalAnd_:iU});function aU(r){let t={x:k(r,"x","logicalNot","bool")};return A.runKernel(Hl,t)}var wa=S({logicalNot_:aU});function lU(r,e){let t=k(r,"a","logicalOr","bool"),n=k(e,"b","logicalOr","bool");Be(t.shape,n.shape);let o={a:t,b:n};return A.runKernel(ql,o)}var mu=S({logicalOr_:lU});function uU(r,e){let t=k(r,"a","logicalXor","bool"),n=k(e,"b","logicalXor","bool");return Be(t.shape,n.shape),yr(mu(r,e),wa(yr(r,e)))}var c_=S({logicalXor_:uU});function cU(r,e,t,n,o){let s=k(r,"x","maxPool"),a=1,i=s,l=!1;s.rank===3&&(l=!0,i=L(s,[1,s.shape[0],s.shape[1],s.shape[2]])),T(i.rank===4,()=>`Error in maxPool: input must be rank 4 but got rank ${i.rank}.`),T(Cr(t,a),()=>`Error in maxPool: Either strides or dilations must be 1. Got strides ${t} and dilations '${a}'`),o!=null&&T(ot(n),()=>`Error in maxPool: pad must be an integer when using, dimRoundingMode ${o} but got pad ${n}.`);let u={x:i},c={filterSize:e,strides:t,pad:n,dimRoundingMode:o},p=A.runKernel(Vo,u,c);return l?L(p,[p.shape[1],p.shape[2],p.shape[3]]):p}var _a=S({maxPool_:cU});function pU(r,e=[1,1,1],t,n,o,s="NDHWC"){let a=k(r,"x","maxPool3d"),i=a,l=!1;a.rank===4&&(l=!0,i=L(a,[1,a.shape[0],a.shape[1],a.shape[2],a.shape[3]])),T(i.rank===5,()=>`Error in maxPool3d: x must be rank 5 but got rank ${i.rank}.`),T(s==="NDHWC",()=>`Error in maxPool3d: Only NDHWC is currently supported, but got dataFormat of ${s}`),o!=null&&T(ot(n),()=>`Error in maxPool3d: pad must be an integer when using, dimRoundingMode ${o} but got pad ${n}.`);let u={x:i},c={filterSize:e,strides:t,pad:n,dimRoundingMode:o,dataFormat:s},p=A.runKernel(ol,u,c);return l?L(p,[p.shape[1],p.shape[2],p.shape[3],p.shape[4]]):p}var cf=S({maxPool3d_:pU});function mU(r,e,t,n,o=!1){let a={x:k(r,"x","maxPoolWithArgmax")},i={filterSize:e,strides:t,pad:n,includeBatchInIndex:o},l=A.runKernel(Bc,a,i);return{result:l[0],indexes:l[1]}}var p_=S({maxPoolWithArgmax_:mU});function fU(r,e){let t=k(r,"a","maximum"),n=k(e,"b","maximum");[t,n]=je(t,n),t.dtype==="bool"&&(t=oe(t,"int32"),n=oe(n,"int32")),Be(t.shape,n.shape);let o={a:t,b:n};return A.runKernel(Bo,o)}var en=S({maximum_:fU});function dU(r,e=null,t=!1){let o={x:k(r,"x","mean")},s={axis:e,keepDims:t};return A.runKernel(Go,o,s)}var dt=S({mean_:dU});function ht(r,e="float32"){if(e==="complex64"){let n=ht(r,"float32"),o=ht(r,"float32");return Nn(n,o)}let t=bc(nt(r),e);return A.makeTensor(t,r,e)}function er(r,e="float32"){if(e==="complex64"){let n=er(r,"float32"),o=ht(r,"float32");return Nn(n,o)}let t=Sm(nt(r),e);return A.makeTensor(t,r,e)}function hU(r,e,{indexing:t="xy"}={}){if(t!=="xy"&&t!=="ij")throw new TypeError(`${t} is not a valid third argument to meshgrid`);if(r===void 0)return[];let n=k(r,"x","meshgrid",r instanceof Pe?r.dtype:"float32");if(e===void 0)return[n];let o=k(e,"y","meshgrid",e instanceof Pe?e.dtype:"float32"),s=nt(n.shape),a=nt(o.shape);return t==="xy"?(n=L(n,[1,-1]),o=L(o,[-1,1]),[ze(er([a,1],n.dtype),n),ze(o,er([1,s],o.dtype))]):(n=L(n,[-1,1]),o=L(o,[1,-1]),[ze(n,er([1,a],n.dtype)),ze(er([s,1],o.dtype),o)])}function gU(r,e=null,t=!1){let o={x:k(r,"x","min")},s={axis:e,keepDims:t};return A.runKernel(Wo,o,s)}var oi=S({min_:gU});function xU(r,e){let t=k(r,"a","minimum"),n=k(e,"b","minimum");[t,n]=je(t,n),t.dtype==="bool"&&(t=oe(t,"int32"),n=oe(n,"int32")),Be(t.shape,n.shape);let o={a:t,b:n};return A.runKernel(jo,o)}var _s=S({minimum_:xU});function yU(r,e,t){T(t==="reflect"||t==="symmetric",()=>`Invalid mode. Mode must be either reflect or symmetric. Got ${t}.`);let n=k(r,"x","mirrorPad");if(n.rank===0)throw new Error("mirrorPad(scalar) is not defined. Pass non-scalar to mirrorPad");T(e.length===n.rank,()=>`Padding doesn't match input. Must be ${n.rank}. Got ${e.length}.`);let o=t==="reflect"?1:0;for(let i=0;i<n.rank;i++)T(e[i].length===2,()=>"Invalid number of paddings. Must be length of 2 each."),T(e[i][0]>=0&&e[i][0]<=n.shape[i]-o&&e[i][1]>=0&&e[i][1]<=n.shape[i]-o,()=>`Padding in dimension ${i} cannot be greater than or equal to ${n.shape[i]-o} or less than 0 for input of shape ${n.shape}`);let s={paddings:e,mode:t},a={x:n};return A.runKernel(Uo,a,s)}var pf=S({mirrorPad_:yU});function bU(r,e){let t=k(r,"a","mod"),n=k(e,"b","mod");[t,n]=je(t,n);let o={a:t,b:n};return A.runKernel(Ki,o)}var mf=S({mod_:bU});function wU(r){let e=k(r,"x","square"),t={};return A.runKernel("Square",{x:e},t)}var Me=S({square_:wU});function _U(r,e=null,t=!1){r=k(r,"x","moments");let n=nr(e,r.shape),o=dt(r,n,t),s=o.shape;t||(s=to(o.shape,n));let a=Me(le(oe(r,"float32"),L(o,s))),i=dt(a,n,t);return{mean:o,variance:i}}var ap=S({moments_:_U});function kU(r,e,t,n){let o=k(e,"data","multiRNNCell"),s=ca(t,"c","multiRNNCell"),a=ca(n,"h","multiRNNCell"),i=o,l=[];for(let p=0;p<r.length;p++){let m=r[p](i,s[p],a[p]);l.push(m[0]),l.push(m[1]),i=m[1]}let u=[],c=[];for(let p=0;p<l.length;p+=2)u.push(l[p]),c.push(l[p+1]);return[u,c]}var vU=S({multiRNNCell_:kU});function CU(r,e,t,n=!1){let o=k(r,"logits","multinomial"),s=o.size,a=o.rank;if(s<2)throw new Error(`Error in multinomial: you need at least 2 outcomes, but got ${s}.`);if(a>2)throw new Error(`Rank of probabilities must be 1 or 2, but is ${a}`);t=t||Math.random();let l={logits:a===1?L(o,[1,-1]):o},u={numSamples:e,seed:t,normalized:n},c=A.runKernel(Vc,l,u);return a===1?L(c,[c.size]):c}var m_=S({multinomial_:CU});function IU(r,e){let t=k(r,"a","notEqual"),n=k(e,"b","notEqual");[t,n]=je(t,n),Be(t.shape,n.shape);let o={a:t,b:n};return A.runKernel(Xi,o)}var ro=S({notEqual_:IU});function NU(r){let t={x:k(r,"x","onesLike")};return A.runKernel(Us,t)}var sr=S({onesLike_:NU});function SU(r,e){let t=k(r,"v1","outerProduct"),n=k(e,"v2","outerProduct");T(t.rank===1&&n.rank===1,()=>`Error in outerProduct: inputs must be rank 1, but got ranks ${t.rank} and ${n.rank}.`);let o=L(t,[-1,1]),s=L(n,[1,-1]);return ze(o,s)}var TU=S({outerProduct_:SU});function AU(r,e,t=0){let n=k(r,"x","pad");if(n.rank===0)throw new Error("pad(scalar) is not defined. Pass non-scalar to pad");let o={paddings:e,constantValue:t},s={x:n};return A.runKernel(Ko,s,o)}var Lr=S({pad_:AU});function EU(r,e,t=0){return T(e.length===2,()=>"Invalid number of paddings. Must be length of 2."),Lr(r,[e],t)}var DU=S({pad1d_:EU});function $U(r,e,t=0){return T(e.length===2&&e[0].length===2&&e[1].length===2,()=>"Invalid number of paddings. Must be length of 2 each."),Lr(r,e,t)}var RU=S({pad2d_:$U});function FU(r,e,t=0){return T(e.length===3&&e[0].length===2&&e[1].length===2&&e[2].length===2,()=>"Invalid number of paddings. Must be length of 2 each."),Lr(r,e,t)}var OU=S({pad3d_:FU});function PU(r,e,t=0){return T(e.length===4&&e[0].length===2&&e[1].length===2&&e[2].length===2&&e[3].length===2,()=>"Invalid number of paddings. Must be length of 2 each."),Lr(r,e,t)}var MU=S({pad4d_:PU});function LU(r,e,t){let n=k(r,"x","spaceToBatchND");T(n.rank>=1+e.length,()=>`input rank ${n.rank} should be > than [blockShape] ${e.length}`),T(t.length===e.length,()=>`paddings.shape[0] ${t.length} must be equal to [blockShape] ${e.length}`),T(n.shape.reduce((a,i,l)=>l>0&&l<=e.length?a&&(i+t[l-1][0]+t[l-1][1])%e[l-1]==0:a,!0),()=>`input spatial dimensions ${n.shape.slice(1)} with paddings ${t.toString()} must be divisible by blockShapes ${e.toString()}`);let o={x:n},s={blockShape:e,paddings:t};return A.runKernel(al,o,s)}var ka=S({spaceToBatchND_:LU});function zU(r,e,t,n,o,s){o==null&&(o=[1,1]),s==null&&(s=1),n===0&&(n="valid");let a=k(r,"x","maxPool"),i=a,l=!1;a.rank===3&&(l=!0,i=L(a,[1,a.shape[0],a.shape[1],a.shape[2]])),T(Cr(s,o),()=>`Error in pool: Either strides or dilations must be 1. Got strides ${s} and dilations '${o}'`);let u=Ww(i.shape,e,s,o,n),c=[u.dilationHeight,u.dilationWidth],p;n==="same"?p=VU([u.filterHeight,u.filterWidth],c):p=[[0,0],[0,0]];let m=c[0]===1&&c[1]===1,[f,d]=BU([u.inHeight,u.inWidth],c,p),h=m?n:"valid",g=m?i:ka(i,c,f),b=(t==="avg"?()=>ha(g,e,s,h):()=>_a(g,e,s,h))(),w=m?b:ga(b,c,d);return l?L(w,[w.shape[1],w.shape[2],w.shape[3]]):w}function BU(r,e,t){let n=t.map(c=>c[0]),o=t.map(c=>c[1]),s=r.concat(n,o),a=e.map((c,p)=>(c-s[p]%c)%c),i=o.map((c,p)=>c+a[p]),l=e.map((c,p)=>[n[p],i[p]]),u=e.map((c,p)=>[0,a[p]]);return[l,u]}function VU(r,e){let n=r.map((a,i)=>a+(a-1)*(e[i]-1)).map(a=>a-1),o=n.map(a=>Math.floor(a/2)),s=n.map((a,i)=>a-o[i]);return n.map((a,i)=>[o[i],s[i]])}var f_=S({pool_:zU});function GU(r,e){let t=k(r,"base","pow"),n=k(e,"exp","pow");[t,n]=je(t,n);let o={a:t,b:n};return A.runKernel(Xo,o)}var zr=S({pow_:GU});function WU(r,e){let t=k(r,"x","prelu"),n=k(e,"alpha","prelu"),o={x:t,alpha:n};return A.runKernel(Yo,o)}var va=S({prelu_:WU});function jU(r,e=null,t=!1){let n=k(r,"x","prod");n.dtype==="bool"&&(n=oe(n,"int32"));let o={x:n},s={axis:e,keepDims:t};return A.runKernel(Qi,o,s)}var fu=S({prod_:jU});function UU(r,e,t){let n=nt(r),o=null;if(t==null||t==="float32")o=new Float32Array(n);else if(t==="int32")o=new Int32Array(n);else if(t==="bool")o=new Uint8Array(n);else throw new Error(`Unknown data type ${t}`);for(let s=0;s<n;s++)o[s]=e();return A.makeTensor(o,r,t)}var HU=S({rand_:UU});var _g=hc(__());var lp=class{constructor(e,t,n,o,s){this.mean=e,this.stdDev=t,this.dtype=n,this.nextVal=NaN,this.truncated=o,this.truncated&&(this.upper=this.mean+this.stdDev*2,this.lower=this.mean-this.stdDev*2);let a=s||Math.random();this.random=_g.alea(a.toString())}nextValue(){if(!isNaN(this.nextVal)){let o=this.nextVal;return this.nextVal=NaN,o}let e,t,n=!1;for(;!n;){let o,s,a;do o=2*this.random()-1,s=2*this.random()-1,a=o*o+s*s;while(a>=1||a===0);let i=Math.sqrt(-2*Math.log(a)/a);e=this.mean+this.stdDev*o*i,t=this.mean+this.stdDev*s*i,(!this.truncated||this.isValidTruncated(e))&&(n=!0)}return(!this.truncated||this.isValidTruncated(t))&&(this.nextVal=this.convertValue(t)),this.convertValue(e)}convertValue(e){return this.dtype==null||this.dtype==="float32"?e:Math.round(e)}isValidTruncated(e){return e<=this.upper&&e>=this.lower}},k_=class{constructor(e,t,n,o){this.alpha=e,this.beta=1/t,this.dtype=n;let s=o||Math.random();this.randu=_g.alea(s.toString()),this.randn=new lp(0,1,n,!1,this.randu()),e<1?this.d=e+2/3:this.d=e-1/3,this.c=1/Math.sqrt(9*this.d)}nextValue(){let e,t,n,o,s,a;for(;;){do o=this.randn.nextValue(),a=1+this.c*o;while(a<=0);if(a*=a*a,e=o*o,t=1-.331*e*e,n=.5*e+this.d*(1-a+Math.log(a)),s=this.randu(),s<t||Math.log(s)<n)break}return a=1/this.beta*this.d*a,this.alpha<1&&(a*=Math.pow(this.randu(),1/this.alpha)),this.convertValue(a)}convertValue(e){return this.dtype==="float32"?e:Math.round(e)}},v_=class{constructor(e=0,t=1,n,o){if(this.canReturnFloat=()=>this.dtype==null||this.dtype==="float32",this.min=e,this.range=t-e,this.dtype=n,o==null&&(o=Math.random()),typeof o=="number"&&(o=o.toString()),!this.canReturnFloat()&&this.range<=1)throw new Error(`The difference between ${e} - ${t} <= 1 and dtype is not float`);this.random=_g.alea(o)}convertValue(e){return this.canReturnFloat()?e:Math.round(e)}nextValue(){return this.convertValue(this.min+this.range*this.random())}};function QU(r,e,t=1,n="float32",o){if(t==null&&(t=1),n==null&&(n="float32"),n!=="float32"&&n!=="int32")throw new Error(`Unsupported data type ${n}`);let s=new k_(e,t,n,o),a=Ce(r,n);for(let i=0;i<a.values.length;i++)a.values[i]=s.nextValue();return a.toTensor()}var eH=S({randomGamma_:QU});function tH(r,e=0,t=1,n,o){if(n!=null&&n==="bool")throw new Error(`Unsupported data type ${n}`);let s=new lp(e,t,n,!1,o),a=Ce(r,n);for(let i=0;i<a.values.length;i++)a.values[i]=s.nextValue();return a.toTensor()}var kg=S({randomNormal_:tH});function rH(r,e=0,t=1,n="float32",o){let s=Ce(r,n),a=new v_(e,t,null,o);for(let i=0;i<s.values.length;i++)s.values[i]=a.nextValue();return s.toTensor()}var ks=S({randomUniform_:rH});function Ca(r,e,t=1,n="float32"){if(t===0)throw new Error("Cannot have a step of zero");let o={start:r,stop:e,step:t,dtype:n};return A.runKernel(sl,{},o)}function nH(r){let t={input:k(r,"input","real")};return A.runKernel(Gc,t)}var gl=S({real_:nH});function oH(r){let t={x:k(r,"x","reciprocal")};return A.runKernel(ea,t)}var ff=S({reciprocal_:oH});function sH(r){let t={x:k(r,"x","relu")};return A.runKernel(Zo,t)}var $r=S({relu_:sH});function iH(r){let t={x:k(r,"x","relu6")};return A.runKernel(Qo,t)}var hu=S({relu6_:iH});function aH(r,e){let n={x:k(r,"x","reverse")},o={dims:e};return A.runKernel(es,n,o)}var Kt=S({reverse_:aH});function lH(r){let e=k(r,"x","reverse");return T(e.rank===1,()=>`Error in reverse1D: x must be rank 1 but got rank ${e.rank}.`),Kt(e,0)}var uH=S({reverse1d_:lH});function cH(r,e){let t=k(r,"x","reverse");return T(t.rank===2,()=>`Error in reverse2D: x must be rank 2 but got rank ${t.rank}.`),Kt(t,e)}var pH=S({reverse2d_:cH});function mH(r,e){let t=k(r,"x","reverse");return T(t.rank===3,()=>`Error in reverse3D: x must be rank 3 but got rank ${t.rank}.`),Kt(t,e)}var fH=S({reverse3d_:mH});function dH(r,e){let t=k(r,"x","reverse");return T(t.rank===4,()=>`Error in reverse4D: x must be rank 4 but got rank ${t.rank}.`),Kt(t,e)}var hH=S({reverse4d_:dH});function gH(r){let t={x:k(r,"x","round")};return A.runKernel(ts,t)}var gu=S({round_:gH});function xH(r){let t={x:k(r,"x","rsqrt")};return A.runKernel(rs,t)}var xu=S({rsqrt_:xH});function ue(r,e){if((ar(r)&&e!=="string"||Array.isArray(r))&&e!=="complex64")throw new Error("Error creating a new Scalar: value must be a primitive (number|boolean|string)");if(e==="string"&&ar(r)&&!(r instanceof Uint8Array))throw new Error("When making a scalar from encoded string, the value must be `Uint8Array`.");return Yr(r,[],[],e)}function yH(r){let t={x:k(r,"x","selu")};return A.runKernel(ra,t)}var yu=S({selu_:yH});function bH(r,e,t,n,o,s=[1,1],a="NHWC"){let i=k(r,"x","separableConv2d"),l=k(e,"depthwiseFilter","separableConv2d"),u=k(t,"pointwiseFilter","separableConv2d"),c=i,p=!1;if(i.rank===3&&(p=!0,c=L(i,[1,i.shape[0],i.shape[1],i.shape[2]])),a==="NCHW")throw new Error("separableConv2d currently does not support dataFormat NCHW; only NHWC is supported");T(c.rank===4,()=>`Error in separableConv2d: input must be rank 4, but got rank ${c.rank}.`),T(l.rank===4,()=>`Error in separableConv2d: depthwise filter must be rank 4, but got rank ${l.rank}.`),T(u.rank===4,()=>`Error in separableConv2d: pointwise filter must be rank 4, but got rank ${l.rank}.`),T(u.shape[0]===1,()=>`Error in separableConv2d: the first dimension of pointwise filter must be 1, but got ${u.shape[0]}.`),T(u.shape[1]===1,()=>`Error in separableConv2d: the second dimension of pointwise filter must be 1, but got ${u.shape[1]}.`);let m=l.shape[2],f=l.shape[3];T(u.shape[2]===m*f,()=>`Error in separableConv2d: the third dimension of pointwise filter must be ${m*f}, but got ${u.shape[2]}.`);let d=xs(c,l,n,o,a,s),g=Jr(d,u,1,"valid",a);return p?L(g,[g.shape[1],g.shape[2],g.shape[3]]):g}var df=S({separableConv2d_:bH});async function wH(r,e){let t=k(r,"x","setdiff1d"),n=k(e,"y","setdiff1d");T(t.dtype===n.dtype,()=>`x and y should have the same dtype, but got x (${t.dtype}) and y (${n.dtype}).`),T(t.rank===1,()=>`x should be 1D tensor, but got x (${t.shape}).`),T(n.rank===1,()=>`y should be 1D tensor, but got y (${n.shape}).`);let o=await t.data(),s=await n.data(),a=new Set(s),i=0;for(let c=0;c<o.length;c++)a.has(o[c])||i++;let l=new ut([i],t.dtype),u=new ut([i],"int32");for(let c=0,p=0;c<o.length;c++)a.has(o[c])||(l.values[p]=o[c],u.values[p]=c,p++);return[l.toTensor(),u.toTensor()]}var C_=wH;function _H(r){let t={x:k(r,"x","sign")};return A.runKernel(oa,t)}var hf=S({sign_:_H});function kH(r){let t={x:k(r,"x","sin")};return A.runKernel(ns,t)}var bu=S({sin_:kH});function vH(r){let t={x:k(r,"x","sinh")};return A.runKernel(na,t)}var wu=S({sinh_:vH});function CH(r,e,t){let n=k(r,"x","slice1d");return T(n.rank===1,()=>`slice1d expects a rank-1 tensor, but got a rank-${n.rank} tensor`),Re(n,[e],[t])}var gf=S({slice1d_:CH});function IH(r,e,t){let n=k(r,"x","slice2d");return T(n.rank===2,()=>`slice2d expects a rank-2 tensor, but got a rank-${n.rank} tensor`),Re(n,e,t)}var vg=S({slice2d_:IH});function NH(r,e,t){let n=k(r,"x","slice3d");return T(n.rank===3,()=>`slice3d expects a rank-3 tensor, but got a rank-${n.rank} tensor`),Re(n,e,t)}var xf=S({slice3d_:NH});function SH(r,e,t){let n=k(r,"x","slice4d");return T(n.rank===4,()=>`slice4d expects a rank-4 tensor, but got a rank-${n.rank} tensor`),Re(n,e,t)}var up=S({slice4d_:SH});function TH(r,e=-1){let t=k(r,"logits","softmax","float32");if(e===-1&&(e=t.rank-1),e!==t.rank-1)throw Error(`Softmax along a non-last dimension is not yet supported. Logits was rank ${t.rank} and dim was ${e}`);let n={logits:t},o={dim:e};return A.runKernel(as,n,o)}var Ia=S({softmax_:TH});function AH(r){T(r.dtype==="complex64",()=>`The dtype for tf.spectral.fft() must be complex64 but got ${r.dtype}.`);let e={input:r};return A.runKernel(Rc,e)}var Na=S({fft_:AH});function EH(r){T(r.dtype==="complex64",()=>`The dtype for tf.spectral.ifft() must be complex64 but got ${r.dtype}.`);let e={input:r};return A.runKernel(Fc,e)}var si=S({ifft_:EH});function DH(r){let e=r.shape[r.shape.length-1],t=r.size/e,n;if(e<=2){let o=L(r,[t,e]);n=si(o)}else{let o=[t,2*(e-1)],s=L(gl(r),[t,e]),a=L(lu(r),[t,e]),i=Kt(Re(s,[0,1],[t,e-2]),1),l=O(Kt(Re(a,[0,1],[t,e-2]),1),ue(-1)),u=Qe([s,i],1),c=Qe([a,l],1),p=L(Nn(u,c),[o[0],o[1]]);n=si(p)}if(n=gl(n),r.rank===3&&r.shape[0]!==0){let o=n,s=r.shape[0];n=L(n,[s,n.shape[0]/s,n.shape[1]]),o.dispose()}return n}var _u=S({irfft_:DH});function $H(r,e,t=0){let o={x:k(r,"x","split")},s={numOrSizeSplits:e,axis:t};return A.runKernel(Ys,o,s)}var tr=S({split_:$H});function RH(r,e){T(r.dtype==="float32",()=>`The dtype for rfft() must be real value but got ${r.dtype}`);let t=r.shape[r.shape.length-1],n=r.size/t,o;if(e!=null&&e<t){let d=r.shape.map(g=>0),h=r.shape.map(g=>g);h[r.shape.length-1]=e,o=Re(r,d,h),t=e}else if(e!=null&&e>t){let d=r.shape.map(h=>h);d[r.shape.length-1]=e-t,o=Qe([r,ht(d)],r.shape.length-1),t=e}else o=r;let s=Ie(o),a=L(Nn(o,s),[n,t]),i=Na(a),l=Math.floor(t/2)+1,u=gl(i),c=lu(i),p=tr(u,[l,t-l],u.shape.length-1),m=tr(c,[l,t-l],c.shape.length-1),f=o.shape.slice();return f[o.shape.length-1]=l,L(Nn(p[0],m[0]),f)}var Sa=S({rfft_:RH});function FH(r){let t={x:k(r,"x","sqrt")};return A.runKernel(ss,t)}var gt=S({sqrt_:FH});function OH(r,e){let t=k(r,"a","squaredDifference"),n=k(e,"b","squaredDifference");[t,n]=je(t,n),Be(t.shape,n.shape);let o={a:t,b:n},s={};return A.runKernel(ls,o,s)}var ku=S({squaredDifference_:OH});function PH(r,e){let t=k(r,"x","squeeze");return L(t,ew(t.shape,e).newShape)}var Tn=S({squeeze_:PH});function MH(r,e=0){let t=ca(r,"tensors","stack","string_or_numeric");T(t.length>=1,()=>"Pass at least one tensor to tf.stack"),t.length>0&&T(e<=t[0].rank,()=>"Axis must be <= rank of the tensor");let n=t,o={axis:e};return A.runKernel(Hs,n,o)}var Gt=S({stack_:MH});function LH(r,e=0){let n={x:k(r,"x","step")},o={alpha:e};return A.runKernel(Yn,n,o)}var vs=S({step_:LH});function zH(r,e,t,n,o=0,s=0,a=0,i=0,l=0){let c={x:k(r,"x","stridedSlice")},p={begin:e,end:t,strides:n,beginMask:o,endMask:s,ellipsisMask:a,newAxisMask:i,shrinkAxisMask:l};return A.runKernel(ia,c,p)}var yf=S({stridedSlice_:zH});function BH(r){let t={x:k(r,"x","tan")};return A.runKernel(cs,t)}var bf=S({tan_:BH});function Ct(r,e){On(r);let t=Ar(r,e);if(t.length!==1)throw new Error("tensor1d() requires values to be a flat/TypedArray");return Yr(r,null,t,e)}function ii(r,e,t){if(On(r),e!=null&&e.length!==2)throw new Error("tensor2d() requires shape to have two numbers");let n=Ar(r,t);if(n.length!==2&&n.length!==1)throw new Error("tensor2d() requires values to be number[][] or flat/TypedArray");if(n.length===1&&e==null)throw new Error("tensor2d() requires shape to be provided when `values` are a flat/TypedArray");return Yr(r,e,n,t)}function VH(r,e,t){if(On(r),e!=null&&e.length!==4)throw new Error("tensor4d() requires shape to have four numbers");let n=Ar(r,t);if(n.length!==4&&n.length!==1)throw new Error("tensor4d() requires values to be number[][][][] or flat/TypedArray");if(n.length===1&&e==null)throw new Error("tensor4d() requires shape to be provided when `values` are a flat array");return Yr(r,e,n,t)}function GH(r,e,t){if(On(r),e!=null&&e.length!==5)throw new Error("tensor5d() requires shape to have five numbers");let n=Ar(r,t);if(n.length!==5&&n.length!==1)throw new Error("tensor5d() requires values to be number[][][][][] or flat/TypedArray");if(n.length===1&&e==null)throw new Error("tensor5d() requires shape to be provided when `values` are a flat array");return Yr(r,e,n,t)}function WH(r,e,t){if(On(r),e!=null&&e.length!==6)throw new Error("tensor6d() requires shape to have six numbers");let n=Ar(r,t);if(n.length!==6&&n.length!==1)throw new Error("tensor6d() requires values to be number[][][][][][] or flat/TypedArray");if(n.length===1&&e==null)throw new Error("tensor6d() requires shape to be provided when `values` are a flat array");return e=e||n,Yr(r,e,n,t)}function jH(r,e=1,t=!0){let n=k(r,"x","topk");if(n.rank===0)throw new Error("topk() expects the input to be of rank 1 or higher");let o=n.shape[n.shape.length-1];if(e>o)throw new Error(`'k' passed to topk() must be <= the last dimension (${o}) but got ${e}`);let s={x:n},a={k:e,sorted:t},[i,l]=A.runKernel(aa,s,a);return{values:i,indices:l}}var wf=S({topk_:jH});function UH(r,e=0,t=1,n,o){if(n!=null&&n==="bool")throw new Error("Unsupported data type $ { dtype }");let s=new lp(e,t,n,!0,o),a=Ce(r,n);for(let i=0;i<a.values.length;i++)a.values[i]=s.nextValue();return a.toTensor()}var vu=S({truncatedNormal_:UH});function HH(r,e=0){let t=k(r,"x","unique","string_or_numeric");T(t.rank>0,()=>"The input tensor must be at least 1D");let n={x:t},o={axis:e},[s,a]=A.runKernel(Kc,n,o);return{values:s,indices:a}}var cp=S({unique_:HH});function qH(r,e,t){let n=k(r,"x","unsortedSegmentSum"),o=k(e,"segmentIds","unsortedSegmentSum","int32");T(ot(t),()=>"numSegments must be of dtype int");let s={x:n,segmentIds:o},a={numSegments:t};return A.runKernel(ul,s,a)}var _f=S({unsortedSegmentSum_:qH});function KH(r,e=0){let t=k(r,"x","unstack","string_or_numeric");T(e>=-t.shape.length&&e<t.shape.length,()=>`Axis = ${e} is not in [-${t.shape.length}, ${t.shape.length})`);let n={value:t},o={axis:e};return A.runKernel(Zs,n,o)}var fr=S({unstack_:KH});function I_(r,e=!0,t,n){return A.makeVariable(r,e,t,n)}function Cg(r,e){let t=[];for(let s=0;s<e.length;s++)e[s]&&t.push(s);let n=Ce(r,"int32"),o=Ce([t.length,r.length],"int32");for(let s=0;s<t.length;s++){let a=n.indexToLoc(t[s]),i=s*r.length;o.values.set(a,i)}return o.toTensor()}async function XH(r){let e=k(r,"condition","whereAsync","bool"),t=await e.data(),n=Cg(e.shape,t);return r!==e&&e.dispose(),n}var kf=XH;async function YH(r,e,t){let n=k(r,"tensor","boolMask"),o=k(e,"mask","boolMask","bool"),s=t==null?0:t,a=o.rank,i=n.shape;T(a>0,()=>"mask cannot be scalar"),Dt(i.slice(s,s+a),o.shape,"mask's shape must match the first K dimensions of tensor's shape,");let l=1;for(let h=s;h<s+a;h++)l*=i[h];let u=i.slice(0,s).concat([l],i.slice(s+a)),c=L(n,u),p=L(o,[-1]),m=await kf(p),f=Tn(m,[1]),d=Qn(c,f,s);return r!==n&&n.dispose(),e!==o&&o.dispose(),f.dispose(),c.dispose(),p.dispose(),m.dispose(),d}var IIe=YH;function ZH(r,e="euclidean",t=null,n=!1){r=k(r,"x","norm");let o=n1(r,e,t),s=o.shape;if(n){let a=nr(t,r.shape);s=to(o.shape,a)}return L(o,s)}function n1(r,e,t=null){if(r.rank===0)return Tt(r);if(r.rank!==1&&t===null)return n1(L(r,[-1]),e,t);if(r.rank===1||typeof t=="number"||Array.isArray(t)&&t.length===1){if(e===1)return fe(Tt(r),t);if(e===Infinity)return mr(Tt(r),t);if(e===-Infinity)return oi(Tt(r),t);if(e==="euclidean"||e===2)return gt(fe(zr(Tt(r),ue(2,"int32")),t));throw new Error(`Error in norm: invalid ord value: ${e}`)}if(Array.isArray(t)&&t.length===2){if(e===1)return mr(fe(Tt(r),t[0]),t[1]-1);if(e===Infinity)return mr(fe(Tt(r),t[1]),t[0]);if(e===-Infinity)return oi(fe(Tt(r),t[1]),t[0]);if(e==="fro"||e==="euclidean")return gt(fe(Me(r),t));throw new Error(`Error in norm: invalid ord value: ${e}`)}throw new Error(`Error in norm: invalid axis: ${t}`)}var pp=S({norm_:ZH});function JH(r,e,t,n,o=!0){let s=k(r,"v","movingAverage"),a=k(e,"x","movingAverage"),i=k(t,"decay","movingAverage");yw(s,a),T(Xr(s.shape,a.shape),()=>"Shape mismatch in v and x");let l=ue(1),u=le(l,i),c=O(le(a,s),u);if(o){T(n!=null,()=>"When using zeroDebias: true, step is required.");let p=k(n,"step","movingAverage");c=pe(c,le(l,zr(i,p)))}return J(s,c)}var ZIe=S({movingAverage_:JH});function QH(r,e,t){let n=k(r,"indices","scatterND","int32"),o=k(e,"updates","scatterND");mg(o,n,t);let s={indices:n,updates:o},a={shape:t};return A.runKernel(ta,s,a)}var o1=S({scatterND_:QH});function s1(r,e,t,n){if(r.dtype!=="int32")throw new Error(`tf.sparseToDense() expects the indices to be int32 type, but the dtype was ${r.dtype}.`);if(r.rank>2)throw new Error(`sparseIndices should be a scalar, vector, or matrix, but got shape ${r.shape}.`);let o=r.rank>0?r.shape[0]:1,s=r.rank>1?r.shape[1]:1;if(t.length!==s)throw new Error(`outputShape has incorrect number of elements:, ${t.length}, should be: ${s}.`);let a=e.size;if(!(e.rank===0||e.rank===1&&a===o))throw new Error(`sparseValues has incorrect shape ${e.shape}, should be [] or [${o}]`);if(e.dtype!==n.dtype)throw new Error("sparseValues.dtype must match defaultValues.dtype")}function eq(r,e,t,n=0){let o=k(r,"sparseIndices","sparseToDense","int32"),s=k(e,"sparseValues","sparseToDense"),a=k(n,"defaultValue","sparseToDense",s.dtype);s1(o,s,t,a);let i={sparseIndices:o,sparseValues:s,defaultValue:a},l={outputShape:t};return A.runKernel(qc,i,l)}var Ig=S({sparseToDense_:eq});function tq(r,e){let t=k(e,"indices","gatherND","int32"),o={params:k(r,"x","gatherND"),indices:t};return A.runKernel(zi,o)}var i1=S({gatherND_:tq});function a1(r,e){if(e==null)return r.shape.slice();if(Xr(r.shape,e))return e;if(r.shape.length===e.length){let t=[];for(let n=0;n<r.shape.length;n++)e[n]==null&&r.shape[n]!=null?t.push(r.shape[n]):t.push(e[n]);return t}return e}function rq(r,e,t,n){let o=k(r,"x","dropout");if(T(o.dtype==="float32",()=>`x has to be a floating point tensor since it's going to be scaled, but got a ${o.dtype} tensor instead.`),T(e>=0&&e<1,()=>`rate must be a float in the range [0, 1), but got ${e}.`),e===0)return r instanceof Pe?o.clone():o;let s=a1(o,t),a=1-e,i=pe(ws(J(ks(s,0,1,"float32",n),a)),a);return O(o,i)}var l1=S({dropout_:rq});function u1(r){return Math.floor(Math.pow(2,Math.ceil(Math.log(r)/Math.log(2))))}function Ng(r,e,t){let n=1-r%2,o=new Float32Array(r);for(let s=0;s<r;++s){let a=2*Math.PI*s/(r+n-1);o[s]=e-t*Math.cos(a)}return Ct(o,"float32")}async function nq(r,e,t=1){let n=k(r,"predictions","inTopK"),o=k(e,"targets","inTopK");T(n.rank>1,()=>`inTopK() expects the predictions to be of rank 2 or higher, but got ${n.rank}`),T(n.rank-1===o.rank,()=>`predictions rank should be 1 larger than targets rank, but got predictions rank ${n.rank} and targets rank ${o.rank}`),Dt(n.shape.slice(0,n.shape.length-1),o.shape,"predictions's shape should be align with the targets' shape, except the last dimension.");let s=n.shape[n.shape.length-1];T(t>0&&t<=s,()=>`'k' passed to inTopK() must be > 0 && <= the predictions last dimension (${s}), but got ${t}`);let a=await n.data(),i=await o.data(),[l,u]=[a.length/s,s],c=tw("bool",l);for(let p=0;p<l;p++){let m=p*u,f=a.subarray(m,m+u),d=[];for(let h=0;h<f.length;h++)d.push({value:f[h],index:h});d.sort((h,g)=>g.value-h.value),c[p]=0;for(let h=0;h<t;h++)if(d[h].index===i[p]){c[p]=1;break}}return r!==n&&n.dispose(),e!==o&&o.dispose(),vr(c,o.shape,"bool")}var ONe=nq;var no={};We(no,{conv2d:()=>c1,depthwiseConv2d:()=>p1,matMul:()=>m1});function oq(r,e,t,n,o,s="NHWC",a){let i=r;r.rank===3&&(i=L(r,[1,r.shape[0],r.shape[1],r.shape[2]]));let l=e;l.rank===3&&(l=L(e,[1,e.shape[0],e.shape[1],e.shape[2]])),T(i.rank===4,()=>`Error in conv2dDerFilter: input must be rank 4, but got shape ${i.shape}.`),T(l.rank===4,()=>`Error in conv2dDerFilter: dy must be rank 4, but got shape ${l.shape}.`),T(t.length===4,()=>`Error in conv2dDerFilter: filterShape must be length 4, but got ${t}.`);let u=s==="NHWC"?i.shape[3]:i.shape[1],c=s==="NHWC"?l.shape[3]:l.shape[1];T(u===t[2],()=>`Error in conv2dDerFilter: depth of input ${u}) must match input depth in filter (${t[2]}.`),T(c===t[3],()=>`Error in conv2dDerFilter: depth of dy (${c}) must match output depth for filter (${t[3]}).`),a!=null&&T(ot(o),()=>`Error in conv2dDerFilter: pad must be an integer when using, dimRoundingMode ${a} but got pad ${o}.`);let p={x:i,dy:l},m={strides:n,pad:o,dataFormat:s,dimRoundingMode:a,filterShape:t};return A.runKernel(Cc,p,m)}var mp=S({conv2DBackpropFilter_:oq});function Cu(r,e,t){if(t==null||t==="linear")return r;if(t==="relu")return O(r,vs(e));throw new Error(`Cannot compute gradient for fused activation ${t}.`)}function Iu(r,e){let t=e,n=kt(r.shape,e.shape);return n.length>0&&(t=fe(t,n)),L(t,r.shape)}function Nu(r,e,t,n){if(e==="linear")return r;if(e==="relu")return $r(r);if(e==="elu")return ys(r);if(e==="relu6")return hu(r);if(e==="prelu")return va(r,t);if(e==="leakyrelu")return ba(r,n);if(e==="sigmoid")return Dr(r);throw new Error(`Unknown fused activation ${e}.`)}var Su=(r,e)=>!(r>0)||e==="linear";function sq({x:r,filter:e,strides:t,pad:n,dataFormat:o="NHWC",dilations:s=[1,1],dimRoundingMode:a,bias:i,activation:l="linear",preluActivationWeights:u,leakyreluAlpha:c}){if(l=l||"linear",Su(A.state.gradientDepth,l)===!1){let I=Jr(r,e,t,n,o,s,a);return i!=null&&(I=J(I,i)),Nu(I,l,u,c)}let p=k(r,"x","conv2d"),m=k(e,"filter","conv2d"),f=p,d=!1;p.rank===3&&(d=!0,f=L(p,[1,p.shape[0],p.shape[1],p.shape[2]])),T(f.rank===4,()=>`Error in fused conv2d: input must be rank 4, but got rank ${f.rank}.`),T(m.rank===4,()=>`Error in fused conv2d: filter must be rank 4, but got rank ${m.rank}.`),a!=null&&T(ot(n),()=>`Error in fused conv2d: pad must be an integer when using, dimRoundingMode ${a} but got pad ${n}.`),T(f.shape[3]===m.shape[2],()=>`Error in conv2d: depth of input (${f.shape[3]}) must match input depth for filter ${m.shape[2]}.`),T(Cr(t,s),()=>`Error in conv2D: Either strides or dilations must be 1. Got strides ${t} and dilations '${s}'`),T(o==="NHWC",()=>`Error in conv2d: got dataFormat of ${o} but only NHWC is currently supported.`);let h=ru(f.shape,m.shape,t,s,n,a),g;i!=null&&(g=k(i,"bias","fused conv2d"),[g]=je(g,p),Be(h.outShape,g.shape));let x;u!=null&&(x=k(u,"prelu weights","fused conv2d"));let b=(I,E)=>{let[D,$,R,M]=E,G=Cu(I,R,l);T(zn(s),()=>`Error in gradient of fused conv2D: dilation rates greater than 1 are not yet supported in gradients. Got dilations '${s}'`);let j=sp($.shape,G,D,t,n),U=mp($,G,D.shape,t,n),H=[j,U];if(M!=null){let q=Iu(M,G);H.push(q)}return H},w={x:f,filter:m,bias:g,preluActivationWeights:x},_={strides:t,pad:n,dataFormat:o,dilations:s,dimRoundingMode:a,activation:l,leakyreluAlpha:c};return i==null?Qr((E,D,$)=>{let R=A.runKernel(ei,w,_);return $([D,E,R]),d&&(R=L(R,[R.shape[1],R.shape[2],R.shape[3]])),{value:R,gradFunc:b}})(f,m):Qr((E,D,$,R)=>{let M=A.runKernel(ei,w,_);return R([D,E,M,$]),d&&(M=L(M,[M.shape[1],M.shape[2],M.shape[3]])),{value:M,gradFunc:b}})(f,m,g)}var c1=S({fusedConv2d_:sq});function iq(r,e,t,n,o,s=[1,1],a){let i=r;r.rank===3&&(i=L(r,[1,r.shape[0],r.shape[1],r.shape[2]]));let l=e;l.rank===3&&(l=L(e,[1,e.shape[0],e.shape[1],e.shape[2]]));let u={x:i,dy:l},c={strides:n,pad:o,dimRoundingMode:a,dilations:s,filterShape:t};return A.runKernel(Tc,u,c)}var Sg=S({depthwiseConv2dNativeBackpropFilter_:iq});function aq(r,e,t,n,o,s=[1,1],a){let i=e,l=!1;e.rank===3&&(l=!0,i=L(e,[1,e.shape[0],e.shape[1],e.shape[2]]));let u={dy:i,filter:t},c={strides:n,pad:o,dimRoundingMode:a,dilations:s,inputShape:r},p=A.runKernel(Ac,u,c);return l?L(p,[p.shape[1],p.shape[2],p.shape[3]]):p}var Tg=S({depthwiseConv2dNativeBackpropInput_:aq});function lq({x:r,filter:e,strides:t,pad:n,dataFormat:o="NHWC",dilations:s=[1,1],dimRoundingMode:a,bias:i,activation:l="linear",preluActivationWeights:u,leakyreluAlpha:c}){if(Su(A.state.gradientDepth,l)===!1){let I=xs(r,e,t,n,o,s,a);return i!=null&&(I=J(I,i)),Nu(I,l,u,c)}let p=k(r,"x","depthwiseConv2d"),m=k(e,"filter","depthwiseConv2d"),f=p,d=!1;p.rank===3&&(d=!0,f=L(p,[1,p.shape[0],p.shape[1],p.shape[2]])),T(f.rank===4,()=>`Error in fused depthwiseConv2d: input must be rank 4, but got rank ${f.rank}.`),T(m.rank===4,()=>`Error in fused depthwiseConv2d: filter must be rank 4, but got rank ${m.rank}.`),T(f.shape[3]===m.shape[2],()=>`Error in fused depthwiseConv2d: number of input channels (${f.shape[3]}) must match the inChannels dimension in filter ${m.shape[2]}.`),s==null&&(s=[1,1]),T(Cr(t,s),()=>`Error in fused depthwiseConv2d: Either strides or dilations must be 1. Got strides ${t} and dilations '${s}'`),a!=null&&T(ot(n),()=>`Error in fused depthwiseConv2d: pad must be an integer when using dimRoundingMode ${a} but got pad ${n}.`);let h=ru(f.shape,m.shape,t,s,n,a,!0),g;i!=null&&(g=k(i,"bias","fused conv2d"),[g]=je(g,p),Be(h.outShape,g.shape));let x;u!=null&&(x=k(u,"prelu weights","fused depthwiseConv2d"));let b=(I,E)=>{T(zn(s),()=>`Error in gradient of fused depthwiseConv2d: dilation rates greater than 1 are not yet supported. Got dilations '${s}'`);let[D,$,R,M]=E,G=Cu(I,R,l),j=Tg($.shape,G,D,t,n,s,a),U=Sg($,G,D.shape,t,n,s,a);if(M!=null){let H=Iu(g,G);return[j,U,H]}return[j,U]},w={x:f,filter:m,bias:g,preluActivationWeights:x},_={strides:t,pad:n,dataFormat:o,dilations:s,dimRoundingMode:a,activation:l,leakyreluAlpha:c};return i==null?Qr((E,D,$)=>{let R=A.runKernel(ti,w,_);return $([D,E,R]),d&&(R=L(R,[R.shape[1],R.shape[2],R.shape[3]])),{value:R,gradFunc:b}})(f,m):Qr((E,D,$,R)=>{let M=A.runKernel(ti,w,_);return R([D,E,M,$]),d&&(M=L(M,[M.shape[1],M.shape[2],M.shape[3]])),{value:M,gradFunc:b}})(f,m,g)}var p1=S({fusedDepthwiseConv2d_:lq});function uq({a:r,b:e,transposeA:t=!1,transposeB:n=!1,bias:o,activation:s="linear",preluActivationWeights:a,leakyreluAlpha:i}){if(Su(A.state.gradientDepth,s)===!1){let M=ze(r,e,t,n);return o!=null&&(M=J(M,o)),Nu(M,s,a,i)}let l=k(r,"a","fused matMul"),u=k(e,"b","fused matMul");[l,u]=je(l,u);let c=t?l.shape[l.rank-2]:l.shape[l.rank-1],p=n?u.shape[u.rank-1]:u.shape[u.rank-2],m=t?l.shape[l.rank-1]:l.shape[l.rank-2],f=n?u.shape[u.rank-2]:u.shape[u.rank-1],d=l.shape.slice(0,-2),h=u.shape.slice(0,-2),g=nt(d),x=nt(h);T(l.rank>=2&&u.rank>=2&&l.rank===u.rank,()=>`Error in fused matMul: inputs must have the same rank of at least 2, got ranks ${l.rank} and ${u.rank}.`),T(Xr(d,h),()=>`Error in fused matMul: outer dimensions (${d}) and (${h}) of Tensors with shapes ${l.shape} and ${u.shape} must match.`),T(c===p,()=>`Error in fused matMul: inner shapes (${c}) and (${p}) of Tensors with shapes ${l.shape} and ${u.shape} and transposeA=${t} and transposeB=${n} must match.`);let b=l.shape.slice(0,-2).concat([m,f]),w=t?L(l,[g,c,m]):L(l,[g,m,c]),_=n?L(u,[x,f,p]):L(u,[x,p,f]),I;o!=null&&(I=k(o,"bias","fused matMul"),[I]=je(I,l),Be(b,I.shape));let E;a!=null&&(E=k(a,"prelu weights","fused matMul"));let D=(M,G)=>{let[j,U,H,q]=G,X=Cu(L(M,H.shape),H,s),ne,Y;if(!t&&!n?(ne=ze(X,U,!1,!0),Y=ze(j,X,!0,!1)):!t&&n?(ne=ze(X,U,!1,!1),Y=ze(X,j,!0,!1)):t&&!n?(ne=ze(U,X,!1,!0),Y=ze(j,X,!1,!1)):(ne=ze(U,X,!0,!0),Y=ze(X,j,!0,!0)),o!=null){let re=Iu(q,X);return[ne,Y,re]}else return[ne,Y]},$={a:w,b:_,bias:I,preluActivationWeights:E},R={transposeA:t,transposeB:n,activation:s,leakyreluAlpha:i};return o==null?Qr((G,j,U)=>{let H=A.runKernel(Qs,$,R);return U([G,j,H]),{value:L(H,b),gradFunc:D}})(w,_):Qr((G,j,U,H)=>{let q=A.runKernel(Qs,$,R);return H([G,j,q,U]),{value:L(q,b),gradFunc:D}})(w,_,I)}var m1=S({fusedMatMul_:uq});function cq(r){return Ng(r,.54,.46)}var f1=S({hammingWindow_:cq});function pq(r){return Ng(r,.5,.5)}var Ag=S({hannWindow_:pq});function mq(r,e,t,n=!1,o=0){let s=0,a=[];for(;s+e<=r.size;)a.push(Re(r,s,e)),s+=t;if(n)for(;s<r.size;){let i=s+e-r.size,l=Qe([Re(r,s,e-i),bs([i],o)]);a.push(l),s+=t}return a.length===0?ii([],[0,e]):L(Qe(a),[a.length,e])}var Eg=S({frame_:mq});function fq(r,e,t,n,o=Ag){n==null&&(n=u1(e));let s=Eg(r,e,t),a=O(s,o(e));return Sa(a,n)}var d1=S({stft_:fq});function dq(r,e,t,n,o="bilinear",s=0){let a=k(r,"image","cropAndResize"),i=k(e,"boxes","cropAndResize","float32"),l=k(t,"boxInd","cropAndResize","int32"),u=i.shape[0];T(a.rank===4,()=>`Error in cropAndResize: image must be rank 4,but got rank ${a.rank}.`),T(i.rank===2&&i.shape[1]===4,()=>`Error in cropAndResize: boxes must be have size [${u},4] but had shape ${i.shape}.`),T(l.rank===1&&l.shape[0]===u,()=>`Error in cropAndResize: boxInd must be have size [${u}] but had shape ${i.shape}.`),T(n.length===2,()=>`Error in cropAndResize: cropSize must be of length 2, but got length ${n.length}.`),T(n[0]>=1&&n[1]>=1,()=>`cropSize must be atleast [1,1], but was ${n}`),T(o==="bilinear"||o==="nearest",()=>`method must be bilinear or nearest, but was ${o}`);let c={image:a,boxes:i,boxInd:l},p={method:o,extrapolationValue:s,cropSize:n};return A.runKernel($i,c,p)}var h1=S({cropAndResize_:dq});function hq(r){let e=k(r,"image","flipLeftRight","float32");T(e.rank===4,()=>`Error in flipLeftRight: image must be rank 4,but got rank ${e.rank}.`);let t={image:e};return A.runKernel(Li,t,{})}var g1=S({flipLeftRight_:hq});function gq(r,e,t=0,n=.5){let o=k(r,"image","rotateWithOffset","float32");T(o.rank===4,()=>`Error in rotateWithOffset: image must be rank 4,but got rank ${o.rank}.`);let s={image:o},a={radians:e,fillValue:t,center:n};return A.runKernel(ua,s,a)}var x1=S({rotateWithOffset_:gq});function oo(r,e,t,n,o,s){n==null&&(n=.5),o==null&&(o=Number.NEGATIVE_INFINITY),s==null&&(s=0);let a=r.shape[0];return t=Math.min(t,a),T(0<=n&&n<=1,()=>`iouThreshold must be in [0, 1], but was '${n}'`),T(r.rank===2,()=>`boxes must be a 2D tensor, but was of rank '${r.rank}'`),T(r.shape[1]===4,()=>`boxes must have 4 columns, but 2nd dimension was ${r.shape[1]}`),T(e.rank===1,()=>"scores must be a 1D tensor"),T(e.shape[0]===a,()=>`scores has incompatible shape with boxes. Expected ${a}, but was ${e.shape[0]}`),T(0<=s&&s<=1,()=>`softNmsSigma must be in [0, 1], but was '${s}'`),{maxOutputSize:t,iouThreshold:n,scoreThreshold:o,softNmsSigma:s}}function xq(r,e,t,n=.5,o=Number.NEGATIVE_INFINITY){let s=k(r,"boxes","nonMaxSuppression"),a=k(e,"scores","nonMaxSuppression"),i=oo(s,a,t,n,o);t=i.maxOutputSize,n=i.iouThreshold,o=i.scoreThreshold;let l={maxOutputSize:t,iouThreshold:n,scoreThreshold:o};return A.runKernel(Yi,{boxes:s,scores:a},l)}var y1=S({nonMaxSuppression_:xq});function b1(r,e,t){let n=yq(r,e,t),o=n<0?-(n+1):n;r.splice(o,0,e)}function yq(r,e,t){return wq(r,e,t||bq)}function bq(r,e){return r>e?1:r<e?-1:0}function wq(r,e,t){let n=0,o=r.length,s=0,a=!1;for(;n<o;){s=n+(o-n>>>1);let i=t(e,r[s]);i>0?n=s+1:(o=s,a=!i)}return a?n:-n-1}function Dg(r,e,t,n,o){return N_(r,e,t,n,o,0)}function $g(r,e,t,n,o,s){return N_(r,e,t,n,o,0,!1,s,!0)}function Rg(r,e,t,n,o,s){return N_(r,e,t,n,o,s,!0)}function N_(r,e,t,n,o,s,a=!1,i=!1,l=!1){let u=[];for(let g=0;g<e.length;g++)e[g]>o&&u.push({score:e[g],boxIndex:g,suppressBeginIndex:0});u.sort(w1);let c=s>0?-.5/s:0,p=[],m=[];for(;p.length<t&&u.length>0;){let g=u.pop(),{score:x,boxIndex:b,suppressBeginIndex:w}=g;if(x<o)break;let _=!1;for(let I=p.length-1;I>=w;--I){let E=_q(r,b,p[I]);if(E>=n){_=!0;break}if(g.score=g.score*kq(n,c,E),g.score<=o)break}g.suppressBeginIndex=p.length,_||(g.score===x?(p.push(b),m.push(g.score)):g.score>o&&b1(u,g,w1))}let f=p.length,d=t-f;i&&d>0&&(p.push(...new Array(d).fill(0)),m.push(...new Array(d).fill(0)));let h={selectedIndices:p};return a&&(h.selectedScores=m),l&&(h.validOutputs=f),h}function _q(r,e,t){let n=r.subarray(e*4,e*4+4),o=r.subarray(t*4,t*4+4),s=Math.min(n[0],n[2]),a=Math.min(n[1],n[3]),i=Math.max(n[0],n[2]),l=Math.max(n[1],n[3]),u=Math.min(o[0],o[2]),c=Math.min(o[1],o[3]),p=Math.max(o[0],o[2]),m=Math.max(o[1],o[3]),f=(i-s)*(l-a),d=(p-u)*(m-c);if(f<=0||d<=0)return 0;let h=Math.max(s,u),g=Math.max(a,c),x=Math.min(i,p),b=Math.min(l,m),w=Math.max(x-h,0)*Math.max(b-g,0);return w/(f+d-w)}function kq(r,e,t){let n=Math.exp(e*t*t);return t<=r?n:0}function w1(r,e){return r.score-e.score||r.score===e.score&&e.boxIndex-r.boxIndex}async function vq(r,e,t,n=.5,o=Number.NEGATIVE_INFINITY){let s=k(r,"boxes","nonMaxSuppressionAsync"),a=k(e,"scores","nonMaxSuppressionAsync"),i=oo(s,a,t,n,o);t=i.maxOutputSize,n=i.iouThreshold,o=i.scoreThreshold;let l=await Promise.all([s.data(),a.data()]),u=l[0],c=l[1],{selectedIndices:p}=Dg(u,c,t,n,o);return s!==r&&s.dispose(),a!==e&&a.dispose(),Ct(p,"int32")}var _1=vq;function Cq(r,e,t,n=.5,o=Number.NEGATIVE_INFINITY,s=0){let a=k(r,"boxes","nonMaxSuppression"),i=k(e,"scores","nonMaxSuppression"),l=oo(a,i,t,n,o,s);t=l.maxOutputSize,n=l.iouThreshold,o=l.scoreThreshold,s=l.softNmsSigma;let u={boxes:a,scores:i},c={maxOutputSize:t,iouThreshold:n,scoreThreshold:o,softNmsSigma:s},p=A.runKernel(Ji,u,c);return{selectedIndices:p[0],selectedScores:p[1]}}var k1=S({nonMaxSuppressionWithScore_:Cq});async function Iq(r,e,t,n=.5,o=Number.NEGATIVE_INFINITY,s=0){let a=k(r,"boxes","nonMaxSuppressionAsync"),i=k(e,"scores","nonMaxSuppressionAsync"),l=oo(a,i,t,n,o,s);t=l.maxOutputSize,n=l.iouThreshold,o=l.scoreThreshold,s=l.softNmsSigma;let u=await Promise.all([a.data(),i.data()]),c=u[0],p=u[1],{selectedIndices:m,selectedScores:f}=Rg(c,p,t,n,o,s);return a!==r&&a.dispose(),i!==e&&i.dispose(),{selectedIndices:Ct(m,"int32"),selectedScores:Ct(f)}}var v1=Iq;function Nq(r,e,t,n=.5,o=Number.NEGATIVE_INFINITY,s=!1){let a=k(r,"boxes","nonMaxSuppression"),i=k(e,"scores","nonMaxSuppression"),l=oo(a,i,t,n,o,null),u=l.maxOutputSize,c=l.iouThreshold,p=l.scoreThreshold,m={boxes:a,scores:i},f={maxOutputSize:u,iouThreshold:c,scoreThreshold:p,padToMaxOutputSize:s},d=A.runKernel(Zi,m,f);return{selectedIndices:d[0],validOutputs:d[1]}}var C1=S({nonMaxSuppressionPadded_:Nq});async function Sq(r,e,t,n=.5,o=Number.NEGATIVE_INFINITY,s=!1){let a=k(r,"boxes","nonMaxSuppressionAsync"),i=k(e,"scores","nonMaxSuppressionAsync"),l=oo(a,i,t,n,o,null),u=l.maxOutputSize,c=l.iouThreshold,p=l.scoreThreshold,[m,f]=await Promise.all([a.data(),i.data()]),{selectedIndices:d,validOutputs:h}=$g(m,f,u,c,p,s);return a!==r&&a.dispose(),i!==e&&i.dispose(),{selectedIndices:Ct(d,"int32"),validOutputs:ue(h,"int32")}}var I1=Sq;function Tq(r,e,t=!1,n=!1){let o=k(r,"images","resizeBilinear");T(o.rank===3||o.rank===4,()=>`Error in resizeBilinear: x must be rank 3 or 4, but got rank ${o.rank}.`),T(e.length===2,()=>`Error in resizeBilinear: new shape must 2D, but got shape ${e}.`),T(n===!1||t===!1,()=>"Error in resizeBilinear: If halfPixelCenters is true, alignCorners must be false.");let s=o,a=!1;o.rank===3&&(a=!0,s=L(o,[1,o.shape[0],o.shape[1],o.shape[2]]));let[]=e,i={images:s},l={alignCorners:t,halfPixelCenters:n,size:e},u=A.runKernel(Jo,i,l);return a?L(u,[u.shape[1],u.shape[2],u.shape[3]]):u}var Fg=S({resizeBilinear_:Tq});function Aq(r,e,t=!1,n=!1){let o=k(r,"images","resizeNearestNeighbor");T(o.rank===3||o.rank===4,()=>`Error in resizeNearestNeighbor: x must be rank 3 or 4, but got rank ${o.rank}.`),T(e.length===2,()=>`Error in resizeNearestNeighbor: new shape must 2D, but got shape ${e}.`),T(o.dtype==="float32"||o.dtype==="int32",()=>"`images` must have `int32` or `float32` as dtype"),T(n===!1||t===!1,()=>"Error in resizeNearestNeighbor: If halfPixelCenters is true, alignCorners must be false.");let s=o,a=!1;o.rank===3&&(a=!0,s=L(o,[1,o.shape[0],o.shape[1],o.shape[2]]));let[]=e,i={images:s},l={alignCorners:t,halfPixelCenters:n,size:e},u=A.runKernel(il,i,l);return a?L(u,[u.shape[1],u.shape[2],u.shape[3]]):u}var Og=S({resizeNearestNeighbor_:Aq});function Eq(r,e="binary",t=!1,n=.5){let o=k(r,"image","threshold"),s=.2989,a=.587,i=.114,l=o.shape[0]*o.shape[1],u=O(Ct([n]),255),c,p,m,f;if(T(o.rank===3,()=>`Error in threshold: image must be rank 3,but got rank ${o.rank}.`),T(o.shape[2]===3||o.shape[2]===1,()=>`Error in threshold: image color channel must be equal to 3 or 1but got ${o.shape[2]}.`),T(o.dtype==="int32"||o.dtype==="float32",()=>`Error in dtype: image dtype must be int32 or float32,but got dtype ${o.dtype}.`),T(e==="otsu"||e==="binary",()=>`Method must be binary or otsu, but was ${e}`),o.shape[2]===3){[c,p,m]=tr(o,[1,1,1],-1);let g=O(c,s),x=O(p,a),b=O(m,i);f=J(J(g,x),b)}else f=r;if(e==="otsu"){let g=Zm(oe(gu(f),"int32"),vr([]),256);u=Dq(g,l)}let d=t?gn(f,u):Vt(f,u);return oe(O(d,255),"int32")}function Dq(r,e){let t=Ct([-1]),n=Ct([0]),o=Ct([0]),s,a,i,l,u,c;for(let p=0;p<r.size-1;p++){s=Re(r,0,p+1),a=Re(r,p+1),u=pe(fe(s),e),c=pe(fe(a),e);let m=fe(O(s,Ca(0,s.size)));i=pe(m,fe(s));let f=bs(a.shape,s.size),d=J(Ca(0,a.size),f),h=O(a,d);l=pe(fe(h),fe(a));let g=le(i,l),x=le(i,l),b=O(u,c);o=O(O(b,g),x);let w=Vt(o,n);n=vt(w,o,n),t=vt(w,Ct([p]),t)}return t}var N1=S({threshold_:Eq});function $q(r,e,t="nearest",n="constant",o=0,s){let a=k(r,"image","transform","float32"),i=k(e,"transforms","transform","float32");T(a.rank===4,()=>`Error in transform: image must be rank 4,but got rank ${a.rank}.`),T(i.rank===2&&(i.shape[0]===a.shape[0]||i.shape[0]===1)&&i.shape[1]===8,()=>"Error in transform: Input transform should be batch x 8 or 1 x 8"),T(s==null||s.length===2,()=>`Error in transform: outputShape must be [height, width] or null, but got ${s}.`);let l={image:a,transforms:i},u={interpolation:t,fillMode:n,fillValue:o,outputShape:s};return A.runKernel(la,l,u)}var S1=S({transform_:$q});function Rq(r,e,t){T(e%1==0,()=>`bandPart(): numLower must be an integer, got ${e}.`),T(t%1==0,()=>`bandPart(): numUpper must be an integer, got ${t}.`);let n=k(r,"a","bandPart");T(n.rank>=2,()=>`bandPart(): Rank must be at least 2, got ${n.rank}.`);let o=n.shape,[s,a]=n.shape.slice(-2);if(!(e<=s))throw new Error(`bandPart(): numLower (${e}) must not be greater than the number of rows (${s}).`);if(!(t<=a))throw new Error(`bandPart(): numUpper (${t}) must not be greater than the number of columns (${a}).`);e<0&&(e=s),t<0&&(t=a);let i=L(Ca(0,s,1,"int32"),[-1,1]),l=Ca(0,a,1,"int32"),u=le(i,l),c=yr(gn(u,ue(+e,"int32")),hn(u,ue(-t,"int32"))),p=ht([s,a],n.dtype);return L(Gt(fr(L(n,[-1,s,a])).map(m=>vt(c,m,p))),o)}var T1=S({bandPart_:Rq});function Fq(r){let e;if(Array.isArray(r)){e=!1,T(r!=null&&r.length>0,()=>"Gram-Schmidt process: input must not be null, undefined, or empty");let o=r[0].shape[0];for(let s=1;s<r.length;++s)T(r[s].shape[0]===o,()=>`Gram-Schmidt: Non-unique lengths found in the input vectors: (${r[s].shape[0]} vs. ${o})`)}else e=!0,r=tr(r,r.shape[0],0).map(o=>Tn(o,[0]));T(r.length<=r[0].shape[0],()=>`Gram-Schmidt: Number of vectors (${r.length}) exceeds number of dimensions (${r[0].shape[0]}).`);let t=[],n=r;for(let o=0;o<r.length;++o)t.push(A.tidy(()=>{let s=n[o];if(o>0)for(let a=0;a<o;++a){let i=O(fe(O(t[a],s)),t[a]);s=le(s,i)}return pe(s,pp(s,"euclidean"))}));return e?Gt(t,0):t}var A1=S({gramSchmidt_:Fq});function Oq(r,e=!1){if(T(r.rank>=2,()=>`qr() requires input tensor to have a rank >= 2, but got rank ${r.rank}`),r.rank===2)return E1(r,e);{let t=r.shape.slice(0,r.shape.length-2).reduce((l,u)=>l*u),n=fr(L(r,[t,r.shape[r.shape.length-2],r.shape[r.shape.length-1]]),0),o=[],s=[];n.forEach(l=>{let[u,c]=E1(l,e);o.push(u),s.push(c)});let a=L(Gt(o,0),r.shape),i=L(Gt(s,0),r.shape);return[a,i]}}function E1(r,e=!1){return A.tidy(()=>{T(r.shape.length===2,()=>`qr2d() requires a 2D Tensor, but got a ${r.shape.length}D Tensor.`);let t=r.shape[0],n=r.shape[1],o=ip(t),s=Ln(r),a=ii([[1]],[1,1]),i=Ln(a),l=t>=n?n:t;for(let u=0;u<l;++u){let c=s,p=i,m=o;[i,s,o]=A.tidy(()=>{let f=Re(s,[u,u],[t-u,1]),d=pp(f),h=Re(s,[u,u],[1,1]),g=vt(Vt(h,0),ii([[-1]]),ii([[1]])),x=le(h,O(g,d)),b=pe(f,x);b.shape[0]===1?i=Ln(a):i=Qe([a,Re(b,[1,0],[b.shape[0]-1,b.shape[1]])],0);let w=qe(pe(ze(g,x),d)),_=Re(s,[u,0],[t-u,n]),I=O(w,i),E=Ue(i);if(u===0)s=le(_,ze(I,ze(E,_)));else{let R=le(_,ze(I,ze(E,_)));s=Qe([Re(s,[0,0],[u,n]),R],0)}let D=Ue(I),$=Re(o,[0,u],[t,o.shape[1]-u]);if(u===0)o=le($,ze(ze($,i),D));else{let R=le($,ze(ze($,i),D));o=Qe([Re(o,[0,0],[t,u]),R],1)}return[i,s,o]}),Ee([c,p,m])}return!e&&t>n&&(o=Re(o,[0,0],[t,n]),s=Re(s,[0,0],[n,n])),[o,s]})}var D1=S({qr_:Oq});var Wt;(function(r){r[r.NONE=0]="NONE",r[r.MEAN=1]="MEAN",r[r.SUM=2]="SUM",r[r.SUM_BY_NONZERO_WEIGHTS=3]="SUM_BY_NONZERO_WEIGHTS"})(Wt||(Wt={}));function Pq(r,e,t=Wt.SUM_BY_NONZERO_WEIGHTS){let n=k(r,"losses","computeWeightedLoss"),o=null;e!=null&&(o=k(e,"weights","computeWeightedLoss"));let s=o==null?n:O(n,o);if(t===Wt.NONE)return s;if(t===Wt.SUM)return fe(s);if(t===Wt.MEAN){if(o==null)return dt(s);{let a=n.size/o.size,i=pe(fe(s),fe(o));return a>1?pe(i,ue(a)):i}}if(t===Wt.SUM_BY_NONZERO_WEIGHTS){if(o==null)return pe(fe(s),ue(n.size));{let a=O(o,er(n.shape)),i=oe(fe(ro(a,ue(0))),"float32");return pe(fe(s),i)}}throw Error(`Unknown reduction: ${t}`)}var Rr=S({computeWeightedLoss_:Pq});function Mq(r,e,t,n=Wt.SUM_BY_NONZERO_WEIGHTS){let o=k(r,"labels","absoluteDifference"),s=k(e,"predictions","absoluteDifference"),a=null;t!=null&&(a=k(t,"weights","absoluteDifference")),Dt(o.shape,s.shape,"Error in absoluteDifference: ");let i=Tt(le(o,s));return Rr(i,a,n)}var $1=S({absoluteDifference_:Mq});function Lq(r,e,t,n,o=Wt.SUM_BY_NONZERO_WEIGHTS){let s=k(r,"labels","cosineDistance"),a=k(e,"predictions","cosineDistance"),i=null;n!=null&&(i=k(n,"weights","cosineDistance")),Dt(s.shape,a.shape,"Error in cosineDistance: ");let l=ue(1),u=le(l,fe(O(s,a),t,!0));return Rr(u,i,o)}var R1=S({cosineDistance_:Lq});function zq(r,e,t,n=Wt.SUM_BY_NONZERO_WEIGHTS){let o=k(r,"labels","hingeLoss"),s=k(e,"predictions","hingeLoss"),a=null;t!=null&&(a=k(t,"weights","hingeLoss")),Dt(o.shape,s.shape,"Error in hingeLoss: ");let i=ue(1);o=le(O(ue(2),o),i);let l=$r(le(i,O(o,s)));return Rr(l,a,n)}var F1=S({hingeLoss_:zq});function Bq(r,e,t,n=1,o=Wt.SUM_BY_NONZERO_WEIGHTS){let s=k(r,"labels","huberLoss"),a=k(e,"predictions","huberLoss"),i=null;t!=null&&(i=k(t,"weights","huberLoss")),Dt(s.shape,a.shape,"Error in huberLoss: ");let l=ue(n),u=Tt(le(a,s)),c=_s(u,l),p=le(u,c),m=J(O(ue(.5),Me(c)),O(l,p));return Rr(m,i,o)}var O1=S({huberLoss_:Bq});function Vq(r,e,t,n=1e-7,o=Wt.SUM_BY_NONZERO_WEIGHTS){let s=k(r,"labels","logLoss"),a=k(e,"predictions","logLoss"),i=null;t!=null&&(i=k(t,"weights","logLoss")),Dt(s.shape,a.shape,"Error in logLoss: ");let l=ue(1),u=ue(n),c=qe(O(s,pr(J(a,u)))),p=O(le(l,s),pr(J(le(l,a),u))),m=le(c,p);return Rr(m,i,o)}var P1=S({logLoss_:Vq});function Gq(r,e,t,n=Wt.SUM_BY_NONZERO_WEIGHTS){let o=k(r,"labels","meanSquaredError"),s=k(e,"predictions","meanSquaredError"),a=null;t!=null&&(a=k(t,"weights","meanSquaredError")),Dt(o.shape,s.shape,"Error in meanSquaredError: ");let i=ku(o,s);return Rr(i,a,n)}var M1=S({meanSquaredError_:Gq});function Wq(r,e){let t=k(r,"labels","sigmoidCrossEntropyWithLogits"),n=k(e,"logits","sigmoidCrossEntropyWithLogits");Dt(t.shape,n.shape,"Error in sigmoidCrossEntropyWithLogits: ");let o=$r(n),s=O(n,t),a=cu(Qt(qe(Tt(n))));return J(le(o,s),a)}function jq(r,e,t,n=0,o=Wt.SUM_BY_NONZERO_WEIGHTS){let s=k(r,"multiClassLabels","sigmoidCrossEntropy"),a=k(e,"logits","sigmoidCrossEntropy"),i=null;if(t!=null&&(i=k(t,"weights","sigmoidCrossEntropy")),Dt(s.shape,a.shape,"Error in sigmoidCrossEntropy: "),n>0){let u=ue(n),c=ue(1),p=ue(.5);s=J(O(s,le(c,u)),O(p,u))}let l=Wq(s,a);return Rr(l,i,o)}var L1=S({sigmoidCrossEntropy_:jq});function Uq(r,e,t=-1){if(t===-1&&(t=e.rank-1),t!==e.rank-1)throw Error(`Softmax cross entropy along a non-last dimension is not yet supported. Labels / logits was rank ${e.rank} and dim was ${t}`);return Qr((o,s,a)=>{let l=uf(s,[t],!0),u=le(oe(s,"float32"),l);a([o,u]);let c=qe(O(u,o));return{value:fe(c,[t]),gradFunc:(f,d)=>{let[h,g]=d,x=to(f.shape,[t]);return[O(L(f,x),le(oe(h,"float32"),Qt(g))),O(L(f,x),le(Qt(g),oe(h,"float32")))]}}})(r,e)}function Hq(r,e,t,n=0,o=Wt.SUM_BY_NONZERO_WEIGHTS){let s=k(r,"onehotLabels","softmaxCrossEntropy"),a=k(e,"logits","softmaxCrossEntropy"),i=null;if(t!=null&&(i=k(t,"weights","softmaxCrossEntropy")),Dt(s.shape,a.shape,"Error in softmaxCrossEntropy: "),n>0){let u=ue(n),c=ue(1),p=ue(s.shape[1]);s=J(O(s,le(c,u)),pe(u,p))}let l=Uq(s,a);return Rr(l,i,o)}var z1=S({softmaxCrossEntropy_:Hq});function qq(r,e,t,n){let o=k(r,"indices","sparseFillEmptyRows"),s=k(e,"values","sparseFillEmptyRows"),a=k(t,"denseShape","sparseFillEmptyRows"),i=k(n,"defaultValue","sparseFillEmptyRows",s.dtype);if(o.rank!==2)throw new Error(`Indices should be Tensor2D but received shape
|
|
${o.shape}`);if(s.rank!==1)throw new Error(`Values should be Tensor1D but received shape ${s.shape}`);if(a.rank!==1)throw new Error(`Dense shape should be Tensor1D but received shape ${a.shape}`);if(i.rank!==0)throw new Error(`Default value should be a scalar but received shape ${i.shape}`);let l={indices:o,values:s,denseShape:a,defaultValue:i},u=A.runKernel(Uc,l);return{outputIndices:u[0],outputValues:u[1],emptyRowIndicator:u[2],reverseIndexMap:u[3]}}var B1=S({sparseFillEmptyRows_:qq});function Kq(r,e,t){let n=k(r,"inputIndices","sparseReshape"),o=k(e,"inputShape","sparseReshape"),s=k(t,"newShape","sparseReshape");if(n.rank!==2)throw new Error(`Input indices should be Tensor2D but received shape
|
|
${n.shape}`);if(o.rank!==1)throw new Error(`Input shape should be Tensor1D but received shape ${o.shape}`);if(s.rank!==1)throw new Error(`New shape should be Tensor1D but received shape ${s.shape}`);let a={inputIndices:n,inputShape:o,newShape:s},i=A.runKernel(Hc,a);return{outputIndices:i[0],outputShape:i[1]}}var V1=S({sparseReshape_:Kq});var vDe={fft:Na,ifft:si,rfft:Sa,irfft:_u},TDe={hammingWindow:f1,hannWindow:Ag,frame:Eg,stft:d1},ai={flipLeftRight:g1,resizeNearestNeighbor:Og,resizeBilinear:Fg,rotateWithOffset:x1,cropAndResize:h1,nonMaxSuppression:y1,nonMaxSuppressionAsync:_1,nonMaxSuppressionWithScore:k1,nonMaxSuppressionWithScoreAsync:v1,nonMaxSuppressionPadded:C1,nonMaxSuppressionPaddedAsync:I1,threshold:N1,transform:S1},G1={bandPart:T1,gramSchmidt:A1,qr:D1},e$e={absoluteDifference:$1,computeWeightedLoss:Rr,cosineDistance:R1,hingeLoss:F1,huberLoss:O1,logLoss:P1,meanSquaredError:M1,sigmoidCrossEntropy:L1,softmaxCrossEntropy:z1},W1={sparseFillEmptyRows:B1,sparseReshape:V1};var Br=class extends hg{minimize(e,t=!1,n){let{value:o,grads:s}=this.computeGradients(e,n);if(n!=null){let a=n.map(i=>({name:i.name,tensor:s[i.name]}));this.applyGradients(a)}else this.applyGradients(s);return Ee(s),t?o:(o.dispose(),null)}get iterations(){return this.iterations_==null&&(this.iterations_=0),this.iterations_}incrementIterations(){this.iterations_=this.iterations+1}computeGradients(e,t){return yg(e,t)}dispose(){this.iterations_!=null&&Ee(this.iterations_)}async saveIterations(){return this.iterations_==null&&(this.iterations_=0),{name:"iter",tensor:ue(this.iterations_,"int32")}}async getWeights(){throw new Error("getWeights() is not implemented for this optimizer yet.")}async setWeights(e){throw new Error(`setWeights() is not implemented for this optimizer class ${this.getClassName()}`)}async extractIterations(e){return this.iterations_=(await e[0].tensor.data())[0],e.slice(1)}};Object.defineProperty(Br,Symbol.hasInstance,{value:r=>r.minimize!=null&&r.computeGradients!=null&&r.applyGradients!=null});var fp=class extends Br{constructor(e,t,n=null){super();this.learningRate=e,this.rho=t,this.epsilon=n,this.accumulatedGrads=[],this.accumulatedUpdates=[],n==null&&(this.epsilon=A.backend.epsilon())}applyGradients(e){(Array.isArray(e)?e.map(n=>n.name):Object.keys(e)).forEach((n,o)=>{let s=A.registeredVariables[n],a=!1;this.accumulatedGrads[o]==null&&(this.accumulatedGrads[o]={originalName:`${n}/accum_grad`,variable:V(()=>Ie(s).variable(a))}),this.accumulatedUpdates[o]==null&&(this.accumulatedUpdates[o]={originalName:`${n}/accum_var`,variable:V(()=>Ie(s).variable(a))});let i=Array.isArray(e)?e[o].tensor:e[n];if(i==null)return;let l=this.accumulatedGrads[o].variable,u=this.accumulatedUpdates[o].variable;V(()=>{let c=J(O(l,this.rho),O(Me(i),1-this.rho)),p=O(pe(gt(J(u,this.epsilon)),gt(J(l,this.epsilon))),i),m=J(O(u,this.rho),O(Me(p),1-this.rho));l.assign(c),u.assign(m);let f=J(O(p,-this.learningRate),s);s.assign(f)})}),this.incrementIterations()}dispose(){this.accumulatedUpdates!=null&&(Ee(this.accumulatedGrads.map(e=>e.variable)),Ee(this.accumulatedUpdates.map(e=>e.variable)))}async getWeights(){let e=[...this.accumulatedGrads,...this.accumulatedUpdates];return[await this.saveIterations()].concat(e.map(t=>({name:t.originalName,tensor:t.variable})))}async setWeights(e){e=await this.extractIterations(e);let t=e.length/2,n=!1;this.accumulatedGrads=e.slice(0,t).map(o=>({originalName:o.name,variable:o.tensor.variable(n)})),this.accumulatedUpdates=e.slice(t,t*2).map(o=>({originalName:o.name,variable:o.tensor.variable(n)}))}getConfig(){return{learningRate:this.learningRate,rho:this.rho,epsilon:this.epsilon}}static fromConfig(e,t){return new e(t.learningRate,t.rho,t.epsilon)}};fp.className="Adadelta";dn(fp);var dp=class extends Br{constructor(e,t=.1){super();this.learningRate=e,this.initialAccumulatorValue=t,this.accumulatedGrads=[]}applyGradients(e){(Array.isArray(e)?e.map(n=>n.name):Object.keys(e)).forEach((n,o)=>{let s=A.registeredVariables[n];if(this.accumulatedGrads[o]==null){let l=!1;this.accumulatedGrads[o]={originalName:`${n}/accumulator`,variable:V(()=>bs(s.shape,this.initialAccumulatorValue).variable(l))}}let a=Array.isArray(e)?e[o].tensor:e[n];if(a==null)return;let i=this.accumulatedGrads[o].variable;V(()=>{let l=J(i,Me(a));i.assign(l);let u=J(O(pe(a,gt(J(l,A.backend.epsilon()))),-this.learningRate),s);s.assign(u)})}),this.incrementIterations()}dispose(){this.accumulatedGrads!=null&&Ee(this.accumulatedGrads.map(e=>e.variable))}async getWeights(){return[await this.saveIterations()].concat(this.accumulatedGrads.map(e=>({name:e.originalName,tensor:e.variable})))}async setWeights(e){e=await this.extractIterations(e);let t=!1;this.accumulatedGrads=e.map(n=>({originalName:n.name,variable:n.tensor.variable(t)}))}getConfig(){return{learningRate:this.learningRate,initialAccumulatorValue:this.initialAccumulatorValue}}static fromConfig(e,t){return new e(t.learningRate,t.initialAccumulatorValue)}};dp.className="Adagrad";dn(dp);var hp=class extends Br{constructor(e,t,n,o=null){super();this.learningRate=e,this.beta1=t,this.beta2=n,this.epsilon=o,this.accumulatedFirstMoment=[],this.accumulatedSecondMoment=[],V(()=>{this.accBeta1=ue(t).variable(),this.accBeta2=ue(n).variable()}),o==null&&(this.epsilon=A.backend.epsilon())}applyGradients(e){let t=Array.isArray(e)?e.map(n=>n.name):Object.keys(e);V(()=>{let n=le(1,this.accBeta1),o=le(1,this.accBeta2);t.forEach((s,a)=>{let i=A.registeredVariables[s],l=!1;this.accumulatedFirstMoment[a]==null&&(this.accumulatedFirstMoment[a]={originalName:`${s}/m`,variable:V(()=>Ie(i).variable(l))}),this.accumulatedSecondMoment[a]==null&&(this.accumulatedSecondMoment[a]={originalName:`${s}/v`,variable:V(()=>Ie(i).variable(l))});let u=Array.isArray(e)?e[a].tensor:e[s];if(u==null)return;let c=this.accumulatedFirstMoment[a].variable,p=this.accumulatedSecondMoment[a].variable,m=J(O(c,this.beta1),O(u,1-this.beta1)),f=J(O(p,this.beta2),O(Me(u),1-this.beta2)),d=pe(m,n),h=pe(f,o);c.assign(m),p.assign(f);let g=J(O(pe(d,J(gt(h),this.epsilon)),-this.learningRate),i);i.assign(g)}),this.accBeta1.assign(O(this.accBeta1,this.beta1)),this.accBeta2.assign(O(this.accBeta2,this.beta2))}),this.incrementIterations()}dispose(){this.accBeta1.dispose(),this.accBeta2.dispose(),this.accumulatedFirstMoment!=null&&Ee(this.accumulatedFirstMoment.map(e=>e.variable)),this.accumulatedSecondMoment!=null&&Ee(this.accumulatedSecondMoment.map(e=>e.variable))}async getWeights(){let e=[...this.accumulatedFirstMoment,...this.accumulatedSecondMoment];return[await this.saveIterations()].concat(e.map(t=>({name:t.originalName,tensor:t.variable})))}async setWeights(e){e=await this.extractIterations(e),V(()=>{this.accBeta1.assign(zr(this.beta1,this.iterations_+1)),this.accBeta2.assign(zr(this.beta2,this.iterations_+1))});let t=e.length/2,n=!1;this.accumulatedFirstMoment=e.slice(0,t).map(o=>({originalName:o.name,variable:o.tensor.variable(n)})),this.accumulatedSecondMoment=e.slice(t,t*2).map(o=>({originalName:o.name,variable:o.tensor.variable(n)}))}getConfig(){return{learningRate:this.learningRate,beta1:this.beta1,beta2:this.beta2,epsilon:this.epsilon}}static fromConfig(e,t){return new e(t.learningRate,t.beta1,t.beta2,t.epsilon)}};hp.className="Adam";dn(hp);var gp=class extends Br{constructor(e,t,n,o=null,s=0){super();this.learningRate=e,this.beta1=t,this.beta2=n,this.epsilon=o,this.decay=s,this.accumulatedFirstMoment=[],this.accumulatedWeightedInfNorm=[],V(()=>{this.iteration=ue(0).variable(),this.accBeta1=ue(t).variable()}),o==null&&(this.epsilon=A.backend.epsilon())}applyGradients(e){let t=Array.isArray(e)?e.map(n=>n.name):Object.keys(e);V(()=>{let n=le(1,this.accBeta1),o=pe(-this.learningRate,J(O(this.iteration,this.decay),1));t.forEach((s,a)=>{let i=A.registeredVariables[s],l=!1;this.accumulatedFirstMoment[a]==null&&(this.accumulatedFirstMoment[a]={originalName:`${s}/m`,variable:Ie(i).variable(l)}),this.accumulatedWeightedInfNorm[a]==null&&(this.accumulatedWeightedInfNorm[a]={originalName:`${s}/v`,variable:Ie(i).variable(l)});let u=Array.isArray(e)?e[a].tensor:e[s];if(u==null)return;let c=this.accumulatedFirstMoment[a].variable,p=this.accumulatedWeightedInfNorm[a].variable,m=J(O(c,this.beta1),O(u,1-this.beta1)),f=O(p,this.beta2),d=Tt(u),h=en(f,d);c.assign(m),p.assign(h);let g=J(O(pe(o,n),pe(m,J(h,this.epsilon))),i);i.assign(g)}),this.iteration.assign(J(this.iteration,1)),this.accBeta1.assign(O(this.accBeta1,this.beta1))}),this.incrementIterations()}dispose(){this.accBeta1.dispose(),this.iteration.dispose(),this.accumulatedFirstMoment!=null&&Ee(this.accumulatedFirstMoment.map(e=>e.variable)),this.accumulatedWeightedInfNorm!=null&&Ee(this.accumulatedWeightedInfNorm.map(e=>e.variable))}async getWeights(){throw new Error("getWeights() is not implemented for Adamax yet.")}async setWeights(e){throw new Error("setWeights() is not implemented for Adamax yet.")}getConfig(){return{learningRate:this.learningRate,beta1:this.beta1,beta2:this.beta2,epsilon:this.epsilon,decay:this.decay}}static fromConfig(e,t){return new e(t.learningRate,t.beta1,t.beta2,t.epsilon,t.decay)}};gp.className="Adamax";dn(gp);var xl=class extends Br{constructor(e){super();this.learningRate=e,this.setLearningRate(e)}applyGradients(e){(Array.isArray(e)?e.map(n=>n.name):Object.keys(e)).forEach((n,o)=>{let s=Array.isArray(e)?e[o].tensor:e[n];if(s==null)return;let a=A.registeredVariables[n];V(()=>{let i=J(O(this.c,s),a);a.assign(i)})}),this.incrementIterations()}setLearningRate(e){this.learningRate=e,this.c!=null&&this.c.dispose(),this.c=$t(ue(-e))}dispose(){this.c.dispose()}async getWeights(){return[await this.saveIterations()]}async setWeights(e){if(e=await this.extractIterations(e),e.length!==0)throw new Error("SGD optimizer does not have settable weights.")}getConfig(){return{learningRate:this.learningRate}}static fromConfig(e,t){return new e(t.learningRate)}};xl.className="SGD";dn(xl);var xp=class extends xl{constructor(e,t,n=!1){super(e);this.learningRate=e,this.momentum=t,this.useNesterov=n,this.accumulations=[],this.m=ue(this.momentum)}applyGradients(e){(Array.isArray(e)?e.map(n=>n.name):Object.keys(e)).forEach((n,o)=>{let s=A.registeredVariables[n];if(this.accumulations[o]==null){let l=!1;this.accumulations[o]={originalName:`${n}/momentum`,variable:V(()=>Ie(s).variable(l))}}let a=this.accumulations[o].variable,i=Array.isArray(e)?e[o].tensor:e[n];i!=null&&V(()=>{let l,u=J(O(this.m,a),i);this.useNesterov?l=J(O(this.c,J(i,O(u,this.m))),s):l=J(O(this.c,u),s),a.assign(u),s.assign(l)})}),this.incrementIterations()}dispose(){this.m.dispose(),this.accumulations!=null&&Ee(this.accumulations.map(e=>e.variable))}setMomentum(e){this.momentum=e}async getWeights(){return[await this.saveIterations()].concat(this.accumulations.map(e=>({name:e.originalName,tensor:e.variable})))}async setWeights(e){e=await this.extractIterations(e);let t=!1;this.accumulations=e.map(n=>({originalName:n.name,variable:n.tensor.variable(t)}))}getConfig(){return{learningRate:this.learningRate,momentum:this.momentum,useNesterov:this.useNesterov}}static fromConfig(e,t){return new e(t.learningRate,t.momentum,t.useNesterov)}};xp.className="Momentum";dn(xp);var yp=class extends Br{constructor(e,t=.9,n=0,o=null,s=!1){super();if(this.learningRate=e,this.decay=t,this.momentum=n,this.epsilon=o,this.accumulatedMeanSquares=[],this.accumulatedMoments=[],this.accumulatedMeanGrads=[],this.centered=s,o==null&&(this.epsilon=A.backend.epsilon()),e==null)throw new Error("learningRate for RMSPropOptimizer must be defined.")}applyGradients(e){(Array.isArray(e)?e.map(n=>n.name):Object.keys(e)).forEach((n,o)=>{let s=A.registeredVariables[n],a=!1;this.accumulatedMeanSquares[o]==null&&(this.accumulatedMeanSquares[o]={originalName:`${n}/rms`,variable:V(()=>Ie(s).variable(a))}),this.accumulatedMoments[o]==null&&(this.accumulatedMoments[o]={originalName:`${n}/momentum`,variable:V(()=>Ie(s).variable(a))}),this.accumulatedMeanGrads[o]==null&&this.centered&&(this.accumulatedMeanGrads[o]={originalName:`${n}/mg`,variable:V(()=>Ie(s).variable(a))});let i=Array.isArray(e)?e[o].tensor:e[n];if(i==null)return;let l=this.accumulatedMeanSquares[o].variable,u=this.accumulatedMoments[o].variable;V(()=>{let c=J(O(l,this.decay),O(Me(i),1-this.decay));if(this.centered){let p=this.accumulatedMeanGrads[o].variable,m=J(O(p,this.decay),O(i,1-this.decay)),f=pe(O(i,this.learningRate),gt(le(c,J(Me(m),this.epsilon)))),d=J(O(u,this.momentum),f);l.assign(c),p.assign(m),u.assign(d);let h=le(s,d);s.assign(h)}else{let p=J(O(l,this.decay),O(Me(i),1-this.decay)),m=J(O(u,this.momentum),pe(O(i,this.learningRate),gt(J(p,this.epsilon))));l.assign(p),u.assign(m);let f=le(s,m);s.assign(f)}})}),this.incrementIterations()}dispose(){this.accumulatedMeanSquares!=null&&Ee(this.accumulatedMeanSquares.map(e=>e.variable)),this.accumulatedMeanGrads!=null&&this.centered&&Ee(this.accumulatedMeanGrads.map(e=>e.variable)),this.accumulatedMoments!=null&&Ee(this.accumulatedMoments.map(e=>e.variable))}async getWeights(){let e=[...this.accumulatedMeanSquares,...this.accumulatedMoments];return this.centered&&e.push(...this.accumulatedMeanGrads),[await this.saveIterations()].concat(e.map(t=>({name:t.originalName,tensor:t.variable})))}async setWeights(e){e=await this.extractIterations(e);let t=this.centered?e.length/3:e.length/2,n=!1;this.accumulatedMeanSquares=e.slice(0,t).map(o=>({originalName:o.name,variable:o.tensor.variable(n)})),this.accumulatedMoments=e.slice(t,t*2).map(o=>({originalName:o.name,variable:o.tensor.variable(n)})),this.centered&&(this.accumulatedMeanGrads=e.slice(t*2,t*3).map(o=>({originalName:o.name,variable:o.tensor.variable(n)})))}getConfig(){return{learningRate:this.learningRate,decay:this.decay,momentum:this.momentum,epsilon:this.epsilon,centered:this.centered}}static fromConfig(e,t){return new e(t.learningRate,t.decay,t.momentum,t.epsilon,t.centered)}};yp.className="RMSProp";dn(yp);var Ta=class{static sgd(e){return new xl(e)}static momentum(e,t,n=!1){return new xp(e,t,n)}static rmsprop(e,t=.9,n=0,o=null,s=!1){return new yp(e,t,n,o,s)}static adam(e=.001,t=.9,n=.999,o=null){return new hp(e,t,n,o)}static adadelta(e=.001,t=.95,n=null){return new fp(e,t,n)}static adamax(e=.002,t=.9,n=.999,o=null,s=0){return new gp(e,t,n,o,s)}static adagrad(e,t=.1){return new dp(e,t)}};var Tu={sgd:Ta.sgd,momentum:Ta.momentum,adadelta:Ta.adadelta,adagrad:Ta.adagrad,rmsprop:Ta.rmsprop,adamax:Ta.adamax,adam:Ta.adam};var Xq=(()=>typeof requestAnimationFrame!="undefined"?requestAnimationFrame:typeof setImmediate!="undefined"?setImmediate:r=>r())();function vf(){return new Promise(r=>Xq(()=>r()))}var C={};We(C,{ERF_A1:()=>iK,ERF_A2:()=>aK,ERF_A3:()=>lK,ERF_A4:()=>uK,ERF_A5:()=>cK,ERF_P:()=>sK,PARALLELIZE_THRESHOLD:()=>Pg,SELU_SCALE:()=>T_,SELU_SCALEALPHA:()=>S_,applyActivation:()=>Nu,assertAndGetBroadcastShape:()=>Be,assertAxesAreInnerMostDims:()=>nU,assertParamsConsistent:()=>Yq,assignToTypedArray:()=>yK,axesAreInnerMostDims:()=>a_,calculateShapes:()=>CT,checkEinsumDimSizes:()=>CK,combineLocations:()=>GT,complexWithEvenIndex:()=>hK,complexWithOddIndex:()=>gK,computeConv2DInfo:()=>ru,computeConv3DInfo:()=>zT,computeDefaultPad:()=>jw,computeDilation2DInfo:()=>Tj,computeOptimalWindowSize:()=>Jq,computeOutAndReduceShapes:()=>l_,computeOutShape:()=>Zq,computePool2DInfo:()=>Ww,computePool3DInfo:()=>Aj,convertConv2DDataFormat:()=>BT,decodeEinsumEquation:()=>kK,eitherStridesOrDilationsAreOne:()=>Cr,expandShapeToKeepDim:()=>to,exponent:()=>wK,exponents:()=>bK,fromStringArrayToUint8:()=>RK,fromUint8ToStringArray:()=>$K,getAxesPermutation:()=>u_,getBroadcastDims:()=>y4,getComplexWithIndex:()=>xK,getEinsumComputePath:()=>IK,getEinsumPermutation:()=>vK,getFusedBiasGradient:()=>Iu,getFusedDyActivation:()=>Cu,getImageCenter:()=>Qq,getInnerMostAxes:()=>oU,getPermuted:()=>tK,getReductionAxes:()=>kt,getReshaped:()=>eK,getReshapedPermuted:()=>rK,getSliceBeginCoords:()=>nK,getSliceSize:()=>oK,getUndoAxesPermutation:()=>lf,isIdentityPermutation:()=>NK,log:()=>mK,mergeRealAndImagArrays:()=>fK,prepareAndValidate:()=>vT,prepareSplitSize:()=>TK,segment_util:()=>E_,shouldFuse:()=>Su,slice_util:()=>or,splitRealAndImagArrays:()=>dK,tupleValuesAreOne:()=>zn,upcastType:()=>lr,validateInput:()=>mg,validateUpdateShape:()=>Pw,warn:()=>pK});function Yq(r,e){let t=r[0].length;r.forEach((o,s)=>{T(o.length===t,()=>`Error in concat${t}D: rank of tensors[${s}] must be the same as the rank of the rest (${t})`)}),T(e>=0&&e<t,()=>`Error in concat${t}D: axis must be between 0 and ${t-1}.`);let n=r[0];r.forEach((o,s)=>{for(let a=0;a<t;a++)T(a===e||o[a]===n[a],()=>`Error in concat${t}D: Shape of tensors[${s}] (${o}) does not match the shape of the rest (${n}) along the non-concatenated axis ${s}.`)})}function Zq(r,e){let t=r[0].slice();for(let n=1;n<r.length;n++)t[e]+=r[n][e];return t}var Pg=30;function Jq(r){return r<=Pg?r:yc(r,Math.floor(Math.sqrt(r)))}function Qq(r,e,t){let n=t*(typeof r=="number"?r:r[0]),o=e*(typeof r=="number"?r:r[1]);return[n,o]}function eK(r,e,t,n=!0){let o=[];if(n)o=o.concat(e.slice(0)),o.push(r[0]/t),o=o.concat(r.slice(1));else{o=o.concat(r[0]);let s=e.length;for(let a=0;a<s;++a)o=o.concat([r[a+1]/e[a],e[a]]);o=o.concat(r.slice(s+1))}return o}function tK(r,e,t=!0){let n=[];if(t){n.push(e);for(let o=e+1;o<r;++o)o<=2*e?(n.push(o),n.push(o-(e+1))):n.push(o)}else{let o=[],s=[];for(let a=1;a<r;++a)a>=e*2+1||a%2==1?s.push(a):o.push(a);n.push(...o),n.push(0),n.push(...s)}return n}function rK(r,e,t,n=!0){let o=[];n?o.push(r[0]/t):o.push(r[0]*t);for(let s=1;s<r.length;++s)s<=e.length?n?o.push(e[s-1]*r[s]):o.push(r[s]/e[s-1]):o.push(r[s]);return o}function nK(r,e){let t=[0];for(let n=0;n<e;++n)t.push(r[n][0]);return t}function oK(r,e,t){let n=r.slice(0,1);for(let o=0;o<t;++o)n.push(r[o+1]-e[o][0]-e[o][1]);return n}var S_=1.7580993408473768,T_=1.0507009873554805;var sK=.3275911,iK=.254829592,aK=-.284496736,lK=1.421413741,uK=-1.453152027,cK=1.061405429;function pK(...r){W().getBool("IS_TEST")||console.warn(...r)}function mK(...r){W().getBool("IS_TEST")||console.log(...r)}function fK(r,e){if(r.length!==e.length)throw new Error(`Cannot merge real and imag arrays of different lengths. real:${r.length}, imag: ${e.length}.`);let t=new Float32Array(r.length*2);for(let n=0;n<t.length;n+=2)t[n]=r[n/2],t[n+1]=e[n/2];return t}function dK(r){let e=new Float32Array(r.length/2),t=new Float32Array(r.length/2);for(let n=0;n<r.length;n+=2)e[n/2]=r[n],t[n/2]=r[n+1];return{real:e,imag:t}}function hK(r){let e=Math.ceil(r.length/4),t=new Float32Array(e),n=new Float32Array(e);for(let o=0;o<r.length;o+=4)t[Math.floor(o/4)]=r[o],n[Math.floor(o/4)]=r[o+1];return{real:t,imag:n}}function gK(r){let e=Math.floor(r.length/4),t=new Float32Array(e),n=new Float32Array(e);for(let o=2;o<r.length;o+=4)t[Math.floor(o/4)]=r[o],n[Math.floor(o/4)]=r[o+1];return{real:t,imag:n}}function xK(r,e){let t=r[e*2],n=r[e*2+1];return{real:t,imag:n}}function yK(r,e,t,n){r[n*2]=e,r[n*2+1]=t}function bK(r,e){let t=new Float32Array(r/2),n=new Float32Array(r/2);for(let o=0;o<Math.ceil(r/2);o++){let s=(e?2:-2)*Math.PI*(o/r);t[o]=Math.cos(s),n[o]=Math.sin(s)}return{real:t,imag:n}}function wK(r,e,t){let n=(t?2:-2)*Math.PI*(r/e),o=Math.cos(n),s=Math.sin(n);return{real:o,imag:s}}var A_="->",_K=/->/g,j1=",",U1="...";function kK(r,e){r=r.replace(/\s/g,"");let t=(r.length-r.replace(_K,"").length)/A_.length;if(t<1)throw new Error("Equations without an arrow are not supported.");if(t>1)throw new Error(`Equation must contain exactly one arrow ("${A_}").`);let[n,o]=r.split(A_);T(n.indexOf(U1)===-1,()=>`The ellipsis notation ("${U1}") is not supported yet.`);let s=n.split(j1),a=s.length;if(e!==a)throw new Error(`Expected ${a} input tensors, received ${e}`);if(a>2)throw new Error("Support for more than 2 input tensors is not implemented yet.");let i=[];for(let m=0;m<o.length;++m){let f=o[m];if(!s.some(d=>d.indexOf(f)!==-1))throw new Error(`Output subscripts contain the label ${f} not present in the input subscripts.`);i.indexOf(f)===-1&&i.push(f)}for(let m=0;m<n.length;++m){let f=n[m];i.indexOf(f)===-1&&f!==j1&&i.push(f)}let l=new Array(s.length);for(let m=0;m<a;++m){if(new Set(s[m].split("")).size!==s[m].length)throw new Error(`Found duplicate axes in input component ${s[m]}. Support for duplicate axes in input is not implemented yet.`);l[m]=[];for(let f=0;f<s[m].length;++f)l[m].push(i.indexOf(s[m][f]))}let u=i.length,c=o.length,p=[];for(let m=c;m<u;++m)p.push(m);return{allDims:i,summedDims:p,idDims:l}}function vK(r,e){let t=new Array(r);t.fill(-1);for(let o=0;o<e.length;++o)t[e[o]]=o;let n=[];for(let o=0;o<r;++o)t[o]===-1&&n.push(o);return t=t.filter(o=>o!==-1),{permutationIndices:t,expandDims:n}}function CK(r,e,t){let n=new Array(r);for(let o=0;o<t.length;++o){let s=t[o].shape;for(let a=0;a<e[o].length;++a)n[e[o][a]]===void 0?n[e[o][a]]=s[a]:T(n[e[o][a]]===s[a],()=>`Expected dimension ${n[e[o][a]]} at axis ${a} of input shaped ${JSON.stringify(s)}, but got dimension ${s[a]}`)}}function IK(r,e){let t=r,n=[],o=0;r.length===0&&t.push(-1),o=r.length+1;for(let a=0;a<o;++a)n.push([]);let s=[];for(let a=0;a<t.length;++a){let i=t[a],l=SK(e,i);for(let u of l)s.indexOf(u)===-1&&(n[a].push(u),s.push(u))}return{path:t,steps:n}}function NK(r){return r.every((e,t)=>e===t)}function SK(r,e){let t=[];for(let n=0;n<r.length;++n)(r[n].length===0||r[n].indexOf(e)!==-1||e===-1)&&t.push(n);return t}function TK(r,e,t=0){let n=[];if(typeof e=="number")T(r.shape[t]%e==0,()=>"Number of splits must evenly divide the axis."),n=new Array(e).fill(r.shape[t]/e);else{let o=e.reduce((a,i)=>(i===-1&&(a+=1),a),0);T(o<=1,()=>"There should be only one negative value in split array.");let s=e.indexOf(-1);if(s!==-1){let a=e.reduce((i,l)=>l>0?i+l:i);e[s]=r.shape[t]-a}T(r.shape[t]===e.reduce((a,i)=>a+i),()=>"The sum of sizes must match the size of the axis dimension."),n=e}return n}var E_={};We(E_,{collectGatherOpShapeInfo:()=>DK,computeOutShape:()=>EK,segOpComputeOptimalWindowSize:()=>AK});function AK(r,e){let t=!1,n;for(r<=Pg?(n=r,t=!0):n=yc(r,Math.floor(Math.sqrt(r)));!t;)n>e||n===r?t=!0:n=yc(r,n+1);return n}function EK(r,e,t){let n=[],o=r.length;for(let s=0;s<o;s++)s!==e?n.push(r[s]):n.push(t);return n}function DK(r,e,t,n){let o=e.shape.length,s=r.shape.length;if(n!==0&&(n<-o||n>o))throw new Error(`Expect batchDims in the range of [-${o}, ${o}], but got ${n}`);if(n<0&&(n+=o),n>s)throw new Error(`batchDims (${n}) must be less than rank(x) (
|
|
${s}).`);if(t<n)throw new Error(`batchDims (${n}) must be less than or equal to axis (${t}).`);for(let p=0;p<n;++p)if(r.shape[p]!==e.shape[p])throw new Error(`x.shape[${p}]: ${r.shape[p]} should be equal to indices.shape[${p}]: ${e.shape[p]}.`);let a=r.shape[t],i=[],l=1,u=1,c=1;for(let p=0;p<n;++p)i.push(r.shape[p]),l*=r.shape[p];for(let p=n;p<t;p++)i.push(r.shape[p]),u*=r.shape[p];for(let p=n;p<o;p++)i.push(e.shape[p]);for(let p=t+1;p<s;p++)i.push(r.shape[p]),c*=r.shape[p];return{batchSize:l,sliceSize:c,outerSize:u,dimSize:a,outputShape:i}}function $K(r){try{return r.map(e=>Zc(e))}catch(e){throw new Error(`Failed to decode encoded string bytes into utf-8, error: ${e}`)}}function RK(r){return r.map(e=>cl(e))}var Fr={};We(Fr,{nonMaxSuppressionV3Impl:()=>Dg,nonMaxSuppressionV4Impl:()=>$g,nonMaxSuppressionV5Impl:()=>Rg,whereImpl:()=>Cg});function te(r,e){Array.isArray(r)||(r=[r]),r.forEach(t=>{t!=null&&y.assert(t.dtype!=="complex64",()=>`${e} does not support complex64 tensors in the CPU backend.`)})}var FK=Fr.whereImpl,Au=class extends Ms{constructor(){super();this.blockSize=48,this.firstUse=!0,this.data=new Xa(this,hs())}nextDataId(){return Au.nextDataId++}write(e,t,n){this.firstUse&&(this.firstUse=!1,W().get("IS_NODE")&&C.warn(`
|
|
============================
|
|
Hi there \u{1F44B}. Looks like you are running TensorFlow.js in Node.js. To speed things up dramatically, install our node backend, which binds to TensorFlow C++, by running npm i @tensorflow/tfjs-node, or npm i @tensorflow/tfjs-node-gpu if you have CUDA. Then call require('@tensorflow/tfjs-node'); (-gpu suffix for CUDA) at the start of your program. Visit https://github.com/tensorflow/tfjs-node for more details.
|
|
============================`));let o={id:this.nextDataId()};return this.data.set(o,{values:e,dtype:n,refCount:1}),o}makeTensorInfo(e,t,n){let o;if(t==="string"&&n!=null&&n.length>0&&y.isString(n[0])){let s=n.map(a=>y.encodeString(a));o=this.write(s,e,t)}else o=this.write(n,e,t);return{dataId:o,shape:e,dtype:t}}refCount(e){return this.data.has(e)?this.data.get(e).refCount:0}incRef(e){let t=this.data.get(e);t.refCount++}decRef(e){if(this.data.has(e)){let t=this.data.get(e);t.refCount--}}move(e,t,n,o,s){this.data.set(e,{values:t,dtype:o,refCount:s})}numDataIds(){return this.data.numDataIds()}async read(e){return this.readSync(e)}readSync(e){let{dtype:t,complexTensorInfos:n}=this.data.get(e);if(t==="complex64"){let o=this.readSync(n.real.dataId),s=this.readSync(n.imag.dataId);return C.mergeRealAndImagArrays(o,s)}return this.data.get(e).values}bufferSync(e){let t=this.readSync(e.dataId),n=t;if(e.dtype==="string")try{n=t.map(o=>y.decodeString(o))}catch(o){throw new Error("Failed to decode encoded string bytes into utf-8")}return Ce(e.shape,e.dtype,n)}makeOutput(e,t,n){let o=this.write(e,t,n);return hs().makeTensorFromDataId(o,t,n,this)}disposeData(e,t=!1){if(this.data.has(e)){if(this.data.get(e).refCount--,!t&&this.data.get(e).refCount>0)return!1;let{complexTensorInfos:n}=this.data.get(e);n!=null&&(this.disposeData(n.real.dataId,!0),this.disposeData(n.imag.dataId,!0)),this.data.delete(e)}return!0}disposeIntermediateTensorInfo(e){this.disposeData(e.dataId)}async time(e){let t=y.now();return e(),{kernelMs:y.now()-t}}memory(){return{unreliable:!0,reasons:["The reported memory is an upper bound. Due to automatic garbage collection, the true allocated memory may be less."]}}where(e){te([e],"where");let t=this.readSync(e.dataId);return FK(e.shape,t)}dispose(){}floatPrecision(){return 32}epsilon(){return super.epsilon()}};Au.nextDataId=0;var qg={};We(qg,{addImpl:()=>$_,bincountImpl:()=>_p,bincountReduceImpl:()=>Mg,ceilImpl:()=>R_,concatImpl:()=>Eu,expImpl:()=>F_,expm1Impl:()=>P_,floorImpl:()=>M_,gatherV2Impl:()=>Lg,greaterImpl:()=>L_,lessImpl:()=>z_,linSpaceImpl:()=>zg,logImpl:()=>B_,maxImpl:()=>Bg,maximumImpl:()=>V_,minimumImpl:()=>G_,multiplyImpl:()=>Cf,negImpl:()=>W_,notEqualImpl:()=>j_,prodImpl:()=>U_,rangeImpl:()=>$u,rsqrtImpl:()=>H_,simpleAbsImpl:()=>D_,sliceImpl:()=>Ru,sparseFillEmptyRowsImpl:()=>Vg,sparseReshapeImpl:()=>Gg,squaredDifferenceImpl:()=>q_,stridedSliceImpl:()=>Wg,subImpl:()=>K_,tileImpl:()=>jg,topKImpl:()=>Ug,transposeImpl:()=>kp,uniqueImpl:()=>Hg});function D_(r){let e=new Float32Array(r.length);for(let t=0;t<r.length;++t)e[t]=Math.abs(r[t]);return e}var OK=r=>{let{x:e}=r.inputs,t=r.backend;te(e,"abs");let n=new Float32Array(y.sizeFromShape(e.shape)),o=t.data.get(e.dataId).values;return n=D_(o),t.makeOutput(n,e.shape,"float32")},H1={kernelName:Bs,backendName:"cpu",kernelFunc:OK};function Ye(r){return(e,t,n,o,s)=>{let a=C.assertAndGetBroadcastShape(e,t),i=a.length,l=y.computeStrides(a),u=y.sizeFromShape(a),c=y.getTypedArrayFromDType(s,u),p=e.length,m=t.length,f=y.computeStrides(e),d=y.computeStrides(t),h=C.getBroadcastDims(e,a),g=C.getBroadcastDims(t,a);if(h.length+g.length===0)for(let x=0;x<c.length;++x)c[x]=r(n[x%n.length],o[x%o.length]);else for(let x=0;x<c.length;++x){let b=y.indexToLoc(x,i,l),w=b.slice(-p);h.forEach(D=>w[D]=0);let _=y.locToIndex(w,p,f),I=b.slice(-m);g.forEach(D=>I[D]=0);let E=y.locToIndex(I,m,d);c[x]=r(n[_],o[E])}return[c,a]}}function dr(r){let{inputs:e,backend:t}=r,{real:n,imag:o}=e,s=t.data.get(n.dataId).values,a=t.data.get(o.dataId).values,i=t.makeTensorInfo(n.shape,"complex64"),l=t.data.get(i.dataId);return l.complexTensorInfos={real:t.makeTensorInfo(n.shape,"float32",s),imag:t.makeTensorInfo(o.shape,"float32",a)},i}var q1={kernelName:vc,backendName:"cpu",kernelFunc:dr};function bp(r,e,t="float32"){if(t==="complex64"){let o=bp(r,e,"float32"),s=bp(r,e,"float32");return dr({inputs:{real:o,imag:s},backend:r})}let n=y.makeZerosTypedArray(y.sizeFromShape(e),t);return r.makeTensorInfo(e,t,n)}function Or(r){let{inputs:e,backend:t}=r,{x:n}=e;return t.incRef(n.dataId),{dataId:n.dataId,shape:n.shape,dtype:n.dtype}}var K1={kernelName:Xn,backendName:"cpu",kernelFunc:Or};function so(r){let{inputs:e,backend:t}=r,{input:n}=e,o=t.data.get(n.dataId).complexTensorInfos.real,s=t.data.get(o.dataId).values;return t.makeTensorInfo(o.shape,o.dtype,s)}var X1={kernelName:Gc,backendName:"cpu",kernelFunc:so};function io(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{dtype:s}=n;if(s==="complex64"){if(o.dtype==="complex64")return Or({inputs:{x:o},backend:t});let a=bp(t,o.shape,o.dtype),i=io({inputs:{x:o},backend:t,attrs:{dtype:"float32"}}),l=dr({inputs:{real:i,imag:a},backend:t});return t.disposeIntermediateTensorInfo(a),t.disposeIntermediateTensorInfo(i),l}if(o.dtype==="complex64"){let a=so({inputs:{input:o},backend:t}),i=io({inputs:{x:a},backend:t,attrs:{dtype:s}});return t.disposeIntermediateTensorInfo(a),i}if(!y.hasEncodingLoss(o.dtype,s)){let a=Or({inputs:{x:o},backend:t});return{dataId:a.dataId,shape:a.shape,dtype:s}}if(s==="int32"){let a=t.data.get(o.dataId).values,i=Int32Array.from(a);return t.makeTensorInfo(o.shape,"int32",i)}if(s==="bool"){let a=t.data.get(o.dataId).values,i=y.toTypedArray([0],o.dtype),[l,u]=Ye((c,p)=>c!==p?1:0)(o.shape,[],a,i,"bool");return t.makeTensorInfo(u,"bool",l)}throw new Error(`Error in Cast: failed to cast ${o.dtype} to ${s}`)}var Y1={kernelName:qn,backendName:"cpu",kernelFunc:io};function et(r,e,t,n){return t==null?({inputs:o,backend:s})=>{let{a,b:i}=o,l=s;te([a,i],r);let u=l.data.get(a.dataId).values,c=l.data.get(i.dataId).values,p=n||a.dtype,[m,f]=e(a.shape,i.shape,u,c,p);return l.makeTensorInfo(f,p,m)}:({inputs:o,backend:s})=>{let{a,b:i}=o,l=s;if(a.dtype==="complex64"||i.dtype==="complex64"){let u=io({inputs:{x:a},backend:l,attrs:{dtype:"complex64"}}),c=l.data.get(u.dataId),p=c.complexTensorInfos.real,m=c.complexTensorInfos.imag,f=l.data.get(p.dataId).values,d=l.data.get(m.dataId).values,h=io({inputs:{x:i},backend:l,attrs:{dtype:"complex64"}}),g=l.data.get(h.dataId),x=g.complexTensorInfos.real,b=g.complexTensorInfos.imag,w=l.data.get(x.dataId).values,_=l.data.get(b.dataId).values,[I,E,D]=t(a.shape,i.shape,f,d,w,_),$=l.makeTensorInfo(D,"float32",I),R=l.makeTensorInfo(D,"float32",E),M=dr({inputs:{real:$,imag:R},backend:l});return l.disposeIntermediateTensorInfo(u),l.disposeIntermediateTensorInfo(h),l.disposeIntermediateTensorInfo($),l.disposeIntermediateTensorInfo(R),M}else{let u=l.data.get(a.dataId).values,c=l.data.get(i.dataId).values,p=n||a.dtype,[m,f]=e(a.shape,i.shape,u,c,p);return l.makeTensorInfo(f,p,m)}}}function wp(r){return(e,t,n,o,s,a)=>{let i=C.assertAndGetBroadcastShape(e,t),l=y.sizeFromShape(i),u=i.length,c=y.computeStrides(i),p=y.getTypedArrayFromDType("float32",l),m=y.getTypedArrayFromDType("float32",l),f=C.getBroadcastDims(e,i),d=C.getBroadcastDims(t,i),h=C.mergeRealAndImagArrays(n,o),g=C.mergeRealAndImagArrays(s,a),x=e.length,b=y.computeStrides(e),w=t.length,_=y.computeStrides(t);if(f.length+d.length===0)for(let I=0;I<p.length;I++){let E=I%h.length,D=I%g.length,$=r(h[E*2],h[E*2+1],g[D*2],g[D*2+1]);p[I]=$.real,m[I]=$.imag}else for(let I=0;I<p.length;I++){let E=y.indexToLoc(I,u,c),D=E.slice(-x);f.forEach(j=>D[j]=0);let $=y.locToIndex(D,x,b),R=E.slice(-w);d.forEach(j=>R[j]=0);let M=y.locToIndex(R,w,_),G=r(h[$*2],h[$*2+1],g[M*2],g[M*2+1]);p[I]=G.real,m[I]=G.imag}return[p,m,i]}}var $_=Ye((r,e)=>r+e),PK=wp((r,e,t,n)=>({real:r+t,imag:e+n})),Aa=et(Pn,$_,PK),Z1={kernelName:Pn,backendName:"cpu",kernelFunc:Aa};function _p(r,e,t,n,o){let s=y.sizeFromShape(n),a=y.makeZerosTypedArray(o,t);for(let i=0;i<r.length;i++){let l=r[i];if(l<0)throw new Error("Input x must be non-negative!");l>=o||(s>0?a[l]+=e[i]:a[l]+=1)}return a}function Mg(r,e,t,n=!1){let o=r.shape[0],s=r.shape[1],a=Ce([o,t],e.dtype);for(let i=0;i<o;i++)for(let l=0;l<s;l++){let u=r.get(i,l);if(u<0)throw new Error("Input x must be non-negative!");u>=t||(n?a.set(1,i,u):e.size>0?a.set(a.get(i,u)+e.get(i,l),i,u):a.set(a.get(i,u)+1,i,u))}return a}function ao(r){return(e,t,n)=>{let o=y.getTypedArrayFromDType(t,e.length);for(let s=0;s<e.length;++s)o[s]=r(e[s],n);return o}}function $e(r,e,t){return({inputs:n,attrs:o,backend:s})=>{let{x:a}=n;if(te(a,r),a.dtype==="string"||t==="string")throw new Error("unaryKernelFunc does not support string input/output");let i=s,l=i.data.get(a.dataId).values,u=y.sizeFromShape(a.shape),c=t||a.dtype,p=y.getArrayFromDType(c,u);for(let m=0;m<u;++m)p[m]=e(l[m],o);return i.makeTensorInfo(a.shape,c,p)}}function lo(r,e,t){return({inputs:n,attrs:o,backend:s})=>{let{x:a}=n;if(te(a,r),a.dtype==="string"||t==="string")throw new Error("unaryKernelFunc does not support string input/output");let i=s,l=i.data.get(a.dataId).values,u=t||a.dtype,c=e(l,u,o);return i.makeTensorInfo(a.shape,u,c)}}var R_=ao(r=>Math.ceil(r)),MK=lo(Io,R_),J1={kernelName:Io,backendName:"cpu",kernelFunc:MK};function Eu(r,e,t,n){let o=y.getArrayFromDType(t,y.sizeFromShape(e));if(n&&t!=="string"){let s=0;r.forEach(a=>{let i=y.sizeFromShape(a.shape);o.set(a.vals,s),s+=i})}else{let s=0;r.forEach(a=>{let i=t==="string"?C.fromUint8ToStringArray(a.vals):a.vals,l=0;for(let u=0;u<a.shape[0];++u){let c=u*e[1]+s;for(let p=0;p<a.shape[1];++p)o[c+p]=i[l++]}s+=a.shape[1]})}return o}var F_=ao(r=>Math.exp(r)),O_=lo($o,F_),Q1={kernelName:$o,backendName:"cpu",kernelFunc:O_};var P_=ao(r=>Math.expm1(r)),LK=lo(Mi,P_),eA={kernelName:Mi,backendName:"cpu",kernelFunc:LK};var M_=ao(r=>Math.floor(r)),zK=lo(Ro,M_),tA={kernelName:Ro,backendName:"cpu",kernelFunc:zK};function Lg(r,e,t){let n=Ce(t,r.dtype);for(let o=0;o<n.size;++o){let a=n.indexToLoc(o).slice(),i=a[0],l=a[2],u=e.locToIndex([i,l]);a[2]=e.values[u];let c=r.locToIndex(a);n.values[o]=r.values[c]}return n}var L_=Ye((r,e)=>r>e?1:0),BK=et(Bi,L_,null,"bool"),rA={kernelName:Bi,backendName:"cpu",kernelFunc:BK};var z_=Ye((r,e)=>r<e?1:0),VK=et(ji,z_,null,"bool"),nA={kernelName:ji,backendName:"cpu",kernelFunc:VK};function zg(r,e,t){let n=(e-r)/(t-1),o=y.makeZerosTypedArray(t,"float32");o[0]=r;for(let s=1;s<o.length;s++)o[s]=o[s-1]+n;return o}var B_=ao(r=>Math.log(r)),GK=lo(Lo,B_),oA={kernelName:Lo,backendName:"cpu",kernelFunc:GK};function Bg(r,e,t,n){let o=y.getTypedArrayFromDType(n,y.sizeFromShape(t));for(let s=0;s<o.length;++s){let a=s*e,i=r[a];for(let l=0;l<e;++l){let u=r[a+l];u>i&&(i=u)}o[s]=i}return o}var V_=Ye((r,e)=>Math.max(r,e)),WK=et(Bo,V_),sA={kernelName:Bo,backendName:"cpu",kernelFunc:WK};var G_=Ye((r,e)=>Math.min(r,e)),jK=et(jo,G_),iA={kernelName:jo,backendName:"cpu",kernelFunc:jK};var Cf=Ye((r,e)=>r*e),UK=wp((r,e,t,n)=>({real:r*t-e*n,imag:r*n+e*t})),Du=et(Ho,Cf,UK),aA={kernelName:Ho,backendName:"cpu",kernelFunc:Du};function W_(r,e,t){let n=y.createScalarValue(-1,t);return Cf([],e,n,r,t)}function HK(r){let{inputs:e,backend:t}=r,{x:n}=e;te(n,"neg");let o=t.data.get(n.dataId).values,[s,a]=W_(o,n.shape,n.dtype);return t.makeTensorInfo(a,n.dtype,s)}var lA={kernelName:js,backendName:"cpu",kernelFunc:HK};var j_=Ye((r,e)=>r!==e?1:0),qK=et(Xi,j_,null,"bool"),uA={kernelName:Xi,backendName:"cpu",kernelFunc:qK};function kp(r,e,t,n,o){let s=e.length,a=y.sizeFromShape(e),i=y.computeStrides(e),l=y.computeStrides(o),u=y.getTypedArrayFromDType(t,y.sizeFromShape(o));for(let c=0;c<a;++c){let p=y.indexToLoc(c,s,i),m=new Array(p.length);for(let d=0;d<m.length;d++)m[d]=p[n[d]];let f=y.locToIndex(m,s,l);u[f]=r[c]}return u}function Xt(r){let{inputs:e,attrs:t,backend:n}=r,{x:o}=e,{perm:s}=t;te(o,"transpose");let a=o.shape.length,i=new Array(a);for(let p=0;p<i.length;p++)i[p]=o.shape[s[p]];let l=n.data.get(o.dataId).values,u=kp(l,o.shape,o.dtype,s,i);return{dataId:n.write(u,i,o.dtype),shape:i,dtype:o.dtype}}var cA={kernelName:ms,backendName:"cpu",kernelFunc:Xt};function U_(r,e,t,n){let[o,s]=C.computeOutAndReduceShapes(r,n),a=lr(e,"int32"),i=y.makeZerosTypedArray(y.sizeFromShape(o),a),l=y.sizeFromShape(s);for(let u=0;u<i.length;++u){let c=u*l,p=1;for(let m=0;m<l;++m)p*=t[c+m];i[u]=p}return{outVals:i,outShape:o,outDtype:a}}function KK(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n;te(o,"prod");let i=o.shape.length,l=y.parseAxisParam(s,o.shape),u=C.getAxesPermutation(l,i),c=l,p=o,m=[];u!=null&&(p=Xt({inputs:{x:o},backend:t,attrs:{perm:u}}),m.push(p),c=C.getInnerMostAxes(c.length,i));let f=t.data.get(p.dataId).values,{outVals:d,outShape:h,outDtype:g}=U_(p.shape,p.dtype,f,c),x=h;return a&&(x=C.expandShapeToKeepDim(h,l)),m.forEach(b=>t.disposeIntermediateTensorInfo(b)),t.makeTensorInfo(x,g,d)}var pA={kernelName:Qi,backendName:"cpu",kernelFunc:KK};function $u(r,e,t,n){let o=r===e,s=r<e&&t<0,a=e<r&&t>1;if(o||s||a)return y.makeZerosTypedArray(0,n);let i=Math.abs(Math.ceil((e-r)/t)),l=y.makeZerosTypedArray(i,n);e<r&&t===1&&(t=-1),l[0]=r;for(let u=1;u<l.length;u++)l[u]=l[u-1]+t;return l}var H_=ao(r=>1/Math.sqrt(r)),XK=lo(rs,H_),mA={kernelName:rs,backendName:"cpu",kernelFunc:XK};function Ru(r,e,t,n,o){let s=or.isSliceContinous(n,e,t),a=y.sizeFromShape(t),i=y.computeStrides(n);if(s){let p=or.computeFlatOffset(e,i);return o==="string"?r.slice(p,p+a):r.subarray(p,p+a)}let l=o==="string"?C.fromUint8ToStringArray(r):r,u=Ce(n,o,l),c=Ce(t,o);for(let p=0;p<c.size;++p){let m=c.indexToLoc(p),f=m.map((d,h)=>d+e[h]);c.set(u.get(...f),...m)}return o==="string"?C.fromStringArrayToUint8(c.values):c.values}function uo(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{begin:s,size:a}=n;te(o,"slice");let[i,l]=or.parseSliceParams(o,s,a);or.assertParamsValid(o,i,l);let u=t.data.get(o.dataId).values,c=Ru(u,i,l,o.shape,o.dtype);return t.makeTensorInfo(l,o.dtype,c)}var fA={kernelName:Xs,backendName:"cpu",kernelFunc:uo};function Vg(r,e,t,n,o,s,a){let i=e[0],l=s[0],u=new Array(l),c=new Array(i),p=e[1];if(l===0){if(i!==0)throw new Error(`Received SparseTensor with denseShape[0] = 0 but
|
|
indices.shape[0] = ${i}`);let g=y.getArrayFromDType(t,0),x=y.getArrayFromDType(o,0);return[g,[0,p],x,u,c]}let m=!0,f=0,d=new Array(l).fill(0);for(let g=0;g<i;++g){let x=r[g*p];if(x<0)throw new Error(`indices(${g}, 0) is invalid: ${x} < 0`);if(x>=l)throw new Error(`indices(${g}, 0) is invalid: ${x} >= ${l}`);++d[x],m=m&&x>=f,f=x}let h=!0;for(let g=0;g<l;++g){let x=d[g]===0;u[g]=x,h=h&&!x,d[g]=Math.max(d[g],1),g>0&&(d[g]+=d[g-1])}if(h&&m){let g=r,x=n;for(let b=0;b<i;++b)c[b]=b;return[g,[i,p],x,u,c]}else{let g=d[l-1],x=y.getArrayFromDType(t,g*p),b=y.getArrayFromDType(o,g),w=new Array(l).fill(0);for(let _=0;_<i;++_){let I=r[_*p],E=w[I],D=(I===0?0:d[I-1])+E;w[I]++;for(let $=0;$<p;++$)x[D*p+$]=r[_*p+$];b[D]=n[_],c[_]=D}for(let _=0;_<l;++_)if(w[_]===0){let E=_===0?0:d[_-1];x[E*p+0]=_;for(let D=1;D<p;++D)x[E*p+D]=0;b[E]=a}return[x,[i,p],b,u,c]}}function Gg(r,e,t,n,o){let s=y.sizeFromShape(n),a=e[0],i=o.length,l=[],u=1,c=-1;for(let g=0;g<i;++g){let x=o[g];if(x===-1){if(c!==-1)throw new Error(`only one output dimension may be -1, not both ${c} and ${g}`);c=g,l.push(1)}else{if(x<0)throw new Error(`size ${g} must be non-negative, not ${x}`);u*=x,l.push(x)}}if(c!==-1){if(u<=0)throw new Error("reshape cannot infer the missing input size for an empty tensor unless all specified input sizes are non-zero");let g=Math.trunc(s/u);if(u*g!==s)throw new Error(`Input to reshape is a SparseTensor with ${s}
|
|
dense values, but the requested shape requires a multiple of ${u}. inputShape=${n} outputShape= ${l}`);l[c]=g}let p=y.sizeFromShape(l);if(p!==s)throw new Error(`Input to reshape is a tensor with ${s} dense values, but the requested shape has ${p}. inputShape=${n} outputShape=${l}`);let m=n.length,f=[];if(m>0){f[m-1]=1;for(let g=m-2;g>=0;--g)f[g]=f[g+1]*n[g+1]}let d=[];if(i>0){d[i-1]=1;for(let g=i-2;g>=0;--g)d[g]=d[g+1]*l[g+1]}let h=y.getArrayFromDType(t,a*i);for(let g=0;g<a;++g){let x=0;for(let b=0;b<m;++b)x+=r[g*m+b]*f[b];for(let b=0;b<i;++b)h[g*i+b]=Math.trunc(x/d[b]),x%=d[b]}return[h,[a,i],l]}var q_=Ye((r,e)=>{let t=r-e;return t*t}),YK=et(ls,q_),dA={kernelName:ls,backendName:"cpu",kernelFunc:YK};function Wg(r,e,t,n){let o=Ce(r,e.dtype);for(let s=0;s<o.size;s++){let a=o.indexToLoc(s),i=new Array(a.length);for(let l=0;l<i.length;l++)i[l]=a[l]*t[l]+n[l];o.set(e.get(...i),...a)}return o}var K_=Ye((r,e)=>r-e),ZK=wp((r,e,t,n)=>({real:r-t,imag:e-n})),If=et(us,K_,ZK),hA={kernelName:us,backendName:"cpu",kernelFunc:If};function jg(r,e){let t=new Array(r.rank);for(let o=0;o<t.length;o++)t[o]=r.shape[o]*e[o];let n=Ce(t,r.dtype);for(let o=0;o<n.values.length;++o){let s=n.indexToLoc(o),a=new Array(r.rank);for(let l=0;l<a.length;l++)a[l]=s[l]%r.shape[l];let i=r.locToIndex(a);n.values[o]=r.values[i]}return n}function Ug(r,e,t,n,o){let s=e[e.length-1],[a,i]=[r.length/s,s],l=y.getTypedArrayFromDType(t,a*n),u=y.getTypedArrayFromDType("int32",a*n);for(let p=0;p<a;p++){let m=p*i,f=r.subarray(m,m+i),d=[];for(let b=0;b<f.length;b++)d.push({value:f[b],index:b});d.sort((b,w)=>w.value-b.value);let h=p*n,g=l.subarray(h,h+n),x=u.subarray(h,h+n);for(let b=0;b<n;b++)g[b]=d[b].value,x[b]=d[b].index}let c=e.slice();return c[c.length-1]=n,[Ce(c,t,l),Ce(c,"int32",u)]}function Hg(r,e,t,n){let o=y.parseAxisParam(e,t)[0],s=[1,t[0],1];for(let d=0;d<o;d++)s[0]*=t[d];s[1]=t[o];for(let d=o+1;d<t.length;d++)s[2]*=t[d];let a={},i=new Int32Array(t[o]),l=new ut(s,n,r),u=[],c=s[0]===1&&s[2]===1;for(let d=0;d<t[o];d++){let h;if(c)h=r[d].toString();else{let g=[];for(let x=0;x<s[0];x++)for(let b=0;b<s[2];b++)g.push(l.get(x,d,b));h=g.join(",")}if(a[h]!==void 0)i[d]=a[h];else{let g=Object.keys(a).length;a[h]=g,i[d]=g,u.push(d)}}let p=s.slice();p[1]=Object.keys(a).length;let m=new ut(p,n);u.forEach((d,h)=>{for(let g=0;g<s[0];g++)for(let x=0;x<s[2];x++)m.set(l.get(g,d,x),g,h,x)});let f=t.slice();return f[o]=p[1],{outputValues:m.values,outputShape:f,indices:i}}var X_="3.6.0";np("cpu",()=>new Au,1);var Y_=$e(Fi,r=>r>=0?r:Math.exp(r)-1),gA={kernelName:Fi,backendName:"cpu",kernelFunc:Y_};function Z_(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{alpha:s}=n;te([o],"leakyRelu");let a=y.sizeFromShape(o.shape),i=t.data.get(o.dataId).values,l=y.getTypedArrayFromDType("float32",a);for(let u=0;u<i.length;u++)l[u]=i[u]<0?s*i[u]:i[u];return t.makeTensorInfo(o.shape,"float32",l)}var xA={kernelName:Mo,backendName:"cpu",kernelFunc:Z_};var JK=Ye((r,e)=>r<0?e*r:r);function J_(r){let{inputs:e,backend:t}=r,{x:n,alpha:o}=e;te([n,o],"prelu");let s=t.data.get(n.dataId).values,a=t.data.get(o.dataId).values,[i,l]=JK(n.shape,o.shape,s,a,n.dtype);return t.makeTensorInfo(l,n.dtype,i)}var yA={kernelName:Yo,backendName:"cpu",kernelFunc:J_};var Q_=$e(Zo,r=>Math.max(0,r)),bA={kernelName:Zo,backendName:"cpu",kernelFunc:Q_};var ek=$e(Qo,r=>Math.min(Math.max(0,r),6)),wA={kernelName:Qo,backendName:"cpu",kernelFunc:ek};var tk=$e(os,r=>1/(1+Math.exp(-r))),_A={kernelName:os,backendName:"cpu",kernelFunc:tk};function vp(r,e,t,n,o){if(t==="linear")return Or({inputs:{x:e},backend:r});if(t==="relu")return Q_({inputs:{x:e},backend:r});if(t==="elu")return Y_({inputs:{x:e},backend:r});if(t==="relu6")return ek({inputs:{x:e},backend:r});if(t==="prelu")return J_({inputs:{x:e,alpha:n},backend:r});if(t==="leakyrelu")return Z_({inputs:{x:e},backend:r,attrs:{alpha:o}});if(t==="sigmoid")return tk({inputs:{x:e},backend:r});throw new Error(`Activation ${t} has not been implemented for the CPU backend.`)}function Ze(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{shape:s}=n,a=y.sizeFromShape(o.shape),i=y.inferFromImplicitShape(s,a),l=y.sizeFromShape(i);y.assert(a===l,()=>`The new shape (${i}) has ${l} elements and the old shape (${o.shape}) has ${a} elements. The new shape and old shape must have the same number of elements.`),t.incRef(o.dataId);let u=t.data.get(o.dataId);if(u.complexTensorInfos!=null){let c=u.complexTensorInfos.real,p=u.complexTensorInfos.imag;c.shape=i,p.shape=i}return{dataId:o.dataId,shape:i,dtype:o.dtype}}var kA={kernelName:qs,backendName:"cpu",kernelFunc:Ze};function rk(r){let{inputs:e,backend:t,attrs:n}=r,{a:o,b:s}=e,{transposeA:a,transposeB:i}=n;te([o,s],"matMul");let l=o.shape.length,u=s.shape.length,c=a?o.shape[l-2]:o.shape[l-1],p=i?s.shape[u-1]:s.shape[u-2],m=a?o.shape[l-1]:o.shape[l-2],f=i?s.shape[u-2]:s.shape[u-1],d=o.shape.slice(0,-2),h=s.shape.slice(0,-2),g=y.sizeFromShape(d),x=y.sizeFromShape(h),b=g===x||g===1||x===1;y.assert(l>=2&&u>=2&&b,()=>`Error in matMul: the input batch dimensions must either be the same or at least one input batch dimension must be 1. Got input batch dimensions of (${d}) and (${h}).`);let _=(g>x?o.shape.slice(0,-2):s.shape.slice(0,-2)).concat([m,f]);y.assert(c===p,()=>`Error in matMul: inner shapes (${c}) and (${p}) of Tensors with shapes ${o.shape} and ${s.shape} and transposeA=${a} and transposeB=${i} must match.`);let I=a?[g,c,m]:[g,m,c],E=i?[x,f,p]:[x,p,f],D=Ze({inputs:{x:o},backend:t,attrs:{shape:I}}),$=Ze({inputs:{x:s},backend:t,attrs:{shape:E}}),R=a?D.shape[1]:D.shape[2],M=a?D.shape[2]:D.shape[1],G=i?$.shape[1]:$.shape[2],j=Math.max(g,x),U=t.data.get(D.dataId).values,H=t.data.get($.dataId).values,q=y.computeStrides(D.shape),X=y.computeStrides($.shape),[ne,Y,re]=a?[q[0],1,q[1]]:[q[0],q[1],1],[ee,ie,me]=i?[1,X[1],X[0]]:[X[1],1,X[0]],ae=M*G,de=Ce([j,M,G],D.dtype),he=de.values,xe=t.blockSize;for(let we=0;we<j;we++)for(let De=0;De<M;De+=xe)for(let ve=0;ve<G;ve+=xe)for(let Ge=0;Ge<R;Ge+=xe){let Ke=Math.min(De+xe,M),at=Math.min(ve+xe,G),At=Math.min(Ge+xe,R);for(let Et=De;Et<Ke;Et++)for(let He=ve;He<at;He++){let ct=0;for(let mt=Ge;mt<At;mt++){let zt=Math.min(we,g-1)*ne,vn=Math.min(we,x-1)*me,Zt=U[zt+Et*Y+mt*re],cn=H[mt*ee+He*ie+vn];ct+=Zt*cn}he[we*ae+(Et*G+He)]+=ct}}return t.disposeIntermediateTensorInfo(D),t.disposeIntermediateTensorInfo($),t.makeTensorInfo(_,de.dtype,de.values)}var vA={kernelName:Co,backendName:"cpu",kernelFunc:rk};function QK(r){let{inputs:e,backend:t,attrs:n}=r,{a:o,b:s,bias:a,preluActivationWeights:i}=e,{transposeA:l,transposeB:u,activation:c,leakyreluAlpha:p}=n,m,f,d,h=[];m=rk({inputs:{a:o,b:s},attrs:{transposeA:l,transposeB:u},backend:t}),a&&(f=Aa({inputs:{a:m,b:a},backend:t}),h.push(m),m=f),c&&(d=vp(t,m,c,i,p),h.push(m),m=d);for(let x of h)t.disposeIntermediateTensorInfo(x);return m}var CA={kernelName:Qs,backendName:"cpu",kernelFunc:QK};var e6=$e(ki,r=>Math.acos(r)),IA={kernelName:ki,backendName:"cpu",kernelFunc:e6};var t6=$e(vi,r=>Math.acosh(r)),NA={kernelName:vi,backendName:"cpu",kernelFunc:t6};function r6(r){let{inputs:e,backend:t}=r,n=e;te(e,"addN");let o=n.map(i=>t.data.get(i.dataId).values),s=Ce(n[0].shape,n[0].dtype),a=s.values;for(let i=0;i<n.length;i++){let l=o[i];for(let u=0;u<a.length;u++)a[u]+=l[u]}return t.makeTensorInfo(s.shape,s.dtype,s.values)}var SA={kernelName:_o,backendName:"cpu",kernelFunc:r6};function n6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n;te(o,"all");let i=y.parseAxisParam(s,o.shape),l=i,u=C.getAxesPermutation(l,o.shape.length),c=o;u!=null&&(c=Xt({inputs:{x:o},backend:t,attrs:{perm:u}}),l=C.getInnerMostAxes(l.length,o.shape.length)),C.assertAxesAreInnerMostDims("all",l,c.shape.length);let[p,m]=C.computeOutAndReduceShapes(c.shape,l),f=y.sizeFromShape(m),d=y.makeZerosTypedArray(y.sizeFromShape(p),c.dtype),h=t.data.get(c.dataId).values;for(let x=0;x<d.length;++x){let b=x*f,w=h[b];for(let _=0;_<f;++_){let I=h[b+_];w=w&&I}d[x]=w}u!=null&&t.disposeIntermediateTensorInfo(c);let g=t.makeTensorInfo(p,c.dtype,d);if(a){let x=C.expandShapeToKeepDim(p,i),b=Ze({inputs:{x:g},backend:t,attrs:{shape:x}});return t.disposeIntermediateTensorInfo(g),b}return g}var TA={kernelName:Ci,backendName:"cpu",kernelFunc:n6};function o6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n;te(o,"any");let i=y.parseAxisParam(s,o.shape),l=i,u=C.getAxesPermutation(l,o.shape.length),c=o;u!=null&&(c=Xt({inputs:{x:o},backend:t,attrs:{perm:u}}),l=C.getInnerMostAxes(l.length,o.shape.length)),C.assertAxesAreInnerMostDims("any",l,c.shape.length);let[p,m]=C.computeOutAndReduceShapes(c.shape,l),f=y.sizeFromShape(m),d=y.makeZerosTypedArray(y.sizeFromShape(p),c.dtype),h=t.data.get(c.dataId).values;for(let x=0;x<d.length;++x){let b=x*f,w=h[b];for(let _=0;_<f;++_){let I=h[b+_];w=w||I}d[x]=w}u!=null&&t.disposeIntermediateTensorInfo(c);let g=t.makeTensorInfo(p,c.dtype,d);if(a){let x=C.expandShapeToKeepDim(p,i),b=Ze({inputs:{x:g},backend:t,attrs:{shape:x}});return t.disposeIntermediateTensorInfo(g),b}return g}var AA={kernelName:Ii,backendName:"cpu",kernelFunc:o6};function s6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s}=n;te(o,"argMax");let a=y.parseAxisParam(s,o.shape),i=C.getAxesPermutation(a,o.shape.length),l=o,u=[];i!=null&&(l=Xt({inputs:{x:o},backend:t,attrs:{perm:i}}),u.push(l),a=C.getInnerMostAxes(a.length,l.shape.length)),a=[a[0]],C.assertAxesAreInnerMostDims("argMax",a,l.shape.length);let[c,p]=C.computeOutAndReduceShapes(l.shape,a),m=y.sizeFromShape(c),f=y.makeZerosTypedArray(m,"int32"),d=y.sizeFromShape(p),h=t.data.get(l.dataId).values;for(let g=0;g<f.length;++g){let x=g*d,b=h[x],w=0;for(let _=0;_<d;++_){let I=h[x+_];I>b&&(b=I,w=_)}f[g]=w}return u.forEach(g=>t.disposeIntermediateTensorInfo(g)),t.makeTensorInfo(c,"int32",f)}var EA={kernelName:ko,backendName:"cpu",kernelFunc:s6};function i6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s}=n;te(o,"argMin");let a=y.parseAxisParam(s,o.shape),i=C.getAxesPermutation(a,o.shape.length),l=o,u=[];i!=null&&(l=Xt({inputs:{x:o},backend:t,attrs:{perm:i}}),u.push(l),a=C.getInnerMostAxes(a.length,l.shape.length)),a=[a[0]],C.assertAxesAreInnerMostDims("argMin",a,l.shape.length);let[c,p]=C.computeOutAndReduceShapes(l.shape,a),m=y.sizeFromShape(c),f=y.makeZerosTypedArray(m,"int32"),d=y.sizeFromShape(p),h=t.data.get(l.dataId).values;for(let g=0;g<f.length;++g){let x=g*d,b=h[x],w=0;for(let _=0;_<d;++_){let I=h[x+_];I<b&&(b=I,w=_)}f[g]=w}return u.forEach(g=>t.disposeIntermediateTensorInfo(g)),t.makeTensorInfo(c,"int32",f)}var DA={kernelName:Ya,backendName:"cpu",kernelFunc:i6};var a6=$e(Ni,r=>Math.asin(r)),$A={kernelName:Ni,backendName:"cpu",kernelFunc:a6};var l6=$e(Si,r=>Math.asinh(r)),RA={kernelName:Si,backendName:"cpu",kernelFunc:l6};var u6=$e(Ti,r=>Math.atan(r)),FA={kernelName:Ti,backendName:"cpu",kernelFunc:u6};var c6=Ye((r,e)=>Math.atan2(r,e)),p6=et(Ei,c6),OA={kernelName:Ei,backendName:"cpu",kernelFunc:p6};var m6=$e(Ai,r=>Math.atanh(r)),PA={kernelName:Ai,backendName:"cpu",kernelFunc:m6};function Cp(r,e,t,n,o,s){let a=o.strideHeight,i=o.strideWidth,l=o.dilationHeight,u=o.dilationWidth,c=o.effectiveFilterHeight,p=o.effectiveFilterWidth,m=o.padInfo.top,f=o.padInfo.left,d=s==="max"?Number.NEGATIVE_INFINITY:Number.POSITIVE_INFINITY,h=Ce(o.outShape,t),g=h.values,x=o.outShape[1]*o.outShape[2]*o.outShape[3],b=o.outShape[2]*o.outShape[3],w=o.outShape[3];for(let _=0;_<o.batchSize;++_){let I=_*x,E=_*n[0];for(let D=0;D<o.inChannels;++D)for(let $=0;$<o.outHeight;++$){let R=$*a-m,M=Math.max(0,R),G=Math.min(o.inHeight,c+R),j=I+$*b;for(let U=0;U<o.outWidth;++U){let H=U*i-f,q=Math.max(0,H),X=Math.min(o.inWidth,p+H),ne=d,Y=0,re=0;for(let ie=M;ie<G;ie+=l){let me=E+ie*n[1];for(let ae=q;ae<X;ae+=u){let de=me+ae*n[2],he=r[de+D];s==="max"&&he>ne?ne=he:s==="avg"&&(Y+=he,re++)}if(isNaN(ne))break}let ee=j+U*w+D;g[ee]=s==="avg"?Y/re:ne}}}return h}function Kg(r,e,t,n,o=!1,s=!1){let a=Ce(n.outShape,"int32"),i=n.strideHeight,l=n.strideWidth,u=n.dilationHeight,c=n.dilationWidth,p=n.effectiveFilterHeight,m=n.effectiveFilterWidth,f=n.padInfo.top,d=n.padInfo.left,h=Ce(e,t,r);for(let g=0;g<n.batchSize;++g)for(let x=0;x<n.inChannels;++x)for(let b=0;b<n.outHeight;++b){let w=b*i-f,_=w;for(;_<0;)_+=u;let I=Math.min(n.inHeight,p+w);for(let E=0;E<n.outWidth;++E){let D=E*l-d,$=D;for(;$<0;)$+=c;let R=Math.min(n.inWidth,m+D),M=Number.NEGATIVE_INFINITY,G=-1;for(let j=_;j<I;j+=u){let U=j-w;for(let H=$;H<R;H+=c){let q=H-D,X=h.get(g,j,H,x);X>M&&(M=X,o?G=s?((g*n.inHeight+j)*n.inWidth+H)*n.inChannels+x:(j*n.inWidth+H)*n.inChannels+x:G=U*m+q)}}a.set(G,g,b,E,x)}}return a}function Xg(r,e,t,n,o,s){let a=o.strideDepth,i=o.strideHeight,l=o.strideWidth,u=o.dilationDepth,c=o.dilationHeight,p=o.dilationWidth,m=o.effectiveFilterDepth,f=o.effectiveFilterHeight,d=o.effectiveFilterWidth,h=o.padInfo.front,g=o.padInfo.top,x=o.padInfo.left,b=s==="max"?Number.NEGATIVE_INFINITY:Number.POSITIVE_INFINITY,w=Ce(o.outShape,t),_=w.values,I=o.outShape[1]*o.outShape[2]*o.outShape[3]*o.outShape[4],E=o.outShape[2]*o.outShape[3]*o.outShape[4],D=o.outShape[3]*o.outShape[4],$=o.outShape[4];for(let R=0;R<o.batchSize;++R){let M=R*I,G=R*n[0];for(let j=0;j<o.inChannels;++j)for(let U=0;U<o.outDepth;++U){let H=U*a-h,q=H;for(;q<0;)q+=u;let X=Math.min(o.inDepth,m+H),ne=M+U*E;for(let Y=0;Y<o.outHeight;++Y){let re=Y*i-g,ee=re;for(;ee<0;)ee+=c;let ie=Math.min(o.inHeight,f+re),me=ne+Y*D;for(let ae=0;ae<o.outWidth;++ae){let de=ae*l-x,he=de;for(;he<0;)he+=p;let xe=Math.min(o.inWidth,d+de),we=me+ae*$,De=b,ve=0,Ge=0;for(let at=q;at<X;at+=u){let At=G+at*n[1];for(let Et=ee;Et<ie;Et+=c){let He=At+Et*n[2];for(let ct=he;ct<xe;ct+=p){let mt=He+ct*n[3],zt=r[mt+j];if(s==="max"&&zt>De?De=zt:s==="avg"&&(ve+=zt,Ge++),isNaN(De))break}if(isNaN(De))break}if(isNaN(De))break}let Ke=we+j;_[Ke]=s==="avg"?ve/Ge:De}}}}return w}function MA(r,e){let t=Ce(e.outShape,"int32"),n=e.strideDepth,o=e.strideHeight,s=e.strideWidth,a=e.dilationDepth,i=e.dilationHeight,l=e.dilationWidth,u=e.effectiveFilterDepth,c=e.effectiveFilterHeight,p=e.effectiveFilterWidth,m=e.padInfo.front,f=e.padInfo.top,d=e.padInfo.left;for(let h=0;h<e.batchSize;++h)for(let g=0;g<e.inChannels;++g)for(let x=0;x<e.outDepth;++x){let b=x*n-m,w=b;for(;w<0;)w+=a;let _=Math.min(e.inDepth,u+b);for(let I=0;I<e.outHeight;++I){let E=I*o-f,D=E;for(;D<0;)D+=i;let $=Math.min(e.inHeight,c+E);for(let R=0;R<e.outWidth;++R){let M=R*s-d,G=M;for(;G<0;)G+=l;let j=Math.min(e.inWidth,p+M),U=Number.NEGATIVE_INFINITY,H=-1;for(let q=w;q<_;q+=a){let X=q-b;for(let ne=D;ne<$;ne+=i){let Y=ne-E;for(let re=G;re<j;re+=l){let ee=re-M,ie=r.get(h,q,ne,re,g);ie>=U&&(U=ie,H=X*c*p+Y*c+ee)}}}t.set(H,h,x,I,R,g)}}}return t}function f6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e;te(o,"avgPool");let{filterSize:s,strides:a,pad:i,dimRoundingMode:l}=n,u=1;y.assert(C.eitherStridesOrDilationsAreOne(a,u),()=>`Error in avgPool: Either strides or dilations must be 1. Got strides ${a} and dilations '${u}'`);let c=C.computePool2DInfo(o.shape,s,a,u,i,l),p;if(c.filterWidth===1&&c.filterHeight===1&&y.arraysEqual(c.inShape,c.outShape))p=Or({inputs:{x:o},backend:t});else{let m=t.data.get(o.dataId).values,f=y.computeStrides(o.shape),d=Cp(m,o.shape,o.dtype,f,c,"avg");p=t.makeTensorInfo(c.outShape,o.dtype,d.values)}return p}var LA={kernelName:vo,backendName:"cpu",kernelFunc:f6};function d6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{filterSize:s,strides:a,pad:i,dimRoundingMode:l,dataFormat:u}=n;te(o,"avgPool3d");let c=C.computePool3DInfo(o.shape,s,a,1,i,l,u),p=t.data.get(o.dataId).values,m=Xg(p,o.shape,o.dtype,y.computeStrides(o.shape),c,"avg");return t.makeTensorInfo(m.shape,"float32",m.values)}var zA={kernelName:Za,backendName:"cpu",kernelFunc:d6};function h6(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,input:s}=e,{filterSize:a,strides:i,pad:l,dimRoundingMode:u}=n;te([o,s],"avgPool3DGrad");let c=C.computePool3DInfo(s.shape,a,i,1,l,u),p=c.strideDepth,m=c.strideHeight,f=c.strideWidth,d=c.filterDepth,h=c.filterHeight,g=c.filterWidth,x=c.dilationDepth,b=c.dilationHeight,w=c.dilationWidth,_=c.effectiveFilterDepth,I=c.effectiveFilterHeight,E=c.effectiveFilterWidth,D=_-1-c.padInfo.front,$=E-1-c.padInfo.left,R=I-1-c.padInfo.top,M=Ce(s.shape,"float32"),G=1/(d*h*g),j=t.bufferSync(o);for(let U=0;U<c.batchSize;++U)for(let H=0;H<c.inChannels;++H)for(let q=0;q<c.inDepth;++q)for(let X=0;X<c.inHeight;++X)for(let ne=0;ne<c.inWidth;++ne){let Y=q-D,re=X-R,ee=ne-$,ie=0;for(let me=0;me<_;me+=x){let ae=(Y+me)/p;if(!(ae<0||ae>=c.outDepth||Math.floor(ae)!==ae))for(let de=0;de<I;de+=b){let he=(re+de)/m;if(!(he<0||he>=c.outHeight||Math.floor(he)!==he))for(let xe=0;xe<E;xe+=w){let we=(ee+xe)/f;if(we<0||we>=c.outWidth||Math.floor(we)!==we)continue;ie+=j.get(U,ae,he,we,H)}}}M.set(ie*G,U,q,X,ne,H)}return t.makeTensorInfo(M.shape,M.dtype,M.values)}var BA={kernelName:_c,backendName:"cpu",kernelFunc:h6};function g6(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,input:s}=e,a=s;te([o,s],"avgPoolGrad");let{filterSize:i,strides:l,pad:u}=n,c=C.computePool2DInfo(a.shape,i,l,1,u),p=c.strideHeight,m=c.strideWidth,f=c.filterHeight,d=c.filterWidth,h=c.dilationHeight,g=c.dilationWidth,x=c.effectiveFilterHeight,b=c.effectiveFilterWidth,w=b-1-c.padInfo.left,_=x-1-c.padInfo.top,I=Ce(a.shape,"float32"),E=1/(f*d),D=t.data.get(o.dataId).values,$=Ce(o.shape,"float32",D);for(let R=0;R<c.batchSize;++R)for(let M=0;M<c.inChannels;++M)for(let G=0;G<c.inHeight;++G)for(let j=0;j<c.inWidth;++j){let U=G-_,H=j-w,q=0;for(let X=0;X<x;X+=h){let ne=(U+X)/p;if(!(ne<0||ne>=c.outHeight||Math.floor(ne)!==ne))for(let Y=0;Y<b;Y+=g){let re=(H+Y)/m;if(re<0||re>=c.outWidth||Math.floor(re)!==re)continue;q+=$.get(R,ne,re,M)}}I.set(q*E,R,G,j,M)}return t.makeTensorInfo(I.shape,I.dtype,I.values)}var VA={kernelName:wc,backendName:"cpu",kernelFunc:g6};function x6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,scale:s,offset:a,mean:i,variance:l}=e;y.assert(i.shape.length===l.shape.length,()=>"Batch normalization gradient requires mean and variance to have equal ranks."),y.assert(a==null||i.shape.length===a.shape.length,()=>"Batch normalization gradient requires mean and offset to have equal ranks."),y.assert(s==null||i.shape.length===s.shape.length,()=>"Batch normalization gradient requires mean and scale to have equal ranks."),te([o,i,l,s,a],"batchNorm");let{varianceEpsilon:u}=n;u==null&&(u=.001);let c=t.data.get(o.dataId).values,p=t.data.get(i.dataId).values,m=t.data.get(l.dataId).values,f=s?t.data.get(s.dataId).values:new Float32Array([1]),d=a?t.data.get(a.dataId).values:new Float32Array([0]),h=new Float32Array(c.length),g=d.length,x=f.length,b=m.length,w=p.length,_=0,I=0,E=0,D=0;for(let $=0;$<c.length;++$)h[$]=d[_++]+(c[$]-p[I++])*f[E++]/Math.sqrt(m[D++]+u),_>=g&&(_=0),I>=w&&(I=0),E>=x&&(E=0),D>=b&&(D=0);return t.makeTensorInfo(o.shape,o.dtype,h)}var GA={kernelName:Oo,backendName:"cpu",kernelFunc:x6};function y6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{blockShape:s,crops:a}=n;te([o],"batchToSpaceND");let i=s.reduce((x,b)=>x*b),l=C.getReshaped(o.shape,s,i),u=C.getPermuted(l.length,s.length),c=C.getReshapedPermuted(o.shape,s,i),p=C.getSliceBeginCoords(a,s.length),m=C.getSliceSize(c,a,s.length),f=Ze({inputs:{x:o},backend:t,attrs:{shape:l}}),d=Xt({inputs:{x:f},backend:t,attrs:{perm:u}}),h=Ze({inputs:{x:d},backend:t,attrs:{shape:c}}),g=uo({inputs:{x:h},backend:t,attrs:{begin:p,size:m}});return t.disposeIntermediateTensorInfo(f),t.disposeIntermediateTensorInfo(d),t.disposeIntermediateTensorInfo(h),g}var WA={kernelName:Ja,backendName:"cpu",kernelFunc:y6};function b6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,weights:s}=e,{size:a}=n,i=t.data.get(o.dataId).values,l=t.data.get(s.dataId).values,u=_p(i,l,s.dtype,s.shape,a);return t.makeTensorInfo([a],s.dtype,u)}var jA={kernelName:kc,backendName:"cpu",kernelFunc:b6};var w6=$e(Kn,(r,e)=>{let t=e;return r>t.clipValueMax?t.clipValueMax:r<t.clipValueMin?t.clipValueMin:r}),UA={kernelName:Kn,backendName:"cpu",kernelFunc:w6};var _6=r=>{let{x:e}=r.inputs,t=r.backend,n=new Float32Array(y.sizeFromShape(e.shape)),o=t.data.get(e.dataId),s=o.complexTensorInfos.real,a=o.complexTensorInfos.imag,i=t.data.get(s.dataId).values,l=t.data.get(a.dataId).values;for(let u=0;u<i.length;u++){let c=i[u],p=l[u];n[u]=Math.hypot(c,p)}return t.makeOutput(n,e.shape,"float32")},HA={kernelName:Qa,backendName:"cpu",kernelFunc:_6};function li(r){let{inputs:e,backend:t}=r,{input:n}=e,o=t.data.get(n.dataId).complexTensorInfos.imag,s=t.data.get(o.dataId).values;return t.makeTensorInfo(o.shape,o.dtype,s)}var qA={kernelName:Oc,backendName:"cpu",kernelFunc:li};function yl(r){let{inputs:e,backend:t,attrs:n}=r,{axis:o}=n,s=y.parseAxisParam(o,e[0].shape)[0],a=C.computeOutShape(e.map(h=>h.shape),s);if(y.sizeFromShape(a)===0)return t.makeTensorInfo(a,e[0].dtype,[]);let i=e.filter(h=>y.sizeFromShape(h.shape)>0);if(i.length===1)return Or({inputs:{x:i[0]},backend:t});let l=i.map(h=>h.shape);if(C.assertParamsConsistent(l,s),i[0].dtype==="complex64"){let h=i.map(_=>so({inputs:{input:_},backend:t})),g=i.map(_=>li({inputs:{input:_},backend:t})),x=yl({inputs:h,backend:t,attrs:{axis:s}}),b=yl({inputs:g,backend:t,attrs:{axis:s}}),w=dr({inputs:{real:x,imag:b},backend:t});return h.forEach(_=>t.disposeIntermediateTensorInfo(_)),g.forEach(_=>t.disposeIntermediateTensorInfo(_)),t.disposeIntermediateTensorInfo(x),t.disposeIntermediateTensorInfo(b),w}let u=i.map(h=>{let g=y.sizeFromShape(h.shape.slice(s));return Ze({inputs:{x:h},backend:t,attrs:{shape:[-1,g]}})}),c=u.map(h=>({vals:t.data.get(h.dataId).values,shape:h.shape}));a=C.computeOutShape(u.map(h=>h.shape),1);let p=u[0].shape[0]===1,m=Eu(c,a,e[0].dtype,p),f=C.computeOutShape(i.map(h=>h.shape),s),d=t.makeTensorInfo(f,e[0].dtype,m);return u.forEach(h=>t.disposeIntermediateTensorInfo(h)),d}var KA={kernelName:Vs,backendName:"cpu",kernelFunc:yl};function nk(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s}=e,{strides:a,pad:i,dataFormat:l,dilations:u,dimRoundingMode:c}=n;te([o,s],"conv2d");let p=C.convertConv2DDataFormat(l),m=C.computeConv2DInfo(o.shape,s.shape,a,u,i,c,!1,p),f=m.filterHeight,d=m.filterWidth,h=m.dilationHeight,g=m.dilationWidth,x=m.padInfo.left,b=m.padInfo.top,w=m.dataFormat==="channelsLast",_=new ut(m.outShape,o.dtype),I=y.computeStrides(o.shape),E=y.computeStrides(s.shape),D=I[0],$=w?I[1]:I[2],R=w?I[2]:1,M=w?1:I[1],G=_.strides[0],j=w?_.strides[1]:_.strides[2],U=w?_.strides[2]:1,H=w?1:_.strides[1],q=t.data.get(o.dataId).values,X=t.data.get(s.dataId).values,ne=_.values;for(let Y=0;Y<m.batchSize;++Y){let re=Y*D,ee=Y*G;for(let ie=0;ie<m.outHeight;++ie){let me=ee+ie*j,ae=ie*m.strideHeight-b;for(let de=0;de<f;++de){let he=ae+de*h;if(he<0||he>=m.inHeight)continue;let xe=de*E[0],we=re+he*$;for(let De=0;De<m.outWidth;++De){let ve=me+De*U,Ge=De*m.strideWidth-x;for(let Ke=0;Ke<d;++Ke){let at=Ge+Ke*g;if(at<0||at>=m.inWidth)continue;let At=xe+Ke*E[1],Et=we+at*R,He=At;for(let ct=0;ct<m.inChannels;++ct){let mt=q[Et+ct*M];for(let zt=0;zt<m.outChannels;++zt)ne[ve+zt*H]+=mt*X[He+zt];He+=m.outChannels}}}}}}return t.makeTensorInfo(_.shape,_.dtype,ne)}var XA={kernelName:No,backendName:"cpu",kernelFunc:nk};function k6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,dy:s}=e,{strides:a,pad:i,dataFormat:l,dimRoundingMode:u,filterShape:c}=n;te([o,s],"conv2dBackpropFilter");let p=C.convertConv2DDataFormat(l),m=C.computeConv2DInfo(o.shape,c,a,1,i,u,!1,p),{strideHeight:f,strideWidth:d,filterHeight:h,filterWidth:g}=m,x=m.dataFormat==="channelsLast",b=new ut(m.filterShape,"float32"),w=m.padInfo.left,_=m.padInfo.top,I=t.data.get(o.dataId).values,E=t.data.get(s.dataId).values,D=new ut(o.shape,o.dtype,I),$=new ut(s.shape,s.dtype,E);for(let R=0;R<h;++R){let M=Math.max(0,Math.ceil((_-R)/f)),G=Math.min(m.outHeight,(m.inHeight+_-R)/f);for(let j=0;j<g;++j){let U=Math.max(0,Math.ceil((w-j)/d)),H=Math.min(m.outWidth,(m.inWidth+w-j)/d);for(let q=0;q<m.inChannels;++q)for(let X=0;X<m.outChannels;++X){let ne=0;for(let Y=0;Y<m.batchSize;++Y)for(let re=M;re<G;++re){let ee=R+re*f-_;for(let ie=U;ie<H;++ie){let me=j+ie*d-w;x?ne+=D.get(Y,ee,me,q)*$.get(Y,re,ie,X):ne+=D.get(Y,q,ee,me)*$.get(Y,X,re,ie)}}b.set(ne,R,j,q,X)}}}return t.makeTensorInfo(b.shape,b.dtype,b.values)}var YA={kernelName:Cc,backendName:"cpu",kernelFunc:k6};function v6(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,filter:s}=e,{inputShape:a,strides:i,pad:l,dataFormat:u,dimRoundingMode:c}=n;te([o,s],"conv2dBackpropInput");let p=y.computeStrides(s.shape),m=y.computeStrides(o.shape),f=C.convertConv2DDataFormat(u),d=C.computeConv2DInfo(a,s.shape,i,1,l,c,!1,f),h=new ut(d.inShape,"float32"),g=h.values,x=t.data.get(o.dataId).values,b=t.data.get(s.dataId).values,[w,_,I]=p,{batchSize:E,filterHeight:D,filterWidth:$,inChannels:R,inHeight:M,inWidth:G,outChannels:j,outHeight:U,outWidth:H,strideHeight:q,strideWidth:X}=d;f=d.dataFormat;let ne=D-1-d.padInfo.top,Y=$-1-d.padInfo.left,re=f==="channelsLast",ee=h.strides[0],ie=re?h.strides[1]:h.strides[2],me=re?h.strides[2]:1,ae=re?1:h.strides[1],de=m[0],he=re?m[1]:m[2],xe=re?m[2]:1,we=re?1:m[1];for(let De=0;De<E;++De)for(let ve=0;ve<R;++ve)for(let Ge=0;Ge<M;++Ge){let Ke=Ge-ne,at=Math.max(0,Math.ceil(Ke/q)),At=Math.min(U,(D+Ke)/q);for(let Et=0;Et<G;++Et){let He=Et-Y,ct=Math.max(0,Math.ceil(He/X)),mt=Math.min(H,($+He)/X),zt=0;for(let Zt=at;Zt<At;++Zt){let cn=Zt*q-Ke;for(let Mr=ct;Mr<mt;++Mr){let Un=Mr*X-He,ir=de*De+he*Zt+xe*Mr,Cn=w*(D-1-cn)+_*($-1-Un)+I*ve;for(let Ur=0;Ur<j;++Ur){let _r=x[ir+we*Ur],pn=b[Cn+Ur];zt+=_r*pn}}}let vn=ee*De+ie*Ge+me*Et+ae*ve;g[vn]=zt}}return t.makeTensorInfo(h.shape,h.dtype,h.values)}var ZA={kernelName:So,backendName:"cpu",kernelFunc:v6};function C6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s}=e,{strides:a,pad:i,dilations:l}=n;te([o,s],"conv3d");let u=C.computeConv3DInfo(o.shape,s.shape,a,l,i),{filterDepth:c,filterHeight:p,filterWidth:m,dilationDepth:f,dilationHeight:d,dilationWidth:h,padInfo:g}=u,x=g.front,b=g.left,w=g.top,_=new ut(u.outShape,o.dtype),I=t.data.get(o.dataId).values,E=t.data.get(s.dataId).values,D=_.values,$=y.computeStrides(o.shape),R=y.computeStrides(s.shape);for(let M=0;M<u.batchSize;++M){let G=M*$[0],j=M*_.strides[0];for(let U=0;U<u.outDepth;++U){let H=j+U*_.strides[1],q=U*u.strideDepth-x;for(let X=0;X<c;++X){let ne=q+X*f;if(ne<0||ne>=u.inDepth)continue;let Y=X*R[0],re=G+ne*$[1];for(let ee=0;ee<u.outHeight;++ee){let ie=H+ee*_.strides[2],me=ee*u.strideHeight-w;for(let ae=0;ae<p;++ae){let de=me+ae*d;if(de<0||de>=u.inHeight)continue;let he=Y+ae*R[1],xe=re+de*$[2];for(let we=0;we<u.outWidth;++we){let De=ie+we*u.outChannels,ve=we*u.strideWidth-b;for(let Ge=0;Ge<m;++Ge){let Ke=ve+Ge*h;if(Ke<0||Ke>=u.inWidth)continue;let at=he+Ge*R[2],At=xe+Ke*u.inChannels,Et=at;for(let He=0;He<u.inChannels;++He){let ct=I[At+He];for(let mt=0;mt<u.outChannels;++mt)D[De+mt]+=ct*E[Et+mt];Et+=u.outChannels}}}}}}}}return t.makeTensorInfo(_.shape,_.dtype,_.values)}var JA={kernelName:el,backendName:"cpu",kernelFunc:C6};function I6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,dy:s}=e,{strides:a,pad:i,filterShape:l}=n;te([o,s],"conv3dBackpropFilterV2");let u=y.computeStrides(o.shape),c=y.computeStrides(s.shape),p=C.computeConv3DInfo(o.shape,l,a,1,i),m=p.strideDepth,f=p.strideHeight,d=p.strideWidth,h=p.filterDepth,g=p.filterHeight,x=p.filterWidth,b=new ut(p.filterShape,"float32"),w=b.values,[_,I,E,D]=b.strides,$=t.data.get(s.dataId).values,[R,M,G,j]=c,U=t.data.get(o.dataId).values,[H,q,X,ne]=u,Y=p.padInfo.front,re=p.padInfo.left,ee=p.padInfo.top;for(let ie=0;ie<h;++ie){let me=Math.max(0,Math.ceil((Y-ie)/m)),ae=Math.min(p.outDepth,(p.inDepth+Y-ie)/m),de=ie*_;for(let he=0;he<g;++he){let xe=Math.max(0,Math.ceil((ee-he)/f)),we=Math.min(p.outHeight,(p.inHeight+ee-he)/f),De=he*I+de;for(let ve=0;ve<x;++ve){let Ge=Math.max(0,Math.ceil((re-ve)/d)),Ke=Math.min(p.outWidth,(p.inWidth+re-ve)/d),at=ve*E+De;for(let At=0;At<p.inChannels;++At){let Et=At*D+at;for(let He=0;He<p.outChannels;++He){let ct=0;for(let mt=0;mt<p.batchSize;++mt){let zt=mt*H,vn=mt*R;for(let Zt=me;Zt<ae;++Zt){let Mr=(ie+Zt*m-Y)*q+zt,Un=Zt*M+vn;for(let ir=xe;ir<we;++ir){let Ur=(he+ir*f-ee)*X+Mr,_r=ir*G+Un;for(let pn=Ge;pn<Ke;++pn){let zl=(ve+pn*d-re)*ne+Ur,Wa=pn*j+_r;ct+=U[zl+At]*$[Wa+He]}}}}w[Et+He]=ct}}}}}return t.makeTensorInfo(b.shape,b.dtype,b.values)}var QA={kernelName:Ic,backendName:"cpu",kernelFunc:I6};function N6(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,filter:s}=e,{pad:a,strides:i,inputShape:l}=n;te([o],"conv3dBackpropInputV2");let u=y.computeStrides(o.shape),c=y.computeStrides(s.shape),p=C.computeConv3DInfo(l,s.shape,i,1,a),m=new ut(p.inShape,"float32"),f=m.values,[d,h,g,x]=m.strides,b=t.data.get(o.dataId).values,[w,_,I,E]=u,D=t.data.get(s.dataId).values,[$,R,M,G]=c,{batchSize:j,filterDepth:U,filterHeight:H,filterWidth:q,inChannels:X,inDepth:ne,inHeight:Y,inWidth:re,outChannels:ee,outDepth:ie,outHeight:me,outWidth:ae,strideDepth:de,strideHeight:he,strideWidth:xe}=p,we=U-1-p.padInfo.front,De=H-1-p.padInfo.top,ve=q-1-p.padInfo.left;for(let Ge=0;Ge<j;++Ge)for(let Ke=0;Ke<X;++Ke)for(let at=0;at<ne;++at){let At=at-we,Et=Math.max(0,Math.ceil(At/de)),He=Math.min(ie,(U+At)/de);for(let ct=0;ct<Y;++ct){let mt=ct-De,zt=Math.max(0,Math.ceil(mt/he)),vn=Math.min(me,(H+mt)/he);for(let Zt=0;Zt<re;++Zt){let cn=Zt-ve,Mr=Math.max(0,Math.ceil(cn/xe)),Un=Math.min(ae,(q+cn)/xe),ir=0;for(let Cn=Et;Cn<He;++Cn){let Ur=Cn*de-At;for(let _r=zt;_r<vn;++_r){let pn=_r*he-mt;for(let Fn=Mr;Fn<Un;++Fn){let zl=Fn*xe-cn,Wa=w*Ge+_*Cn+I*_r+E*Fn,Ps=$*(U-1-Ur)+R*(H-1-pn)+M*(q-1-zl)+G*Ke;for(let xi=0;xi<ee;++xi){let pm=b[Wa+xi],Bl=D[Ps+xi];ir+=pm*Bl}}}}f[d*Ge+h*at+g*ct+x*Zt+Ke]=ir}}}return t.makeTensorInfo(m.shape,m.dtype,m.values)}var eE={kernelName:Nc,backendName:"cpu",kernelFunc:N6};var S6=$e(To,r=>Math.cos(r)),tE={kernelName:To,backendName:"cpu",kernelFunc:S6};var T6=$e(Di,r=>Math.cosh(r)),rE={kernelName:Di,backendName:"cpu",kernelFunc:T6};function A6(r){let{inputs:e,backend:t,attrs:n}=r,{image:o,boxes:s,boxInd:a}=e,{cropSize:i,method:l,extrapolationValue:u}=n,[c,p,m,f]=o.shape,d=s.shape[0],[h,g]=i,x=Ce([d,h,g,f],"float32"),b=t.data.get(s.dataId).values,w=t.data.get(a.dataId).values,_=t.data.get(o.dataId).values,I=y.computeStrides(o.shape),E=y.computeStrides(x.shape);for(let D=0;D<d;D++){let $=D*4,R=b[$],M=b[$+1],G=b[$+2],j=b[$+3],U=w[D];if(U>=c)continue;let H=h>1?(G-R)*(p-1)/(h-1):0,q=g>1?(j-M)*(m-1)/(g-1):0;for(let X=0;X<h;X++){let ne=h>1?R*(p-1)+X*H:.5*(R+G)*(p-1);if(ne<0||ne>p-1){for(let Y=0;Y<g;Y++)for(let re=0;re<f;re++){let ee=re+Y*E[2]+X*E[1]+D*E[0];x.values[ee]=u}continue}if(l==="bilinear"){let Y=Math.floor(ne),re=Math.ceil(ne),ee=ne-Y;for(let ie=0;ie<g;ie++){let me=g>1?M*(m-1)+ie*q:.5*(M+j)*(m-1);if(me<0||me>m-1){for(let xe=0;xe<f;xe++){let we=xe+ie*E[2]+X*E[1]+D*E[0];x.values[we]=u}continue}let ae=Math.floor(me),de=Math.ceil(me),he=me-ae;for(let xe=0;xe<f;xe++){let we=xe+ae*I[2]+Y*I[1]+U*I[0],De=_[we];we=xe+de*I[2]+Y*I[1]+U*I[0];let ve=_[we];we=xe+ae*I[2]+re*I[1]+U*I[0];let Ge=_[we];we=xe+de*I[2]+re*I[1]+U*I[0];let Ke=_[we],at=De+(ve-De)*he,At=Ge+(Ke-Ge)*he;we=xe+ie*E[2]+X*E[1]+D*E[0],x.values[we]=at+(At-at)*ee}}}else for(let Y=0;Y<g;++Y){let re=g>1?M*(m-1)+Y*q:.5*(M+j)*(m-1);if(re<0||re>m-1){for(let me=0;me<f;me++){let ae=me+Y*E[2]+X*E[1]+D*E[0];x.values[ae]=u}continue}let ee=Math.round(re),ie=Math.round(ne);for(let me=0;me<f;me++){let ae=me+ee*I[2]+ie*I[1]+U*I[0],de=me+Y*E[2]+X*E[1]+D*E[0];x.values[de]=_[ae]}}}}return t.makeTensorInfo(x.shape,x.dtype,x.values)}var nE={kernelName:$i,backendName:"cpu",kernelFunc:A6};function E6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,exclusive:a,reverse:i}=n;te(o,"cumsum");let l=C.getAxesPermutation([s],o.shape.length),u=o;l!=null&&(u=Xt({inputs:{x:o},backend:t,attrs:{perm:l}}));let c=C.getInnerMostAxes(1,o.shape.length)[0];if(c!==u.shape.length-1)throw new Error(`backend.cumsum in CPU expects an inner-most axis=${u.shape.length-1} but got axis=${c}`);let p=lr(u.dtype,"int32"),m=y.makeZerosTypedArray(y.sizeFromShape(u.shape),p),f=t.data.get(u.dataId).values,d=u.shape[u.shape.length-1],h=i?(x,b)=>x+d-b-1:(x,b)=>x+b;for(let x=0;x<f.length;x+=d)for(let b=0;b<d;b++){let w=h(x,b);if(b===0)m[w]=a?0:f[w];else{let _=h(x,b-1);m[w]=a?f[_]+m[_]:f[w]+m[_]}}let g=t.makeTensorInfo(u.shape,p,m);if(l!=null){let x=C.getUndoAxesPermutation(l),b=Xt({inputs:{x:g},backend:t,attrs:{perm:x}});return t.disposeIntermediateTensorInfo(g),t.disposeIntermediateTensorInfo(u),b}return g}var oE={kernelName:Ao,backendName:"cpu",kernelFunc:E6};function D6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,weights:s}=e,{size:a,binaryOutput:i}=n;if(o.shape.length===1){let l=t.data.get(o.dataId).values,u=t.data.get(s.dataId).values,c=_p(l,u,s.dtype,s.shape,a);return t.makeTensorInfo([a],s.dtype,c)}else if(o.shape.length===2){let l=t.bufferSync(o),u=t.bufferSync(s),c=Mg(l,u,a,i);return t.makeTensorInfo(c.shape,s.dtype,c.values)}throw new Error(`Error in denseBincount: input must be at most rank 2, but got rank${o.shape.length}.`)}var sE={kernelName:Sc,backendName:"cpu",kernelFunc:D6};function $6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{blockSize:s,dataFormat:a}=n;y.assert(a==="NHWC",()=>`Only NHWC dataFormat supported on CPU for depthToSpace. Got ${a}`),y.assert(s>1,()=>`blockSize should be > 1 for depthToSpace, but was: ${s}`);let i=o.shape[0],l=o.shape[1],u=o.shape[2],c=o.shape[3],p=l*s,m=u*s,f=c/(s*s),d=t.data.get(o.dataId).values,h=new Float32Array(i*p*m*f),g=0;for(let x=0;x<i;++x)for(let b=0;b<p;++b){let w=Math.floor(b/s),_=b%s;for(let I=0;I<m;++I){let E=Math.floor(I/s),D=I%s,$=(_*s+D)*f;for(let R=0;R<f;++R){let G=R+$+c*(E+u*(w+l*x));h[g++]=d[G]}}}return t.makeTensorInfo([i,p,m,f],o.dtype,h)}var iE={kernelName:Ri,backendName:"cpu",kernelFunc:$6};function ok(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s}=e,{strides:a,pad:i,dilations:l,dimRoundingMode:u}=n;te([o,s],"depthwiseConv2DNative");let c=y.computeStrides(o.shape),p=y.computeStrides(s.shape),m=l;m==null&&(m=[1,1]),y.assert(C.eitherStridesOrDilationsAreOne(a,m),()=>`Error in depthwiseConv2d: Either strides or dilations must be 1. Got strides ${a} and dilations '${m}'`);let f=C.computeConv2DInfo(o.shape,s.shape,a,m,i,u,!0),{filterHeight:d,filterWidth:h,dilationHeight:g,dilationWidth:x,padInfo:b}=f,w=b.left,_=b.top,I=f.outChannels/f.inChannels,E=new ut(f.outShape,o.dtype),D=t.data.get(o.dataId).values,$=t.data.get(s.dataId).values,R=E.values;for(let M=0;M<f.batchSize;++M){let G=M*c[0],j=M*E.strides[0];for(let U=0;U<f.outHeight;++U){let H=j+U*E.strides[1],q=U*f.strideHeight-_;for(let X=0;X<d;++X){let ne=q+X*g;if(ne<0||ne>=f.inHeight)continue;let Y=X*p[0],re=G+ne*c[1];for(let ee=0;ee<f.outWidth;++ee){let ie=H+ee*E.strides[2],me=ee*f.strideWidth-w;for(let ae=0;ae<h;++ae){let de=me+ae*x;if(de<0||de>=f.inWidth)continue;let he=Y+ae*p[1],xe=re+de*f.inChannels,we=ie,De=he;for(let ve=0;ve<f.inChannels;++ve){let Ge=D[xe+ve];for(let Ke=0;Ke<I;++Ke)R[we+Ke]+=Ge*$[De+Ke];we+=I,De+=I}}}}}}return t.makeTensorInfo(E.shape,E.dtype,E.values)}var aE={kernelName:Eo,backendName:"cpu",kernelFunc:ok};function R6(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,dy:s}=e,{strides:a,dilations:i,pad:l,dimRoundingMode:u,filterShape:c}=n;te([o,s],"depthwiseConv2dNativeBackpropFilter");let p=C.computeConv2DInfo(o.shape,c,a,i,l,u,!0),{strideHeight:m,strideWidth:f,filterHeight:d,filterWidth:h}=p,g=new ut(p.filterShape,"float32"),x=p.padInfo.left,b=p.padInfo.top,w=p.outChannels/p.inChannels,_=t.data.get(o.dataId).values,I=new ut(o.shape,o.dtype,_),E=t.data.get(s.dataId).values,D=new ut(s.shape,s.dtype,E);for(let $=0;$<d;++$){let R=Math.max(0,Math.ceil((b-$)/m)),M=Math.min(p.outHeight,(p.inHeight+b-$)/m);for(let G=0;G<h;++G){let j=Math.max(0,Math.ceil((x-G)/f)),U=Math.min(p.outWidth,(p.inWidth+x-G)/f);for(let H=0;H<p.outChannels;++H){let q=Math.trunc(H/w),X=H%w,ne=0;for(let Y=0;Y<p.batchSize;++Y)for(let re=R;re<M;++re){let ee=$+re*m-b;for(let ie=j;ie<U;++ie){let me=G+ie*f-x;ne+=I.get(Y,ee,me,q)*D.get(Y,re,ie,H)}}g.set(ne,$,G,q,X)}}}return t.makeTensorInfo(g.shape,g.dtype,g.values)}var lE={kernelName:Tc,backendName:"cpu",kernelFunc:R6};function F6(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,filter:s}=e,{strides:a,dilations:i,pad:l,dimRoundingMode:u,inputShape:c}=n;te([o,s],"depthwiseConv2DNativeBackpropInput");let p=y.computeStrides(o.shape),m=y.computeStrides(s.shape),f=C.computeConv2DInfo(c,s.shape,a,i,l,u,!0),d=new ut(f.inShape,"float32"),h=d.values,[g,x,b]=d.strides,w=t.data.get(o.dataId).values,[_,I,E]=p,D=t.data.get(s.dataId).values,[$,R,M]=m,{batchSize:G,filterHeight:j,filterWidth:U,inChannels:H,inHeight:q,inWidth:X,outChannels:ne,outHeight:Y,outWidth:re,strideHeight:ee,strideWidth:ie}=f,me=j-1-f.padInfo.top,ae=U-1-f.padInfo.left,de=ne/H;for(let he=0;he<G;++he)for(let xe=0;xe<H;++xe)for(let we=0;we<q;++we){let De=we-me,ve=Math.max(0,Math.ceil(De/ee)),Ge=Math.min(Y,(j+De)/ee);for(let Ke=0;Ke<X;++Ke){let at=Ke-ae,At=Math.max(0,Math.ceil(at/ie)),Et=Math.min(re,(U+at)/ie),He=0;for(let ct=ve;ct<Ge;++ct){let mt=ct*ee-De;for(let zt=At;zt<Et;++zt){let vn=zt*ie-at,Zt=_*he+I*ct+E*zt,cn=$*(j-1-mt)+R*(U-1-vn)+M*xe;for(let Mr=0;Mr<de;++Mr){let Un=xe*de+Mr,ir=w[Zt+Un],Cn=D[cn+Mr];He+=ir*Cn}}}h[g*he+x*we+b*Ke+xe]=He}}return t.makeTensorInfo(d.shape,d.dtype,d.values)}var uE={kernelName:Ac,backendName:"cpu",kernelFunc:F6};function O6(r){let{inputs:e,backend:t}=r,{x:n}=e,o=y.sizeFromShape(n.shape),s=t.data.get(n.dataId).values,a=Ce([o,o],n.dtype),i=a.values;for(let u=0;u<s.length;u++)i[u*o+u]=s[u];let l=[...n.shape,...n.shape];return t.makeTensorInfo(l,a.dtype,a.values)}var cE={kernelName:Ec,backendName:"cpu",kernelFunc:O6};var pE={kernelName:tl,backendName:"cpu",kernelFunc:({inputs:r,backend:e,attrs:t})=>{let{x:n,filter:o}=r,{strides:s,pad:a,dilations:i}=t,l=e,u=l.data.get(n.dataId).values,c=n.shape.length,p=l.data.get(o.dataId).values,m=o.shape.length,{batchSize:f,inHeight:d,inWidth:h,inChannels:g,outHeight:x,outWidth:b,padInfo:w,strideHeight:_,strideWidth:I,filterHeight:E,filterWidth:D,dilationHeight:$,dilationWidth:R,outShape:M}=C.computeDilation2DInfo(n.shape,o.shape,s,a,"NHWC",i),G=y.sizeFromShape(M),j=M.length,U=y.getArrayFromDType(n.dtype,G);for(let q=0;q<f;++q)for(let X=0;X<x;++X){let ne=X*_-w.top;for(let Y=0;Y<b;++Y){let re=Y*I-w.left;for(let ee=0;ee<g;++ee){let ie=Number.MIN_SAFE_INTEGER;for(let ae=0;ae<E;++ae){let de=ne+ae*$;if(de>=0&&de<d)for(let he=0;he<D;++he){let xe=re+he*R;if(xe>=0&&xe<h){let we=y.locToIndex([q,de,xe,ee],c,y.computeStrides(n.shape)),De=y.locToIndex([ae,he,ee],m,y.computeStrides(o.shape)),ve=u[we]+p[De];ve>ie&&(ie=ve)}}}let me=y.locToIndex([q,X,Y,ee],j,y.computeStrides(M));U[me]=ie}}}return{dataId:l.write(y.toTypedArray(U,n.dtype),M,n.dtype),shape:M,dtype:n.dtype}}};var mE={kernelName:$m,backendName:"cpu",kernelFunc:({inputs:r,backend:e,attrs:t})=>{let{x:n,filter:o,dy:s}=r,{strides:a,pad:i,dilations:l}=t,u=e,c=y.toNestedArray(n.shape,u.data.get(n.dataId).values),p=y.toNestedArray(o.shape,u.data.get(o.dataId).values),{batchSize:m,inHeight:f,inWidth:d,inChannels:h,outHeight:g,outWidth:x,padInfo:b,strideHeight:w,strideWidth:_,filterHeight:I,filterWidth:E,dilationHeight:D,dilationWidth:$,outShape:R}=C.computeDilation2DInfo(n.shape,o.shape,a,i,"NHWC",l);y.assert(s.rank===R.length,()=>`Error in ${$m}, dy must have the same rank as output ${R.length}, but got ${s.rank}`);let M=y.toNestedArray(R,u.data.get(s.dataId).values),G=y.makeZerosNestedTypedArray(o.shape,o.dtype);for(let U=0;U<m;++U)for(let H=0;H<g;++H){let q=H*w-b.top;for(let X=0;X<x;++X){let ne=X*_-b.left;for(let Y=0;Y<h;++Y){let re=Number.MIN_SAFE_INTEGER,ee=0,ie=0;for(let me=0;me<I;++me){let ae=q+me*D;if(ae>=0&&ae<f)for(let de=0;de<E;++de){let he=ne+de*$;if(he>=0&&he<d){let xe=c[U][ae][he][Y]+p[me][de][Y];xe>re&&(re=xe,ee=me,ie=de)}}}G[ee][ie][Y]+=M[U][H][X][Y]}}}return{dataId:u.write(y.toTypedArray(G,n.dtype),o.shape,o.dtype),shape:o.shape,dtype:o.dtype}}};var fE={kernelName:Dm,backendName:"cpu",kernelFunc:({inputs:r,backend:e,attrs:t})=>{let{x:n,filter:o,dy:s}=r,{strides:a,pad:i,dilations:l}=t,u=e,c=y.toNestedArray(n.shape,u.data.get(n.dataId).values),p=y.toNestedArray(o.shape,u.data.get(o.dataId).values),{batchSize:m,inHeight:f,inWidth:d,inChannels:h,outHeight:g,outWidth:x,padInfo:b,strideHeight:w,strideWidth:_,filterHeight:I,filterWidth:E,dilationHeight:D,dilationWidth:$,outShape:R}=C.computeDilation2DInfo(n.shape,o.shape,a,i,"NHWC",l);y.assert(s.rank===R.length,()=>`Error in ${Dm}, dy must have the same rank as output ${R.length}, but got ${s.rank}`);let M=y.toNestedArray(R,u.data.get(s.dataId).values),G=y.makeZerosNestedTypedArray(n.shape,n.dtype);for(let U=0;U<m;++U)for(let H=0;H<g;++H){let q=H*w-b.top;for(let X=0;X<x;++X){let ne=X*_-b.left;for(let Y=0;Y<h;++Y){let re=Number.MIN_SAFE_INTEGER,ee=q<0?0:q,ie=ne<0?0:ne;for(let me=0;me<I;++me){let ae=q+me*D;if(ae>=0&&ae<f)for(let de=0;de<E;++de){let he=ne+de*$;if(he>=0&&he<d){let xe=c[U][ae][he][Y]+p[me][de][Y];xe>re&&(re=xe,ee=ae,ie=he)}}}G[U][ee][ie][Y]+=M[U][H][X][Y]}}}return{dataId:u.write(y.toTypedArray(G,n.dtype),n.shape,n.dtype),shape:n.shape,dtype:n.dtype}}};function Ea(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n;te(o,"sum");let i;o.dtype==="bool"?i=io({inputs:{x:o},backend:t,attrs:{dtype:"int32"}}):i=Or({inputs:{x:o},backend:t});let l=i.shape.length,u=y.parseAxisParam(s,i.shape),c=C.getAxesPermutation(u,l),p=u,m=i;c!=null&&(m=Xt({inputs:{x:i},backend:t,attrs:{perm:c}}),p=C.getInnerMostAxes(p.length,l)),C.assertAxesAreInnerMostDims("sum",p,m.shape.length);let[f,d]=C.computeOutAndReduceShapes(m.shape,p),h=C.upcastType(m.dtype,"int32"),g=bp(t,f,h),x=y.sizeFromShape(d),b=t.data.get(g.dataId).values,w=t.data.get(m.dataId).values;for(let _=0;_<b.length;++_){let I=_*x,E=0;for(let D=0;D<x;++D)E+=w[I+D];b[_]=E}if(a){let _=C.expandShapeToKeepDim(g.shape,u),I=g;g=Ze({inputs:{x:g},backend:t,attrs:{shape:_}}),t.disposeIntermediateTensorInfo(I)}return t.disposeIntermediateTensorInfo(i),c!=null&&t.disposeIntermediateTensorInfo(m),g}var dE={kernelName:is,backendName:"cpu",kernelFunc:Ea};function P6(r){let{inputs:e,backend:t,attrs:n}=r,{equation:o}=n,s=e,{allDims:a,summedDims:i,idDims:l}=C.decodeEinsumEquation(o,s.length);C.checkEinsumDimSizes(a.length,l,s);let{path:u,steps:c}=C.getEinsumComputePath(i,l),p=c.length,m=null,f=a.length,d=[];for(let h=0;h<p;++h){for(let g of c[h]){let{permutationIndices:x,expandDims:b}=C.getEinsumPermutation(f,l[g]),w;C.isIdentityPermutation(x)?w=s[g]:(w=Xt({inputs:{x:s[g]},backend:t,attrs:{perm:x}}),d.push(w));let _=w.shape.slice();for(let I=0;I<b.length;++I)_.splice(b[I],0,1);y.arraysEqual(w.shape,_)||(w=Ze({inputs:{x:w},backend:t,attrs:{shape:_}}),d.push(w)),m===null?m=w:(m=Du({inputs:{a:w,b:m},backend:t}),d.push(m))}h<p-1&&(u[h]>=0&&(m=Ea({inputs:{x:m},backend:t,attrs:{axis:u[h]-(a.length-f),keepDims:!1}}),d.push(m)),f--)}for(let h of d)h!==m&&t.disposeIntermediateTensorInfo(h);return m}var hE={kernelName:Dc,backendName:"cpu",kernelFunc:P6};function M6(r){let{inputs:e,backend:t}=r,{dy:n,y:o}=e;te([n,o],"eluGrad");let s=new Float32Array(y.sizeFromShape(o.shape)),a=t.data.get(o.dataId).values,i=t.data.get(n.dataId).values;for(let l=0;l<a.length;++l){let u=a[l];u>=1?s[l]=i[l]:s[l]=i[l]*(u+1)}return t.makeTensorInfo(o.shape,"float32",s)}var gE={kernelName:$c,backendName:"cpu",kernelFunc:M6};var L6=Ye((r,e)=>r===e?1:0),sk=et(Pi,L6,null,"bool"),xE={kernelName:Pi,backendName:"cpu",kernelFunc:sk};var z6=C.ERF_P,B6=C.ERF_A1,V6=C.ERF_A2,G6=C.ERF_A3,W6=C.ERF_A4,j6=C.ERF_A5,U6=$e(Oi,r=>{let e=Math.sign(r),t=Math.abs(r),n=1/(1+z6*t);return e*(1-((((j6*n+W6)*n+G6)*n+V6)*n+B6)*n*Math.exp(-t*t))}),yE={kernelName:Oi,backendName:"cpu",kernelFunc:U6};function Ip(r){let{inputs:e,backend:t,attrs:n}=r,{input:o}=e,{dim:s}=n,a=o.shape.length,i=o.shape.slice(),l=s;return s<0&&(y.assert(-(a+1)<=s,()=>`Axis must be in the interval [${-(a+1)}, ${a}]`),l=a+s+1),i.splice(l,0,1),Ze({inputs:{x:o},backend:t,attrs:{shape:i}})}var bE={kernelName:Gs,backendName:"cpu",kernelFunc:Ip};var H6=Ye((r,e)=>r/e),Nf=et(Do,H6),Sf={kernelName:Do,backendName:"cpu",kernelFunc:Nf};function Yg(r,e,t){let n=r.shape,o=n[0],s=n[1],a=t.data.get(r.dataId),i=a.complexTensorInfos.real,l=a.complexTensorInfos.imag,u=[o,s],c=y.sizeFromShape(u),p=y.getTypedArrayFromDType("float32",c),m=y.getTypedArrayFromDType("float32",c);for(let g=0;g<o;g++){let x=uo({inputs:{x:i},backend:t,attrs:{begin:[g,0],size:[1,s]}}),b=uo({inputs:{x:l},backend:t,attrs:{begin:[g,0],size:[1,s]}}),w=dr({inputs:{real:x,imag:b},backend:t}),{real:_,imag:I}=q6(w,e,t),E=C.mergeRealAndImagArrays(_,I);for(let D=0;D<s;D++){let $=C.getComplexWithIndex(E,D);p[g*s+D]=$.real,m[g*s+D]=$.imag}t.disposeIntermediateTensorInfo(x),t.disposeIntermediateTensorInfo(b),t.disposeIntermediateTensorInfo(w)}let f=t.makeTensorInfo(u,"float32",p),d=t.makeTensorInfo(u,"float32",m),h=dr({inputs:{real:f,imag:d},backend:t});return t.disposeIntermediateTensorInfo(f),t.disposeIntermediateTensorInfo(d),h}function q6(r,e,t){let n=y.sizeFromShape(r.shape),o=t.data.get(r.dataId),s=t.data.get(o.complexTensorInfos.real.dataId).values,a=t.data.get(o.complexTensorInfos.imag.dataId).values;if(K6(n)){let i=ik(s,a,n,e,t),l=[r.shape[0],r.shape[1]];if(e){let u=t.makeTensorInfo(l,"float32",i.real),c=t.makeTensorInfo(l,"float32",i.imag),p=t.makeTensorInfo([],"float32",y.createScalarValue(n,"float32")),m=Or({inputs:{x:p},backend:t}),f=Sf.kernelFunc({inputs:{a:u,b:p},backend:t}),d=Sf.kernelFunc({inputs:{a:c,b:m},backend:t}),h=t.data.get(f.dataId).values,g=t.data.get(d.dataId).values;return t.disposeIntermediateTensorInfo(u),t.disposeIntermediateTensorInfo(c),t.disposeIntermediateTensorInfo(p),t.disposeIntermediateTensorInfo(m),t.disposeIntermediateTensorInfo(f),t.disposeIntermediateTensorInfo(d),{real:h,imag:g}}return i}else{let i=C.mergeRealAndImagArrays(s,a),l=X6(i,n,e);return C.splitRealAndImagArrays(l)}}function K6(r){return(r&r-1)==0}function ik(r,e,t,n,o){if(t===1)return{real:r,imag:e};let s=C.mergeRealAndImagArrays(r,e),a=t/2,i=C.complexWithEvenIndex(s),l=i.real,u=i.imag,c=[l.length],p=o.makeTensorInfo(c,"float32",l),m=o.makeTensorInfo(c,"float32",u),f=dr({inputs:{real:p,imag:m},backend:o}),d=C.complexWithOddIndex(s),h=d.real,g=d.imag,x=[h.length],b=o.makeTensorInfo(x,"float32",h),w=o.makeTensorInfo(x,"float32",g),_=dr({inputs:{real:b,imag:w},backend:o}),I=ik(l,u,a,n,o),E=I.real,D=I.imag,$=[E.length],R=o.makeTensorInfo($,"float32",E),M=o.makeTensorInfo($,"float32",D),G=dr({inputs:{real:R,imag:M},backend:o}),j=ik(h,g,a,n,o),U=j.real,H=j.imag,q=[U.length],X=o.makeTensorInfo(q,"float32",U),ne=o.makeTensorInfo(q,"float32",H),Y=dr({inputs:{real:X,imag:ne},backend:o}),re=C.exponents(t,n),ee=[re.real.length],ie=o.makeTensorInfo(ee,"float32",re.real),me=o.makeTensorInfo(ee,"float32",re.imag),ae=dr({inputs:{real:ie,imag:me},backend:o}),de=Du({inputs:{a:ae,b:Y},backend:o}),he=Aa({inputs:{a:G,b:de},backend:o}),xe=If({inputs:{a:G,b:de},backend:o}),we=so({inputs:{input:he},backend:o}),De=so({inputs:{input:xe},backend:o}),ve=li({inputs:{input:he},backend:o}),Ge=li({inputs:{input:xe},backend:o}),Ke=yl({inputs:[we,De],backend:o,attrs:{axis:0}}),at=yl({inputs:[ve,Ge],backend:o,attrs:{axis:0}}),At=o.data.get(Ke.dataId).values,Et=o.data.get(at.dataId).values;return o.disposeIntermediateTensorInfo(p),o.disposeIntermediateTensorInfo(m),o.disposeIntermediateTensorInfo(f),o.disposeIntermediateTensorInfo(b),o.disposeIntermediateTensorInfo(w),o.disposeIntermediateTensorInfo(_),o.disposeIntermediateTensorInfo(R),o.disposeIntermediateTensorInfo(M),o.disposeIntermediateTensorInfo(G),o.disposeIntermediateTensorInfo(X),o.disposeIntermediateTensorInfo(ne),o.disposeIntermediateTensorInfo(Y),o.disposeIntermediateTensorInfo(ie),o.disposeIntermediateTensorInfo(me),o.disposeIntermediateTensorInfo(ae),o.disposeIntermediateTensorInfo(de),o.disposeIntermediateTensorInfo(he),o.disposeIntermediateTensorInfo(xe),o.disposeIntermediateTensorInfo(we),o.disposeIntermediateTensorInfo(ve),o.disposeIntermediateTensorInfo(De),o.disposeIntermediateTensorInfo(Ge),o.disposeIntermediateTensorInfo(Ke),o.disposeIntermediateTensorInfo(at),{real:At,imag:Et}}function X6(r,e,t){let n=new Float32Array(e*2);for(let o=0;o<e;o++){let s=0,a=0;for(let i=0;i<e;i++){let l=C.exponent(o*i,e,t),u=C.getComplexWithIndex(r,i);s+=u.real*l.real-u.imag*l.imag,a+=u.real*l.imag+u.imag*l.real}t&&(s/=e,a/=e),C.assignToTypedArray(n,s,a,o)}return n}function Y6(r){let{inputs:e,backend:t}=r,{input:n}=e,o=y.sizeFromShape(n.shape),s=n.shape[n.shape.length-1],a=o/s,i=Ze({inputs:{x:n},backend:t,attrs:{shape:[a,s]}}),l=Yg(i,!1,t),u=Ze({inputs:{x:l},backend:t,attrs:{shape:n.shape}});return t.disposeIntermediateTensorInfo(i),t.disposeIntermediateTensorInfo(l),u}var wE={kernelName:Rc,backendName:"cpu",kernelFunc:Y6};function Tf(r){let{backend:e,attrs:t}=r,{shape:n,value:o,dtype:s}=t,a=s||y.inferDtype(o),i=y.getArrayFromDType(a,y.sizeFromShape(n));return Z6(i,o,a),e.makeTensorInfo(n,a,i)}var _E={kernelName:rl,backendName:"cpu",kernelFunc:Tf};function Z6(r,e,t){r.fill(e)}var kE={kernelName:Li,backendName:"cpu",kernelFunc:({inputs:r,attrs:e,backend:t})=>{let{image:n}=r,o=t,s=y.getTypedArrayFromDType(n.dtype,y.sizeFromShape(n.shape)),[a,i,l,u]=n.shape,c=o.data.get(n.dataId).values;for(let m=0;m<a;m++){let f=m*l*i*u;for(let d=0;d<i;d++){let h=d*(l*u);for(let g=0;g<l;g++){let x=g*u;for(let b=0;b<u;b++){let _=[a,d,g,b][2],I=Math.round(l-_),E=f+h+x+b,D=c[E];if(I>=0&&I<l){let $=I*u,R=f+h+$+b;D=c[R]}s[E]=D}}}}return{dataId:o.write(s,n.shape,n.dtype),shape:n.shape,dtype:n.dtype}}};var J6=Ye((r,e)=>Math.floor(r/e)),Q6=et(Fo,J6,null,"int32"),vE={kernelName:Fo,backendName:"cpu",kernelFunc:Q6};function e5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s,bias:a,preluActivationWeights:i}=e,{strides:l,pad:u,dataFormat:c,dilations:p,dimRoundingMode:m,activation:f,leakyreluAlpha:d}=n,h=nk({inputs:{x:o,filter:s},backend:t,attrs:{strides:l,pad:u,dataFormat:c,dilations:p,dimRoundingMode:m}});if(a){let g=h;h=Aa({inputs:{a:h,b:a},backend:t}),t.disposeIntermediateTensorInfo(g)}if(f){let g=h;h=vp(t,h,f,i,d),t.disposeIntermediateTensorInfo(g)}return h}var CE={kernelName:ei,backendName:"cpu",kernelFunc:e5};function t5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s,bias:a,preluActivationWeights:i}=e,{strides:l,pad:u,dataFormat:c,dilations:p,dimRoundingMode:m,activation:f,leakyreluAlpha:d}=n,h=ok({inputs:{x:o,filter:s},backend:t,attrs:{strides:l,pad:u,dataFormat:c,dilations:p,dimRoundingMode:m}});if(a){let g=h;h=Aa({inputs:{a:h,b:a},backend:t}),t.disposeIntermediateTensorInfo(g)}if(f){let g=h;h=vp(t,h,f,i,d),t.disposeIntermediateTensorInfo(g)}return h}var IE={kernelName:ti,backendName:"cpu",kernelFunc:t5};function r5(r){let{inputs:e,backend:t}=r,{params:n,indices:o}=e,s=y.sizeFromShape(n.shape),a=o.shape,i=a[a.length-1],[l,u,c,p]=C.prepareAndValidate(n,o);if(u===0)return t.makeTensorInfo(l,n.dtype,[]);let m=Ce([u,c],n.dtype),f=t.data.get(o.dataId).values,d=t.data.get(n.dataId).values;for(let h=0;h<u;h++){let g=[],x=0;for(let b=0;b<i;b++){let w=f[h*i+b];x+=w*p[b],g.push(w)}if(x<0||x>=s/c)throw new Error(`Invalid indices: ${g} does not index into ${n.shape}`);for(let b=0;b<c;b++)m.values[h*c+b]=d[x*c+b]}return t.makeTensorInfo(l,m.dtype,m.values)}var NE={kernelName:zi,backendName:"cpu",kernelFunc:r5};function n5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,indices:s}=e,{axis:a,batchDims:i}=n;te([o,s],"gatherV2");let l=i;i==null&&(l=0);let u=y.sizeFromShape(s.shape),c=y.parseAxisParam(a,o.shape)[0],p=C.segment_util.collectGatherOpShapeInfo(o,s,c,l),m=Ze({inputs:{x:o},backend:t,attrs:{shape:[p.batchSize,p.outerSize,p.dimSize,p.sliceSize]}}),f=Ze({inputs:{x:s},backend:t,attrs:{shape:[p.batchSize,u/p.batchSize]}}),d=[p.batchSize,p.outerSize,u/p.batchSize,p.sliceSize],h=t.bufferSync(f),g=t.bufferSync(m),x=Lg(g,h,d);return t.disposeIntermediateTensorInfo(m),t.disposeIntermediateTensorInfo(f),t.makeTensorInfo(p.outputShape,x.dtype,x.values)}var SE={kernelName:Ws,backendName:"cpu",kernelFunc:n5};var o5=Ye((r,e)=>r>=e?1:0),s5=et(Po,o5,null,"bool"),TE={kernelName:Po,backendName:"cpu",kernelFunc:s5};function i5(r){let{inputs:e,backend:t}=r,{input:n}=e,o=y.sizeFromShape(n.shape),s=n.shape[n.shape.length-1],a=o/s,i=Ze({inputs:{x:n},backend:t,attrs:{shape:[a,s]}}),l=Yg(i,!0,t),u=Ze({inputs:{x:l},backend:t,attrs:{shape:n.shape}});return t.disposeIntermediateTensorInfo(i),t.disposeIntermediateTensorInfo(l),u}var AE={kernelName:Fc,backendName:"cpu",kernelFunc:i5};var a5=$e(Vi,r=>Number.isFinite(r)?1:0,"bool"),EE={kernelName:Vi,backendName:"cpu",kernelFunc:a5};var l5=$e(Gi,r=>Math.abs(r)===Infinity?1:0,"bool"),DE={kernelName:Gi,backendName:"cpu",kernelFunc:l5};var u5=$e(Wi,r=>Number.isNaN(r)?1:0,"bool"),$E={kernelName:Wi,backendName:"cpu",kernelFunc:u5};var c5=Ye((r,e)=>r<=e?1:0),p5=et(Ui,c5,null,"bool"),RE={kernelName:Ui,backendName:"cpu",kernelFunc:p5};function m5(r){let{backend:e,attrs:t}=r,{start:n,stop:o,num:s}=t,a=zg(n,o,s);return e.makeTensorInfo([a.length],"float32",a)}var FE={kernelName:Pc,backendName:"cpu",kernelFunc:m5};var f5=$e(Hi,r=>Math.log1p(r)),OE={kernelName:Hi,backendName:"cpu",kernelFunc:f5};var d5=Ye((r,e)=>r&&e),h5=et(qi,d5,null,"bool"),PE={kernelName:qi,backendName:"cpu",kernelFunc:h5};var g5=$e(Hl,r=>r?0:1,"bool"),ME={kernelName:Hl,backendName:"cpu",kernelFunc:g5};var x5=Ye((r,e)=>r||e),y5=et(ql,x5,null,"bool"),LE={kernelName:ql,backendName:"cpu",kernelFunc:y5};function b5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{depthRadius:s,bias:a,alpha:i,beta:l}=n;te(o,"LRN");let u=o.shape[3],c=u-1,p=t.data.get(o.dataId).values,m=y.sizeFromShape(o.shape),f=new Float32Array(m);function d(h){let g=h%u,x=h-g+Math.max(0,g-s),b=h-g+Math.min(g+s,c),w=0;for(;x<=b;x++){let _=p[x];w+=_*_}return w}for(let h=0;h<m;h++){let g=d(h),x=p[h]*Math.pow(a+i*g,-l);f[h]=x}return t.makeTensorInfo(o.shape,o.dtype,f)}var zE={kernelName:nl,backendName:"cpu",kernelFunc:b5};function w5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,y:s,dy:a}=e,{depthRadius:i,bias:l,alpha:u,beta:c}=n;te(a,"LRNGrad");let p=y.sizeFromShape(a.shape),m=a.shape[3],f=t.data.get(a.dataId).values,d=t.data.get(o.dataId).values,h=t.data.get(s.dataId).values,g=new Float32Array(p),x=p;for(let b=0;b<x;b++){let w=b%m,_=b-w+Math.max(0,w-i),I=b-w+Math.min(m,w+i+1),E=0;for(let D=_;D<I;D++)E+=Math.pow(d[D],2);E=u*E+l;for(let D=_;D<I;D++){let $=-2*u*c*d[D]*h[b]/E;b===D&&($+=Math.pow(E,-c)),$*=f[b],g[D]+=$}}return t.makeTensorInfo(a.shape,o.dtype,g)}var BE={kernelName:Mc,backendName:"cpu",kernelFunc:w5};function ak(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{reductionIndices:s,keepDims:a}=n,i=t,l=o.shape,u=l.length,c=y.parseAxisParam(s,l),p=c,m=C.getAxesPermutation(p,u),f=i.data.get(o.dataId).values;if(m!=null){let _=new Array(u);for(let I=0;I<_.length;I++)_[I]=l[m[I]];f=kp(f,l,o.dtype,m,_),p=C.getInnerMostAxes(p.length,u),l=_}te(o,"max"),C.assertAxesAreInnerMostDims("max",p,u);let[d,h]=C.computeOutAndReduceShapes(l,p),g=y.sizeFromShape(h),x=Bg(f,g,d,o.dtype),b=i.write(x,d,o.dtype),w=d;return a&&(w=C.expandShapeToKeepDim(d,c)),{dataId:b,shape:w,dtype:o.dtype}}var VE={kernelName:zo,backendName:"cpu",kernelFunc:ak};function _5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e;te(o,"maxPool");let{filterSize:s,strides:a,pad:i,dimRoundingMode:l}=n,u=1;y.assert(C.eitherStridesOrDilationsAreOne(a,u),()=>`Error in maxPool: Either strides or dilations must be 1. Got strides ${a} and dilations '${u}'`);let c=C.computePool2DInfo(o.shape,s,a,u,i,l),p;if(c.filterWidth===1&&c.filterHeight===1&&y.arraysEqual(c.inShape,c.outShape))p=Or({inputs:{x:o},backend:t});else{let m=t.data.get(o.dataId).values,f=y.computeStrides(o.shape),d=Cp(m,o.shape,o.dtype,f,c,"max");p=t.makeTensorInfo(c.outShape,o.dtype,d.values)}return p}var GE={kernelName:Vo,backendName:"cpu",kernelFunc:_5};function k5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{filterSize:s,strides:a,pad:i,dimRoundingMode:l,dataFormat:u}=n;te(o,"maxPool3d");let c=C.computePool3DInfo(o.shape,s,a,1,i,l,u),p=t.data.get(o.dataId).values,m=Xg(p,o.shape,o.dtype,y.computeStrides(o.shape),c,"max");return t.makeTensorInfo(m.shape,"float32",m.values)}var WE={kernelName:ol,backendName:"cpu",kernelFunc:k5};function v5(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,input:s}=e,{filterSize:a,strides:i,pad:l,dimRoundingMode:u}=n;te([o,s],"maxPool3DGrad");let c=C.computePool3DInfo(s.shape,a,i,1,l,u),p=t.bufferSync(s),m=MA(p,c),f=c.strideDepth,d=c.strideHeight,h=c.strideWidth,g=c.dilationDepth,x=c.dilationHeight,b=c.dilationWidth,w=c.effectiveFilterDepth,_=c.effectiveFilterHeight,I=c.effectiveFilterWidth,E=w-1-c.padInfo.front,D=I-1-c.padInfo.left,$=_-1-c.padInfo.top,R=Ce(s.shape,"float32"),M=t.bufferSync(o);for(let G=0;G<c.batchSize;++G)for(let j=0;j<c.inChannels;++j)for(let U=0;U<c.inDepth;++U)for(let H=0;H<c.inHeight;++H)for(let q=0;q<c.inWidth;++q){let X=U-E,ne=H-$,Y=q-D,re=0;for(let ee=0;ee<w;ee+=g){let ie=(X+ee)/f;if(!(ie<0||ie>=c.outDepth||Math.floor(ie)!==ie))for(let me=0;me<_;me+=x){let ae=(ne+me)/d;if(!(ae<0||ae>=c.outHeight||Math.floor(ae)!==ae))for(let de=0;de<I;de+=b){let he=(Y+de)/h;if(he<0||he>=c.outWidth||Math.floor(he)!==he)continue;let xe=w*_*I-1-m.get(G,ie,ae,he,j),we=ee*_*I+me*I+de,De=xe===we?1:0;if(De===0)continue;re+=M.get(G,ie,ae,he,j)*De}}}R.set(re,G,U,H,q,j)}return t.makeTensorInfo(R.shape,R.dtype,R.values)}var jE={kernelName:zc,backendName:"cpu",kernelFunc:v5};function C5(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,input:s,output:a}=e,i=s;te([s,a],"maxPoolGrad");let{filterSize:l,strides:u,pad:c,dimRoundingMode:p}=n,m=C.computePool2DInfo(i.shape,l,u,1,c,p),f=t.data.get(i.dataId).values,d=Ce(m.outShape,i.dtype,Kg(f,i.shape,i.dtype,m).values),h=m.strideHeight,g=m.strideWidth,x=m.dilationHeight,b=m.dilationWidth,w=m.effectiveFilterHeight,_=m.effectiveFilterWidth,I=_-1-m.padInfo.left,E=w-1-m.padInfo.top,D=Ce(i.shape,"float32"),$=t.data.get(o.dataId).values,R=Ce(o.shape,"float32",$);for(let M=0;M<m.batchSize;++M)for(let G=0;G<m.inChannels;++G)for(let j=0;j<m.inHeight;++j)for(let U=0;U<m.inWidth;++U){let H=j-E,q=U-I,X=0;for(let ne=0;ne<w;ne+=x){let Y=(H+ne)/h;if(!(Y<0||Y>=m.outHeight||Math.floor(Y)!==Y))for(let re=0;re<_;re+=b){let ee=(q+re)/g;if(ee<0||ee>=m.outWidth||Math.floor(ee)!==ee)continue;let ie=w*_-1-d.get(M,Y,ee,G),me=ne*_+re,ae=ie===me?1:0;if(ae===0)continue;X+=R.get(M,Y,ee,G)*ae}}D.set(X,M,j,U,G)}return t.makeTensorInfo(D.shape,D.dtype,D.values)}var UE={kernelName:Lc,backendName:"cpu",kernelFunc:C5};function HE(r,e,t,n,o){let s=y.computeStrides(e),a=Cp(r,e,t,s,o,"max"),i=Kg(r,e,t,o,!0,n);return[a.values,i.values]}var qE={kernelName:Bc,backendName:"cpu",kernelFunc:({inputs:r,attrs:e,backend:t})=>{let{x:n}=r,{filterSize:o,strides:s,pad:a,includeBatchInIndex:i}=e,l=t;te(n,"MaxPoolWithArgmax");let u=l.data.get(n.dataId).values,c=C.computePool2DInfo(n.shape,o,s,[1,1],a),[p,m]=HE(u,n.shape,n.dtype,i,c),f=l.write(p,c.outShape,n.dtype),d=l.write(m,c.outShape,n.dtype);return[{dataId:f,shape:c.outShape,dtype:n.dtype},{dataId:d,shape:c.outShape,dtype:"int32"}]}};function I5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n,i=y.parseAxisParam(s,o.shape),u=C.computeOutAndReduceShapes(o.shape,i)[1],c=y.sizeFromShape(u),p=[],m=t.makeTensorInfo([],"float32",new Float32Array([c]));p.push(m);let f=io({inputs:{x:o},backend:t,attrs:{dtype:"float32"}});p.push(f);let d=Nf({inputs:{a:f,b:m},backend:t});p.push(d);let h=Ea({inputs:{x:d},backend:t,attrs:{axis:s,keepDims:a}});return p.forEach(g=>t.disposeIntermediateTensorInfo(g)),h}var KE={kernelName:Go,backendName:"cpu",kernelFunc:I5};function N5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n;te(o,"min");let i=y.parseAxisParam(s,o.shape),l=i,u=C.getAxesPermutation(l,o.shape.length),c=o;u!=null&&(c=Xt({inputs:{x:o},backend:t,attrs:{perm:u}}),l=C.getInnerMostAxes(l.length,o.shape.length)),C.assertAxesAreInnerMostDims("min",l,c.shape.length);let[p,m]=C.computeOutAndReduceShapes(c.shape,l),f=y.sizeFromShape(m),d=y.makeZerosTypedArray(y.sizeFromShape(p),c.dtype),h=t.data.get(c.dataId).values;for(let x=0;x<d.length;++x){let b=x*f,w=h[b];for(let _=0;_<f;++_){let I=h[b+_];I<w&&(w=I)}d[x]=w}u!=null&&t.disposeIntermediateTensorInfo(c);let g=t.makeTensorInfo(p,c.dtype,d);if(a){let x=C.expandShapeToKeepDim(p,i),b=Ze({inputs:{x:g},backend:t,attrs:{shape:x}});return t.disposeIntermediateTensorInfo(g),b}return g}var XE={kernelName:Wo,backendName:"cpu",kernelFunc:N5};function S5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{paddings:s,mode:a}=n;te(o,"mirrorPad");let i=s.map((w,_)=>w[0]+o.shape[_]+w[1]),l=s.map(w=>w[0]),u=s.map((w,_)=>w[0]+o.shape[_]),c=a==="reflect"?0:1,p=t.data.get(o.dataId).values,m=o.shape.length,f=y.computeStrides(o.shape),d=y.sizeFromShape(i),h=i.length,g=y.computeStrides(i),x=y.getTypedArrayFromDType(o.dtype,d);for(let w=0;w<d;w++){let _=y.indexToLoc(w,h,g);for(let E=0;E<h;E++)_[E]<l[E]?_[E]=l[E]*2-_[E]-c:_[E]>=u[E]&&(_[E]=(u[E]-1)*2-_[E]+c);_=_.map((E,D)=>E-l[D]);let I=y.locToIndex(_,m,f);x[w]=p[I]}return{dataId:t.write(x,i,o.dtype),shape:i,dtype:o.dtype}}var YE={kernelName:Uo,backendName:"cpu",kernelFunc:S5};var T5=Ye((r,e)=>{let t=r%e;return r<0&&e<0||r>=0&&e>=0?t:(t+e)%e}),A5=et(Ki,T5),ZE={kernelName:Ki,backendName:"cpu",kernelFunc:A5};var QE=hc(__());function lk(r){let{inputs:e,backend:t,attrs:n}=r,{logits:o}=e,{dim:s}=n,a=o.shape.length,i=s;if(i===-1&&(i=a-1),i!==a-1)throw Error(`Softmax along a non-last dimension is not yet supported. Logits was rank ${a} and dim was ${i}`);let l=y.parseAxisParam([i],o.shape),u=ak({inputs:{x:o},backend:t,attrs:{reductionIndices:l,keepDims:!1}}),c=C.expandShapeToKeepDim(u.shape,l),p=Ze({inputs:{x:u},backend:t,attrs:{shape:c}}),m=If({inputs:{a:o,b:p},backend:t}),f=O_({inputs:{x:m},backend:t}),d=Ea({inputs:{x:f},backend:t,attrs:{axis:l,keepDims:!1}}),h=Ze({inputs:{x:d},backend:t,attrs:{shape:c}}),g=Nf({inputs:{a:f,b:h},backend:t});return t.disposeIntermediateTensorInfo(u),t.disposeIntermediateTensorInfo(p),t.disposeIntermediateTensorInfo(m),t.disposeIntermediateTensorInfo(f),t.disposeIntermediateTensorInfo(d),t.disposeIntermediateTensorInfo(h),g}var JE={kernelName:as,backendName:"cpu",kernelFunc:lk};function E5(r){let{inputs:e,backend:t,attrs:n}=r,{logits:o}=e,{numSamples:s,seed:a,normalized:i}=n;te(o,"multinomial");let l=i?o:lk({inputs:{logits:o},backend:t,attrs:{dim:-1}}),u=l.shape[0],c=l.shape[1],p=t.data.get(l.dataId).values,m=[u,s],f=y.makeZerosTypedArray(y.sizeFromShape(m),"int32");for(let d=0;d<u;++d){let h=d*c,g=new Float32Array(c-1);g[0]=p[h];for(let w=1;w<g.length;++w)g[w]=g[w-1]+p[h+w];let x=QE.alea(a.toString()),b=d*s;for(let w=0;w<s;++w){let _=x();f[b+w]=g.length;for(let I=0;I<g.length;I++)if(_<g[I]){f[b+w]=I;break}}}return i||t.disposeIntermediateTensorInfo(l),t.makeTensorInfo(m,"int32",f)}var e2={kernelName:Vc,backendName:"cpu",kernelFunc:E5};var D5=Fr.nonMaxSuppressionV3Impl;function $5(r){let{inputs:e,backend:t,attrs:n}=r,{boxes:o,scores:s}=e,{maxOutputSize:a,iouThreshold:i,scoreThreshold:l}=n;te(o,"NonMaxSuppression");let u=t.data.get(o.dataId).values,c=t.data.get(s.dataId).values,{selectedIndices:p}=D5(u,c,a,i,l);return t.makeTensorInfo([p.length],"int32",new Int32Array(p))}var t2={kernelName:Yi,backendName:"cpu",kernelFunc:$5};var R5=Fr.nonMaxSuppressionV4Impl;function F5(r){let{inputs:e,backend:t,attrs:n}=r,{boxes:o,scores:s}=e,{maxOutputSize:a,iouThreshold:i,scoreThreshold:l,padToMaxOutputSize:u}=n;te(o,"NonMaxSuppressionPadded");let c=t.data.get(o.dataId).values,p=t.data.get(s.dataId).values,{selectedIndices:m,validOutputs:f}=R5(c,p,a,i,l,u);return[t.makeTensorInfo([m.length],"int32",new Int32Array(m)),t.makeTensorInfo([],"int32",new Int32Array([f]))]}var r2={kernelName:Zi,backendName:"cpu",kernelFunc:F5};var O5=Fr.nonMaxSuppressionV5Impl;function P5(r){let{inputs:e,backend:t,attrs:n}=r,{boxes:o,scores:s}=e,{maxOutputSize:a,iouThreshold:i,scoreThreshold:l,softNmsSigma:u}=n;te(o,"NonMaxSuppressionWithScore");let c=t.data.get(o.dataId).values,p=t.data.get(s.dataId).values,m=a,f=i,d=l,h=u,{selectedIndices:g,selectedScores:x}=O5(c,p,m,f,d,h);return[t.makeTensorInfo([g.length],"int32",new Int32Array(g)),t.makeTensorInfo([x.length],"float32",new Float32Array(x))]}var n2={kernelName:Ji,backendName:"cpu",kernelFunc:P5};function M5(r){let{inputs:e,backend:t,attrs:n}=r,{indices:o}=e,{depth:s,onValue:a,offValue:i}=n;te(o,"oneHot");let l=y.sizeFromShape(o.shape),u=new Float32Array(l*s);u.fill(i);let c=t.data.get(o.dataId).values;for(let p=0;p<l;++p)c[p]>=0&&c[p]<s&&(u[p*s+c[p]]=a);return t.makeTensorInfo([...o.shape,s],"int32",u)}var o2={kernelName:qo,backendName:"cpu",kernelFunc:M5};function Af(r){let{inputs:e,backend:t}=r,{x:n}=e;if(n.dtype==="string")throw new Error("zerosLike is not supported for string tensors");if(n.dtype==="complex64"){let o=so({inputs:{input:n},backend:t}),s=Af({inputs:{x:o},backend:t}),a=li({inputs:{input:n},backend:t}),i=Af({inputs:{x:a},backend:t}),l=dr({inputs:{real:s,imag:i},backend:t});return t.disposeIntermediateTensorInfo(o),t.disposeIntermediateTensorInfo(s),t.disposeIntermediateTensorInfo(a),t.disposeIntermediateTensorInfo(i),l}else return Tf({backend:t,attrs:{shape:n.shape,value:0,dtype:n.dtype}})}var s2={kernelName:Js,backendName:"cpu",kernelFunc:Af};function i2(r){let{inputs:e,backend:t}=r,{x:n}=e;if(n.dtype==="string")throw new Error("onesLike is not supported for string tensors");if(n.dtype==="complex64"){let o=so({inputs:{input:n},backend:t}),s=i2({inputs:{x:o},backend:t}),a=li({inputs:{input:n},backend:t}),i=Af({inputs:{x:a},backend:t}),l=dr({inputs:{real:s,imag:i},backend:t});return t.disposeIntermediateTensorInfo(o),t.disposeIntermediateTensorInfo(s),t.disposeIntermediateTensorInfo(a),t.disposeIntermediateTensorInfo(i),l}else return Tf({backend:t,attrs:{shape:n.shape,value:1,dtype:n.dtype}})}var a2={kernelName:Us,backendName:"cpu",kernelFunc:i2};function uk(r){let{inputs:e,backend:t,attrs:n}=r,{axis:o}=n;if(e.length===1)return Ip({inputs:{input:e[0]},backend:t,attrs:{dim:o}});let s=e[0].shape,a=e[0].dtype;e.forEach(c=>{y.assertShapesMatch(s,c.shape,"All tensors passed to stack must have matching shapes"),y.assert(a===c.dtype,()=>"All tensors passed to stack must have matching dtypes")});let i=[],l=e.map(c=>{let p=Ip({inputs:{input:c},backend:t,attrs:{dim:o}});return i.push(p),p}),u=yl({inputs:l,backend:t,attrs:{axis:o}});return i.forEach(c=>t.disposeIntermediateTensorInfo(c)),u}var l2={kernelName:Hs,backendName:"cpu",kernelFunc:uk};function L5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{paddings:s,constantValue:a}=n;te(o,"pad");let i=s.map((b,w)=>b[0]+o.shape[w]+b[1]),l=s.map(b=>b[0]),u=t.data.get(o.dataId).values,c=y.sizeFromShape(o.shape),p=o.shape.length,m=y.computeStrides(o.shape),f=y.sizeFromShape(i),d=i.length,h=y.computeStrides(i),g=y.getTypedArrayFromDType(o.dtype,f);a!==0&&g.fill(a);for(let b=0;b<c;b++){let _=y.indexToLoc(b,p,m).map((E,D)=>E+l[D]),I=y.locToIndex(_,d,h);g[I]=u[b]}return{dataId:t.write(g,i,o.dtype),shape:i,dtype:o.dtype}}var Zg={kernelName:Ko,backendName:"cpu",kernelFunc:L5};var z5=Ye((r,e)=>Math.pow(r,e)),B5=et(Xo,z5),u2={kernelName:Xo,backendName:"cpu",kernelFunc:B5};function V5(r){let{backend:e,attrs:t}=r,{start:n,stop:o,dtype:s,step:a}=t,i=$u(n,o,a,s);return e.makeTensorInfo([i.length],s,i)}var c2={kernelName:sl,backendName:"cpu",kernelFunc:V5};var G5=$e(ea,r=>1/r),p2={kernelName:ea,backendName:"cpu",kernelFunc:G5};function W5(r){let{inputs:e,backend:t,attrs:n}=r,{images:o}=e,{alignCorners:s,halfPixelCenters:a,size:i}=n;te(o,"resizeBilinear");let l=y.computeStrides(o.shape),[u,c]=i,[p,m,f,d]=o.shape,h=t.data.get(o.dataId).values,g=new Float32Array(y.sizeFromShape([p,u,c,d])),x=[s&&u>1?m-1:m,s&&c>1?f-1:f],b=[s&&u>1?u-1:u,s&&c>1?c-1:c],w=0,_=x[0]/b[0],I=x[1]/b[1];for(let E=0;E<p;E++)for(let D=0;D<u;D++){let $;a?$=_*(D+.5)-.5:$=_*D;let R=Math.max(0,Math.floor($)),M=$-R,G=Math.min(m-1,Math.ceil($)),j=E*l[0]+R*l[1],U=E*l[0]+G*l[1];for(let H=0;H<c;H++){let q;a?q=I*(H+.5)-.5:q=I*H;let X=Math.max(0,Math.floor(q)),ne=q-X,Y=Math.min(f-1,Math.ceil(q)),re=j+X*l[2],ee=U+X*l[2],ie=j+Y*l[2],me=U+Y*l[2];for(let ae=0;ae<d;ae++){let de=h[re+ae],he=h[ee+ae],xe=h[ie+ae],we=h[me+ae],De=de+(xe-de)*ne,ve=he+(we-he)*ne,Ge=De+(ve-De)*M;g[w++]=Ge}}}return t.makeTensorInfo([p,u,c,d],"float32",g)}var m2={kernelName:Jo,backendName:"cpu",kernelFunc:W5};function j5(r){let{inputs:e,backend:t,attrs:n}=r,{images:o,dy:s}=e,{alignCorners:a}=n;te([s,o],"resizeBilinearGrad");let i=y.computeStrides(o.shape),[l,u,c,p]=o.shape,[,m,f]=s.shape,d=new Float32Array(l*u*c*p),h=[a&&m>1?u-1:u,a&&f>1?c-1:c],g=[a&&m>1?m-1:m,a&&f>1?f-1:f],x=h[0]/g[0],b=h[1]/g[1],w=t.data.get(s.dataId).values,_=0;for(let I=0;I<l;I++){let E=I*i[0];for(let D=0;D<m;D++){let $=D*x,R=Math.floor($),M=Math.min(Math.ceil($),u-1),G=E+R*i[1],j=E+M*i[1],U=$-R,H=1-U;for(let q=0;q<f;q++){let X=q*b,ne=Math.floor(X),Y=Math.min(Math.ceil(X),c-1),re=X-ne,ee=1-re,ie=G+ne*i[2],me=G+Y*i[2],ae=j+ne*i[2],de=j+Y*i[2],he=H*ee,xe=H*re,we=U*ee,De=U*re;for(let ve=0;ve<p;ve++){let Ge=w[_++];d[ie+ve]+=Ge*he,d[me+ve]+=Ge*xe,d[ae+ve]+=Ge*we,d[de+ve]+=Ge*De}}}}return t.makeTensorInfo([l,c,u,p],"float32",d)}var f2={kernelName:jc,backendName:"cpu",kernelFunc:j5};function U5(r){let{inputs:e,backend:t,attrs:n}=r,{images:o}=e,{alignCorners:s,halfPixelCenters:a,size:i}=n;te(o,"resizeNearestNeighbor");let l=y.computeStrides(o.shape),[u,c]=i,[p,m,f,d]=o.shape,h=t.data.get(o.dataId).values,g=new Float32Array(p*u*c*d),x=[s&&u>1?m-1:m,s&&c>1?f-1:f],b=[s&&u>1?u-1:u,s&&c>1?c-1:c],w=x[0]/b[0],_=x[1]/b[1],I=0;for(let E=0;E<p;E++){let D=E*l[0];for(let $=0;$<u;$++){let R=a?w*($+.5):w*$,M=Math.min(m-1,s?Math.round(R):Math.floor(R));a&&(M=Math.max(0,M));let G=D+M*l[1];for(let j=0;j<c;j++){let U=a?_*(j+.5):_*j,H=Math.min(f-1,s?Math.round(U):Math.floor(U));a&&(H=Math.max(0,H));let q=G+H*l[2];for(let X=0;X<d;X++){let ne=h[q+X];g[I++]=ne}}}}return t.makeTensorInfo([p,u,c,d],o.dtype,g)}var d2={kernelName:il,backendName:"cpu",kernelFunc:U5};function H5(r){let{inputs:e,backend:t,attrs:n}=r,{images:o,dy:s}=e,{alignCorners:a}=n;te([s,o],"resizeNearestNeighborGrad");let i=y.computeStrides(o.shape),l=y.computeStrides(s.shape),[u,c,p,m]=o.shape,[,f,d]=s.shape,h=new Float32Array(u*c*p*m),g=t.data.get(s.dataId).values,x=[a&&f>1?c-1:c,a&&d>1?p-1:p],b=[a&&f>1?f-1:f,a&&d>1?d-1:d],w=x[0]/b[0],_=x[1]/b[1],I=1/w,E=1/_,D=Math.ceil(I)*2+2,$=Math.ceil(E)*2+2;for(let R=0;R<u;R++){let M=R*i[0];for(let G=0;G<c;G++){let j=M+G*i[1],U=Math.floor(G*I),H=Math.floor(U-D/2);for(let q=0;q<p;q++){let X=j+q*i[2],ne=Math.floor(q*E),Y=Math.floor(ne-$/2);for(let re=0;re<m;re++){let ee=0;for(let ie=0;ie<D;ie++){let me=ie+H;if(me<0||me>=f)continue;let ae=M+me*l[1],de=me*w,he=Math.min(c-1,a?Math.round(de):Math.floor(de));if(G===he)for(let xe=0;xe<$;xe++){let we=xe+Y;if(we<0||we>=d)continue;let De=ae+we*l[2],ve=we*_,Ge=Math.min(p-1,a?Math.round(ve):Math.floor(ve));q===Ge&&(ee+=g[De+re])}}h[X+re]=ee}}}}return t.makeTensorInfo(o.shape,o.dtype,h)}var h2={kernelName:Wc,backendName:"cpu",kernelFunc:H5};function q5(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{dims:s}=n;te(o,"reverse");let a=o.shape.length,i=y.parseAxisParam(s,o.shape);if(a===0)return Or({inputs:{x:o},backend:t});let l=new ut(o.shape,o.dtype),u=t.bufferSync(o);for(let c=0;c<l.size;c++){let p=l.indexToLoc(c),m=p.slice();i.forEach(f=>m[f]=o.shape[f]-1-m[f]),l.set(u.get(...m),...p)}return t.makeTensorInfo(l.shape,l.dtype,l.values)}var g2={kernelName:es,backendName:"cpu",kernelFunc:q5};var x2={kernelName:ua,backendName:"cpu",kernelFunc:({inputs:r,attrs:e,backend:t})=>{let{image:n}=r,{radians:o,fillValue:s,center:a}=e,i=t,l=y.getTypedArrayFromDType(n.dtype,y.sizeFromShape(n.shape)),[u,c,p,m]=n.shape,[f,d]=C.getImageCenter(a,c,p),h=255,g=Math.sin(o),x=Math.cos(o),b=i.data.get(n.dataId).values;for(let _=0;_<u;_++){let I=_*p*c*m;for(let E=0;E<c;E++){let D=E*(p*m);for(let $=0;$<p;$++){let R=$*m;for(let M=0;M<m;M++){let G=[u,E,$,M],j=G[2],U=G[1],H=(j-f)*x-(U-d)*g,q=(j-f)*g+(U-d)*x;H=Math.round(H+f),q=Math.round(q+d);let X=s;if(typeof s!="number"&&(M===3?X=h:X=s[M]),H>=0&&H<p&&q>=0&&q<c){let Y=q*(p*m),re=H*m,ee=I+Y+re+M;X=b[ee]}let ne=I+D+R+M;l[ne]=X}}}}return{dataId:i.write(l,n.shape,n.dtype),shape:n.shape,dtype:n.dtype}}};var K5=$e(ts,r=>{let e=Math.floor(r);return r-e<.5?Math.floor(r):r-e>.5?Math.ceil(r):e%2==0?e:e+1}),y2={kernelName:ts,backendName:"cpu",kernelFunc:K5};function Jg(r,e,t,n,o,s,a,i,l,u){let c=[n/o,o],p=r.values,m=e.values;if(n===0)return Ce(t,e.dtype);let f=Ce(c,e.dtype);f.values.fill(l);for(let d=0;d<s;d++){let h=[],g=0;for(let x=0;x<a;x++){let b=p[d*a+x];h.push(b),g+=b*i[x]}if(g<0||g>=n/o)throw new Error(`Invalid indices: ${h} does not index into ${t}`);for(let x=0;x<o;x++)u?f.values[g*o+x]+=m[d*o+x]:f.values[g*o+x]=e.rank===0?m[0]:m[d*o+x]}return f}function X5(r){let{inputs:e,backend:t,attrs:n}=r,{indices:o,updates:s}=e,{shape:a}=n,{sliceRank:i,numUpdates:l,sliceSize:u,strides:c,outputSize:p}=C.calculateShapes(s,o,a),m=!0,f=t.bufferSync(o),d=t.bufferSync(s),h=Jg(f,d,a,p,u,l,i,c,0,m);return t.makeTensorInfo(a,h.dtype,h.values)}var b2={kernelName:ta,backendName:"cpu",kernelFunc:X5};function Y5(r){let{inputs:e,backend:t}=r,{condition:n,t:o,e:s}=e;te([n,o,s],"select");let a=n.shape.length,i=t.data.get(n.dataId).values,l=t.data.get(o.dataId).values,u=t.data.get(s.dataId).values,c=lr(o.dtype,s.dtype),p=y.makeZerosTypedArray(y.sizeFromShape(o.shape),c),m=0,f=a===0||a>1||o.shape.length===1?1:y.sizeFromShape(o.shape.slice(1));for(let d=0;d<i.length;d++)for(let h=0;h<f;h++)i[d]===1?p[m++]=l[d]:p[m++]=u[d];return t.makeTensorInfo(o.shape,c,p)}var w2={kernelName:Ks,backendName:"cpu",kernelFunc:Y5};var Z5=C.SELU_SCALEALPHA,J5=C.SELU_SCALE,Q5=$e(ra,r=>r>=0?J5*r:Z5*(Math.exp(r)-1)),_2={kernelName:ra,backendName:"cpu",kernelFunc:Q5};var e8=$e(oa,r=>r<0?-1:r>0?1:0),k2={kernelName:oa,backendName:"cpu",kernelFunc:e8};var t8=$e(ns,r=>Math.sin(r)),v2={kernelName:ns,backendName:"cpu",kernelFunc:t8};var r8=$e(na,r=>Math.sinh(r)),C2={kernelName:na,backendName:"cpu",kernelFunc:r8};var n8=11920928955078125e-23,I2=Math.log(n8)+2,o8=$e(sa,r=>{let e=r>-I2,t=r<I2,n=Math.exp(r),o;return t?o=n:e?o=r:o=Math.log(1+n),o}),N2={kernelName:sa,backendName:"cpu",kernelFunc:o8};function s8(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{blockShape:s,paddings:a}=n;te([o],"spaceToBatchND");let i=y.sizeFromShape(s),l=[[0,0]];l.push(...a);for(let E=1+s.length;E<o.shape.length;++E)l.push([0,0]);let u=Zg.kernelFunc({inputs:{x:o},backend:t,attrs:{paddings:l,constantValue:0}}),c=C.getReshaped(u.shape,s,i,!1),p=C.getPermuted(c.length,s.length,!1),m=C.getReshapedPermuted(u.shape,s,i,!1),h=Ze({inputs:{x:u},backend:t,attrs:{shape:c}}),b=Xt({inputs:{x:h},backend:t,attrs:{perm:p}}),I=Ze({inputs:{x:b},backend:t,attrs:{shape:m}});return t.disposeIntermediateTensorInfo(u),t.disposeIntermediateTensorInfo(h),t.disposeIntermediateTensorInfo(b),I}var S2={kernelName:al,backendName:"cpu",kernelFunc:s8};function i8(r){let{inputs:e,backend:t}=r,{indices:n,values:o,denseShape:s,defaultValue:a}=e;if(s.shape.length!==1)throw new Error(`Dense shape must be a vector, saw:
|
|
${s.shape}`);if(n.shape.length!==2)throw new Error(`Indices must be a matrix, saw:
|
|
${n.shape}`);if(o.shape.length!==1)throw new Error(`Values must be a vector, saw:
|
|
${o.shape}`);if(a.shape.length!==0)throw new Error(`Default value must be a scalar, saw:
|
|
${a.shape}`);let i=t.data.get(n.dataId).values,l=t.data.get(o.dataId).values,u=t.data.get(s.dataId).values,c=t.data.get(a.dataId).values[0],[p,m,f,d,h]=Vg(i,n.shape,n.dtype,l,o.dtype,u,c);return[t.makeTensorInfo(m,n.dtype,p),t.makeTensorInfo([m[0]],o.dtype,f),t.makeTensorInfo([d.length],"bool",new Uint8Array(d.map(g=>Number(g)))),t.makeTensorInfo([h.length],n.dtype,new Int32Array(h))]}var T2={kernelName:Uc,backendName:"cpu",kernelFunc:i8};function a8(r){let{inputs:e,backend:t}=r,{inputIndices:n,inputShape:o,newShape:s}=e;if(n.shape.length!==2)throw new Error(`Input indices should be a matrix but received shape
|
|
${n.shape}`);if(o.shape.length!==1)throw new Error(`Input shape should be a vector but received shape
|
|
${o.shape}`);if(s.shape.length!==1)throw new Error(`Target shape should be a vector but received shape ${s.shape}`);let a=Array.from(t.data.get(o.dataId).values),i=t.data.get(n.dataId).values,l=Array.from(t.data.get(s.dataId).values),[u,c,p]=Gg(i,n.shape,n.dtype,a,l);return[t.makeTensorInfo(c,n.dtype,u),t.makeTensorInfo([p.length],s.dtype,new Int32Array(p))]}var A2={kernelName:Hc,backendName:"cpu",kernelFunc:a8};function l8(r){let{inputs:e,backend:t,attrs:n}=r,{sparseIndices:o,sparseValues:s,defaultValue:a}=e,{outputShape:i}=n,{sliceRank:l,numUpdates:u,sliceSize:c,strides:p,outputSize:m}=C.calculateShapes(s,o,i),f=!1,d=t.bufferSync(o),h=t.bufferSync(s),g=t.data.get(a.dataId).values[0],x=Jg(d,h,i,m,c,u,l,p,g,f);return t.makeTensorInfo(i,x.dtype,x.values)}var E2={kernelName:qc,backendName:"cpu",kernelFunc:l8};function u8(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{numOrSizeSplits:s,axis:a}=n,i=y.parseAxisParam(a,o.shape)[0],l=C.prepareSplitSize(o,s,i),u=new Array(o.shape.length).fill(0),c=o.shape.slice();return l.map(p=>{let m=[...c];m[i]=p;let f=uo({inputs:{x:o},backend:t,attrs:{begin:u,size:m}});return u[i]+=p,f})}var D2={kernelName:Ys,backendName:"cpu",kernelFunc:u8};var c8=$e(ss,r=>Math.sqrt(r)),$2={kernelName:ss,backendName:"cpu",kernelFunc:c8};var R2={kernelName:ll,backendName:"cpu",kernelFunc:({inputs:r,backend:e})=>{let{x:t}=r,n=e;te(t,"square");let o=n.data.get(t.dataId).values,s=new Float32Array(o.length);for(let i=0;i<o.length;++i){let l=o[i];s[i]=l*l}return{dataId:n.write(s,t.shape,t.dtype),shape:t.shape,dtype:t.dtype}}};var p8=$e(Yn,(r,e)=>{let t=e;return isNaN(r)?NaN:r>0?1:t.alpha}),F2={kernelName:Yn,backendName:"cpu",kernelFunc:p8};function m8(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{begin:s,end:a,strides:i,beginMask:l,endMask:u,ellipsisMask:c,newAxisMask:p,shrinkAxisMask:m}=n;te(o,"stridedSlice");let{nonStrided:f,$begin:d,$strides:h,size:g,newShape:x,outShape:b}=or.sliceInfo(o.shape,s,a,i,l,u,c,p,m),w=Ze({inputs:{x:o},backend:t,attrs:{shape:x}}),_;if(f){let E=uo({inputs:{x:w},backend:t,attrs:{begin:d,size:g}});_=Ze({inputs:{x:E},backend:t,attrs:{shape:b}}),t.disposeIntermediateTensorInfo(E)}else if(b.some(E=>E===0))_=t.makeTensorInfo(b,o.dtype,[]);else{let E=t.bufferSync(w),D=Wg(b,E,h,d);_=t.makeTensorInfo(D.shape,D.dtype,D.values)}let I=Ze({inputs:{x:_},backend:t,attrs:{shape:b}});return t.disposeIntermediateTensorInfo(w),t.disposeIntermediateTensorInfo(_),I}var O2={kernelName:ia,backendName:"cpu",kernelFunc:m8};var f8=$e(cs,r=>Math.tan(r)),P2={kernelName:cs,backendName:"cpu",kernelFunc:f8};var d8=$e(ps,r=>Math.tanh(r)),M2={kernelName:ps,backendName:"cpu",kernelFunc:d8};function h8(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{reps:s}=n;te(o,"tile");let a=jg(t.bufferSync(o),s);return t.makeTensorInfo(a.shape,a.dtype,a.values)}var L2={kernelName:Mn,backendName:"cpu",kernelFunc:h8};function g8(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{k:s,sorted:a}=n;te(o,"topk");let i=t.data.get(o.dataId).values,[l,u]=Ug(i,o.shape,o.dtype,s,a);return[t.makeTensorInfo(l.shape,l.dtype,l.values),t.makeTensorInfo(u.shape,u.dtype,u.values)]}var z2={kernelName:aa,backendName:"cpu",kernelFunc:g8};function x8(r){let{inputs:e,attrs:t,backend:n}=r,{image:o,transforms:s}=e,{interpolation:a,fillMode:i,fillValue:l,outputShape:u}=t,[c,p,m,f]=o.shape,[d,h]=u!=null?u:[p,m],g=[c,d,h,f],x=y.computeStrides(o.shape),b=x[0],w=x[1],_=x[2],I=y.getTypedArrayFromDType(o.dtype,y.sizeFromShape(g));I.fill(l);let E=n.data.get(o.dataId).values,D=n.data.get(s.dataId).values;for(let R=0;R<c;++R){let M=s.shape[0]===1?D:D.subarray(R*8,R*8+8);for(let G=0;G<d;++G)for(let j=0;j<h;++j)for(let U=0;U<f;++U){let H,q=M[6]*j+M[7]*G+1;if(q===0)continue;let X=(M[0]*j+M[1]*G+M[2])/q,ne=(M[3]*j+M[4]*G+M[5])/q,Y=V2(X,m,i),re=V2(ne,p,i);switch(a){case"nearest":H=k8(E,p,m,b,w,_,R,re,Y,U,l);break;case"bilinear":H=v8(E,p,m,b,w,_,R,re,Y,U,l);break;default:throw new Error(`Error in Transform: Expect 'nearest' or 'bilinear', but got ${a}`)}let ee=R*b+G*w+j*_+U;I[ee]=H}return n.makeTensorInfo(g,o.dtype,I)}return{dataId:n.write(I,g,o.dtype),shape:o.shape,dtype:o.dtype}}var B2={kernelName:la,backendName:"cpu",kernelFunc:x8};function V2(r,e,t){switch(t){case"reflect":return y8(r,e);case"wrap":return b8(r,e);case"nearest":return _8(r,e);case"constant":default:return w8(r,e)}}function y8(r,e){let t=r;if(t<0)if(e<=1)t=0;else{let n=2*e;t<n&&(t=n*Math.trunc(-t/n)+t),t=t<-e?t+n:-t-1}else if(t>e-1)if(e<=1)t=0;else{let n=2*e;t-=n*Math.trunc(t/n),t>=e&&(t=n-t-1)}return y.clamp(0,t,e-1)}function b8(r,e){let t=r;if(t<0)if(e<=1)t=0;else{let n=e-1;t+=e*(Math.trunc(-t/n)+1)}else if(t>e-1)if(e<=1)t=0;else{let n=e-1;t-=e*Math.trunc(t/n)}return y.clamp(0,t,e-1)}function w8(r,e){return r}function _8(r,e){return y.clamp(0,r,e-1)}function Ef(r,e,t,n,o,s,a,i,l,u,c){let p=a*n+i*o+l*s+u;return 0<=i&&i<e&&0<=l&&l<t?r[p]:c}function k8(r,e,t,n,o,s,a,i,l,u,c){let p=Math.round(i),m=Math.round(l);return Ef(r,e,t,n,o,s,a,p,m,u,c)}function v8(r,e,t,n,o,s,a,i,l,u,c){let p=Math.floor(i),m=Math.floor(l),f=p+1,d=m+1,h=(d-l)*Ef(r,e,t,n,o,s,a,p,m,u,c)+(l-m)*Ef(r,e,t,n,o,s,a,p,d,u,c),g=(d-l)*Ef(r,e,t,n,o,s,a,f,m,u,c)+(l-m)*Ef(r,e,t,n,o,s,a,f,d,u,c);return(f-i)*h+(i-p)*g}function C8(r){let{inputs:e,attrs:t,backend:n}=r,{axis:o}=t,{x:s}=e;te(s,"unique");let a=n.data.get(s.dataId).values,{outputValues:i,outputShape:l,indices:u}=Hg(a,o,s.shape,s.dtype);return[n.makeTensorInfo(l,s.dtype,i),n.makeTensorInfo([u.length],"int32",u)]}var G2={kernelName:Kc,backendName:"cpu",kernelFunc:C8};function I8(r){let{inputs:e,backend:t,attrs:n}=r,{value:o}=e,{axis:s}=n;s<0&&(s+=o.shape.length);let a=o.shape.length,i=o.shape[s],l=new Array(a-1),u=0;for(let f=0;f<a;f++)f!==s&&(l[u++]=o.shape[f]);let c=new Array(a).fill(0),p=o.shape.slice();p[s]=1;let m=new Array(i);for(let f=0;f<m.length;f++){c[s]=f;let d=uo({inputs:{x:o},backend:t,attrs:{begin:c,size:p}});m[f]=Ze({inputs:{x:d},backend:t,attrs:{shape:l}}),t.disposeIntermediateTensorInfo(d)}return m}var W2={kernelName:Zs,backendName:"cpu",kernelFunc:I8};function N8(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,segmentIds:s}=e,{numSegments:a}=n;te(o,"unsortedSegmentSum");let i=o.shape.length,l=s.shape.length,u=[],c=[],p=i-l,m=s;for(let d=0;d<p;++d){let h=Ip({inputs:{input:m},backend:t,attrs:{dim:d+1}});m=h,c.push(h)}for(let d=0;d<a;++d){let h=y.createScalarValue(d,"int32"),g=t.makeTensorInfo([],"int32",h),x=sk({inputs:{a:g,b:m},backend:t}),b=io({inputs:{x},backend:t,attrs:{dtype:"float32"}}),w=Du({inputs:{a:b,b:o},backend:t}),_=Ea({inputs:{x:w},backend:t,attrs:{axis:0,keepDims:!1}});u.push(_),c.push(g),c.push(x),c.push(b),c.push(w),c.push(_)}let f=uk({inputs:u,backend:t,attrs:{axis:0}});return c.forEach(d=>t.disposeIntermediateTensorInfo(d)),f}var j2={kernelName:ul,backendName:"cpu",kernelFunc:N8};var S8=[CA,H1,IA,NA,Z1,SA,TA,AA,EA,DA,$A,RA,FA,OA,PA,LA,zA,BA,VA,vA,GA,WA,jA,Y1,J1,UA,q1,HA,KA,YA,ZA,XA,QA,eE,JA,tE,rE,nE,oE,sE,iE,aE,lE,uE,cE,pE,fE,mE,Sf,hE,gA,gE,xE,yE,Q1,bE,eA,wE,_E,kE,tA,vE,CE,IE,NE,SE,rA,TE,K1,AE,qA,EE,DE,$E,xA,nA,RE,FE,oA,OE,PE,ME,LE,zE,BE,sA,GE,WE,jE,UE,qE,VE,KE,XE,iA,YE,ZE,e2,aA,lA,t2,r2,n2,uA,o2,a2,l2,Zg,u2,yA,pA,c2,X1,p2,bA,wA,kA,m2,f2,d2,h2,g2,x2,y2,mA,b2,w2,_2,_A,k2,v2,C2,fA,JE,N2,S2,T2,A2,E2,D2,$2,R2,dA,F2,O2,hA,dE,P2,M2,L2,z2,cA,B2,G2,W2,j2,s2];for(let r of S8)Kl(r);var Z2={};We(Z2,{assertNotComplex:()=>Cs,bindCanvasToFramebuffer:()=>M8,bindColorTextureToFramebuffer:()=>Rf,bindTextureToProgramUniformSampler:()=>Ck,bindTextureUnit:()=>K2,bindVertexBufferToProgramAttribute:()=>Qg,callAndCheck:()=>Ne,canBeRepresented:()=>mk,createFragmentShader:()=>dk,createFramebuffer:()=>_k,createProgram:()=>hk,createStaticIndexBuffer:()=>yk,createStaticVertexBuffer:()=>xk,createTexture:()=>bk,createVertexShader:()=>fk,getBatchDim:()=>$a,getExtensionOrThrow:()=>Np,getFramebufferErrorMessage:()=>X2,getMaxTexturesInShader:()=>Sk,getNumChannels:()=>O8,getProgramUniformLocation:()=>vk,getProgramUniformLocationOrThrow:()=>kk,getRowsCols:()=>Ra,getShapeAs3D:()=>Ff,getTextureShapeFromLogicalShape:()=>Ik,getWebGLDisjointQueryTimerVersion:()=>Tk,getWebGLErrorMessage:()=>q2,getWebGLMaxTextureSize:()=>Nk,hasExtension:()=>An,isCapableOfRenderingToFloatTexture:()=>Ak,isDownloadFloatTextureEnabled:()=>Ek,isReshapeFree:()=>_l,isWebGLFenceEnabled:()=>$k,isWebGLVersionEnabled:()=>ox,linkProgram:()=>gk,resetMaxTextureSize:()=>L8,resetMaxTexturesInShader:()=>z8,unbindColorTextureFromFramebuffer:()=>ex,unbindTextureUnit:()=>P8,validateFramebuffer:()=>Sp,validateProgram:()=>$f,validateTextureSize:()=>wk});var Fu={},ck={alpha:!1,antialias:!1,premultipliedAlpha:!1,preserveDrawingBuffer:!1,depth:!1,stencil:!1,failIfMajorPerformanceCaveat:!0};function pk(r,e){Fu[r]=e}function Vn(r){if(!(r in Fu)){let t=A8(r);if(t!==null)Fu[r]=t;else return console.log("Could not get context for WebGL version",r),null}let e=Fu[r];return e.isContextLost()?(delete Fu[r],Vn(r)):(e.disable(e.DEPTH_TEST),e.disable(e.STENCIL_TEST),e.disable(e.BLEND),e.disable(e.DITHER),e.disable(e.POLYGON_OFFSET_FILL),e.disable(e.SAMPLE_COVERAGE),e.enable(e.SCISSOR_TEST),e.enable(e.CULL_FACE),e.cullFace(e.BACK),Fu[r])}function T8(r){if(typeof OffscreenCanvas!="undefined"&&r===2)return new OffscreenCanvas(300,150);if(typeof document!="undefined")return document.createElement("canvas");throw new Error("Cannot create a canvas in this context")}function A8(r){if(r!==1&&r!==2)throw new Error("Cannot get WebGL rendering context, WebGL is disabled.");let e=T8(r);return e.addEventListener("webglcontextlost",t=>{t.preventDefault(),delete Fu[r]},!1),r===1?e.getContext("webgl",ck)||e.getContext("experimental-webgl",ck):e.getContext("webgl2",ck)}var bl;(function(r){r[r.DENSE=0]="DENSE",r[r.SHARED_BATCH=1]="SHARED_BATCH"})(bl||(bl={}));var Pr;(function(r){r[r.RENDER=0]="RENDER",r[r.UPLOAD=1]="UPLOAD",r[r.PIXELS=2]="PIXELS",r[r.DOWNLOAD=3]="DOWNLOAD"})(Pr||(Pr={}));var Ir;(function(r){r[r.UNPACKED_FLOAT16=0]="UNPACKED_FLOAT16",r[r.UNPACKED_FLOAT32=1]="UNPACKED_FLOAT32",r[r.PACKED_4X1_UNSIGNED_BYTE=2]="PACKED_4X1_UNSIGNED_BYTE",r[r.PACKED_2X2_FLOAT32=3]="PACKED_2X2_FLOAT32",r[r.PACKED_2X2_FLOAT16=4]="PACKED_2X2_FLOAT16"})(Ir||(Ir={}));function Ou(r,e){return[e,r]}function U2(r,e){return r*e}function wl(r){let e=y.sizeFromShape(r),t=Math.ceil(e/4);return y.sizeToSquarishShape(t)}function ui(r,e){return[Math.max(1,Math.ceil(e/2)),Math.max(1,Math.ceil(r/2))]}function H2(r,e){let[t,n]=ui(r,e);return t*n*4}function Df(r,e){let t=r,n,o,s,a,i,l,u,c,p,m;return W().getNumber("WEBGL_VERSION")===2?(n=t.R32F,o=t.R16F,s=t.RGBA16F,a=t.RGBA32F,i=t.RED,u=4,c=1,p=t.HALF_FLOAT,m=t.FLOAT):(n=r.RGBA,o=r.RGBA,s=r.RGBA,a=t.RGBA,i=r.RGBA,u=4,c=4,p=e!=null?e.HALF_FLOAT_OES:null,m=r.FLOAT),l=r.RGBA,{internalFormatFloat:n,internalFormatHalfFloat:o,internalFormatPackedHalfFloat:s,internalFormatPackedFloat:a,textureFormatFloat:i,downloadTextureFormat:l,downloadUnpackNumChannels:u,defaultNumChannels:c,textureTypeHalfFloat:p,textureTypeFloat:m}}function Ne(r,e){let t=e();return W().getBool("DEBUG")&&E8(r),t}function E8(r){let e=r.getError();if(e!==r.NO_ERROR)throw new Error("WebGL Error: "+q2(r,e))}var D8=596e-10,$8=65504;function mk(r){return!!(W().getBool("WEBGL_RENDER_FLOAT32_ENABLED")||r===0||D8<Math.abs(r)&&Math.abs(r)<$8)}function q2(r,e){switch(e){case r.NO_ERROR:return"NO_ERROR";case r.INVALID_ENUM:return"INVALID_ENUM";case r.INVALID_VALUE:return"INVALID_VALUE";case r.INVALID_OPERATION:return"INVALID_OPERATION";case r.INVALID_FRAMEBUFFER_OPERATION:return"INVALID_FRAMEBUFFER_OPERATION";case r.OUT_OF_MEMORY:return"OUT_OF_MEMORY";case r.CONTEXT_LOST_WEBGL:return"CONTEXT_LOST_WEBGL";default:return`Unknown error code ${e}`}}function Np(r,e){return Da(r,()=>r.getExtension(e),'Extension "'+e+'" not supported on this browser.')}function fk(r,e){let t=Da(r,()=>r.createShader(r.VERTEX_SHADER),"Unable to create vertex WebGLShader.");if(Ne(r,()=>r.shaderSource(t,e)),Ne(r,()=>r.compileShader(t)),r.getShaderParameter(t,r.COMPILE_STATUS)===!1)throw console.log(r.getShaderInfoLog(t)),new Error("Failed to compile vertex shader.");return t}function dk(r,e){let t=Da(r,()=>r.createShader(r.FRAGMENT_SHADER),"Unable to create fragment WebGLShader.");if(Ne(r,()=>r.shaderSource(t,e)),Ne(r,()=>r.compileShader(t)),r.getShaderParameter(t,r.COMPILE_STATUS)===!1)throw F8(e,r.getShaderInfoLog(t)),new Error("Failed to compile fragment shader.");return t}var R8=/ERROR: [0-9]+:([0-9]+):/g;function F8(r,e){let t=R8.exec(e);if(t==null){console.log(`Couldn't parse line number in error: ${e}`),console.log(r);return}let n=+t[1],o=r.split(`
|
|
`),s=o.length.toString().length+2,a=o.map((p,m)=>y.rightPad((m+1).toString(),s)+p),i=0;for(let p=0;p<a.length;p++)i=Math.max(a[p].length,i);let l=a.slice(0,n-1),u=a.slice(n-1,n),c=a.slice(n);console.log(l.join(`
|
|
`)),console.log(e.split(`
|
|
`)[0]),console.log(`%c ${y.rightPad(u[0],i)}`,"border:1px solid red; background-color:#e3d2d2; color:#a61717"),console.log(c.join(`
|
|
`))}function hk(r){return Da(r,()=>r.createProgram(),"Unable to create WebGLProgram.")}function gk(r,e){if(Ne(r,()=>r.linkProgram(e)),r.getProgramParameter(e,r.LINK_STATUS)===!1)throw console.log(r.getProgramInfoLog(e)),new Error("Failed to link vertex and fragment shaders.")}function $f(r,e){if(Ne(r,()=>r.validateProgram(e)),r.getProgramParameter(e,r.VALIDATE_STATUS)===!1)throw console.log(r.getProgramInfoLog(e)),new Error("Shader program validation failed.")}function xk(r,e){let t=Da(r,()=>r.createBuffer(),"Unable to create WebGLBuffer");return Ne(r,()=>r.bindBuffer(r.ARRAY_BUFFER,t)),Ne(r,()=>r.bufferData(r.ARRAY_BUFFER,e,r.STATIC_DRAW)),t}function yk(r,e){let t=Da(r,()=>r.createBuffer(),"Unable to create WebGLBuffer");return Ne(r,()=>r.bindBuffer(r.ELEMENT_ARRAY_BUFFER,t)),Ne(r,()=>r.bufferData(r.ELEMENT_ARRAY_BUFFER,e,r.STATIC_DRAW)),t}function O8(){return W().getNumber("WEBGL_VERSION")===2?1:4}function bk(r){return Da(r,()=>r.createTexture(),"Unable to create WebGLTexture.")}function wk(r,e){let t=W().getNumber("WEBGL_MAX_TEXTURE_SIZE");if(r<=0||e<=0){let n=`[${r}x${e}]`;throw new Error("Requested texture size "+n+" is invalid.")}if(r>t||e>t){let n=`[${r}x${e}]`,o=`[${t}x${t}]`;throw new Error("Requested texture size "+n+" greater than WebGL maximum on this browser / GPU "+o+".")}}function _k(r){return Da(r,()=>r.createFramebuffer(),"Unable to create WebGLFramebuffer.")}function Qg(r,e,t,n,o,s,a){let i=r.getAttribLocation(e,t);return i===-1?!1:(Ne(r,()=>r.bindBuffer(r.ARRAY_BUFFER,n)),Ne(r,()=>r.vertexAttribPointer(i,o,r.FLOAT,!1,s,a)),Ne(r,()=>r.enableVertexAttribArray(i)),!0)}function K2(r,e,t){Y2(r,t),Ne(r,()=>r.activeTexture(r.TEXTURE0+t)),Ne(r,()=>r.bindTexture(r.TEXTURE_2D,e))}function P8(r,e){Y2(r,e),Ne(r,()=>r.activeTexture(r.TEXTURE0+e)),Ne(r,()=>r.bindTexture(r.TEXTURE_2D,null))}function kk(r,e,t){return Da(r,()=>r.getUniformLocation(e,t),'uniform "'+t+'" not present in program.')}function vk(r,e,t){return r.getUniformLocation(e,t)}function Ck(r,e,t,n){Ne(r,()=>K2(r,e,n)),Ne(r,()=>r.uniform1i(t,n))}function M8(r){Ne(r,()=>r.bindFramebuffer(r.FRAMEBUFFER,null)),Ne(r,()=>r.viewport(0,0,r.canvas.width,r.canvas.height)),Ne(r,()=>r.scissor(0,0,r.canvas.width,r.canvas.height))}function Rf(r,e,t){Ne(r,()=>r.bindFramebuffer(r.FRAMEBUFFER,t)),Ne(r,()=>r.framebufferTexture2D(r.FRAMEBUFFER,r.COLOR_ATTACHMENT0,r.TEXTURE_2D,e,0))}function ex(r,e){Ne(r,()=>r.bindFramebuffer(r.FRAMEBUFFER,e)),Ne(r,()=>r.framebufferTexture2D(r.FRAMEBUFFER,r.COLOR_ATTACHMENT0,r.TEXTURE_2D,null,0))}function Sp(r){let e=r.checkFramebufferStatus(r.FRAMEBUFFER);if(e!==r.FRAMEBUFFER_COMPLETE)throw new Error("Error binding framebuffer: "+X2(r,e))}function X2(r,e){switch(e){case r.FRAMEBUFFER_INCOMPLETE_ATTACHMENT:return"FRAMEBUFFER_INCOMPLETE_ATTACHMENT";case r.FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT:return"FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT";case r.FRAMEBUFFER_INCOMPLETE_DIMENSIONS:return"FRAMEBUFFER_INCOMPLETE_DIMENSIONS";case r.FRAMEBUFFER_UNSUPPORTED:return"FRAMEBUFFER_UNSUPPORTED";default:return`unknown error ${e}`}}function Da(r,e,t){let n=Ne(r,()=>e());if(n==null)throw new Error(t);return n}function Y2(r,e){let t=r.MAX_COMBINED_TEXTURE_IMAGE_UNITS-1,n=e+r.TEXTURE0;if(n<r.TEXTURE0||n>t){let o=`[gl.TEXTURE0, gl.TEXTURE${t}]`;throw new Error(`textureUnit must be in ${o}.`)}}function $a(r,e=2){return y.sizeFromShape(r.slice(0,r.length-e))}function Ra(r){if(r.length===0)throw Error("Cannot get rows and columns of an empty shape array.");return[r.length>1?r[r.length-2]:1,r[r.length-1]]}function Ff(r){let e=[1,1,1];return r.length===0||r.length===1&&r[0]===1||(e=[$a(r),...Ra(r)]),e}function Ik(r,e=!1){let t=W().getNumber("WEBGL_MAX_TEXTURE_SIZE");e&&(t=t*2,r=r.map((o,s)=>s>=r.length-2?y.nearestLargerEven(r[s]):r[s]),r.length===1&&(r=[2,r[0]])),r.length!==2&&(r=y.squeezeShape(r).newShape);let n=y.sizeFromShape(r);if(r.length<=1&&n<=t)return[1,n];if(r.length===2&&r[0]<=t&&r[1]<=t)return r;if(r.length===3&&r[0]*r[1]<=t&&r[2]<=t)return[r[0]*r[1],r[2]];if(r.length===3&&r[0]<=t&&r[1]*r[2]<=t)return[r[0],r[1]*r[2]];if(r.length===4&&r[0]*r[1]*r[2]<=t&&r[3]<=t)return[r[0]*r[1]*r[2],r[3]];if(r.length===4&&r[0]<=t&&r[1]*r[2]*r[3]<=t)return[r[0],r[1]*r[2]*r[3]];if(e){let o=$a(r),s=2,a=2;return r.length&&([s,a]=Ra(r)),n=o*(s/2)*(a/2),y.sizeToSquarishShape(n).map(i=>i*2)}return y.sizeToSquarishShape(n)}function tx(r){return r%2==0}function _l(r,e){if(r=r.slice(-2),e=e.slice(-2),y.arraysEqual(r,e)||!r.length||!e.length||r[0]===0||r[1]===0||e[0]===0||e[1]===0)return!0;if(r.length!==e.length){let t=r.slice(-1)[0],n=e.slice(-1)[0];if(t===n||tx(t)&&tx(n)&&(r[0]===1||e[0]===1))return!0}return r[1]===e[1]&&tx(r[0])&&tx(e[0])}var rx,nx;function Nk(r){if(rx==null){let e=Vn(r);rx=e.getParameter(e.MAX_TEXTURE_SIZE)}return rx}function L8(){rx=null}function z8(){nx=null}function Sk(r){if(nx==null){let e=Vn(r);nx=e.getParameter(e.MAX_TEXTURE_IMAGE_UNITS)}return Math.min(16,nx)}function Tk(r){if(r===0)return 0;let e,t=Vn(r);return An(t,"EXT_disjoint_timer_query_webgl2")&&r===2?e=2:An(t,"EXT_disjoint_timer_query")?e=1:e=0,e}function An(r,e){return r.getExtension(e)!=null}function ox(r){try{if(Vn(r)!=null)return!0}catch(e){return console.log("Error when getting WebGL context: ",e),!1}return!1}function Ak(r){if(r===0)return!1;let e=Vn(r);if(r===1){if(!An(e,"OES_texture_float"))return!1}else if(!An(e,"EXT_color_buffer_float"))return!1;return Dk(e)}function Ek(r){if(r===0)return!1;let e=Vn(r);if(r===1){if(!An(e,"OES_texture_float")||!An(e,"WEBGL_color_buffer_float"))return!1}else{if(An(e,"EXT_color_buffer_float"))return Dk(e);let n="EXT_color_buffer_half_float";if(An(e,n)){let o=e.getExtension(n);return B8(e,o)}return!1}return Dk(e)}function Dk(r){let e=Df(r),t=r.createTexture();r.bindTexture(r.TEXTURE_2D,t);let n=1,o=1;r.texImage2D(r.TEXTURE_2D,0,e.internalFormatFloat,n,o,0,e.textureFormatFloat,e.textureTypeFloat,null);let s=r.createFramebuffer();r.bindFramebuffer(r.FRAMEBUFFER,s),r.framebufferTexture2D(r.FRAMEBUFFER,r.COLOR_ATTACHMENT0,r.TEXTURE_2D,t,0);let a=r.checkFramebufferStatus(r.FRAMEBUFFER)===r.FRAMEBUFFER_COMPLETE;return r.bindTexture(r.TEXTURE_2D,null),r.bindFramebuffer(r.FRAMEBUFFER,null),r.deleteTexture(t),r.deleteFramebuffer(s),a}function B8(r,e){let t=Df(r,e),n=r.createTexture();r.bindTexture(r.TEXTURE_2D,n);let o=1,s=1;r.texImage2D(r.TEXTURE_2D,0,t.internalFormatHalfFloat,o,s,0,t.textureFormatFloat,t.textureTypeHalfFloat,null);let a=r.createFramebuffer();r.bindFramebuffer(r.FRAMEBUFFER,a),r.framebufferTexture2D(r.FRAMEBUFFER,r.COLOR_ATTACHMENT0,r.TEXTURE_2D,n,0);let i=r.checkFramebufferStatus(r.FRAMEBUFFER)===r.FRAMEBUFFER_COMPLETE;return r.bindTexture(r.TEXTURE_2D,null),r.bindFramebuffer(r.FRAMEBUFFER,null),r.deleteTexture(n),r.deleteFramebuffer(a),i}function $k(r){return r!==2?!1:Vn(r).fenceSync!=null}function Cs(r,e){Array.isArray(r)||(r=[r]),r.forEach(t=>{t!=null&&y.assert(t.dtype!=="complex64",()=>`${e} does not support complex64 tensors in the WebGL backend.`)})}var Oe=W();Oe.registerFlag("HAS_WEBGL",()=>Oe.getNumber("WEBGL_VERSION")>0);Oe.registerFlag("WEBGL_VERSION",()=>ox(2)?2:ox(1)?1:0);Oe.registerFlag("WEBGL_CHECK_NUMERICAL_PROBLEMS",()=>!1);Oe.registerFlag("WEBGL_BUFFER_SUPPORTED",()=>Oe.get("WEBGL_VERSION")===2);Oe.registerFlag("WEBGL_CPU_FORWARD",()=>!0);Oe.registerFlag("WEBGL_FORCE_F16_TEXTURES",()=>!1);Oe.registerFlag("WEBGL_PACK",()=>Oe.getBool("HAS_WEBGL"));Oe.registerFlag("WEBGL_PACK_NORMALIZATION",()=>Oe.getBool("WEBGL_PACK"));Oe.registerFlag("WEBGL_PACK_CLIP",()=>Oe.getBool("WEBGL_PACK"));Oe.registerFlag("WEBGL_PACK_DEPTHWISECONV",()=>Oe.getBool("WEBGL_PACK"));Oe.registerFlag("WEBGL_PACK_BINARY_OPERATIONS",()=>Oe.getBool("WEBGL_PACK"));Oe.registerFlag("WEBGL_PACK_UNARY_OPERATIONS",()=>Oe.getBool("WEBGL_PACK"));Oe.registerFlag("WEBGL_PACK_ARRAY_OPERATIONS",()=>Oe.getBool("WEBGL_PACK"));Oe.registerFlag("WEBGL_PACK_IMAGE_OPERATIONS",()=>Oe.getBool("WEBGL_PACK"));Oe.registerFlag("WEBGL_PACK_REDUCE",()=>Oe.getBool("WEBGL_PACK"));Oe.registerFlag("WEBGL_LAZILY_UNPACK",()=>Oe.getBool("WEBGL_PACK"));Oe.registerFlag("WEBGL_CONV_IM2COL",()=>Oe.getBool("WEBGL_PACK"));Oe.registerFlag("WEBGL_MAX_TEXTURE_SIZE",()=>Nk(Oe.getNumber("WEBGL_VERSION")));Oe.registerFlag("WEBGL_MAX_TEXTURES_IN_SHADER",()=>Sk(Oe.getNumber("WEBGL_VERSION")));Oe.registerFlag("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION",()=>{let r=Oe.getNumber("WEBGL_VERSION");return r===0?0:Tk(r)});Oe.registerFlag("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE",()=>Oe.getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION")>0&&!Jl.isMobile());Oe.registerFlag("WEBGL_RENDER_FLOAT32_CAPABLE",()=>Ak(Oe.getNumber("WEBGL_VERSION")));Oe.registerFlag("WEBGL_RENDER_FLOAT32_ENABLED",()=>Oe.getBool("WEBGL_FORCE_F16_TEXTURES")?!1:Oe.getBool("WEBGL_RENDER_FLOAT32_CAPABLE"));Oe.registerFlag("WEBGL_DOWNLOAD_FLOAT_ENABLED",()=>Ek(Oe.getNumber("WEBGL_VERSION")));Oe.registerFlag("WEBGL_FENCE_API_ENABLED",()=>$k(Oe.getNumber("WEBGL_VERSION")));Oe.registerFlag("WEBGL_SIZE_UPLOAD_UNIFORM",()=>Oe.getBool("WEBGL_RENDER_FLOAT32_ENABLED")?4:0);Oe.registerFlag("WEBGL_DELETE_TEXTURE_THRESHOLD",()=>-1,r=>{if(r<0&&r!==-1)throw new Error(`WEBGL_DELETE_TEXTURE_THRESHOLD must be -1 (indicating never delete) or at least 0, but got ${r}.`)});Oe.registerFlag("WEBGL_FLUSH_THRESHOLD",()=>Jl.isMobile()&&Oe.getBool("IS_CHROME")?1:-1,r=>{if(r<0&&r!==-1)throw new Error(`WEBGL_FLUSH_THRESHOLD must be -1 (indicating never manual flush) or at least 0, but got ${r}.`)});function Pt(){let r,e,t,n,o,s,a,i,l,u;return W().getNumber("WEBGL_VERSION")===2?(r="#version 300 es",e="in",t="out",n="in",o="texture",s="outputColor",a="out vec4 outputColor;",i=`
|
|
bool isnan_custom(float val) {
|
|
return (val > 0.0 || val < 0.0) ? false : val != 0.0;
|
|
}
|
|
|
|
bvec4 isnan_custom(vec4 val) {
|
|
return bvec4(isnan_custom(val.x),
|
|
isnan_custom(val.y), isnan_custom(val.z), isnan_custom(val.w));
|
|
}
|
|
|
|
#define isnan(value) isnan_custom(value)
|
|
`,l="",u=`
|
|
#define round(value) newRound(value)
|
|
int newRound(float value) {
|
|
return int(floor(value + 0.5));
|
|
}
|
|
|
|
ivec4 newRound(vec4 value) {
|
|
return ivec4(floor(value + vec4(0.5)));
|
|
}
|
|
`):(r="",e="attribute",t="varying",n="varying",o="texture2D",s="gl_FragColor",a="",i=`
|
|
#define isnan(value) isnan_custom(value)
|
|
bool isnan_custom(float val) {
|
|
return (val > 0. || val < 1. || val == 0.) ? false : true;
|
|
}
|
|
bvec4 isnan_custom(vec4 val) {
|
|
return bvec4(isnan(val.x), isnan(val.y), isnan(val.z), isnan(val.w));
|
|
}
|
|
`,l=`
|
|
uniform float INFINITY;
|
|
|
|
bool isinf(float val) {
|
|
return abs(val) == INFINITY;
|
|
}
|
|
bvec4 isinf(vec4 val) {
|
|
return equal(abs(val), vec4(INFINITY));
|
|
}
|
|
`,u=`
|
|
int round(float value) {
|
|
return int(floor(value + 0.5));
|
|
}
|
|
|
|
ivec4 round(vec4 value) {
|
|
return ivec4(floor(value + vec4(0.5)));
|
|
}
|
|
`),{version:r,attribute:e,varyingVs:t,varyingFs:n,texture2D:o,output:s,defineOutput:a,defineSpecialNaN:i,defineSpecialInf:l,defineRound:u}}function Is(r,e,t="index"){let n=y.computeStrides(e);return n.map((o,s)=>{let a=`int ${r[s]} = ${t} / ${o}`,i=s===n.length-1?`int ${r[s+1]} = ${t} - ${r[s]} * ${o}`:`index -= ${r[s]} * ${o}`;return`${a}; ${i};`}).join("")}function Tp(r){let e=y.computeStrides(r).map(t=>t.toString());return`
|
|
int getFlatIndex(ivec3 coords) {
|
|
return coords.x * ${e[0]} + coords.y * ${e[1]} + coords.z;
|
|
}
|
|
`}var sx=`
|
|
const float FLOAT_MAX = 1.70141184e38;
|
|
const float FLOAT_MIN = 1.17549435e-38;
|
|
|
|
lowp vec4 encode_float(highp float v) {
|
|
if (isnan(v)) {
|
|
return vec4(255, 255, 255, 255);
|
|
}
|
|
|
|
highp float av = abs(v);
|
|
|
|
if(av < FLOAT_MIN) {
|
|
return vec4(0.0, 0.0, 0.0, 0.0);
|
|
} else if(v > FLOAT_MAX) {
|
|
return vec4(0.0, 0.0, 128.0, 127.0) / 255.0;
|
|
} else if(v < -FLOAT_MAX) {
|
|
return vec4(0.0, 0.0, 128.0, 255.0) / 255.0;
|
|
}
|
|
|
|
highp vec4 c = vec4(0,0,0,0);
|
|
|
|
highp float e = floor(log2(av));
|
|
highp float m = exp2(fract(log2(av))) - 1.0;
|
|
|
|
c[2] = floor(128.0 * m);
|
|
m -= c[2] / 128.0;
|
|
c[1] = floor(32768.0 * m);
|
|
m -= c[1] / 32768.0;
|
|
c[0] = floor(8388608.0 * m);
|
|
|
|
highp float ebias = e + 127.0;
|
|
c[3] = floor(ebias / 2.0);
|
|
ebias -= c[3] * 2.0;
|
|
c[2] += floor(ebias) * 128.0;
|
|
|
|
c[3] += 128.0 * step(0.0, -v);
|
|
|
|
return c / 255.0;
|
|
}
|
|
`;var Rk=class{constructor(e){this.variableNames=["A"],this.packedInputs=!1,this.packedOutput=!0,this.outPackingScheme=bl.DENSE;let t=wl(e),n=Pt();this.outputShape=e,this.userCode=`
|
|
ivec3 outCoordsFromFlatIndex(int index) {
|
|
${Is(["r","c","d"],e)}
|
|
return ivec3(r, c, d);
|
|
}
|
|
|
|
void main() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${t[0]}, ${t[1]}));
|
|
int index = 4 * (resTexRC.x * ${t[1]} + resTexRC.y);
|
|
|
|
vec4 result = vec4(0.);
|
|
|
|
for (int i=0; i<4; i++) {
|
|
int flatIndex = index + i;
|
|
ivec3 rc = outCoordsFromFlatIndex(flatIndex);
|
|
result[i] = getA(rc.x, rc.y, rc.z);
|
|
}
|
|
|
|
${n.output} = result;
|
|
}
|
|
`}};var Fk=class{constructor(e){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!0,this.outPackingScheme=bl.DENSE;let t=wl(e),n=Pt();this.outputShape=e,this.userCode=`
|
|
ivec3 outCoordsFromFlatIndex(int index) {
|
|
${Is(["r","c","d"],e)}
|
|
return ivec3(r, c, d);
|
|
}
|
|
|
|
void main() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${t[0]}, ${t[1]}));
|
|
int index = 4 * (resTexRC.x * ${t[1]} + resTexRC.y);
|
|
|
|
vec4 result = vec4(0.);
|
|
|
|
for (int i=0; i<4; i++) {
|
|
int flatIndex = index + i;
|
|
ivec3 rc = outCoordsFromFlatIndex(flatIndex);
|
|
result[i] = getChannel(getA(rc.x, rc.y, rc.z), vec2(rc.y, rc.z));
|
|
}
|
|
|
|
${n.output} = result;
|
|
}
|
|
`}};var Ok=class{constructor(e){this.variableNames=["A"],this.outTexUsage=Pr.DOWNLOAD;let t=Pt();this.outputShape=e,this.userCode=`
|
|
${sx}
|
|
|
|
void main() {
|
|
float x = getAAtOutCoords();
|
|
${t.output} = encode_float(x);
|
|
}
|
|
`}};var Pk=class{constructor(e){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!1,this.outTexUsage=Pr.DOWNLOAD;let t=Pt();this.outputShape=e,this.userCode=`
|
|
${sx}
|
|
|
|
void main() {
|
|
ivec3 coords = getOutputCoords();
|
|
float x = getChannel(getAAtOutCoords(), vec2(coords.y, coords.z));
|
|
${t.output} = encode_float(x);
|
|
}
|
|
`}};var Mk=class{constructor(e,t,n=!1){this.variableNames=["A"];let o=Pt(),[s,a]=t;this.outputShape=e;let i="result";n&&(i="floor(result * 255. + 0.5)"),this.userCode=`
|
|
${Tp(e)}
|
|
|
|
void main() {
|
|
ivec3 coords = getOutputCoords();
|
|
|
|
int flatIndex = getFlatIndex(coords);
|
|
int offset = imod(flatIndex, 4);
|
|
|
|
flatIndex = idiv(flatIndex, 4, 1.);
|
|
|
|
int r = flatIndex / ${a};
|
|
int c = imod(flatIndex, ${a});
|
|
vec2 uv = (vec2(c, r) + halfCR) / vec2(${a}.0, ${s}.0);
|
|
vec4 values = ${o.texture2D}(A, uv);
|
|
|
|
float result;
|
|
|
|
if(offset == 0) {
|
|
result = values[0];
|
|
} else if(offset == 1) {
|
|
result = values[1];
|
|
} else if(offset == 2) {
|
|
result = values[2];
|
|
} else {
|
|
result = values[3];
|
|
}
|
|
|
|
${o.output} = vec4(${i}, 0., 0., 0.);
|
|
}
|
|
`}};var Lk=class{constructor(e,t,n=!1){this.variableNames=["A"],this.packedInputs=!1,this.packedOutput=!0;let o=Pt(),[s,a]=t;this.outputShape=e;let i="",l="result";n&&(l="floor(result * 255. + 0.5)");for(let u=0;u<=1;u++)for(let c=0;c<=1;c++){let p=u*2+c;i+=`
|
|
localCoords = coords;
|
|
if(localCoords[2] + ${c} < ${e[2]}) {
|
|
localCoords[2] += ${c};
|
|
if(localCoords[1] + ${u} < ${e[1]}) {
|
|
localCoords[1] += ${u};
|
|
|
|
flatIndex = getFlatIndex(localCoords);
|
|
offset = imod(flatIndex, 4);
|
|
|
|
flatIndex = idiv(flatIndex, 4, 1.);
|
|
|
|
r = flatIndex / ${a};
|
|
c = imod(flatIndex, ${a});
|
|
uv = (vec2(c, r) + halfCR) / vec2(${a}.0, ${s}.0);
|
|
values = ${o.texture2D}(A, uv);
|
|
|
|
if(offset == 0) {
|
|
result[${p}] = values[0];
|
|
} else if(offset == 1) {
|
|
result[${p}] = values[1];
|
|
} else if(offset == 2) {
|
|
result[${p}] = values[2];
|
|
} else {
|
|
result[${p}] = values[3];
|
|
}
|
|
}
|
|
}
|
|
`}this.userCode=`
|
|
${Tp(e)}
|
|
|
|
void main() {
|
|
ivec3 coords = getOutputCoords();
|
|
|
|
vec4 result = vec4(0.);
|
|
int flatIndex, r, c, offset;
|
|
ivec3 localCoords;
|
|
vec2 uv;
|
|
vec4 values;
|
|
|
|
${i}
|
|
|
|
${o.output} = ${l};
|
|
}
|
|
`}};var J2={};We(J2,{bindVertexProgramAttributeStreams:()=>qk,createBufferFromOutputTexture:()=>Yk,createFloat16MatrixTexture:()=>Wk,createFloat16PackedMatrixTexture:()=>Hk,createFloat32MatrixTexture:()=>Gk,createIndexBuffer:()=>Vk,createPackedMatrixTexture:()=>Uk,createUnsignedBytesMatrixTexture:()=>jk,createVertexBuffer:()=>Bk,createVertexShader:()=>zk,downloadByteEncodedFloatMatrixFromOutputTexture:()=>Jk,downloadFloat32MatrixFromBuffer:()=>Zk,downloadMatrixFromPackedOutputTexture:()=>ev,downloadPackedMatrixFromBuffer:()=>Qk,getInternalFormatForFloat16MatrixTexture:()=>ax,getInternalFormatForFloat16PackedMatrixTexture:()=>cx,getInternalFormatForFloat32MatrixTexture:()=>ix,getInternalFormatForPackedMatrixTexture:()=>ux,getInternalFormatForUnsignedBytesMatrixTexture:()=>lx,uploadDenseMatrixToTexture:()=>Kk,uploadPixelDataToTexture:()=>Xk});function zk(r){let e=Pt(),t=`${e.version}
|
|
precision highp float;
|
|
${e.attribute} vec3 clipSpacePos;
|
|
${e.attribute} vec2 uv;
|
|
${e.varyingVs} vec2 resultUV;
|
|
|
|
void main() {
|
|
gl_Position = vec4(clipSpacePos, 1);
|
|
resultUV = uv;
|
|
}`;return fk(r,t)}function Bk(r){let e=new Float32Array([-1,1,0,0,1,-1,-1,0,0,0,1,1,0,1,1,1,-1,0,1,0]);return xk(r,e)}function Vk(r){let e=new Uint16Array([0,1,2,2,1,3]);return yk(r,e)}function Of(r,e,t,n,o,s){wk(e,t);let a=bk(r),i=r.TEXTURE_2D;return Ne(r,()=>r.bindTexture(i,a)),Ne(r,()=>r.texParameteri(i,r.TEXTURE_WRAP_S,r.CLAMP_TO_EDGE)),Ne(r,()=>r.texParameteri(i,r.TEXTURE_WRAP_T,r.CLAMP_TO_EDGE)),Ne(r,()=>r.texParameteri(i,r.TEXTURE_MIN_FILTER,r.NEAREST)),Ne(r,()=>r.texParameteri(i,r.TEXTURE_MAG_FILTER,r.NEAREST)),Ne(r,()=>r.texImage2D(i,0,n,e,t,0,o,s,null)),Ne(r,()=>r.bindTexture(r.TEXTURE_2D,null)),a}function ix(r){return r.internalFormatFloat}function Gk(r,e,t,n){let[o,s]=Ou(e,t);return Of(r,o,s,ix(n),n.textureFormatFloat,r.FLOAT)}function ax(r){return r.internalFormatHalfFloat}function Wk(r,e,t,n){let[o,s]=Ou(e,t);return Of(r,o,s,ax(n),n.textureFormatFloat,n.textureTypeHalfFloat)}function lx(r){return r.downloadTextureFormat}function jk(r,e,t,n){let[o,s]=Ou(e,t);return Of(r,o,s,lx(n),r.RGBA,r.UNSIGNED_BYTE)}function ux(r){return r.internalFormatPackedFloat}function Uk(r,e,t,n){let[o,s]=ui(e,t);return Of(r,o,s,ux(n),r.RGBA,r.FLOAT)}function cx(r){return r.internalFormatPackedHalfFloat}function Hk(r,e,t,n){let[o,s]=ui(e,t);return Of(r,o,s,cx(n),r.RGBA,n.textureTypeHalfFloat)}function qk(r,e,t){let n=0,o=3*4,s=3*4+2*4;return Ne(r,()=>r.bindBuffer(r.ARRAY_BUFFER,t)),Qg(r,e,"clipSpacePos",t,3,s,n)&&Qg(r,e,"uv",t,2,s,o)}function Kk(r,e,t,n,o,s){Ne(r,()=>r.bindTexture(r.TEXTURE_2D,e));let a,i,l;o instanceof Uint8Array?(a=new Uint8Array(t*n*4),i=r.UNSIGNED_BYTE,l=r.RGBA):(a=new Float32Array(t*n*4),i=r.FLOAT,l=s.internalFormatPackedFloat),a.set(o),Ne(r,()=>r.texImage2D(r.TEXTURE_2D,0,l,t,n,0,r.RGBA,i,a)),Ne(r,()=>r.bindTexture(r.TEXTURE_2D,null))}function Xk(r,e,t){Ne(r,()=>r.bindTexture(r.TEXTURE_2D,e)),t.data instanceof Uint8Array?Ne(r,()=>r.texImage2D(r.TEXTURE_2D,0,r.RGBA,t.width,t.height,0,r.RGBA,r.UNSIGNED_BYTE,t.data)):Ne(r,()=>r.texImage2D(r.TEXTURE_2D,0,r.RGBA,r.RGBA,r.UNSIGNED_BYTE,t)),Ne(r,()=>r.bindTexture(r.TEXTURE_2D,null))}function Yk(r,e,t,n){let o=r.createBuffer();Ne(r,()=>r.bindBuffer(r.PIXEL_PACK_BUFFER,o));let i=4*4*e*t;return Ne(r,()=>r.bufferData(r.PIXEL_PACK_BUFFER,i,r.STREAM_READ)),Ne(r,()=>r.readPixels(0,0,t,e,r.RGBA,r.FLOAT,0)),Ne(r,()=>r.bindBuffer(r.PIXEL_PACK_BUFFER,null)),o}function Zk(r,e,t){let n=r,o=new Float32Array(t);return n.bindBuffer(n.PIXEL_PACK_BUFFER,e),n.getBufferSubData(n.PIXEL_PACK_BUFFER,0,o),n.bindBuffer(n.PIXEL_PACK_BUFFER,null),o}function Jk(r,e,t,n){let[o,s]=Ou(e,t),a=4,i=new Uint8Array(U2(e*t,a));return Ne(r,()=>r.readPixels(0,0,o,s,n.downloadTextureFormat,r.UNSIGNED_BYTE,i)),new Float32Array(i.buffer)}function Qk(r,e,t,n,o,s,a,i){let l=r,u=new Float32Array(H2(s,a));return l.bindBuffer(l.PIXEL_PACK_BUFFER,e),l.getBufferSubData(l.PIXEL_PACK_BUFFER,0,u),l.bindBuffer(l.PIXEL_PACK_BUFFER,null),u}function ev(r,e,t){let n=new Float32Array(e*t*4);return Ne(r,()=>r.readPixels(0,0,t,e,r.RGBA,r.FLOAT,n)),n}var px=class{constructor(e){this.outputTexture=null,this.program=null,this.disposed=!1,this.vertexAttrsAreBound=!1,this.itemsToPoll=[];let t=W().getNumber("WEBGL_VERSION");e!=null?(this.gl=e,pk(t,e)):this.gl=Vn(t);let n="WEBGL_color_buffer_float",o="EXT_color_buffer_half_float";if(W().getNumber("WEBGL_VERSION")===1){let s="OES_texture_float",a="OES_texture_half_float";if(this.textureFloatExtension=Np(this.gl,s),An(this.gl,a))this.textureHalfFloatExtension=Np(this.gl,a);else if(W().get("WEBGL_FORCE_F16_TEXTURES"))throw new Error("GL context does not support half float textures, yet the environment flag WEBGL_FORCE_F16_TEXTURES is set to true.");if(this.colorBufferFloatExtension=this.gl.getExtension(n),An(this.gl,o))this.colorBufferHalfFloatExtension=Np(this.gl,o);else if(W().get("WEBGL_FORCE_F16_TEXTURES"))throw new Error("GL context does not support color renderable half floats, yet the environment flag WEBGL_FORCE_F16_TEXTURES is set to true.")}else if(n="EXT_color_buffer_float",An(this.gl,n))this.colorBufferFloatExtension=this.gl.getExtension(n);else if(An(this.gl,o))this.colorBufferHalfFloatExtension=this.gl.getExtension(o);else throw new Error("GL context does not support color renderable floats");this.vertexBuffer=Bk(this.gl),this.indexBuffer=Vk(this.gl),this.framebuffer=_k(this.gl),this.textureConfig=Df(this.gl,this.textureHalfFloatExtension)}get debug(){return W().getBool("DEBUG")}dispose(){if(this.disposed)return;this.program!=null&&console.warn("Disposing a GPGPUContext that still has a bound WebGLProgram. This is probably a resource leak, delete the program with GPGPUContext.deleteProgram before disposing."),this.outputTexture!=null&&console.warn("Disposing a GPGPUContext that still has a bound output matrix texture. This is probably a resource leak, delete the output matrix texture with GPGPUContext.deleteMatrixTexture before disposing.");let e=this.gl;Ne(e,()=>e.finish()),Ne(e,()=>e.bindFramebuffer(e.FRAMEBUFFER,null)),Ne(e,()=>e.deleteFramebuffer(this.framebuffer)),Ne(e,()=>e.bindBuffer(e.ARRAY_BUFFER,null)),Ne(e,()=>e.bindBuffer(e.ELEMENT_ARRAY_BUFFER,null)),Ne(e,()=>e.deleteBuffer(this.indexBuffer)),this.disposed=!0}createFloat32MatrixTexture(e,t){return this.throwIfDisposed(),Gk(this.gl,e,t,this.textureConfig)}createFloat16MatrixTexture(e,t){return this.throwIfDisposed(),Wk(this.gl,e,t,this.textureConfig)}createUnsignedBytesMatrixTexture(e,t){return this.throwIfDisposed(),jk(this.gl,e,t,this.textureConfig)}uploadPixelDataToTexture(e,t){this.throwIfDisposed(),Xk(this.gl,e,t)}uploadDenseMatrixToTexture(e,t,n,o){this.throwIfDisposed(),Kk(this.gl,e,t,n,o,this.textureConfig)}createFloat16PackedMatrixTexture(e,t){return this.throwIfDisposed(),Hk(this.gl,e,t,this.textureConfig)}createPackedMatrixTexture(e,t){return this.throwIfDisposed(),Uk(this.gl,e,t,this.textureConfig)}deleteMatrixTexture(e){this.throwIfDisposed(),this.outputTexture===e&&(ex(this.gl,this.framebuffer),this.outputTexture=null),Ne(this.gl,()=>this.gl.deleteTexture(e))}downloadByteEncodedFloatMatrixFromOutputTexture(e,t,n){return this.downloadMatrixDriver(e,()=>Jk(this.gl,t,n,this.textureConfig))}downloadPackedMatrixFromBuffer(e,t,n,o,s,a){return Qk(this.gl,e,t,n,o,s,a,this.textureConfig)}downloadFloat32MatrixFromBuffer(e,t){return Zk(this.gl,e,t)}createBufferFromTexture(e,t,n){this.bindTextureToFrameBuffer(e);let o=Yk(this.gl,t,n,this.textureConfig);return this.unbindTextureToFrameBuffer(),o}createAndWaitForFence(){let e=this.createFence(this.gl);return this.pollFence(e)}createFence(e){let t,n;if(W().getBool("WEBGL_FENCE_API_ENABLED")){let o=e,s=o.fenceSync(o.SYNC_GPU_COMMANDS_COMPLETE,0);e.flush(),n=()=>{let a=o.clientWaitSync(s,0,0);return a===o.ALREADY_SIGNALED||a===o.CONDITION_SATISFIED},t=s}else W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION")>0?(t=this.beginQuery(),this.endQuery(),n=()=>this.isQueryAvailable(t,W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION"))):n=()=>!0;return{query:t,isFencePassed:n}}downloadMatrixFromPackedTexture(e,t,n){return this.downloadMatrixDriver(e,()=>ev(this.gl,t,n))}createProgram(e){this.throwIfDisposed();let t=this.gl,n=dk(t,e);this.vertexShader==null&&(this.vertexShader=zk(t));let o=hk(t);return Ne(t,()=>t.attachShader(o,this.vertexShader)),Ne(t,()=>t.attachShader(o,n)),gk(t,o),this.debug&&$f(t,o),this.vertexAttrsAreBound||(this.setProgram(o),this.vertexAttrsAreBound=qk(t,this.program,this.vertexBuffer)),o}deleteProgram(e){this.throwIfDisposed(),e===this.program&&(this.program=null),e!=null&&Ne(this.gl,()=>this.gl.deleteProgram(e))}setProgram(e){this.throwIfDisposed(),this.program=e,this.program!=null&&this.debug&&$f(this.gl,this.program),Ne(this.gl,()=>this.gl.useProgram(e))}getUniformLocation(e,t,n=!0){return this.throwIfDisposed(),n?kk(this.gl,e,t):vk(this.gl,e,t)}getAttributeLocation(e,t){return this.throwIfDisposed(),Ne(this.gl,()=>this.gl.getAttribLocation(e,t))}getUniformLocationNoThrow(e,t){return this.throwIfDisposed(),this.gl.getUniformLocation(e,t)}setInputMatrixTexture(e,t,n){this.throwIfDisposed(),this.throwIfNoProgram(),Ck(this.gl,e,t,n)}setOutputMatrixTexture(e,t,n){this.setOutputMatrixTextureDriver(e,n,t)}setOutputPackedMatrixTexture(e,t,n){this.throwIfDisposed();let[o,s]=ui(t,n);this.setOutputMatrixTextureDriver(e,o,s)}setOutputMatrixWriteRegion(e,t,n,o){this.setOutputMatrixWriteRegionDriver(n,e,o,t)}setOutputPackedMatrixWriteRegion(e,t,n,o){throw new Error("setOutputPackedMatrixWriteRegion not implemented.")}debugValidate(){this.program!=null&&$f(this.gl,this.program),Sp(this.gl)}executeProgram(){this.throwIfDisposed(),this.throwIfNoProgram();let e=this.gl;this.debug&&this.debugValidate(),Ne(e,()=>e.drawElements(e.TRIANGLES,6,e.UNSIGNED_SHORT,0))}blockUntilAllProgramsCompleted(){this.throwIfDisposed(),Ne(this.gl,()=>this.gl.finish())}getQueryTimerExtension(){return this.disjointQueryTimerExtension==null&&(this.disjointQueryTimerExtension=Np(this.gl,W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION")===2?"EXT_disjoint_timer_query_webgl2":"EXT_disjoint_timer_query")),this.disjointQueryTimerExtension}getQueryTimerExtensionWebGL2(){return this.getQueryTimerExtension()}getQueryTimerExtensionWebGL1(){return this.getQueryTimerExtension()}beginQuery(){if(W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION")===2){let n=this.gl,o=this.getQueryTimerExtensionWebGL2(),s=n.createQuery();return n.beginQuery(o.TIME_ELAPSED_EXT,s),s}let e=this.getQueryTimerExtensionWebGL1(),t=e.createQueryEXT();return e.beginQueryEXT(e.TIME_ELAPSED_EXT,t),t}endQuery(){if(W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION")===2){let t=this.gl,n=this.getQueryTimerExtensionWebGL2();t.endQuery(n.TIME_ELAPSED_EXT);return}let e=this.getQueryTimerExtensionWebGL1();e.endQueryEXT(e.TIME_ELAPSED_EXT)}async waitForQueryAndGetTime(e){return await y.repeatedTry(()=>this.disposed||this.isQueryAvailable(e,W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION"))),this.getQueryTime(e,W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_VERSION"))}getQueryTime(e,t){if(t===0)return null;if(t===2){let n=this.gl;return n.getQueryParameter(e,n.QUERY_RESULT)/1e6}else{let n=this.getQueryTimerExtensionWebGL1();return n.getQueryObjectEXT(e,n.QUERY_RESULT_EXT)/1e6}}isQueryAvailable(e,t){if(t===0)return!0;if(t===2){let n=this.gl,o=this.getQueryTimerExtensionWebGL2(),s=n.getQueryParameter(e,n.QUERY_RESULT_AVAILABLE);return this.disjoint==null&&(this.disjoint=this.gl.getParameter(o.GPU_DISJOINT_EXT)),s&&!this.disjoint}else{let n=this.getQueryTimerExtensionWebGL1(),o=n.getQueryObjectEXT(e,n.QUERY_RESULT_AVAILABLE_EXT);return this.disjoint==null&&(this.disjoint=this.gl.getParameter(n.GPU_DISJOINT_EXT)),o&&!this.disjoint}}pollFence(e){return new Promise(t=>{this.addItemToPoll(()=>e.isFencePassed(),()=>t())})}pollItems(){let e=V8(this.itemsToPoll.map(t=>t.isDoneFn));for(let t=0;t<=e;++t){let{resolveFn:n}=this.itemsToPoll[t];n()}this.itemsToPoll=this.itemsToPoll.slice(e+1)}addItemToPoll(e,t){this.itemsToPoll.push({isDoneFn:e,resolveFn:t}),!(this.itemsToPoll.length>1)&&y.repeatedTry(()=>(this.pollItems(),this.itemsToPoll.length===0))}bindTextureToFrameBuffer(e){this.throwIfDisposed(),Rf(this.gl,e,this.framebuffer),this.debug&&Sp(this.gl)}unbindTextureToFrameBuffer(){this.outputTexture!=null?(Rf(this.gl,this.outputTexture,this.framebuffer),this.debug&&Sp(this.gl)):ex(this.gl,this.framebuffer)}downloadMatrixDriver(e,t){this.bindTextureToFrameBuffer(e);let n=t();return this.unbindTextureToFrameBuffer(),n}setOutputMatrixTextureDriver(e,t,n){this.throwIfDisposed();let o=this.gl;Rf(o,e,this.framebuffer),this.debug&&Sp(o),this.outputTexture=e,Ne(o,()=>o.viewport(0,0,t,n)),Ne(o,()=>o.scissor(0,0,t,n))}setOutputMatrixWriteRegionDriver(e,t,n,o){this.throwIfDisposed(),Ne(this.gl,()=>this.gl.scissor(e,t,n,o))}throwIfDisposed(){if(this.disposed)throw new Error("Attempted to use disposed GPGPUContext.")}throwIfNoProgram(){if(this.program==null)throw new Error("No GPU program is currently set.")}};function V8(r){let e=0;for(;e<r.length&&r[e]();++e);return e-1}var{getBroadcastDims:Q2}=C;function eD(r,e,t,n){let o=[];r.forEach(d=>{let h=y.sizeFromShape(d.shapeInfo.logicalShape);d.shapeInfo.isUniform?o.push(`uniform float ${d.name}${h>1?`[${h}]`:""};`):(o.push(`uniform sampler2D ${d.name};`),o.push(`uniform int offset${d.name};`))});let s=o.join(`
|
|
`),a=r.map(d=>G8(d,e,n)).join(`
|
|
`),i=e.texShape,l=Pt(),u=U8(l),c,p,m=K8(l);return e.isPacked?(c=W8(e.logicalShape,i),p=q8(l)):(c=j8(e.logicalShape,i),p=H8(l)),n&&(m+=J8),[m,u,p,s,c,a,t].join(`
|
|
`)}function Ap(r){let e=r.shapeInfo.logicalShape;switch(e.length){case 0:return cX(r);case 1:return mX(r);case 2:return dX(r);case 3:return gX(r);case 4:return yX(r);case 5:return bX(r);case 6:return wX(r);default:throw new Error(`${e.length}-D input sampling is not yet supported`)}}function tD(r){switch(r.shapeInfo.logicalShape.length){case 0:return uX(r);case 1:return pX(r);case 2:return fX(r);case 3:return hX(r);default:return xX(r)}}function G8(r,e,t=!1){let n="";t?n+=tD(r):n+=Ap(r);let o=r.shapeInfo.logicalShape,s=e.logicalShape;return o.length<=s.length&&(t?n+=_X(r,e):n+=kX(r,e)),n}function W8(r,e){switch(r.length){case 0:return rD();case 1:return Q8(r,e);case 2:return aX(r,e);case 3:return tX(r,e);default:return nX(r,e)}}function j8(r,e){switch(r.length){case 0:return rD();case 1:return eX(r,e);case 2:return lX(r,e);case 3:return rX(r,e);case 4:return oX(r,e);case 5:return sX(r,e);case 6:return iX(r,e);default:throw new Error(`${r.length}-D output sampling is not yet supported`)}}function U8(r){return`
|
|
float sampleTexture(sampler2D textureSampler, vec2 uv) {
|
|
return ${r.texture2D}(textureSampler, uv).r;
|
|
}
|
|
`}function H8(r){return`
|
|
void setOutput(float val) {
|
|
${r.output} = vec4(val, 0, 0, 0);
|
|
}
|
|
`}function q8(r){return`
|
|
void setOutput(vec4 val) {
|
|
${r.output} = val;
|
|
}
|
|
`}function K8(r){return`${r.version}
|
|
precision highp float;
|
|
precision highp int;
|
|
precision highp sampler2D;
|
|
${r.varyingFs} vec2 resultUV;
|
|
${r.defineOutput}
|
|
const vec2 halfCR = vec2(0.5, 0.5);
|
|
|
|
struct ivec5
|
|
{
|
|
int x;
|
|
int y;
|
|
int z;
|
|
int w;
|
|
int u;
|
|
};
|
|
|
|
struct ivec6
|
|
{
|
|
int x;
|
|
int y;
|
|
int z;
|
|
int w;
|
|
int u;
|
|
int v;
|
|
};
|
|
|
|
uniform float NAN;
|
|
${r.defineSpecialNaN}
|
|
${r.defineSpecialInf}
|
|
${r.defineRound}
|
|
|
|
int imod(int x, int y) {
|
|
return x - y * (x / y);
|
|
}
|
|
|
|
int idiv(int a, int b, float sign) {
|
|
int res = a / b;
|
|
int mod = imod(a, b);
|
|
if (sign < 0. && mod != 0) {
|
|
res -= 1;
|
|
}
|
|
return res;
|
|
}
|
|
|
|
//Based on the work of Dave Hoskins
|
|
//https://www.shadertoy.com/view/4djSRW
|
|
#define HASHSCALE1 443.8975
|
|
float random(float seed){
|
|
vec2 p = resultUV * seed;
|
|
vec3 p3 = fract(vec3(p.xyx) * HASHSCALE1);
|
|
p3 += dot(p3, p3.yzx + 19.19);
|
|
return fract((p3.x + p3.y) * p3.z);
|
|
}
|
|
|
|
${X8}
|
|
${Y8}
|
|
${Z8}
|
|
`}var X8=`
|
|
vec2 uvFromFlat(int texNumR, int texNumC, int index) {
|
|
int texR = index / texNumC;
|
|
int texC = index - texR * texNumC;
|
|
return (vec2(texC, texR) + halfCR) / vec2(texNumC, texNumR);
|
|
}
|
|
vec2 packedUVfrom1D(int texNumR, int texNumC, int index) {
|
|
int texelIndex = index / 2;
|
|
int texR = texelIndex / texNumC;
|
|
int texC = texelIndex - texR * texNumC;
|
|
return (vec2(texC, texR) + halfCR) / vec2(texNumC, texNumR);
|
|
}
|
|
`,Y8=`
|
|
vec2 packedUVfrom2D(int texelsInLogicalRow, int texNumR,
|
|
int texNumC, int row, int col) {
|
|
int texelIndex = (row / 2) * texelsInLogicalRow + (col / 2);
|
|
int texR = texelIndex / texNumC;
|
|
int texC = texelIndex - texR * texNumC;
|
|
return (vec2(texC, texR) + halfCR) / vec2(texNumC, texNumR);
|
|
}
|
|
`,Z8=`
|
|
vec2 packedUVfrom3D(int texNumR, int texNumC,
|
|
int texelsInBatch, int texelsInLogicalRow, int b,
|
|
int row, int col) {
|
|
int index = b * texelsInBatch + (row / 2) * texelsInLogicalRow + (col / 2);
|
|
int texR = index / texNumC;
|
|
int texC = index - texR * texNumC;
|
|
return (vec2(texC, texR) + halfCR) / vec2(texNumC, texNumR);
|
|
}
|
|
`,J8=`
|
|
float getChannel(vec4 frag, vec2 innerDims) {
|
|
vec2 modCoord = mod(innerDims, 2.);
|
|
return modCoord.x == 0. ?
|
|
(modCoord.y == 0. ? frag.r : frag.g) :
|
|
(modCoord.y == 0. ? frag.b : frag.a);
|
|
}
|
|
float getChannel(vec4 frag, int dim) {
|
|
float modCoord = mod(float(dim), 2.);
|
|
return modCoord == 0. ? frag.r : frag.g;
|
|
}
|
|
`;function rD(){return`
|
|
int getOutputCoords() {
|
|
return 0;
|
|
}
|
|
`}function Q8(r,e){let t=[Math.ceil(e[0]/2),Math.ceil(e[1]/2)];return t[0]===1?`
|
|
int getOutputCoords() {
|
|
return 2 * int(resultUV.x * ${t[1]}.0);
|
|
}
|
|
`:t[1]===1?`
|
|
int getOutputCoords() {
|
|
return 2 * int(resultUV.y * ${t[0]}.0);
|
|
}
|
|
`:`
|
|
int getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${t[0]}, ${t[1]}));
|
|
return 2 * (resTexRC.x * ${t[1]} + resTexRC.y);
|
|
}
|
|
`}function eX(r,e){return e[0]===1?`
|
|
int getOutputCoords() {
|
|
return int(resultUV.x * ${e[1]}.0);
|
|
}
|
|
`:e[1]===1?`
|
|
int getOutputCoords() {
|
|
return int(resultUV.y * ${e[0]}.0);
|
|
}
|
|
`:`
|
|
int getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${e[0]}, ${e[1]}));
|
|
return resTexRC.x * ${e[1]} + resTexRC.y;
|
|
}
|
|
`}function tX(r,e){let t=[Math.ceil(e[0]/2),Math.ceil(e[1]/2)],n=Math.ceil(r[2]/2),o=n*Math.ceil(r[1]/2);return`
|
|
ivec3 getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${t[0]}, ${t[1]}));
|
|
int index = resTexRC.x * ${t[1]} + resTexRC.y;
|
|
|
|
int b = index / ${o};
|
|
index -= b * ${o};
|
|
|
|
int r = 2 * (index / ${n});
|
|
int c = imod(index, ${n}) * 2;
|
|
|
|
return ivec3(b, r, c);
|
|
}
|
|
`}function rX(r,e){let t=Is(["r","c","d"],r);return`
|
|
ivec3 getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${e[0]}, ${e[1]}));
|
|
int index = resTexRC.x * ${e[1]} + resTexRC.y;
|
|
${t}
|
|
return ivec3(r, c, d);
|
|
}
|
|
`}function nX(r,e){let t=[Math.ceil(e[0]/2),Math.ceil(e[1]/2)],n=Math.ceil(r[r.length-1]/2),o=n*Math.ceil(r[r.length-2]/2),s=o,a="",i="b, r, c";for(let l=2;l<r.length-1;l++)s*=r[r.length-l-1],a=`
|
|
int b${l} = index / ${s};
|
|
index -= b${l} * ${s};
|
|
`+a,i=`b${l}, `+i;return`
|
|
ivec${r.length} getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${t[0]}, ${t[1]}));
|
|
int index = resTexRC.x * ${t[1]} + resTexRC.y;
|
|
|
|
${a}
|
|
|
|
int b = index / ${o};
|
|
index -= b * ${o};
|
|
|
|
int r = 2 * (index / ${n});
|
|
int c = imod(index, ${n}) * 2;
|
|
|
|
return ivec${r.length}(${i});
|
|
}
|
|
`}function oX(r,e){let t=Is(["r","c","d","d2"],r);return`
|
|
ivec4 getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${e[0]}, ${e[1]}));
|
|
int index = resTexRC.x * ${e[1]} + resTexRC.y;
|
|
${t}
|
|
return ivec4(r, c, d, d2);
|
|
}
|
|
`}function sX(r,e){let t=Is(["r","c","d","d2","d3"],r);return`
|
|
ivec5 getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx * vec2(${e[0]},
|
|
${e[1]}));
|
|
|
|
int index = resTexRC.x * ${e[1]} + resTexRC.y;
|
|
|
|
${t}
|
|
|
|
ivec5 outShape = ivec5(r, c, d, d2, d3);
|
|
return outShape;
|
|
}
|
|
`}function iX(r,e){let t=Is(["r","c","d","d2","d3","d4"],r);return`
|
|
ivec6 getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${e[0]}, ${e[1]}));
|
|
int index = resTexRC.x * ${e[1]} + resTexRC.y;
|
|
|
|
${t}
|
|
|
|
ivec6 result = ivec6(r, c, d, d2, d3, d4);
|
|
return result;
|
|
}
|
|
`}function aX(r,e){let t=[Math.ceil(e[0]/2),Math.ceil(e[1]/2)];if(y.arraysEqual(r,e))return`
|
|
ivec2 getOutputCoords() {
|
|
return 2 * ivec2(resultUV.yx * vec2(${t[0]}, ${t[1]}));
|
|
}
|
|
`;let n=Math.ceil(r[1]/2);return`
|
|
ivec2 getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${t[0]}, ${t[1]}));
|
|
|
|
int index = resTexRC.x * ${t[1]} + resTexRC.y;
|
|
int r = 2 * (index / ${n});
|
|
int c = imod(index, ${n}) * 2;
|
|
|
|
return ivec2(r, c);
|
|
}
|
|
`}function lX(r,e){return y.arraysEqual(r,e)?`
|
|
ivec2 getOutputCoords() {
|
|
return ivec2(resultUV.yx * vec2(${e[0]}, ${e[1]}));
|
|
}
|
|
`:r[1]===1?`
|
|
ivec2 getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${e[0]}, ${e[1]}));
|
|
int index = resTexRC.x * ${e[1]} + resTexRC.y;
|
|
return ivec2(index, 0);
|
|
}
|
|
`:r[0]===1?`
|
|
ivec2 getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${e[0]}, ${e[1]}));
|
|
int index = resTexRC.x * ${e[1]} + resTexRC.y;
|
|
return ivec2(0, index);
|
|
}
|
|
`:`
|
|
ivec2 getOutputCoords() {
|
|
ivec2 resTexRC = ivec2(resultUV.yx *
|
|
vec2(${e[0]}, ${e[1]}));
|
|
int index = resTexRC.x * ${e[1]} + resTexRC.y;
|
|
int r = index / ${r[1]};
|
|
int c = index - r * ${r[1]};
|
|
return ivec2(r, c);
|
|
}
|
|
`}function Pu(r){return`offset${r}`}function uX(r){let e=r.name,t="get"+e.charAt(0).toUpperCase()+e.slice(1),n=Pt();return`
|
|
vec4 ${t}() {
|
|
return ${n.texture2D}(${e}, halfCR);
|
|
}
|
|
`}function cX(r){let e=r.name,t="get"+e.charAt(0).toUpperCase()+e.slice(1);if(r.shapeInfo.isUniform)return`float ${t}() {return ${e};}`;let[n,o]=r.shapeInfo.texShape;if(n===1&&o===1)return`
|
|
float ${t}() {
|
|
return sampleTexture(${e}, halfCR);
|
|
}
|
|
`;let[s,a]=r.shapeInfo.texShape,i=Pu(e);return`
|
|
float ${t}() {
|
|
vec2 uv = uvFromFlat(${s}, ${a}, ${i});
|
|
return sampleTexture(${e}, uv);
|
|
}
|
|
`}function pX(r){let e=r.name,t="get"+e.charAt(0).toUpperCase()+e.slice(1),n=r.shapeInfo.texShape,o=[Math.ceil(n[0]/2),Math.ceil(n[1]/2)],s=Pt();return`
|
|
vec4 ${t}(int index) {
|
|
vec2 uv = packedUVfrom1D(
|
|
${o[0]}, ${o[1]}, index);
|
|
return ${s.texture2D}(${e}, uv);
|
|
}
|
|
`}function mX(r){let e=r.name,t="get"+e.charAt(0).toUpperCase()+e.slice(1);if(r.shapeInfo.isUniform)return`
|
|
float ${t}(int index) {
|
|
${Ep(r)}
|
|
}
|
|
`;let n=r.shapeInfo.texShape,o=n[0],s=n[1];if(s===1&&o===1)return`
|
|
float ${t}(int index) {
|
|
return sampleTexture(${e}, halfCR);
|
|
}
|
|
`;let a=Pu(e);return s===1?`
|
|
float ${t}(int index) {
|
|
vec2 uv = vec2(0.5, (float(index + ${a}) + 0.5) / ${o}.0);
|
|
return sampleTexture(${e}, uv);
|
|
}
|
|
`:o===1?`
|
|
float ${t}(int index) {
|
|
vec2 uv = vec2((float(index + ${a}) + 0.5) / ${s}.0, 0.5);
|
|
return sampleTexture(${e}, uv);
|
|
}
|
|
`:`
|
|
float ${t}(int index) {
|
|
vec2 uv = uvFromFlat(${o}, ${s}, index + ${a});
|
|
return sampleTexture(${e}, uv);
|
|
}
|
|
`}function fX(r){let e=r.shapeInfo.logicalShape,t=r.name,n="get"+t.charAt(0).toUpperCase()+t.slice(1),o=r.shapeInfo.texShape,s=o[0],a=o[1],i=Pt();if(o!=null&&y.arraysEqual(e,o))return`
|
|
vec4 ${n}(int row, int col) {
|
|
vec2 uv = (vec2(col, row) + halfCR) / vec2(${a}.0, ${s}.0);
|
|
|
|
return ${i.texture2D}(${t}, uv);
|
|
}
|
|
`;let l=[Math.ceil(o[0]/2),Math.ceil(o[1]/2)],u=Math.ceil(e[1]/2);return`
|
|
vec4 ${n}(int row, int col) {
|
|
vec2 uv = packedUVfrom2D(${u}, ${l[0]}, ${l[1]}, row, col);
|
|
return ${i.texture2D}(${t}, uv);
|
|
}
|
|
`}function dX(r){let e=r.shapeInfo.logicalShape,t=r.name,n="get"+t.charAt(0).toUpperCase()+t.slice(1),o=r.shapeInfo.texShape;if(o!=null&&y.arraysEqual(e,o)){let p=o[0],m=o[1];return`
|
|
float ${n}(int row, int col) {
|
|
vec2 uv = (vec2(col, row) + halfCR) / vec2(${m}.0, ${p}.0);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`}let{newShape:s,keptDims:a}=y.squeezeShape(e),i=s;if(i.length<e.length){let p=Dp(r,i),m=["row","col"];return`
|
|
${Ap(p)}
|
|
float ${n}(int row, int col) {
|
|
return ${n}(${$p(m,a)});
|
|
}
|
|
`}if(r.shapeInfo.isUniform)return`
|
|
float ${n}(int row, int col) {
|
|
int index = round(dot(vec2(row, col), vec2(${e[1]}, 1)));
|
|
${Ep(r)}
|
|
}
|
|
`;let l=o[0],u=o[1],c=Pu(t);return u===1?`
|
|
float ${n}(int row, int col) {
|
|
float index = dot(vec3(row, col, ${c}), vec3(${e[1]}, 1, 1));
|
|
vec2 uv = vec2(0.5, (index + 0.5) / ${l}.0);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`:l===1?`
|
|
float ${n}(int row, int col) {
|
|
float index = dot(vec3(row, col, ${c}), vec3(${e[1]}, 1, 1));
|
|
vec2 uv = vec2((index + 0.5) / ${u}.0, 0.5);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`:`
|
|
float ${n}(int row, int col) {
|
|
// Explicitly use integer operations as dot() only works on floats.
|
|
int index = row * ${e[1]} + col + ${c};
|
|
vec2 uv = uvFromFlat(${l}, ${u}, index);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`}function hX(r){let e=r.shapeInfo.logicalShape,t=r.name,n="get"+t.charAt(0).toUpperCase()+t.slice(1),o=r.shapeInfo.texShape,s=[Math.ceil(o[0]/2),Math.ceil(o[1]/2)];if(e[0]===1){let p=e.slice(1),m=[1,2],f=Dp(r,p),d=["b","row","col"];return`
|
|
${tD(f)}
|
|
vec4 ${n}(int b, int row, int col) {
|
|
return ${n}(${$p(d,m)});
|
|
}
|
|
`}let a=s[0],i=s[1],l=Math.ceil(e[2]/2),u=l*Math.ceil(e[1]/2),c=Pt();return`
|
|
vec4 ${n}(int b, int row, int col) {
|
|
vec2 uv = packedUVfrom3D(
|
|
${a}, ${i}, ${u}, ${l}, b, row, col);
|
|
return ${c.texture2D}(${t}, uv);
|
|
}
|
|
`}function gX(r){let e=r.shapeInfo.logicalShape,t=r.name,n="get"+t.charAt(0).toUpperCase()+t.slice(1),o=e[1]*e[2],s=e[2],{newShape:a,keptDims:i}=y.squeezeShape(e),l=a;if(l.length<e.length){let d=Dp(r,l),h=["row","col","depth"];return`
|
|
${Ap(d)}
|
|
float ${n}(int row, int col, int depth) {
|
|
return ${n}(${$p(h,i)});
|
|
}
|
|
`}if(r.shapeInfo.isUniform)return`
|
|
float ${n}(int row, int col, int depth) {
|
|
int index = round(dot(vec3(row, col, depth),
|
|
vec3(${o}, ${s}, 1)));
|
|
${Ep(r)}
|
|
}
|
|
`;let u=r.shapeInfo.texShape,c=u[0],p=u[1],m=r.shapeInfo.flatOffset;if(p===o&&m==null)return`
|
|
float ${n}(int row, int col, int depth) {
|
|
float texR = float(row);
|
|
float texC = dot(vec2(col, depth), vec2(${s}, 1));
|
|
vec2 uv = (vec2(texC, texR) + halfCR) /
|
|
vec2(${p}.0, ${c}.0);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`;if(p===s&&m==null)return`
|
|
float ${n}(int row, int col, int depth) {
|
|
float texR = dot(vec2(row, col), vec2(${e[1]}, 1));
|
|
float texC = float(depth);
|
|
vec2 uv = (vec2(texC, texR) + halfCR) / vec2(${p}.0, ${c}.0);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`;let f=Pu(t);return`
|
|
float ${n}(int row, int col, int depth) {
|
|
// Explicitly use integer operations as dot() only works on floats.
|
|
int index = row * ${o} + col * ${s} + depth + ${f};
|
|
vec2 uv = uvFromFlat(${c}, ${p}, index);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`}function xX(r){let e=r.shapeInfo.logicalShape,t=e.length,n=r.name,o="get"+n.charAt(0).toUpperCase()+n.slice(1),s=r.shapeInfo.texShape,a=[Math.ceil(s[0]/2),Math.ceil(s[1]/2)],i=a[0],l=a[1],u=Math.ceil(e[t-1]/2),c=u*Math.ceil(e[t-2]/2),p="int b, int row, int col",m=`b * ${c} + (row / 2) * ${u} + (col / 2)`;for(let d=2;d<t-1;d++)p=`int b${d}, `+p,c*=e[t-d-1],m=`b${d} * ${c} + `+m;let f=Pt();return`
|
|
vec4 ${o}(${p}) {
|
|
int index = ${m};
|
|
int texR = index / ${l};
|
|
int texC = index - texR * ${l};
|
|
vec2 uv = (vec2(texC, texR) + halfCR) / vec2(${l}, ${i});
|
|
return ${f.texture2D}(${n}, uv);
|
|
}
|
|
`}function yX(r){let e=r.shapeInfo.logicalShape,t=r.name,n="get"+t.charAt(0).toUpperCase()+t.slice(1),o=e[3],s=e[2]*o,a=e[1]*s,{newShape:i,keptDims:l}=y.squeezeShape(e);if(i.length<e.length){let d=Dp(r,i),h=["row","col","depth","depth2"];return`
|
|
${Ap(d)}
|
|
float ${n}(int row, int col, int depth, int depth2) {
|
|
return ${n}(${$p(h,l)});
|
|
}
|
|
`}if(r.shapeInfo.isUniform)return`
|
|
float ${n}(int row, int col, int depth, int depth2) {
|
|
int index = round(dot(vec4(row, col, depth, depth2),
|
|
vec4(${a}, ${s}, ${o}, 1)));
|
|
${Ep(r)}
|
|
}
|
|
`;let u=r.shapeInfo.flatOffset,c=r.shapeInfo.texShape,p=c[0],m=c[1];if(m===a&&u==null)return`
|
|
float ${n}(int row, int col, int depth, int depth2) {
|
|
float texR = float(row);
|
|
float texC =
|
|
dot(vec3(col, depth, depth2),
|
|
vec3(${s}, ${o}, 1));
|
|
vec2 uv = (vec2(texC, texR) + halfCR) /
|
|
vec2(${m}.0, ${p}.0);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`;if(m===o&&u==null)return`
|
|
float ${n}(int row, int col, int depth, int depth2) {
|
|
float texR = dot(vec3(row, col, depth),
|
|
vec3(${e[1]*e[2]}, ${e[2]}, 1));
|
|
float texC = float(depth2);
|
|
vec2 uv = (vec2(texC, texR) + halfCR) /
|
|
vec2(${m}.0, ${p}.0);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`;let f=Pu(t);return`
|
|
float ${n}(int row, int col, int depth, int depth2) {
|
|
// Explicitly use integer operations as dot() only works on floats.
|
|
int index = row * ${a} + col * ${s} +
|
|
depth * ${o} + depth2;
|
|
vec2 uv = uvFromFlat(${p}, ${m}, index + ${f});
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`}function bX(r){let e=r.shapeInfo.logicalShape,t=r.name,n="get"+t.charAt(0).toUpperCase()+t.slice(1),o=e[4],s=e[3]*o,a=e[2]*s,i=e[1]*a,{newShape:l,keptDims:u}=y.squeezeShape(e);if(l.length<e.length){let h=Dp(r,l),g=["row","col","depth","depth2","depth3"];return`
|
|
${Ap(h)}
|
|
float ${n}(int row, int col, int depth, int depth2, int depth3) {
|
|
return ${n}(${$p(g,u)});
|
|
}
|
|
`}if(r.shapeInfo.isUniform)return`
|
|
float ${n}(int row, int col, int depth, int depth2, int depth3) {
|
|
float index = dot(
|
|
vec4(row, col, depth, depth2),
|
|
vec4(${i}, ${a}, ${s}, ${o})) +
|
|
depth3;
|
|
${Ep(r)}
|
|
}
|
|
`;let c=r.shapeInfo.flatOffset,p=r.shapeInfo.texShape,m=p[0],f=p[1];if(f===i&&c==null)return`
|
|
float ${n}(int row, int col, int depth, int depth2, int depth3) {
|
|
int texR = row;
|
|
float texC = dot(vec4(col, depth, depth2, depth3),
|
|
vec4(${a}, ${s}, ${o}, 1));
|
|
vec2 uv = (vec2(texC, texR) + halfCR) /
|
|
vec2(${f}.0, ${m}.0);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`;if(f===o&&c==null)return`
|
|
float ${n}(int row, int col, int depth, int depth2, int depth3) {
|
|
float texR = dot(
|
|
vec4(row, col, depth, depth2),
|
|
vec4(${e[1]*e[2]*e[3]},
|
|
${e[2]*e[3]}, ${e[3]}, 1));
|
|
int texC = depth3;
|
|
vec2 uv = (vec2(texC, texR) + halfCR) /
|
|
vec2(${f}.0, ${m}.0);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`;let d=Pu(t);return`
|
|
float ${n}(int row, int col, int depth, int depth2, int depth3) {
|
|
// Explicitly use integer operations as dot() only works on floats.
|
|
int index = row * ${i} + col * ${a} + depth * ${s} +
|
|
depth2 * ${o} + depth3 + ${d};
|
|
vec2 uv = uvFromFlat(${m}, ${f}, index);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`}function wX(r){let e=r.shapeInfo.logicalShape,t=r.name,n="get"+t.charAt(0).toUpperCase()+t.slice(1),{newShape:o,keptDims:s}=y.squeezeShape(e);if(o.length<e.length){let g=Dp(r,o),x=["row","col","depth","depth2","depth3","depth4"];return`
|
|
${Ap(g)}
|
|
float ${n}(int row, int col, int depth,
|
|
int depth2, int depth3, int depth4) {
|
|
return ${n}(${$p(x,s)});
|
|
}
|
|
`}let a=e[5],i=e[4]*a,l=e[3]*i,u=e[2]*l,c=e[1]*u;if(r.shapeInfo.isUniform)return`
|
|
float ${n}(int row, int col, int depth,
|
|
int depth2, int depth3, int depth4) {
|
|
int index = round(dot(
|
|
vec4(row, col, depth, depth2),
|
|
vec4(${c}, ${u}, ${l}, ${i})) +
|
|
dot(
|
|
vec2(depth3, depth4),
|
|
vec2(${a}, 1)));
|
|
${Ep(r)}
|
|
}
|
|
`;let p=r.shapeInfo.flatOffset,m=r.shapeInfo.texShape,f=m[0],d=m[1];if(d===c&&p==null)return`
|
|
float ${n}(int row, int col, int depth,
|
|
int depth2, int depth3, int depth4) {
|
|
int texR = row;
|
|
float texC = dot(vec4(col, depth, depth2, depth3),
|
|
vec4(${u}, ${l}, ${i}, ${a})) +
|
|
float(depth4);
|
|
vec2 uv = (vec2(texC, texR) + halfCR) /
|
|
vec2(${d}.0, ${f}.0);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`;if(d===a&&p==null)return`
|
|
float ${n}(int row, int col, int depth,
|
|
int depth2, int depth3, int depth4) {
|
|
float texR = dot(vec4(row, col, depth, depth2),
|
|
vec4(${e[1]*e[2]*e[3]*e[4]},
|
|
${e[2]*e[3]*e[4]},
|
|
${e[3]*e[4]},
|
|
${e[4]})) + float(depth3);
|
|
int texC = depth4;
|
|
vec2 uv = (vec2(texC, texR) + halfCR) /
|
|
vec2(${d}.0, ${f}.0);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`;let h=Pu(t);return`
|
|
float ${n}(int row, int col, int depth,
|
|
int depth2, int depth3, int depth4) {
|
|
// Explicitly use integer operations as dot() only works on floats.
|
|
int index = row * ${c} + col * ${u} + depth * ${l} +
|
|
depth2 * ${i} + depth3 * ${a} + depth4 + ${h};
|
|
vec2 uv = uvFromFlat(${f}, ${d}, index);
|
|
return sampleTexture(${t}, uv);
|
|
}
|
|
`}function Ep(r){let e=r.name,t=y.sizeFromShape(r.shapeInfo.logicalShape);return t<2?`return ${e};`:`
|
|
for (int i = 0; i < ${t}; i++) {
|
|
if (i == index) {
|
|
return ${e}[i];
|
|
}
|
|
}
|
|
`}function _X(r,e){let t=r.name,n=t.charAt(0).toUpperCase()+t.slice(1),o="get"+n+"AtOutCoords",s=r.shapeInfo.logicalShape.length,a=e.logicalShape.length,i=Q2(r.shapeInfo.logicalShape,e.logicalShape),l=Ve(a),u=a-s,c,p=["x","y","z","w","u","v"];s===0?c="":a<2&&i.length>=1?c="coords = 0;":c=i.map(b=>`coords.${p[b+u]} = 0;`).join(`
|
|
`);let m="";a<2&&s>0?m="coords":m=r.shapeInfo.logicalShape.map((b,w)=>`coords.${p[w+u]}`).join(", ");let f="return outputValue;",h=y.sizeFromShape(r.shapeInfo.logicalShape)===1,x=y.sizeFromShape(e.logicalShape)===1;if(s===1&&!h&&!x)f=`
|
|
return vec4(outputValue.xy, outputValue.xy);
|
|
`;else if(h&&!x)a===1?f=`
|
|
return vec4(outputValue.x, outputValue.x, 0., 0.);
|
|
`:f=`
|
|
return vec4(outputValue.x);
|
|
`;else if(i.length){let b=s-2,w=s-1;i.indexOf(b)>-1&&i.indexOf(w)>-1?f="return vec4(outputValue.x);":i.indexOf(b)>-1?f="return vec4(outputValue.x, outputValue.y, outputValue.x, outputValue.y);":i.indexOf(w)>-1&&(f="return vec4(outputValue.xx, outputValue.zz);")}return`
|
|
vec4 ${o}() {
|
|
${l} coords = getOutputCoords();
|
|
${c}
|
|
vec4 outputValue = get${n}(${m});
|
|
${f}
|
|
}
|
|
`}function kX(r,e){let t=r.name,n=t.charAt(0).toUpperCase()+t.slice(1),o="get"+n+"AtOutCoords",s=e.texShape,a=r.shapeInfo.texShape,i=r.shapeInfo.logicalShape.length,l=e.logicalShape.length;if(!r.shapeInfo.isUniform&&i===l&&r.shapeInfo.flatOffset==null&&y.arraysEqual(a,s))return`
|
|
float ${o}() {
|
|
return sampleTexture(${t}, resultUV);
|
|
}
|
|
`;let u=Ve(l),c=Q2(r.shapeInfo.logicalShape,e.logicalShape),p=l-i,m,f=["x","y","z","w","u","v"];i===0?m="":l<2&&c.length>=1?m="coords = 0;":m=c.map(h=>`coords.${f[h+p]} = 0;`).join(`
|
|
`);let d="";return l<2&&i>0?d="coords":d=r.shapeInfo.logicalShape.map((h,g)=>`coords.${f[g+p]}`).join(", "),`
|
|
float ${o}() {
|
|
${u} coords = getOutputCoords();
|
|
${m}
|
|
return get${n}(${d});
|
|
}
|
|
`}function Ve(r){if(r<=1)return"int";if(r===2)return"ivec2";if(r===3)return"ivec3";if(r===4)return"ivec4";if(r===5)return"ivec5";if(r===6)return"ivec6";throw Error(`GPU for rank ${r} is not yet supported`)}function Dp(r,e){let t=JSON.parse(JSON.stringify(r));return t.shapeInfo.logicalShape=e,t}function $p(r,e){return e.map(t=>r[t]).join(", ")}function nD(r,e,t,n){let o=e.userCode,s=t.map((f,d)=>{let h={logicalShape:f.shape,texShape:f.isUniform?null:f.texData.texShape,isUniform:f.isUniform,isPacked:f.isUniform?!1:f.texData.isPacked,flatOffset:null};return f.texData!=null&&f.texData.slice!=null&&f.texData.slice.flatOffset>0&&(h.flatOffset=f.texData.slice.flatOffset),{name:e.variableNames[d],shapeInfo:h}}),a=s.map(f=>f.shapeInfo),i={logicalShape:n.shape,texShape:n.texData.texShape,isUniform:!1,isPacked:n.texData.isPacked,flatOffset:null},l=eD(s,i,o,e.packedInputs),u=r.createProgram(l),c=null,p=r.getUniformLocation(u,"NAN",!1);W().getNumber("WEBGL_VERSION")===1&&(c=r.getUniformLocation(u,"INFINITY",!1));let m={};for(let f=0;f<e.variableNames.length;f++){let d=e.variableNames[f],h=!1;m[d]=r.getUniformLocation(u,d,h),m[`offset${d}`]=r.getUniformLocation(u,`offset${d}`,h)}return{program:e,source:l,webGLProgram:u,uniformLocations:m,inShapeInfos:a,outShapeInfo:i,infLoc:c,nanLoc:p}}function oD(r,e){if(r.length!==e.length)throw Error(`Binary was compiled with ${r.length} inputs, but was executed with ${e.length} inputs`);r.forEach((t,n)=>{let o=t.logicalShape,s=e[n],a=s.shape;if(!y.arraysEqual(o,a))throw Error(`Binary was compiled with different shapes than the current args. Shapes ${o} and ${a} must match`);if(t.isUniform&&s.isUniform)return;let i=t.texShape,l=s.isUniform?null:s.texData.texShape;if(!y.arraysEqual(i,l))throw Error(`Binary was compiled with different texture shapes than the current args. Shape ${i} and ${l} must match`)})}function sD(r,e,t,n,o){oD(e.inShapeInfos,t),oD([e.outShapeInfo],[n]);let s=n.texData.texture,a=n.texData.texShape;n.texData.isPacked?r.setOutputPackedMatrixTexture(s,a[0],a[1]):r.setOutputMatrixTexture(s,a[0],a[1]),r.setProgram(e.webGLProgram),W().getNumber("WEBGL_VERSION")===1&&e.infLoc!==null&&r.gl.uniform1f(e.infLoc,Infinity),e.nanLoc!==null&&r.gl.uniform1f(e.nanLoc,NaN),t.forEach((i,l)=>{let u=e.program.variableNames[l],c=e.uniformLocations[u],p=e.uniformLocations[`offset${u}`];if(c!=null){if(i.isUniform){if(y.sizeFromShape(i.shape)<2)r.gl.uniform1f(c,i.uniformValues[0]);else{let m=i.uniformValues;m instanceof Float32Array||(m=new Float32Array(m)),r.gl.uniform1fv(c,m)}return}i.texData.slice!=null&&p!=null&&r.gl.uniform1i(p,i.texData.slice.flatOffset),r.setInputMatrixTexture(i.texData.texture,c,l)}}),o!=null&&o(r,e.webGLProgram),r.executeProgram()}function iD(r,e,t){let n="";e.concat(t).forEach(a=>{let i=a.texData!=null&&a.texData.slice!=null&&a.texData.slice.flatOffset>0,l=a.isUniform?"uniform":a.texData.texShape;n+=`${a.shape}_${l}_${i}`});let o=r.userCode,s=r.constructor.name;return s+="_"+n+"_"+o,s}var{addImpl:aD,bincountImpl:mx,bincountReduceImpl:lD,ceilImpl:uD,concatImpl:cD,expImpl:pD,expm1Impl:mD,floorImpl:fD,gatherV2Impl:dD,greaterImpl:hD,lessImpl:gD,linSpaceImpl:xD,logImpl:yD,maxImpl:bD,maximumImpl:wD,minimumImpl:_D,multiplyImpl:kD,negImpl:vD,prodImpl:CD,rangeImpl:ID,rsqrtImpl:ND,simpleAbsImpl:fx,sliceImpl:SD,sparseFillEmptyRowsImpl:TD,sparseReshapeImpl:AD,stridedSliceImpl:ED,subImpl:DD,tileImpl:$D,topKImpl:RD,transposeImpl:Mu,uniqueImpl:FD}=qg;function tv(r,e){return["x","y","z","w","u","v"].slice(0,e).map(t=>`${r}.${t}`)}function jt(r,e){return e===1?[r]:tv(r,e)}function OD(r,e){if(r===1)return"rc";let t="";for(let n=0;n<r;n++)t+=e[n],n<r-1&&(t+=",");return t}var rv=class{constructor(e){this.variableNames=["A"],this.packedInputs=!1,this.packedOutput=!0,this.outputShape=e;let t=e.length;if(t===0)this.userCode=`
|
|
void main() {
|
|
setOutput(vec4(getA(), 0., 0., 0.));
|
|
}
|
|
`;else{let n=jt("rc",t),o=Ve(t),s=CX(t,e,n),a=IX(t,e[e.length-1],e[e.length-2],n),i=NX(e,n);this.userCode=`
|
|
void main() {
|
|
${o} rc = getOutputCoords();
|
|
|
|
if(${s}) {
|
|
setOutput(vec4(0));
|
|
} else {
|
|
${a}
|
|
|
|
setOutput(vec4(${i}));
|
|
}
|
|
}
|
|
`}}};function vX(r,e){let t=[];for(let n=0;n<=1;n++)for(let o=0;o<=1;o++){let s=`${n===0?"r":"rp1"}, ${o===0?"c":"cp1"}`;for(let a=2;a<r;a++)s=`${e[e.length-1-a]},`+s;t.push(s)}return t}function CX(r,e,t){if(r===1)return`rc > ${e[0]}`;let n="";for(let o=r-2;o<r;o++)n+=`${t[o]} >= ${e[o]}`,o<r-1&&(n+="||");return n}function IX(r,e,t,n){if(r===1)return"";let o=n.slice(-2);return`
|
|
int r = ${o[0]};
|
|
int c = ${o[1]};
|
|
int rp1 = r + 1;
|
|
int cp1 = c + 1;
|
|
|
|
bool cEdge = cp1 >= ${e};
|
|
bool rEdge = rp1 >= ${t};
|
|
`}function NX(r,e){let t=r.length,n=vX(t,e);return t===1?`getA(rc),
|
|
rc + 1 >= ${r[0]} ? 0. : getA(rc + 1),
|
|
0, 0`:`getA(${n[0]}),
|
|
cEdge ? 0. : getA(${n[1]}),
|
|
rEdge ? 0. : getA(${n[2]}),
|
|
rEdge || cEdge ? 0. : getA(${n[3]})`}var Pf=class{constructor(e,t){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=e;let n="";for(let o=0;o<4;o++){let s="thisRC = rc;";o%2==1&&(s+="thisRC.z += 1;"),o>1&&(s+="thisRC.y += 1;"),n+=`
|
|
${s}
|
|
${o>0?"if(thisRC.y < rows && thisRC.z < cols){":""}
|
|
int flatIndex = getFlatIndex(thisRC);
|
|
|
|
ivec3 inputRC = inputCoordsFromReshapedOutCoords(flatIndex);
|
|
vec2 inputRCInnerDims = vec2(float(inputRC.y),float(inputRC.z));
|
|
|
|
result[${o}] =
|
|
getChannel(getA(inputRC.x, inputRC.y, inputRC.z), inputRCInnerDims);
|
|
${o>0?"}":""}
|
|
`}this.userCode=`
|
|
${SX(t)}
|
|
${Tp(e)}
|
|
|
|
void main() {
|
|
ivec3 rc = getOutputCoords();
|
|
|
|
vec4 result = vec4(0.);
|
|
|
|
ivec3 thisRC;
|
|
int rows = ${e[1]};
|
|
int cols = ${e[2]};
|
|
|
|
${n}
|
|
|
|
setOutput(result);
|
|
}
|
|
`}};function SX(r){return`
|
|
ivec3 inputCoordsFromReshapedOutCoords(int index) {
|
|
${Is(["r","c","d"],r)}
|
|
return ivec3(r, c, d);
|
|
}
|
|
`}var nv=class{constructor(e){this.gpgpu=e,this.numUsedTextures=0,this.numFreeTextures=0,this._numBytesAllocated=0,this._numBytesFree=0,this.freeTextures={},this.logEnabled=!1,this.usedTextures={}}acquireTexture(e,t,n){let o=MD(t,n),s=LD(e,o,n);s in this.freeTextures||(this.freeTextures[s]=[]),s in this.usedTextures||(this.usedTextures[s]=[]);let a=PD(e,o,this.gpgpu.gl,this.gpgpu.textureConfig,n);if(this.freeTextures[s].length>0){this.numFreeTextures--,this.numUsedTextures++,this._numBytesFree-=a,this.log();let l=this.freeTextures[s].shift();return this.usedTextures[s].push(l),l}let i;return o===Ir.PACKED_2X2_FLOAT32?i=this.gpgpu.createPackedMatrixTexture(e[0],e[1]):o===Ir.PACKED_2X2_FLOAT16?i=this.gpgpu.createFloat16PackedMatrixTexture(e[0],e[1]):o===Ir.UNPACKED_FLOAT32?i=this.gpgpu.createFloat32MatrixTexture(e[0],e[1]):o===Ir.UNPACKED_FLOAT16?i=this.gpgpu.createFloat16MatrixTexture(e[0],e[1]):o===Ir.PACKED_4X1_UNSIGNED_BYTE&&(i=this.gpgpu.createUnsignedBytesMatrixTexture(e[0],e[1])),this.usedTextures[s].push(i),this.numUsedTextures++,this._numBytesAllocated+=a,this.log(),i}releaseTexture(e,t,n,o){if(this.freeTextures==null)return;let s=MD(n,o),a=LD(t,s,o);a in this.freeTextures||(this.freeTextures[a]=[]);let i=PD(t,s,this.gpgpu.gl,this.gpgpu.textureConfig,o),l=W().get("WEBGL_DELETE_TEXTURE_THRESHOLD");l!==-1&&this._numBytesAllocated>l?(this.gpgpu.deleteMatrixTexture(e),this._numBytesAllocated-=i):(this.freeTextures[a].push(e),this.numFreeTextures++,this._numBytesFree+=i),this.numUsedTextures--;let u=this.usedTextures[a],c=u.indexOf(e);if(c<0)throw new Error("Cannot release a texture that was never provided by this texture manager");u.splice(c,1),this.log()}log(){if(!this.logEnabled)return;let e=this.numFreeTextures+this.numUsedTextures;console.log("Free/Used",`${this.numFreeTextures} / ${this.numUsedTextures}`,`(${e})`);let t=this._numBytesFree/this._numBytesAllocated;console.log(`Bytes allocated: ${this._numBytesAllocated}`),console.log(`Bytes unused: ${this._numBytesFree} (${Math.round(100*t)}%)`)}get numBytesAllocated(){return this._numBytesAllocated}get numBytesFree(){return this._numBytesFree}getNumUsedTextures(){return this.numUsedTextures}getNumFreeTextures(){return this.numFreeTextures}dispose(){if(this.freeTextures!=null){for(let e in this.freeTextures)this.freeTextures[e].forEach(t=>{this.gpgpu.deleteMatrixTexture(t)});for(let e in this.usedTextures)this.usedTextures[e].forEach(t=>{this.gpgpu.deleteMatrixTexture(t)});this.freeTextures=null,this.usedTextures=null,this.numUsedTextures=0,this.numFreeTextures=0,this._numBytesAllocated=0,this._numBytesFree=0}}};function TX(r,e){let t=r;if(e===t.R32F)return 4;if(e===t.R16F)return 2;if(e===t.RGBA32F)return 16;if(e===r.RGBA)return 16;if(e===t.RGBA16F)return 8;throw new Error(`Unknown internal format ${e}`)}function PD(r,e,t,n,o){let s=AX(e,n),a;if(o){let[l,u]=ui(r[0],r[1]);a=l*u}else{let[l,u]=Ou(r[0],r[1]);a=l*u}let i=TX(t,s);return a*i}function AX(r,e){switch(r){case Ir.PACKED_2X2_FLOAT32:return ux(e);case Ir.PACKED_2X2_FLOAT16:return cx(e);case Ir.UNPACKED_FLOAT32:return ix(e);case Ir.UNPACKED_FLOAT16:return ax(e);case Ir.PACKED_4X1_UNSIGNED_BYTE:return lx(e);default:throw new Error(`Unknown physical texture type ${r}`)}}function EX(r){return W().getBool("WEBGL_RENDER_FLOAT32_ENABLED")?r?Ir.PACKED_2X2_FLOAT32:Ir.UNPACKED_FLOAT32:r?Ir.PACKED_2X2_FLOAT16:Ir.UNPACKED_FLOAT16}function MD(r,e){if(r===Pr.UPLOAD)return Ir.PACKED_2X2_FLOAT32;if(r===Pr.RENDER||r==null)return EX(e);if(r===Pr.DOWNLOAD||r===Pr.PIXELS)return Ir.PACKED_4X1_UNSIGNED_BYTE;throw new Error(`Unknown logical texture type ${r}`)}function LD(r,e,t){return`${r[0]}_${r[1]}_${e}_${t}`}var xn=class{constructor(e,t){this.variableNames=["A"],this.outputShape=e,this.userCode=`
|
|
float unaryOperation(float x) {
|
|
${t}
|
|
}
|
|
|
|
void main() {
|
|
float x = getAAtOutCoords();
|
|
float y = unaryOperation(x);
|
|
|
|
setOutput(y);
|
|
}
|
|
`}},br="if (isnan(x)) return x;",zD="return x;",ov="return abs(x);";var BD="return (x >= 0.0) ? x : (exp(x) - 1.0);",VD=br+`
|
|
return (x < 0.0) ? 0.0 : x;
|
|
`,GD=br+`
|
|
return (x < 0.0) ? 0.0 : min(6.0, x);
|
|
`,Mf="return x;",WD="return 1.0 / (1.0 + exp(-1.0 * x));";var jD="return x;",UD=`
|
|
vec4 result;
|
|
|
|
result.r = (x.r >= 0.0) ? x.r : (exp(x.r) - 1.0);
|
|
result.g = (x.g >= 0.0) ? x.g : (exp(x.g) - 1.0);
|
|
result.b = (x.b >= 0.0) ? x.b : (exp(x.b) - 1.0);
|
|
result.a = (x.a >= 0.0) ? x.a : (exp(x.a) - 1.0);
|
|
|
|
return result;
|
|
`,HD=`
|
|
vec4 result = x * vec4(greaterThanEqual(x, vec4(0.0)));
|
|
bvec4 isNaN = isnan(x);
|
|
|
|
result.r = isNaN.r ? x.r : result.r;
|
|
result.g = isNaN.g ? x.g : result.g;
|
|
result.b = isNaN.b ? x.b : result.b;
|
|
result.a = isNaN.a ? x.a : result.a;
|
|
|
|
return result;
|
|
`,qD=`
|
|
vec4 result = min(x, vec4(6.)) * vec4(greaterThanEqual(x, vec4(0.0)));
|
|
bvec4 isNaN = isnan(x);
|
|
|
|
result.r = isNaN.r ? x.r : result.r;
|
|
result.g = isNaN.g ? x.g : result.g;
|
|
result.b = isNaN.b ? x.b : result.b;
|
|
result.a = isNaN.a ? x.a : result.a;
|
|
|
|
return result;
|
|
`,KD="return 1.0 / (1.0 + exp(-1.0 * x));",Ns=class{constructor(e,t){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=e,this.userCode=`
|
|
vec4 unaryOperation(vec4 x) {
|
|
${t}
|
|
}
|
|
|
|
void main() {
|
|
vec4 x = getAAtOutCoords();
|
|
vec4 y = unaryOperation(x);
|
|
|
|
setOutput(y);
|
|
}
|
|
`}};var sv=class{constructor(e){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!1,this.outputShape=e;let t=e.length,n=jt("rc",t),o=Ve(t),s=OD(t,n),a=n.slice(-2),i=t<=1?"rc":`vec2(${a.join(",")})`;this.userCode=`
|
|
void main() {
|
|
${o} rc = getOutputCoords();
|
|
vec4 packedInput = getA(${s});
|
|
|
|
setOutput(getChannel(packedInput, ${i}));
|
|
}
|
|
`}};var DX=Fr.whereImpl,$X=1e-7,RX=1e-4,dx={};function FX(r){return r in dx||(dx[r]={}),dx[r]}var OX=128,PX=600;function MX(){return W().global.screen==null?1024:W().global.screen.height*W().global.screen.width*window.devicePixelRatio*PX/1024/1024}var Lu=class extends Ms{constructor(e){super();if(this.pendingRead=new WeakMap,this.pendingDisposal=new WeakSet,this.dataRefCount=new WeakMap,this.numBytesInGPU=0,this.uploadWaitMs=0,this.downloadWaitMs=0,this.lastGlFlushTime=0,this.warnedAboutMemory=!1,this.pendingDeletes=0,this.disposed=!1,!W().getBool("HAS_WEBGL"))throw new Error("WebGL is not supported on this device");if(e==null){let t=Vn(W().getNumber("WEBGL_VERSION"));this.binaryCache=FX(W().getNumber("WEBGL_VERSION")),this.gpgpu=new px(t),this.canvas=t.canvas,this.gpgpuCreatedLocally=!0}else this.gpgpu=e,this.binaryCache={},this.gpgpuCreatedLocally=!1,this.canvas=e.gl.canvas;this.textureManager=new nv(this.gpgpu),this.numMBBeforeWarning=MX(),this.texData=new Xa(this,hs())}nextDataId(){return Lu.nextDataId++}numDataIds(){return this.texData.numDataIds()-this.pendingDeletes}write(e,t,n){if((W().getBool("WEBGL_CHECK_NUMERICAL_PROBLEMS")||W().getBool("DEBUG"))&&this.checkNumericalProblems(e),n==="complex64"&&e!=null)throw new Error("Cannot write to a complex64 dtype. Please use tf.complex(real, imag).");let o={id:this.nextDataId()};return this.texData.set(o,{shape:t,dtype:n,values:e,usage:Pr.UPLOAD,refCount:1}),o}refCount(e){return this.texData.has(e)?this.texData.get(e).refCount:0}incRef(e){let t=this.texData.get(e);t.refCount++}decRef(e){if(this.texData.has(e)){let t=this.texData.get(e);t.refCount--}}move(e,t,n,o,s){if(W().getBool("DEBUG")&&this.checkNumericalProblems(t),o==="complex64")throw new Error("Cannot write to a complex64 dtype. Please use tf.complex(real, imag).");this.texData.set(e,{shape:n,dtype:o,values:t,usage:Pr.UPLOAD,refCount:s})}disposeIntermediateTensorInfo(e){this.disposeData(e.dataId)}readSync(e){let t=this.texData.get(e),{values:n,dtype:o,complexTensorInfos:s,slice:a,shape:i,isPacked:l}=t;if(a!=null){let m;l?m=new Ns(i,Mf):m=new xn(i,Mf);let f=this.runWebGLProgram(m,[{dataId:e,shape:i,dtype:o}],o),d=this.readSync(f.dataId);return this.disposeIntermediateTensorInfo(f),d}if(n!=null)return this.convertAndCacheOnCPU(e);if(o==="string")return n;let u=this.activeTimers!=null,c;u&&(c=y.now());let p;if(o==="complex64"){let m=this.readSync(s.real.dataId),f=this.readSync(s.imag.dataId);p=C.mergeRealAndImagArrays(m,f)}else p=this.getValuesFromTexture(e);return u&&(this.downloadWaitMs+=y.now()-c),this.convertAndCacheOnCPU(e,p)}async read(e){if(this.pendingRead.has(e)){let d=this.pendingRead.get(e);return new Promise(h=>d.push(h))}let t=this.texData.get(e),{values:n,shape:o,slice:s,dtype:a,complexTensorInfos:i,isPacked:l}=t;if(s!=null){let d;l?d=new Ns(o,Mf):d=new xn(o,Mf);let h=this.runWebGLProgram(d,[{dataId:e,shape:o,dtype:a}],a),g=this.read(h.dataId);return this.disposeIntermediateTensorInfo(h),g}if(n!=null)return this.convertAndCacheOnCPU(e);if(!W().getBool("WEBGL_DOWNLOAD_FLOAT_ENABLED")&&W().getNumber("WEBGL_VERSION")===2)throw new Error("tensor.data() with WEBGL_DOWNLOAD_FLOAT_ENABLED=false and WEBGL_VERSION=2 not yet supported.");let u=null,c;if(a!=="complex64"&&W().get("WEBGL_BUFFER_SUPPORTED")){c=this.decode(e);let d=this.texData.get(c.dataId);u=this.gpgpu.createBufferFromTexture(d.texture,...wl(o))}this.pendingRead.set(e,[]),a!=="complex64"&&await this.gpgpu.createAndWaitForFence();let p;if(a==="complex64"){let d=await Promise.all([this.read(i.real.dataId),this.read(i.imag.dataId)]),h=d[0],g=d[1];p=C.mergeRealAndImagArrays(h,g)}else if(u==null)p=this.getValuesFromTexture(e);else{let d=y.sizeFromShape(o);p=this.gpgpu.downloadFloat32MatrixFromBuffer(u,d)}c!=null&&this.disposeIntermediateTensorInfo(c);let m=this.convertAndCacheOnCPU(e,p),f=this.pendingRead.get(e);return this.pendingRead.delete(e),f.forEach(d=>d(m)),this.pendingDisposal.has(e)&&(this.pendingDisposal.delete(e),this.disposeData(e)&&hs().removeDataId(e,this),this.pendingDeletes--),m}bufferSync(e){let t=this.readSync(e.dataId),n=t;if(e.dtype==="string")try{n=t.map(o=>y.decodeString(o))}catch(o){throw new Error("Failed to decode encoded string bytes into utf-8")}return Ce(e.shape,e.dtype,n)}checkNumericalProblems(e){if(e!=null)for(let t=0;t<e.length;t++){let n=e[t];if(!mk(n))throw W().getBool("WEBGL_RENDER_FLOAT32_CAPABLE")?Error(`The value ${n} cannot be represented with your current settings. Consider enabling float32 rendering: 'tf.env().set('WEBGL_RENDER_FLOAT32_ENABLED', true);'`):Error(`The value ${n} cannot be represented on this device.`)}}getValuesFromTexture(e){let{shape:t,dtype:n,isPacked:o}=this.texData.get(e),s=y.sizeFromShape(t);if(W().getBool("WEBGL_DOWNLOAD_FLOAT_ENABLED")){let m=this.decode(e),f=this.texData.get(m.dataId),d=this.gpgpu.downloadMatrixFromPackedTexture(f.texture,...wl(t)).subarray(0,s);return this.disposeIntermediateTensorInfo(m),d}let a=W().getBool("WEBGL_PACK")&&o===!0,i=a?Ff(t):t,l=a?new Pk(i):new Ok(i),u=this.runWebGLProgram(l,[{shape:i,dtype:n,dataId:e}],"float32"),c=this.texData.get(u.dataId),p=this.gpgpu.downloadByteEncodedFloatMatrixFromOutputTexture(c.texture,c.texShape[0],c.texShape[1]).subarray(0,s);return this.disposeIntermediateTensorInfo(u),p}timerAvailable(){return W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE")>0}async time(e){let t=this.activeTimers,n=[],o=!1;this.programTimersStack==null?(this.programTimersStack=n,o=!0):this.activeTimers.push(n),this.activeTimers=n,e();let s=y.flatten(this.activeTimers.map(l=>l.query)).filter(l=>l!=null),a=y.flatten(this.activeTimers.map(l=>l.name)).filter(l=>l!=null);this.activeTimers=t,o&&(this.programTimersStack=null);let i={uploadWaitMs:this.uploadWaitMs,downloadWaitMs:this.downloadWaitMs,kernelMs:null,wallMs:null};if(W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE")>0){let l=await Promise.all(s);i.kernelMs=y.sum(l),i.getExtraProfileInfo=()=>l.map((u,c)=>({name:a[c],ms:u})).map(u=>`${u.name}: ${u.ms}`).join(", ")}else i.kernelMs={error:"WebGL query timers are not supported in this environment."};return this.uploadWaitMs=0,this.downloadWaitMs=0,i}memory(){return{unreliable:!1,numBytesInGPU:this.numBytesInGPU,numBytesInGPUAllocated:this.textureManager.numBytesAllocated,numBytesInGPUFree:this.textureManager.numBytesFree}}startTimer(){return W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE")>0?this.gpgpu.beginQuery():{startMs:y.now(),endMs:null}}endTimer(e){return W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE")>0?(this.gpgpu.endQuery(),e):(e.endMs=y.now(),e)}async getQueryTime(e){if(W().getNumber("WEBGL_DISJOINT_QUERY_TIMER_EXTENSION_RELIABLE")>0)return this.gpgpu.waitForQueryAndGetTime(e);let t=e;return t.endMs-t.startMs}disposeData(e,t=!1){if(this.pendingDisposal.has(e))return!1;if(!this.texData.has(e))return!0;if(t?this.texData.get(e).refCount=0:this.texData.get(e).refCount--,!t&&this.texData.get(e).refCount>0)return!1;if(this.pendingRead.has(e))return this.pendingDisposal.add(e),this.pendingDeletes++,!1;this.releaseGPUData(e);let{complexTensorInfos:n}=this.texData.get(e);return n!=null&&(this.disposeData(n.real.dataId,t),this.disposeData(n.imag.dataId,t)),this.texData.delete(e),!0}releaseGPUData(e){let{texture:t,dtype:n,texShape:o,usage:s,isPacked:a,slice:i}=this.texData.get(e),l=i&&i.origDataId||e,u=this.dataRefCount.get(l);u>1?this.dataRefCount.set(l,u-1):(this.dataRefCount.delete(l),t!=null&&(this.numBytesInGPU-=this.computeBytes(o,n),this.textureManager.releaseTexture(t,o,s,a)));let c=this.texData.get(e);c.texture=null,c.texShape=null,c.isPacked=!1,c.slice=null}getTexture(e){return this.uploadToGPU(e),this.texData.get(e).texture}getDataInfo(e){return this.texData.get(e)}shouldExecuteOnCPU(e,t=OX){return W().getBool("WEBGL_CPU_FORWARD")&&e.every(n=>this.texData.get(n.dataId).texture==null&&y.sizeFromShape(n.shape)<t)}getGPGPUContext(){return this.gpgpu}where(e){C.warn("tf.where() in webgl locks the UI thread. Call tf.whereAsync() instead");let t=e.dataSync();return DX(e.shape,t)}packedUnaryOp(e,t,n){let o=new Ns(e.shape,t),s=this.compileAndRun(o,[e],n);return hs().makeTensorFromDataId(s.dataId,s.shape,s.dtype)}abs(e){if(this.shouldExecuteOnCPU([e])&&e.dtype!=="complex64"){let o=fx(this.texData.get(e.dataId).values);return this.makeOutput(e.shape,e.dtype,o)}if(W().getBool("WEBGL_PACK_UNARY_OPERATIONS"))return this.packedUnaryOp(e,ov,e.dtype);let t=new xn(e.shape,ov),n=this.compileAndRun(t,[e]);return hs().makeTensorFromDataId(n.dataId,n.shape,n.dtype)}makeTensorInfo(e,t,n){let o;if(t==="string"&&n!=null&&n.length>0&&y.isString(n[0])){let s=n.map(a=>y.encodeString(a));o=this.write(s,e,t)}else o=this.write(n,e,t);return this.texData.get(o).usage=null,{dataId:o,shape:e,dtype:t}}makeOutput(e,t,n){let{dataId:o}=this.makeTensorInfo(e,t,n);return hs().makeTensorFromDataId(o,e,t,this)}unpackTensor(e){let t=new sv(e.shape);return this.runWebGLProgram(t,[e],e.dtype)}packTensor(e){let t=new rv(e.shape),n=!0;return this.runWebGLProgram(t,[e],e.dtype,null,n)}packedReshape(e,t){let n=[$a(e.shape),...Ra(e.shape)],o={dtype:e.dtype,shape:n,dataId:e.dataId},s=[$a(t),...Ra(t)],a=new Pf(s,n),i=!0,l=this.runWebGLProgram(a,[o],e.dtype,null,i);return{dataId:l.dataId,shape:t,dtype:l.dtype}}decode(e){let t=this.texData.get(e),{isPacked:n,shape:o,dtype:s}=t,a=Ff(o),i;n?i=new Fk(a):i=new Rk(a);let l=!0,u=this.runWebGLProgram(i,[{shape:a,dtype:s,dataId:e}],s,null,l);return{dtype:s,shape:o,dataId:u.dataId}}runWebGLProgram(e,t,n,o,s=!1){let a=this.makeTensorInfo(e.outputShape,n),i=this.texData.get(a.dataId);if(e.packedOutput&&(i.isPacked=!0),e.outPackingScheme===bl.DENSE){let g=wl(e.outputShape);i.texShape=g.map(x=>x*2)}if(e.outTexUsage!=null&&(i.usage=e.outTexUsage),y.sizeFromShape(a.shape)===0)return i.values=y.getTypedArrayFromDType(a.dtype,0),a;let l=[],u=t.map(g=>{if(g.dtype==="complex64")throw new Error("GPGPUProgram does not support complex64 input. For complex64 dtypes, please separate the program into real and imaginary parts.");let x=this.texData.get(g.dataId);if(x.texture==null){if(!e.packedInputs&&y.sizeFromShape(g.shape)<=W().getNumber("WEBGL_SIZE_UPLOAD_UNIFORM"))return{shape:g.shape,texData:null,isUniform:!0,uniformValues:x.values};e.packedInputs&&(x.isPacked=!0,x.shape=g.shape)}else if(!!x.isPacked!=!!e.packedInputs)g=x.isPacked?this.unpackTensor(g):this.packTensor(g),l.push(g),x=this.texData.get(g.dataId);else if(x.isPacked&&!_l(x.shape,g.shape)){let b=g,w=g.shape;g.shape=x.shape,g=this.packedReshape(g,w),l.push(g),x=this.texData.get(g.dataId),b.shape=w}return this.uploadToGPU(g.dataId),{shape:g.shape,texData:x,isUniform:!1}});this.uploadToGPU(a.dataId);let c={shape:a.shape,texData:i,isUniform:!1},p=iD(e,u,c),m=this.getAndSaveBinary(p,()=>nD(this.gpgpu,e,u,c)),f=this.activeTimers!=null,d;f&&(d=this.startTimer()),sD(this.gpgpu,m,u,c,o),l.forEach(g=>this.disposeIntermediateTensorInfo(g)),f&&(d=this.endTimer(d),this.activeTimers.push({name:e.constructor.name,query:this.getQueryTime(d)}));let h=W().get("WEBGL_FLUSH_THRESHOLD");if(h>0){let g=y.now();g-this.lastGlFlushTime>h&&(this.gpgpu.gl.flush(),this.lastGlFlushTime=g)}if(!W().getBool("WEBGL_LAZILY_UNPACK")&&i.isPacked&&s===!1){let g=this.unpackTensor(a);return this.disposeIntermediateTensorInfo(a),g}return a}compileAndRun(e,t,n,o,s=!1){return n=n||t[0].dtype,this.runWebGLProgram(e,t,n,o,s)}getAndSaveBinary(e,t){return e in this.binaryCache||(this.binaryCache[e]=t()),this.binaryCache[e]}getTextureManager(){return this.textureManager}dispose(){this.disposed||(W().getBool("IS_TEST")||Object.keys(this.binaryCache).forEach(t=>{this.gpgpu.deleteProgram(this.binaryCache[t].webGLProgram),delete this.binaryCache[t]}),this.textureManager.dispose(),this.canvas!=null&&typeof HTMLCanvasElement!="undefined"&&this.canvas instanceof HTMLCanvasElement?this.canvas.remove():this.canvas=null,this.gpgpuCreatedLocally&&(this.gpgpu.program=null,this.gpgpu.dispose()),this.disposed=!0)}floatPrecision(){return this.floatPrecisionValue==null&&(this.floatPrecisionValue=V(()=>{if(!W().get("WEBGL_RENDER_FLOAT32_ENABLED")){let e=W().getBool("DEBUG");W().set("DEBUG",!1);let t=this.abs(ue(1e-8)).dataSync()[0];if(W().set("DEBUG",e),t>0)return 32}return 16})),this.floatPrecisionValue}epsilon(){return this.floatPrecision()===32?$X:RX}uploadToGPU(e){let t=this.texData.get(e),{shape:n,dtype:o,values:s,texture:a,usage:i,isPacked:l}=t;if(a!=null)return;let u=this.activeTimers!=null,c;u&&(c=y.now());let p=t.texShape;if(p==null&&(p=Ik(n,l),t.texShape=p),s!=null){let m=Ff(n),f,d=p[1],h=p[0],g=s instanceof Uint8Array;l?([d,h]=ui(p[0],p[1]),f=new Lk(m,[h,d],g)):f=new Mk(m,[h,d],g);let x=this.makeTensorInfo([h,d],o);g?this.texData.get(x.dataId).usage=Pr.PIXELS:this.texData.get(x.dataId).usage=Pr.UPLOAD,this.gpgpu.uploadDenseMatrixToTexture(this.getTexture(x.dataId),d,h,s);let b=!0,w=this.runWebGLProgram(f,[x],o,null,b),_=this.texData.get(w.dataId);t.texture=_.texture,t.texShape=_.texShape,t.isPacked=_.isPacked,t.usage=_.usage,this.disposeIntermediateTensorInfo(x),this.texData.delete(w.dataId),t.values=null,u&&(this.uploadWaitMs+=y.now()-c)}else{let m=this.acquireTexture(p,i,o,l);t.texture=m}}convertAndCacheOnCPU(e,t){let n=this.texData.get(e),{dtype:o}=n;return this.releaseGPUData(e),t!=null&&(n.values=LX(t,o)),n.values}acquireTexture(e,t,n,o){if(this.numBytesInGPU+=this.computeBytes(e,n),!this.warnedAboutMemory&&this.numBytesInGPU>this.numMBBeforeWarning*1024*1024){let s=(this.numBytesInGPU/1024/1024).toFixed(2);this.warnedAboutMemory=!0,console.warn(`High memory usage in GPU: ${s} MB, most likely due to a memory leak`)}return this.textureManager.acquireTexture(e,t,o)}computeBytes(e,t){return e[0]*e[1]*y.bytesPerElement(t)}};Lu.nextDataId=0;function LX(r,e){if(e==="float32"||e==="complex64")return r;if(e==="int32"||e==="bool"){let t=e==="int32"?new Int32Array(r.length):new Uint8Array(r.length);for(let n=0;n<t.length;++n)t[n]=Math.round(r[n]);return t}else throw new Error(`Unknown dtype ${e}`)}var iv="3.6.0";function XD(){W().set("WEBGL_FORCE_F16_TEXTURES",!0)}Jl.isBrowser()&&np("webgl",()=>new Lu,2);var Q7e={forceHalfFloat:XD};var hx=`
|
|
if (isnan(a)) return a;
|
|
if (isnan(b)) return b;
|
|
`;var co=class{constructor(e,t,n){this.variableNames=["A","B"],this.outputShape=C.assertAndGetBroadcastShape(t,n),this.userCode=`
|
|
float binaryOperation(float a, float b) {
|
|
${e}
|
|
}
|
|
|
|
void main() {
|
|
float a = getAAtOutCoords();
|
|
float b = getBAtOutCoords();
|
|
setOutput(binaryOperation(a, b));
|
|
}
|
|
`}};var kl=`
|
|
result.r = isNaN.r > 0. ? NAN : result.r;
|
|
result.g = isNaN.g > 0. ? NAN : result.g;
|
|
result.b = isNaN.b > 0. ? NAN : result.b;
|
|
result.a = isNaN.a > 0. ? NAN : result.a;
|
|
`;var Ss=class{constructor(e,t,n,o=!1){this.variableNames=["A","B"],this.supportsBroadcasting=!0,this.packedInputs=!0,this.packedOutput=!0,this.outputShape=C.assertAndGetBroadcastShape(t,n);let s=this.outputShape.length,a="";if(o)if(s===0||y.sizeFromShape(this.outputShape)===1)a=`
|
|
result.y = 0.;
|
|
result.z = 0.;
|
|
result.w = 0.;
|
|
`;else if(a=`
|
|
${Ve(s)} coords = getOutputCoords();
|
|
`,s===1)a+=`
|
|
result.y = (coords + 1) >= ${this.outputShape[0]} ? 0. : result.y;
|
|
result.z = 0.;
|
|
result.w = 0.;
|
|
`;else{let l=jt("coords",s);a+=`
|
|
bool nextRowOutOfBounds =
|
|
(${l[s-2]} + 1) >= ${this.outputShape[s-2]};
|
|
bool nextColOutOfBounds =
|
|
(${l[s-1]} + 1) >= ${this.outputShape[s-1]};
|
|
result.y = nextColOutOfBounds ? 0. : result.y;
|
|
result.z = nextRowOutOfBounds ? 0. : result.z;
|
|
result.w = nextColOutOfBounds || nextRowOutOfBounds ? 0. : result.w;
|
|
`}this.userCode=`
|
|
vec4 binaryOperation(vec4 a, vec4 b) {
|
|
${e}
|
|
}
|
|
|
|
void main() {
|
|
vec4 a = getAAtOutCoords();
|
|
vec4 b = getBAtOutCoords();
|
|
|
|
vec4 result = binaryOperation(a, b);
|
|
${a}
|
|
|
|
setOutput(result);
|
|
}
|
|
`}};function Ut(r){let{inputs:e,backend:t}=r,{x:n}=e;return t.incRef(n.dataId),{dataId:n.dataId,shape:n.shape,dtype:n.dtype}}var YD={kernelName:Xn,backendName:"webgl",kernelFunc:Ut};function yn(r){let{inputs:e,backend:t}=r,{real:n,imag:o}=e,s=t.makeTensorInfo(n.shape,"complex64"),a=t.texData.get(s.dataId),i=Ut({inputs:{x:n},backend:t}),l=Ut({inputs:{x:o},backend:t});return a.complexTensorInfos={real:i,imag:l},s}var ZD={kernelName:vc,backendName:"webgl",kernelFunc:yn};var av="return (a < 0.) ? b * a : a;",lv=`
|
|
vec4 aLessThanZero = vec4(lessThan(a, vec4(0.)));
|
|
return (aLessThanZero * (b * a)) + ((vec4(1.0) - aLessThanZero) * a);
|
|
`;function zX(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{alpha:s}=n,a=t.makeTensorInfo([],"float32",y.createScalarValue(s,"float32")),i=W().getBool("WEBGL_PACK_BINARY_OPERATIONS")?new Ss(lv,o.shape,a.shape):new co(av,o.shape,a.shape),l=t.runWebGLProgram(i,[o,a],o.dtype);return t.disposeIntermediateTensorInfo(a),l}var JD={kernelName:Mo,backendName:"webgl",kernelFunc:zX};var uv="return (a < 0.) ? b * a : a;",cv=`
|
|
vec4 aLessThanZero = vec4(lessThan(a, vec4(0.)));
|
|
return (aLessThanZero * (b * a)) + ((vec4(1.0) - aLessThanZero) * a);
|
|
`;function BX(r){let{inputs:e,backend:t}=r,{x:n,alpha:o}=e,s=W().getBool("WEBGL_PACK_BINARY_OPERATIONS")?new Ss(cv,n.shape,o.shape):new co(uv,n.shape,o.shape);return t.runWebGLProgram(s,[n,o],n.dtype)}var QD={kernelName:Yo,backendName:"webgl",kernelFunc:BX};var gx="if (isnan(x)) return x;",e$=`
|
|
if (isnan(a)) return a;
|
|
if (isnan(b)) return b;
|
|
`,t$=`
|
|
result.r = isNaN.r > 0. ? NAN : result.r;
|
|
result.g = isNaN.g > 0. ? NAN : result.g;
|
|
result.b = isNaN.b > 0. ? NAN : result.b;
|
|
result.a = isNaN.a > 0. ? NAN : result.a;
|
|
`;function _e({opSnippet:r,packedOpSnippet:e,cpuKernelImpl:t,dtype:n}){return({inputs:o,backend:s})=>{let{x:a}=o,i=s,l=n||a.dtype;if(i.shouldExecuteOnCPU([a])&&t!=null){let p=i.texData.get(a.dataId),m=t(p.values,l);return i.makeTensorInfo(a.shape,l,m)}let u=W().getBool("WEBGL_PACK_UNARY_OPERATIONS")&&e!=null,c;return u?c=new Ns(a.shape,e):c=new xn(a.shape,r),i.runWebGLProgram(c,[a],l)}}function st({opSnippet:r,packedOpSnippet:e,checkOutOfBounds:t=!1,supportsComplex:n=!1,cpuKernelImpl:o,dtype:s}){return({inputs:a,backend:i})=>{let{a:l,b:u}=a,c=i;if(n&&l.dtype==="complex64"){let d=c.texData.get(l.dataId),h=c.texData.get(u.dataId),[g,x]=[[d.complexTensorInfos.real,h.complexTensorInfos.real],[d.complexTensorInfos.imag,h.complexTensorInfos.imag]].map(w=>{let[_,I]=w,E={dataId:_.dataId,dtype:_.dtype,shape:l.shape},D={dataId:I.dataId,dtype:I.dtype,shape:u.shape},$=new co(r,l.shape,u.shape);return c.runWebGLProgram($,[E,D],lr(_.dtype,I.dtype))}),b=yn({inputs:{real:g,imag:x},backend:c});return c.disposeIntermediateTensorInfo(g),c.disposeIntermediateTensorInfo(x),b}let p=s||lr(l.dtype,u.dtype);if(c.shouldExecuteOnCPU([l,u])&&o!=null){let d=c.texData.get(l.dataId),h=c.texData.get(u.dataId),[g,x]=o(l.shape,u.shape,d.values,h.values,p),b=c.makeTensorInfo(x,p),w=c.texData.get(b.dataId);return w.values=g,b}let m=W().getBool("WEBGL_PACK_BINARY_OPERATIONS")&&e!=null,f;return m?f=new Ss(e,l.shape,u.shape,t):f=new co(r,l.shape,u.shape),c.runWebGLProgram(f,[l,u],p)}}function vl(r,e=!1){if(r==="linear")return e?jD:zD;if(r==="relu")return e?HD:VD;if(r==="elu")return e?UD:BD;if(r==="relu6")return e?qD:GD;if(r==="prelu")return e?cv:uv;if(r==="leakyrelu")return e?lv:av;if(r==="sigmoid")return e?KD:WD;throw new Error(`Activation ${r} has not been implemented for the WebGL backend.`)}var Lf=class{constructor(e,t,n,o=!1,s=!1,a=!1,i=null,l=!1,u=!1){this.variableNames=["matrixA","matrixB"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=n;let c=o?e[1]:e[2],p=Math.ceil(c/2),m=o?"i * 2, rc.y":"rc.y, i * 2",f=s?"rc.z, i * 2":"i * 2, rc.z",d=o?["a.xxyy","a.zzww"]:["a.xxzz","a.yyww"],h=s?["b.xzxz","b.ywyw"]:["b.xyxy","b.zwzw"],g="",x="";i&&(l?g=`vec4 activation(vec4 a) {
|
|
vec4 b = getPreluActivationWeightsAtOutCoords();
|
|
${i}
|
|
}`:u?g=`vec4 activation(vec4 a) {
|
|
vec4 b = getLeakyreluAlphaAtOutCoords();
|
|
${i}
|
|
}`:g=`vec4 activation(vec4 x) {
|
|
${i}
|
|
}`,x="result = activation(result);");let b=a?"result += getBiasAtOutCoords();":"";a&&this.variableNames.push("bias"),l&&this.variableNames.push("preluActivationWeights"),u&&this.variableNames.push("leakyreluAlpha");let w="rc.x",_="rc.x";e[0]<t[0]?w=`int(min(float(rc.x), ${e[0]-1}.))`:t[0]<e[0]&&(_=`int(min(float(rc.x), ${t[0]-1}.))`),this.userCode=`
|
|
${g}
|
|
|
|
const float sharedDimension = ${p}.0;
|
|
|
|
vec4 dot2x2ARowBCol(ivec3 rc) {
|
|
vec4 result = vec4(0);
|
|
for (int i = 0; i < ${p}; i++) {
|
|
int batchA = ${w};
|
|
int batchB = ${_};
|
|
vec4 a = getMatrixA(batchA, ${m});
|
|
vec4 b = getMatrixB(batchB, ${f});
|
|
|
|
// These swizzled products need to be separately added.
|
|
// See: https://github.com/tensorflow/tfjs/issues/1735
|
|
result += (${d[0]} * ${h[0]});
|
|
result += (${d[1]} * ${h[1]});
|
|
}
|
|
return result;
|
|
}
|
|
|
|
void main() {
|
|
ivec3 rc = getOutputCoords();
|
|
vec4 result = dot2x2ARowBCol(rc);
|
|
|
|
${b}
|
|
|
|
${x}
|
|
|
|
setOutput(result);
|
|
}
|
|
`}};var pv={REAL:"return areal * breal - aimag * bimag;",IMAG:"return areal * bimag + aimag * breal;"},xx=class{constructor(e,t,n){this.variableNames=["AReal","AImag","BReal","BImag"],this.outputShape=C.assertAndGetBroadcastShape(t,n),this.userCode=`
|
|
float binaryOpComplex(
|
|
float areal, float aimag, float breal, float bimag) {
|
|
${e}
|
|
}
|
|
|
|
void main() {
|
|
float areal = getARealAtOutCoords();
|
|
float aimag = getAImagAtOutCoords();
|
|
float breal = getBRealAtOutCoords();
|
|
float bimag = getBImagAtOutCoords();
|
|
setOutput(binaryOpComplex(areal, aimag, breal, bimag));
|
|
}
|
|
`}};var r$="return a * b;";function zf(r){let{inputs:e,backend:t}=r,{a:n,b:o}=e,s=C.upcastType(n.dtype,o.dtype);if(n.dtype==="complex64"){let i=t.texData.get(n.dataId),l=t.texData.get(o.dataId),u=new xx(pv.REAL,n.shape,o.shape),c=new xx(pv.IMAG,n.shape,o.shape),p=[{dataId:i.complexTensorInfos.real.dataId,dtype:i.complexTensorInfos.real.dtype,shape:n.shape},{dataId:i.complexTensorInfos.imag.dataId,dtype:i.complexTensorInfos.imag.dtype,shape:n.shape},{dataId:l.complexTensorInfos.real.dataId,dtype:l.complexTensorInfos.real.dtype,shape:o.shape},{dataId:l.complexTensorInfos.imag.dataId,dtype:l.complexTensorInfos.imag.dtype,shape:o.shape}],m=t.runWebGLProgram(u,p,"float32"),f=t.runWebGLProgram(c,p,"float32"),d=yn({inputs:{real:m,imag:f},backend:t});return t.disposeIntermediateTensorInfo(m),t.disposeIntermediateTensorInfo(f),d}if(t.shouldExecuteOnCPU([n,o])){let i=t.texData.get(n.dataId),l=t.texData.get(o.dataId),[u,c]=kD(n.shape,o.shape,i.values,l.values,s),p=t.makeTensorInfo(c,s),m=t.texData.get(p.dataId);return m.values=u,p}let a;return W().getBool("WEBGL_PACK_BINARY_OPERATIONS")?a=new Ss(r$,n.shape,o.shape):a=new co(r$,n.shape,o.shape),t.runWebGLProgram(a,[n,o],s)}var n$={kernelName:Ho,backendName:"webgl",kernelFunc:zf};function o$(r,e,t){let n=[$a(r.shape),...Ra(r.shape)],o={dtype:r.dtype,shape:n,dataId:r.dataId},s=[$a(e),...Ra(e)],a=new Pf(s,n),i=!0,l=t.runWebGLProgram(a,[o],r.dtype,null,i);return{dataId:l.dataId,shape:e,dtype:l.dtype}}function ce(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{shape:s}=n,a=t,i=y.sizeFromShape(o.shape),l=y.inferFromImplicitShape(s,i),u=y.sizeFromShape(l);y.assert(i===u,()=>`The new shape (${l}) has ${u} elements and the old shape (${o.shape}) has ${i} elements. The new shape and old shape must have the same number of elements.`);let c=a.texData.get(o.dataId);return c.isPacked&&!_l(o.shape,l)&&!(c.texture!==null&&_l(c.shape,l))?o$(o,l,a):(a.incRef(o.dataId),{dataId:o.dataId,shape:l,dtype:o.dtype})}var s$={kernelName:qs,backendName:"webgl",kernelFunc:ce};var yx=class{constructor(e,t){this.variableNames=["x"];let{windowSize:n,batchSize:o,inSize:s,outSize:a}=e;this.outputShape=[o,a];let i=Math.floor(n/4)*4,l=n%4,u="sumValue += dot(values, ones);";if(t!=null){let p=1/t;u=`sumValue += dot(values * ${y.isInt(p)?p.toPrecision(2):p}, ones);`}let c="";s%n>0&&(c=`
|
|
if (inIdx < 0 || inIdx >= ${s}) {
|
|
return 0.0;
|
|
}
|
|
`),this.userCode=`
|
|
const vec4 ones = vec4(1.0, 1.0, 1.0, 1.0);
|
|
|
|
float getValue(int batch, int inIdx) {
|
|
${c}
|
|
return getX(batch, inIdx);
|
|
}
|
|
|
|
void main() {
|
|
ivec2 coords = getOutputCoords();
|
|
int batch = coords[0];
|
|
int outIdx = coords[1];
|
|
int inOffset = outIdx * ${n};
|
|
|
|
float sumValue = 0.0;
|
|
|
|
for (int i = 0; i < ${i}; i += 4) {
|
|
int inIdx = inOffset + i;
|
|
vec4 values = vec4(
|
|
getValue(batch, inIdx),
|
|
getValue(batch, inIdx + 1),
|
|
getValue(batch, inIdx + 2),
|
|
getValue(batch, inIdx + 3)
|
|
);
|
|
|
|
${u}
|
|
}
|
|
|
|
int inIdx = inOffset + ${i};
|
|
if (${l===1}) {
|
|
vec4 values = vec4(getValue(batch, inIdx), 0.0, 0.0, 0.0);
|
|
|
|
${u}
|
|
} else if (${l===2}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, inIdx),
|
|
getValue(batch, inIdx + 1), 0.0, 0.0);
|
|
|
|
${u}
|
|
} else if (${l===3}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, inIdx),
|
|
getValue(batch, inIdx + 1),
|
|
getValue(batch, inIdx + 2), 0.0);
|
|
|
|
${u}
|
|
}
|
|
setOutput(sumValue);
|
|
}
|
|
`}};var mv=class{constructor(e,t){this.variableNames=["x"];let{windowSize:n,batchSize:o,inSize:s,outSize:a}=e;this.outputShape=[o,a];let i="0.0",l="";t==="prod"?i="1.0":t==="min"?(i="1.0 / 1e-20",l="min"):t==="max"&&(i="-1.0 / 1e-20",l="max");let u=`${t}(${t}(${t}(minMaxValue[0], minMaxValue[1]), minMaxValue[2]), minMaxValue[3])`;t==="sum"?u="sumValue":t==="prod"?u="prodValue":t==="all"?u="allValue":t==="any"&&(u="anyValue");let c=Math.floor(n/4)*4,p=n%4,m=`
|
|
if (${t==="sum"}) {
|
|
sumValue += dot(values, ones);
|
|
} else if (${t==="prod"}) {
|
|
vec2 tmp = vec2(values[0], values[1]) * vec2(values[2], values[3]);
|
|
prodValue *= tmp[0] * tmp[1];
|
|
} else {
|
|
minMaxValue = ${l}(values, minMaxValue);
|
|
}
|
|
`,f="vec4";t==="all"?(i="1.0",m=`
|
|
bool reducedAllValue = all(values);
|
|
float floatedReducedAllValue = float(reducedAllValue);
|
|
allValue = float(allValue >= 1.0 && floatedReducedAllValue >= 1.0);
|
|
`,f="bvec4"):t==="any"&&(i="0.0",m=`
|
|
bool reducedAnyValue = any(values);
|
|
float floatedReducedAnyValue = float(reducedAnyValue);
|
|
anyValue = float(anyValue >= 1.0 || floatedReducedAnyValue >= 1.0);
|
|
`,f="bvec4");let d="";s%n>0&&(d=`
|
|
if (inIdx < 0 || inIdx >= ${s}) {
|
|
return initializationValue;
|
|
}
|
|
`),this.userCode=`
|
|
const float initializationValue = ${i};
|
|
const vec4 ones = vec4(1.0, 1.0, 1.0, 1.0);
|
|
|
|
float getValue(int batch, int inIdx) {
|
|
${d}
|
|
return getX(batch, inIdx);
|
|
}
|
|
|
|
void main() {
|
|
ivec2 coords = getOutputCoords();
|
|
int batch = coords[0];
|
|
int outIdx = coords[1];
|
|
int inOffset = outIdx * ${n};
|
|
|
|
vec4 minMaxValue = vec4(${i});
|
|
float prodValue = 1.0;
|
|
float sumValue = 0.0;
|
|
float allValue = 1.0;
|
|
float anyValue = 0.0;
|
|
|
|
for (int i = 0; i < ${c}; i += 4) {
|
|
int inIdx = inOffset + i;
|
|
${f} values = ${f}(
|
|
getValue(batch, inIdx),
|
|
getValue(batch, inIdx + 1),
|
|
getValue(batch, inIdx + 2),
|
|
getValue(batch, inIdx + 3)
|
|
);
|
|
|
|
${m}
|
|
}
|
|
|
|
int inIdx = inOffset + ${c};
|
|
if (${p===1}) {
|
|
${f} values = ${f}(
|
|
getValue(batch, inIdx),
|
|
initializationValue,
|
|
initializationValue,
|
|
initializationValue
|
|
);
|
|
|
|
${m}
|
|
} else if (${p===2}) {
|
|
${f} values = ${f}(
|
|
getValue(batch, inIdx),
|
|
getValue(batch, inIdx + 1),
|
|
initializationValue,
|
|
initializationValue
|
|
);
|
|
|
|
${m}
|
|
} else if (${p===3}) {
|
|
${f} values = ${f}(
|
|
getValue(batch, inIdx),
|
|
getValue(batch, inIdx + 1),
|
|
getValue(batch, inIdx + 2),
|
|
initializationValue
|
|
);
|
|
|
|
${m}
|
|
}
|
|
setOutput(${u});
|
|
}
|
|
`}};function VX(r){let e=[];for(;e.length===0||e[e.length-1].outSize!==1;){let t=e.length?e[e.length-1].outSize:r[1],n=C.computeOptimalWindowSize(t);e.push({inSize:t,windowSize:n,outSize:Math.ceil(t/n)})}return e}function En(r,e,t,n){let o=VX(r.shape),s=r;for(let a=0;a<o.length;a++){let{inSize:i,windowSize:l,outSize:u}=o[a],c,p;t==="mean"?c=a===0?new yx({windowSize:l,inSize:i,batchSize:r.shape[0],outSize:u},i):new yx({windowSize:l,inSize:i,batchSize:r.shape[0],outSize:u}):c=new mv({windowSize:l,inSize:i,batchSize:r.shape[0],outSize:u},t),p=s,s=n.runWebGLProgram(c,[s],e),p.dataId!==r.dataId&&n.disposeIntermediateTensorInfo(p)}return s}var fv=class{constructor(e,t){this.variableNames=["A"];let n=new Array(e.length);for(let a=0;a<n.length;a++)n[a]=e[t[a]];this.outputShape=n,this.rank=n.length;let o=Ve(this.rank),s=GX(t);this.userCode=`
|
|
void main() {
|
|
${o} resRC = getOutputCoords();
|
|
setOutput(getA(${s}));
|
|
}
|
|
`}};function GX(r){let e=r.length;if(e>6)throw Error(`Transpose for rank ${e} is not yet supported`);let t=["resRC.x","resRC.y","resRC.z","resRC.w","resRC.u","resRC.v"],n=new Array(e);for(let o=0;o<r.length;o++)n[r[o]]=t[o];return n.join()}var dv=class{constructor(e,t){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!0;let n=new Array(e.length);for(let c=0;c<n.length;c++)n[c]=e[t[c]];if(this.outputShape=n,this.rank=n.length,this.rank>6)throw Error(`Packed transpose for rank ${this.rank} is not yet supported.`);let o=Ve(this.rank),s=tv("rc",this.rank),a=new Array(this.rank);for(let c=0;c<t.length;c++)a[t[c]]=s[c];let i=`vec2(${a.slice(-2).join()})`,l=`++${s[this.rank-1]} < ${n[this.rank-1]}`,u=`getChannel(getA(${a.join()}), ${i})`;this.userCode=`
|
|
void main() {
|
|
${o} rc = getOutputCoords();
|
|
vec4 result = vec4(0.);
|
|
result[0] = ${u};
|
|
if(${l}) {
|
|
result[1] = ${u};
|
|
}
|
|
--${s[this.rank-1]};
|
|
if(++${s[this.rank-2]} < ${n[this.rank-2]}) {
|
|
result[2] = ${u};
|
|
if(${l}) {
|
|
result[3] = ${u};
|
|
}
|
|
}
|
|
setOutput(result);
|
|
}
|
|
`}};function Cl(r,e,t){let n=W().getBool("WEBGL_PACK_ARRAY_OPERATIONS")?new dv(r.shape,e):new fv(r.shape,e);return t.runWebGLProgram(n,[r],r.dtype)}function i$(r,e,t,n){let o=e,s=r.shape.length,a=y.parseAxisParam(o,r.shape),i=a,l=C.getAxesPermutation(i,s),u=l!=null,c=r;u&&(c=Cl(r,l,n),i=C.getInnerMostAxes(i.length,s)),C.assertAxesAreInnerMostDims("sum",i,s);let[p,m]=C.computeOutAndReduceShapes(c.shape,i),f=p;t&&(f=C.expandShapeToKeepDim(p,a));let d=y.sizeFromShape(m),g=y.sizeFromShape(r.shape)/d,x=ce({inputs:{x:c},attrs:{shape:[g,d]},backend:n}),b=Yl(r.dtype),w=En(x,b,"sum",n),_=ce({inputs:{x:w},attrs:{shape:f},backend:n});return n.disposeIntermediateTensorInfo(x),n.disposeIntermediateTensorInfo(w),u&&n.disposeIntermediateTensorInfo(c),_}function zu(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n;return i$(o,s,a,t)}var a$={kernelName:is,backendName:"webgl",kernelFunc:zu};function Rt(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{perm:s}=n,a=t,i=o.shape.length,l=new Array(i);for(let c=0;c<l.length;c++)l[c]=o.shape[s[c]];let u;if(a.shouldExecuteOnCPU([o])){let p=a.texData.get(o.dataId).values,m=Mu(p,o.shape,o.dtype,s,l);u=a.makeTensorInfo(l,o.dtype);let f=a.texData.get(u.dataId);f.values=m}else u=Cl(o,s,a);return u}var l$={kernelName:ms,backendName:"webgl",kernelFunc:Rt};var hv=1e3;function Bu({a:r,b:e,transposeA:t,transposeB:n,backend:o,bias:s=null,preluActivationWeights:a=null,leakyreluAlpha:i=0,activation:l=null}){let u=r.shape.length,c=e.shape.length,p=t?r.shape[u-2]:r.shape[u-1],m=n?e.shape[c-1]:e.shape[c-2],f=t?r.shape[u-1]:r.shape[u-2],d=n?e.shape[c-2]:e.shape[c-1],h=r.shape.slice(0,-2),g=e.shape.slice(0,-2),x=y.sizeFromShape(h),b=y.sizeFromShape(g),w=x===b||x===1||b===1;y.assert(u>=2&&c>=2&&w,()=>`Error in matMul: the input batch dimensions must either be the same or at least one input batch dimension must be 1. Got input batch dimensions of (${h}) and (${g}).`);let I=(x>b?r.shape.slice(0,-2):e.shape.slice(0,-2)).concat([f,d]);y.assert(p===m,()=>`Error in matMul: inner shapes (${p}) and (${m}) of Tensors with shapes ${r.shape} and ${e.shape} and transposeA=${t} and transposeB=${n} must match.`);let E=t?[x,p,f]:[x,f,p],D=n?[b,d,m]:[b,m,d],$=ce({inputs:{x:r},backend:o,attrs:{shape:E}}),R=ce({inputs:{x:e},backend:o,attrs:{shape:D}}),M=[$,R],G=Math.max(x,b),j=t?$.shape[1]:$.shape[2],U=s!=null,H=a!=null,q=l==="leakyrelu",X=l!=null?vl(l,!0):null,ne=U||H||q||X!=null,Y;if((f===1||d===1)&&j>hv&&ne===!1){let ee=$,ie=R;t&&(ee=Rt({inputs:{x:$},backend:o,attrs:{perm:[0,2,1]}}),M.push(ee)),n&&(ie=Rt({inputs:{x:R},backend:o,attrs:{perm:[0,2,1]}}),M.push(ie));let me=d!==1,ae=d===1,de=ee;me&&(de=ce({inputs:{x:ee},backend:o,attrs:{shape:[G,j,1]}}),M.push(de));let he=d===1?2:1,xe=ie;ae&&(xe=ce({inputs:{x:ie},backend:o,attrs:{shape:[G,1,j]}}),M.push(xe));let we=zf({inputs:{a:de,b:xe},backend:o});Y=zu({inputs:{x:we},backend:o,attrs:{axis:he,keepDims:!0}}),M.push(we)}else{let ee=lr(r.dtype,e.dtype),ie=new Lf(E,D,[G,f,d],t,n,U,X,H,q),me=[$,R];if(s!=null&&me.push(s),H&&me.push(a),q){let ae=o.makeTensorInfo([],"float32",y.createScalarValue(i,"float32"));me.push(ae),M.push(ae)}Y=o.runWebGLProgram(ie,me,ee)}let re=ce({inputs:{x:Y},backend:o,attrs:{shape:I}});M.push(Y);for(let ee of M)o.disposeIntermediateTensorInfo(ee);return re}function WX(r){let{inputs:e,backend:t,attrs:n}=r,{a:o,b:s,bias:a,preluActivationWeights:i}=e,{transposeA:l,transposeB:u,activation:c,leakyreluAlpha:p}=n;return Bu({a:o,b:s,transposeA:l,transposeB:u,backend:t,bias:a,preluActivationWeights:i,leakyreluAlpha:p,activation:c})}var u$={kernelName:Qs,backendName:"webgl",kernelFunc:WX};var c$="return abs(x);";function jX(r){let{inputs:e,backend:t}=r,{x:n}=e;if(t.shouldExecuteOnCPU([n])&&n.dtype!=="complex64"){let s=t.texData.get(n.dataId),a=fx(s.values);return t.makeTensorInfo(n.shape,n.dtype,a)}let o;return W().getBool("WEBGL_PACK_UNARY_OPERATIONS")?o=new Ns(n.shape,c$):o=new xn(n.shape,c$),t.runWebGLProgram(o,[n],n.dtype)}var p$={kernelName:Bs,backendName:"webgl",kernelFunc:jX};var UX=br+`
|
|
if (abs(x) > 1.) {
|
|
return NAN;
|
|
}
|
|
return acos(x);
|
|
`,HX=_e({opSnippet:UX}),m$={kernelName:ki,backendName:"webgl",kernelFunc:HX};var qX=br+`
|
|
if (x < 1.0) return NAN;
|
|
return log(x + sqrt(x * x - 1.0));`,KX=_e({opSnippet:qX}),f$={kernelName:vi,backendName:"webgl",kernelFunc:KX};var d$="return a + b;",XX=st({opSnippet:d$,packedOpSnippet:d$,supportsComplex:!0,cpuKernelImpl:aD}),h$={kernelName:Pn,backendName:"webgl",kernelFunc:XX};var gv=class{constructor(e,t){this.outputShape=[],this.outputShape=e,this.variableNames=t.map((s,a)=>`T${a}`);let n=[];this.variableNames.forEach(s=>{n.push(`float v${s} = get${s}AtOutCoords();`)});let o=this.variableNames.map(s=>`v${s}`).join(" + ");this.userCode=`
|
|
void main() {
|
|
${n.join(`
|
|
`)}
|
|
|
|
float result = ${o};
|
|
setOutput(result);
|
|
}
|
|
`}};var xv=class{constructor(e,t){this.outputShape=[],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=e,this.variableNames=t.map((s,a)=>`T${a}`);let n=[];this.variableNames.forEach(s=>{n.push(`vec4 v${s} = get${s}AtOutCoords();`)});let o=this.variableNames.map(s=>`v${s}`).join(" + ");this.userCode=`
|
|
void main() {
|
|
${n.join(`
|
|
`)}
|
|
|
|
vec4 result = ${o};
|
|
setOutput(result);
|
|
}
|
|
`}};function bx(r){let{inputs:e,backend:t}=r,n=e;if(n.length===1)return Ut({inputs:{x:n[0]},backend:t});if(n.length>W().get("WEBGL_MAX_TEXTURES_IN_SHADER")){let l=Math.floor(n.length/2),u=bx({inputs:n.slice(0,l),backend:t}),c=bx({inputs:n.slice(l),backend:t});return bx({inputs:[u,c],backend:t})}let o=n.map(l=>l.dtype).reduce((l,u)=>lr(l,u)),s=n.map(l=>l.shape),i=W().getBool("WEBGL_PACK")?new xv(n[0].shape,s):new gv(n[0].shape,s);return t.runWebGLProgram(i,n,o)}var g$={kernelName:_o,backendName:"webgl",kernelFunc:bx};function YX(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n,i=o.shape.length,l=y.parseAxisParam(s,o.shape),u=l,c=C.getAxesPermutation(u,i),p=o;c!=null&&(p=Rt({inputs:{x:o},backend:t,attrs:{perm:c}}),u=C.getInnerMostAxes(u.length,i)),C.assertAxesAreInnerMostDims("all",u,i);let[m,f]=C.computeOutAndReduceShapes(p.shape,u),d=y.sizeFromShape(f),h=ce({inputs:{x:p},backend:t,attrs:{shape:[-1,d]}}),g=En(h,h.dtype,"all",t),x;if(a){let b=C.expandShapeToKeepDim(m,l);x=ce({inputs:{x:g},backend:t,attrs:{shape:b}})}else x=ce({inputs:{x:g},backend:t,attrs:{shape:m}});return t.disposeIntermediateTensorInfo(h),t.disposeIntermediateTensorInfo(g),c!=null&&t.disposeIntermediateTensorInfo(p),x}var x$={kernelName:Ci,backendName:"webgl",kernelFunc:YX};function ZX(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n,i=o.shape.length,l=y.parseAxisParam(s,o.shape),u=l,c=C.getAxesPermutation(u,i),p=o;c!=null&&(p=Rt({inputs:{x:o},backend:t,attrs:{perm:c}}),u=C.getInnerMostAxes(u.length,i)),C.assertAxesAreInnerMostDims("any",u,i);let[m,f]=C.computeOutAndReduceShapes(p.shape,u),d=y.sizeFromShape(f),h=ce({inputs:{x:p},backend:t,attrs:{shape:[-1,d]}}),g=En(h,h.dtype,"any",t),x;if(a){let b=C.expandShapeToKeepDim(m,l);x=ce({inputs:{x:g},backend:t,attrs:{shape:b}})}else x=ce({inputs:{x:g},backend:t,attrs:{shape:m}});return t.disposeIntermediateTensorInfo(h),t.disposeIntermediateTensorInfo(g),c!=null&&t.disposeIntermediateTensorInfo(p),x}var y$={kernelName:Ii,backendName:"webgl",kernelFunc:ZX};var yv=class{constructor(e,t,n){this.variableNames=["A"];let{windowSize:o,batchSize:s,outSize:a}=e;n||this.variableNames.push("bestIndicesA"),this.outputShape=[s,a];let i=t==="max"?">":"<",l=n?"inOffset + i;":"round(getBestIndicesA(batch, inOffset + i));";this.userCode=`
|
|
void main() {
|
|
ivec2 coords = getOutputCoords();
|
|
int batch = coords[0];
|
|
int outIdx = coords[1];
|
|
int inOffset = outIdx * ${o};
|
|
|
|
int bestIndex = inOffset;
|
|
float bestValue = getA(batch, bestIndex);
|
|
|
|
for (int i = 0; i < ${o}; i++) {
|
|
int inIdx = ${l};
|
|
float candidate = getA(batch, inIdx);
|
|
if (candidate ${i} bestValue) {
|
|
bestValue = candidate;
|
|
bestIndex = inIdx;
|
|
}
|
|
}
|
|
setOutput(float(bestIndex));
|
|
}
|
|
`}};var bv=class{constructor(e,t,n,o){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!0,y.assert(e.length>2,()=>`Packed arg${n.charAt(0).toUpperCase()+n.slice(1)} supports only inputs with rank above 2.`);let s=e[e.length-1],a=Math.ceil(s/t);this.outputShape=e.slice(0,-1),a>1&&this.outputShape.push(a),o||this.variableNames.push("bestIndicesA");let i=this.outputShape,l=i.length,u=Ve(l),c=jt("coords",l),p,m;if(a===1){m=l+1;let $=Ve(m);p=`
|
|
${$} sourceLocR = ${$}(${c.join()}, 0);
|
|
++${c[l-1]};
|
|
${$} sourceLocG = ${$}(${c.join()}, 0);
|
|
++${c[l-2]};
|
|
${$} sourceLocA = ${$}(${c.join()}, 0);
|
|
--${c[l-1]};
|
|
${$} sourceLocB = ${$}(${c.join()}, 0);
|
|
--${c[l-2]};`}else m=l,p=`
|
|
${u} sourceLocR = coords;
|
|
++${c[l-1]};
|
|
${u} sourceLocG = coords;
|
|
++${c[l-2]};
|
|
${u} sourceLocA = coords;
|
|
--${c[l-1]};
|
|
${u} sourceLocB = coords;
|
|
--${c[l-2]};`;let f=["x","y","z","w","u","v"].slice(0,m),d="."+f[m-1],h=f.map($=>"int "+$),g=jt("sourceLocR",m-1).concat("inIdx.r"),x=jt("sourceLocG",m-1).concat("inIdx.g"),b=jt("sourceLocB",m-1).concat("inIdx.b"),w=jt("sourceLocA",m-1).concat("inIdx.a"),_=n==="max"?"greaterThan":"lessThan",I=o?"":`
|
|
inIdx = round(vec4(getBestIndicesAChannel(${g.join()}),
|
|
getBestIndicesAChannel(${x.join()}),
|
|
getBestIndicesAChannel(${b.join()}),
|
|
getBestIndicesAChannel(${w.join()})));`,E=`vec4(
|
|
getAChannel(${g.join()}),
|
|
hasNextCol ? getAChannel(${x.join()}) : 0.,
|
|
hasNextRow ? getAChannel(${b.join()}) : 0.,
|
|
hasNextRow && hasNextCol ? getAChannel(${w.join()}) : 0.)`,D=o?"":`
|
|
float getBestIndicesAChannel(${h.join()}) {
|
|
return getChannel(getBestIndicesA(${f.join()}),
|
|
vec2(${f.slice(-2).join()}));
|
|
}`;this.userCode=`
|
|
float getAChannel(${h.join()}) {
|
|
return getChannel(getA(${f.join()}),
|
|
vec2(${f.slice(-2).join()}));
|
|
}
|
|
${D}
|
|
void main() {
|
|
${u} coords = getOutputCoords();
|
|
bool hasNextCol = ${c[l-1]} < ${i[l-1]-1};
|
|
bool hasNextRow = ${c[l-2]} < ${i[l-2]-1};
|
|
${p}
|
|
ivec4 srcIdx = ivec4(sourceLocR${d}, sourceLocG${d},
|
|
sourceLocB${d}, sourceLocA${d}) * ${t};
|
|
ivec4 inIdx = srcIdx;
|
|
vec4 bestIndex = vec4(inIdx);
|
|
vec4 bestValue = ${E};
|
|
|
|
for (int i = 0; i < ${t}; i++) {
|
|
inIdx = srcIdx;
|
|
${I}
|
|
vec4 candidate = ${E};
|
|
bvec4 nan = isnan(candidate);
|
|
bvec4 replace = bvec4(
|
|
vec4(${_}(candidate, bestValue)) * (vec4(1.0) - vec4(nan)));
|
|
|
|
bestValue = vec4(replace.x ? candidate.x : bestValue.x,
|
|
replace.y ? candidate.y : bestValue.y,
|
|
replace.z ? candidate.z : bestValue.z,
|
|
replace.w ? candidate.w : bestValue.w);
|
|
bestIndex = mix(bestIndex, vec4(inIdx), vec4(replace));
|
|
srcIdx++;
|
|
}
|
|
setOutput(bestIndex);
|
|
}
|
|
`}};function b$(r,e,t,n=null){let o=e.shape[0],s=e.shape[1];n!=null&&(o=n.shape[0],s=n.shape[1]);let a=C.computeOptimalWindowSize(s),i={windowSize:a,inSize:s,batchSize:o,outSize:Math.ceil(s/a)},l=new yv(i,t,n==null),u=[e];n!=null&&u.push(n);let c=r.runWebGLProgram(l,u,"int32");if(c.shape[1]===1)return c;let p=b$(r,e,t,c);return r.disposeIntermediateTensorInfo(c),p}function w$(r,e,t,n=null){let o=n!=null?n.shape:e.shape,s=o[o.length-1],a=C.computeOptimalWindowSize(s),i=new bv(o,a,t,n==null),l=n==null?[e]:[e,n],u=r.runWebGLProgram(i,l,"int32");if(u.shape.length===e.shape.length){let c=w$(r,e,t,u);return r.disposeIntermediateTensorInfo(u),c}return u}function wx(r,e,t,n){let o=[t];if(C.assertAxesAreInnerMostDims("arg"+n.charAt(0).toUpperCase()+n.slice(1),o,e.shape.length),!W().getBool("WEBGL_PACK_REDUCE")||e.shape.length<=2){let s=[],[a,i]=C.computeOutAndReduceShapes(e.shape,o),l=y.sizeFromShape(i),u=ce({inputs:{x:e},backend:r,attrs:{shape:[-1,l]}});s.push(u);let c=b$(r,u,n);s.push(c);let p=ce({inputs:{x:c},backend:r,attrs:{shape:a}});return s.forEach(m=>r.disposeIntermediateTensorInfo(m)),p}return w$(r,e,n)}function JX(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s}=n,a=y.parseAxisParam(s,o.shape),i=C.getAxesPermutation(a,o.shape.length),l=o,u=[];i!=null&&(l=Rt({inputs:{x:o},backend:t,attrs:{perm:i}}),u.push(l),a=C.getInnerMostAxes(a.length,l.shape.length)),C.assertAxesAreInnerMostDims("argMax",[a[0]],l.shape.length);let c=wx(t,l,a[0],"max");return u.forEach(p=>t.disposeIntermediateTensorInfo(p)),c}var _$={kernelName:ko,backendName:"webgl",kernelFunc:JX};function QX(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s}=n,a=y.parseAxisParam(s,o.shape),i=C.getAxesPermutation(a,o.shape.length),l=o,u=[];i!=null&&(l=Rt({inputs:{x:o},backend:t,attrs:{perm:i}}),u.push(l),a=C.getInnerMostAxes(a.length,l.shape.length)),C.assertAxesAreInnerMostDims("argMin",[a[0]],l.shape.length);let c=wx(t,l,a[0],"min");return u.forEach(p=>t.disposeIntermediateTensorInfo(p)),c}var k$={kernelName:Ya,backendName:"webgl",kernelFunc:QX};var e7=br+`
|
|
if (abs(x) > 1.) {
|
|
return NAN;
|
|
}
|
|
return asin(x);
|
|
`,t7=_e({opSnippet:e7}),v$={kernelName:Ni,backendName:"webgl",kernelFunc:t7};var r7=br+"return log(x + sqrt(x * x + 1.0));",n7=_e({opSnippet:r7}),C$={kernelName:Si,backendName:"webgl",kernelFunc:n7};var o7=br+`
|
|
return atan(x);
|
|
`,s7=_e({opSnippet:o7}),I$={kernelName:Ti,backendName:"webgl",kernelFunc:s7};var i7=e$+`
|
|
return atan(a, b);
|
|
`,a7=`
|
|
vec4 result = atan(a, b);
|
|
vec4 isNaN = min(vec4(isnan(a)) + vec4(isnan(b)), vec4(1.0));
|
|
`+t$+`
|
|
return result;
|
|
`,l7=st({opSnippet:i7,packedOpSnippet:a7}),N$={kernelName:Ei,backendName:"webgl",kernelFunc:l7};var u7=br+`
|
|
if ((x < -1.0) || (x > 1.0)) return NAN;
|
|
return (log(1.0 + x) - log(1.0 - x)) / 2.0;`,c7=_e({opSnippet:u7}),S$={kernelName:Ai,backendName:"webgl",kernelFunc:c7};var ci=class{constructor(e,t,n,o=!1,s=!1){if(this.variableNames=["x"],t==="avg"&&n)throw new Error("Cannot compute positions for average pool.");let a=e.filterWidth,i=e.strideHeight,l=e.strideWidth,u=e.dilationHeight,c=e.dilationWidth,p=e.effectiveFilterHeight,m=e.effectiveFilterWidth,f=e.padInfo.top,d=e.padInfo.left;this.outputShape=e.outShape;let h=t==="avg",g=`((batch * ${e.inHeight} + xR) * ${e.inWidth} + xC) * ${e.inChannels} + d`,x=`(xR * ${e.inWidth} + xC) * ${e.inChannels} + d`,b="0.0";if(h||(b="-1.0 / 1e-20"),n){let $=">=";this.userCode=`
|
|
const ivec2 strides = ivec2(${i}, ${l});
|
|
const ivec2 pads = ivec2(${f}, ${d});
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int batch = coords[0];
|
|
int d = coords[3];
|
|
|
|
ivec2 xRCCorner = coords.yz * strides - pads;
|
|
int xRCorner = xRCCorner.x;
|
|
int xCCorner = xRCCorner.y;
|
|
|
|
// max/min x(?, ?, d) to get y(yR, yC, d).
|
|
// ? = to be determined
|
|
float minMaxValue = 0.0;
|
|
float minMaxValueFound = 0.0;
|
|
int minMaxPosition = 0;
|
|
float avgValue = 0.0;
|
|
|
|
for (int wR = 0; wR < ${p};
|
|
wR += ${u}) {
|
|
int xR = xRCorner + wR;
|
|
|
|
if (xR < 0 || xR >= ${e.inHeight}) {
|
|
continue;
|
|
}
|
|
|
|
for (int wC = 0; wC < ${m};
|
|
wC += ${c}) {
|
|
int xC = xCCorner + wC;
|
|
|
|
if (xC < 0 || xC >= ${e.inWidth}) {
|
|
continue;
|
|
}
|
|
|
|
float value = getX(batch, xR, xC, d);
|
|
|
|
// If a min / max value has already been found, use it. If not,
|
|
// use the current value.
|
|
float currMinMaxValue = mix(
|
|
value, minMaxValue, minMaxValueFound);
|
|
if (value ${$} currMinMaxValue) {
|
|
minMaxValue = value;
|
|
minMaxValueFound = 1.0;
|
|
minMaxPosition = ${o?s?g:x:`wR * ${m} + wC`};
|
|
}
|
|
}
|
|
}
|
|
setOutput(float(minMaxPosition));
|
|
}
|
|
`;return}let w="max",_=`${t}(${t}(${t}(minMaxValue[0], minMaxValue[1]), minMaxValue[2]), minMaxValue[3])`;t==="avg"&&(_="avgValue / count");let I=Math.floor(a/4)*4,E=a%4,D=`
|
|
if (${h}) {
|
|
avgValue += dot(values, ones);
|
|
} else {
|
|
minMaxValue = ${w}(values, minMaxValue);
|
|
}
|
|
`;this.userCode=`
|
|
const ivec2 strides = ivec2(${i}, ${l});
|
|
const ivec2 pads = ivec2(${f}, ${d});
|
|
const float initializationValue = ${b};
|
|
const vec4 ones = vec4(1.0, 1.0, 1.0, 1.0);
|
|
|
|
float count = 0.0;
|
|
|
|
float getValue(int batch, int xR, int xC, int d) {
|
|
if (xC < 0 || xC >= ${e.inWidth}) {
|
|
return initializationValue;
|
|
}
|
|
count += 1.0;
|
|
return getX(batch, xR, xC, d);
|
|
}
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int batch = coords[0];
|
|
int d = coords[3];
|
|
|
|
ivec2 xRCCorner = coords.yz * strides - pads;
|
|
int xRCorner = xRCCorner.x;
|
|
int xCCorner = xRCCorner.y;
|
|
|
|
// max/min x(?, ?, d) to get y(yR, yC, d).
|
|
// ? = to be determined
|
|
vec4 minMaxValue = vec4(${b});
|
|
float avgValue = 0.0;
|
|
count = 0.0;
|
|
|
|
for (int wR = 0; wR < ${p};
|
|
wR += ${u}) {
|
|
int xR = xRCorner + wR;
|
|
|
|
if (xR < 0 || xR >= ${e.inHeight}) {
|
|
continue;
|
|
}
|
|
|
|
for (int wC = 0; wC < ${I}; wC += 4) {
|
|
int xC = xCCorner + wC * ${c};
|
|
|
|
vec4 values = vec4(
|
|
getValue(batch, xR, xC, d),
|
|
getValue(batch, xR, xC + ${c}, d),
|
|
getValue(batch, xR, xC + 2 * ${c}, d),
|
|
getValue(batch, xR, xC + 3 * ${c}, d)
|
|
);
|
|
|
|
${D}
|
|
}
|
|
|
|
int xC = xCCorner + ${I};
|
|
if (${E===1}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, xR, xC, d),
|
|
initializationValue,
|
|
initializationValue,
|
|
initializationValue
|
|
);
|
|
|
|
${D}
|
|
} else if (${E===2}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, xR, xC, d),
|
|
getValue(batch, xR, xC + ${c}, d),
|
|
initializationValue,
|
|
initializationValue
|
|
);
|
|
|
|
${D}
|
|
} else if (${E===3}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, xR, xC, d),
|
|
getValue(batch, xR, xC + ${c}, d),
|
|
getValue(batch, xR, xC + 2 * ${c}, d),
|
|
initializationValue
|
|
);
|
|
|
|
${D}
|
|
}
|
|
}
|
|
setOutput(${_});
|
|
}
|
|
`}},Vu=class{constructor(e,t,n,o=!1,s=!1){if(this.variableNames=["x"],t==="avg"&&n)throw new Error("Cannot compute positions for average pool.");let a=e.filterWidth,i=e.strideDepth,l=e.strideHeight,u=e.strideWidth,c=e.dilationDepth,p=e.dilationHeight,m=e.dilationWidth,f=e.effectiveFilterDepth,d=e.effectiveFilterHeight,h=e.effectiveFilterWidth,g=e.padInfo.front,x=e.padInfo.top,b=e.padInfo.left;this.outputShape=e.outShape;let w=t==="avg",_="0.0";if(w||(_="-1.0 / 1e-20"),n){let M=">=";this.userCode=`
|
|
const ivec3 strides =
|
|
ivec3(${i}, ${l}, ${u});
|
|
const ivec3 pads = ivec3(${g}, ${x}, ${b});
|
|
|
|
void main() {
|
|
ivec5 coords = getOutputCoords();
|
|
int batch = coords.x;
|
|
int ch = coords.u;
|
|
|
|
ivec3 xCorner = ivec3(coords.y, coords.z, coords.w) * strides - pads;
|
|
int xDCorner = xCorner.x;
|
|
int xRCorner = xCorner.y;
|
|
int xCCorner = xCorner.z;
|
|
|
|
// max/min x(?, ?, ?, ch) to get y(yD, yR, yC, ch).
|
|
// ? = to be determined
|
|
float minMaxValue = 0.0;
|
|
float minMaxValueFound = 0.0;
|
|
int minMaxPosition = 0;
|
|
|
|
for (int wD = 0; wD < ${f};
|
|
wD += ${c}) {
|
|
int xD = xDCorner + wD;
|
|
|
|
if (xD < 0 || xD >= ${e.inDepth}) {
|
|
continue;
|
|
}
|
|
|
|
for (int wR = 0; wR < ${d};
|
|
wR += ${p}) {
|
|
int xR = xRCorner + wR;
|
|
|
|
if (xR < 0 || xR >= ${e.inHeight}) {
|
|
continue;
|
|
}
|
|
|
|
for (int wC = 0; wC < ${h};
|
|
wC += ${m}) {
|
|
int xC = xCCorner + wC;
|
|
|
|
if (xC < 0 || xC >= ${e.inWidth}) {
|
|
continue;
|
|
}
|
|
|
|
float value = getX(batch, xD, xR, xC, ch);
|
|
|
|
// If a min / max value has already been found, use it. If not,
|
|
// use the current value.
|
|
float currMinMaxValue = mix(
|
|
value, minMaxValue, minMaxValueFound);
|
|
if (value ${M} currMinMaxValue) {
|
|
minMaxValue = value;
|
|
minMaxValueFound = 1.0;
|
|
minMaxPosition = ${o?s?`(((batch * ${e.inDepth} + xD) * ${e.inHeight} + xR) * ${e.inWidth} + xC) * ${e.inChannels} + ch`:`((xD * ${e.inHeight} + xR) * ${e.inWidth} + xC) * ${e.inChannels} + ch`:`wD * ${d} * ${h} +
|
|
wR * ${h} + wC`};
|
|
}
|
|
}
|
|
}
|
|
}
|
|
setOutput(float(minMaxPosition));
|
|
}
|
|
`;return}let I="max",E=`${t}(${t}(${t}(minMaxValue[0], minMaxValue[1]), minMaxValue[2]), minMaxValue[3])`;t==="avg"&&(E="avgValue / count");let D=Math.floor(a/4)*4,$=a%4,R=`
|
|
if (${w}) {
|
|
avgValue += dot(values, ones);
|
|
} else {
|
|
minMaxValue = ${I}(values, minMaxValue);
|
|
}
|
|
`;this.userCode=`
|
|
const ivec3 strides =
|
|
ivec3(${i}, ${l}, ${u});
|
|
const ivec3 pads = ivec3(${g}, ${x}, ${b});
|
|
const float initializationValue = ${_};
|
|
const vec4 ones = vec4(1.0, 1.0, 1.0, 1.0);
|
|
|
|
float count = 0.0;
|
|
|
|
float getValue(int batch, int xD, int xR, int xC, int ch) {
|
|
if (xC < 0 || xC >= ${e.inWidth}) {
|
|
return initializationValue;
|
|
}
|
|
count += 1.0;
|
|
return getX(batch, xD, xR, xC, ch);
|
|
}
|
|
|
|
void main() {
|
|
ivec5 coords = getOutputCoords();
|
|
int batch = coords.x;
|
|
int ch = coords.u;
|
|
|
|
ivec3 xCorner = ivec3(coords.y, coords.z, coords.w) * strides - pads;
|
|
int xDCorner = xCorner.x;
|
|
int xRCorner = xCorner.y;
|
|
int xCCorner = xCorner.z;
|
|
|
|
// max/min x(?, ?, ?, d) to get y(yD, yR, yC, ch).
|
|
// ? = to be determined
|
|
vec4 minMaxValue = vec4(${_});
|
|
float avgValue = 0.0;
|
|
count = 0.0;
|
|
|
|
for (int wD = 0; wD < ${f};
|
|
wD += ${c}) {
|
|
int xD = xDCorner + wD;
|
|
|
|
if (xD < 0 || xD >= ${e.inDepth}) {
|
|
continue;
|
|
}
|
|
|
|
for (int wR = 0; wR < ${d};
|
|
wR += ${p}) {
|
|
int xR = xRCorner + wR;
|
|
|
|
if (xR < 0 || xR >= ${e.inHeight}) {
|
|
continue;
|
|
}
|
|
|
|
for (int wC = 0; wC < ${D}; wC += 4) {
|
|
int xC = xCCorner + wC * ${m};
|
|
|
|
vec4 values = vec4(
|
|
getValue(batch, xD, xR, xC, ch),
|
|
getValue(batch, xD, xR, xC + ${m}, ch),
|
|
getValue(batch, xD, xR, xC + 2 * ${m}, ch),
|
|
getValue(batch, xD, xR, xC + 3 * ${m}, ch)
|
|
);
|
|
|
|
${R}
|
|
}
|
|
|
|
int xC = xCCorner + ${D};
|
|
if (${$===1}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, xD, xR, xC, ch),
|
|
initializationValue,
|
|
initializationValue,
|
|
initializationValue
|
|
);
|
|
|
|
${R}
|
|
} else if (${$===2}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, xD, xR, xC, ch),
|
|
getValue(batch, xD, xR, xC + ${m}, ch),
|
|
initializationValue,
|
|
initializationValue
|
|
);
|
|
|
|
${R}
|
|
} else if (${$===3}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, xD, xR, xC, ch),
|
|
getValue(batch, xD, xR, xC + ${m}, ch),
|
|
getValue(batch, xD, xR, xC + 2 * ${m}, ch),
|
|
initializationValue
|
|
);
|
|
|
|
${R}
|
|
}
|
|
}
|
|
setOutput(${E});
|
|
}
|
|
}
|
|
`}};function p7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e;Cs(o,"avgPool");let{filterSize:s,strides:a,pad:i,dimRoundingMode:l}=n,u=1;y.assert(C.eitherStridesOrDilationsAreOne(a,u),()=>`Error in avgPool: Either strides or dilations must be 1. Got strides ${a} and dilations '${u}'`);let c=C.computePool2DInfo(o.shape,s,a,u,i,l);if(c.filterWidth===1&&c.filterHeight===1&&y.arraysEqual(c.inShape,c.outShape))return Ut({inputs:{x:o},backend:t});let p=new ci(c,"avg",!1);return t.runWebGLProgram(p,[o],"float32")}var T$={kernelName:vo,backendName:"webgl",kernelFunc:p7};function m7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{filterSize:s,strides:a,pad:i,dimRoundingMode:l,dataFormat:u}=n,c=[1,1,1],p=C.computePool3DInfo(o.shape,s,a,c,i,l,u),m=new Vu(p,"avg",!1);return t.runWebGLProgram(m,[o],"float32")}var A$={kernelName:Za,backendName:"webgl",kernelFunc:m7};var wv=class{constructor(e){this.variableNames=["dy"],this.outputShape=e.inShape;let t=e.filterHeight,n=e.filterWidth,o=e.strideHeight,s=e.strideWidth,a=e.dilationHeight,i=e.dilationWidth,l=e.effectiveFilterHeight,u=e.effectiveFilterWidth,c=l-1-e.padInfo.top,p=u-1-e.padInfo.left,m=1/(t*n);this.userCode=`
|
|
const ivec2 pads = ivec2(${c}, ${p});
|
|
const float avgMultiplier = float(${m});
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int d = coords[3];
|
|
|
|
ivec2 dyRCCorner = coords.yz - pads;
|
|
int dyRCorner = dyRCCorner.x;
|
|
int dyCCorner = dyRCCorner.y;
|
|
|
|
// Convolve dy(?, ?, d) with pos mask(:, :, d) to get dx(xR, xC, d).
|
|
// ? = to be determined. : = across all values in that axis.
|
|
float dotProd = 0.0;
|
|
for (int wR = 0; wR < ${l};
|
|
wR += ${a}) {
|
|
float dyR = float(dyRCorner + wR) / ${o}.0;
|
|
|
|
if (dyR < 0.0 || dyR >= ${e.outHeight}.0 || fract(dyR) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyR = int(dyR);
|
|
|
|
for (int wC = 0; wC < ${u};
|
|
wC+= ${i}) {
|
|
float dyC = float(dyCCorner + wC) / ${s}.0;
|
|
|
|
if (dyC < 0.0 || dyC >= ${e.outWidth}.0 ||
|
|
fract(dyC) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyC = int(dyC);
|
|
|
|
float dyValue = getDy(b, idyR, idyC, d);
|
|
|
|
dotProd += dyValue * avgMultiplier;
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}},_v=class{constructor(e){this.variableNames=["dy"],this.outputShape=e.inShape;let t=e.filterDepth,n=e.filterHeight,o=e.filterWidth,s=e.strideDepth,a=e.strideHeight,i=e.strideWidth,l=e.dilationDepth,u=e.dilationHeight,c=e.dilationWidth,p=e.effectiveFilterDepth,m=e.effectiveFilterHeight,f=e.effectiveFilterWidth,d=p-1-e.padInfo.front,h=m-1-e.padInfo.top,g=f-1-e.padInfo.left,x=1/(t*n*o);this.userCode=`
|
|
const ivec3 pads = ivec3(${d}, ${h}, ${g});
|
|
const float avgMultiplier = float(${x});
|
|
|
|
void main() {
|
|
ivec5 coords = getOutputCoords();
|
|
int batch = coords.x;
|
|
int ch = coords.u;
|
|
|
|
ivec3 dyCorner = ivec3(coords.y, coords.z, coords.w) - pads;
|
|
int dyDCorner = dyCorner.x;
|
|
int dyRCorner = dyCorner.y;
|
|
int dyCCorner = dyCorner.z;
|
|
|
|
// Convolve dy(?, ?, ?, d) with pos mask(:, :, :, ch) to get
|
|
// dx(xD, xR, xC, ch).
|
|
// ? = to be determined. : = across all values in that axis.
|
|
float dotProd = 0.0;
|
|
|
|
for (int wD = 0; wD < ${p};
|
|
wD += ${l}) {
|
|
float dyD = float(dyDCorner + wD) / ${s}.0;
|
|
|
|
if (dyD < 0.0 || dyD >= ${e.outDepth}.0 || fract(dyD) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyD = int(dyD);
|
|
|
|
for (int wR = 0; wR < ${m};
|
|
wR += ${u}) {
|
|
float dyR = float(dyRCorner + wR) / ${a}.0;
|
|
|
|
if (dyR < 0.0 || dyR >= ${e.outHeight}.0 ||
|
|
fract(dyR) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyR = int(dyR);
|
|
|
|
for (int wC = 0; wC < ${f};
|
|
wC += ${c}) {
|
|
float dyC = float(dyCCorner + wC) / ${i}.0;
|
|
|
|
if (dyC < 0.0 || dyC >= ${e.outWidth}.0 ||
|
|
fract(dyC) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyC = int(dyC);
|
|
|
|
float dyValue = getDy(batch, idyD, idyR, idyC, ch);
|
|
|
|
dotProd += dyValue * avgMultiplier;
|
|
}
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}};function f7(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,input:s}=e,a=s,{filterSize:i,strides:l,pad:u,dimRoundingMode:c}=n,p=[1,1,1],m=C.computePool3DInfo(a.shape,i,l,p,u,c),f=new _v(m);return t.runWebGLProgram(f,[o],a.dtype)}var E$={kernelName:_c,backendName:"webgl",kernelFunc:f7};function d7(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,input:s}=e,a=s;Cs([o,s],"avgPoolGrad");let{filterSize:i,strides:l,pad:u}=n,c=C.computePool2DInfo(a.shape,i,l,1,u),p=new wv(c);return t.runWebGLProgram(p,[o],a.dtype)}var D$={kernelName:wc,backendName:"webgl",kernelFunc:d7};function h7(r){let{inputs:e,backend:t,attrs:n}=r,{a:o,b:s}=e,{transposeA:a,transposeB:i}=n;return Bu({a:o,b:s,transposeA:a,transposeB:i,backend:t})}var $$={kernelName:Co,backendName:"webgl",kernelFunc:h7};var kv=class{constructor(e,t,n,o,s,a){this.outputShape=[],this.variableNames=["x","mean","variance"],C.assertAndGetBroadcastShape(e,t),C.assertAndGetBroadcastShape(e,n);let i="0.0";o!=null&&(C.assertAndGetBroadcastShape(e,o),this.variableNames.push("offset"),i="getOffsetAtOutCoords()");let l="1.0";s!=null&&(C.assertAndGetBroadcastShape(e,s),this.variableNames.push("scale"),l="getScaleAtOutCoords()"),this.outputShape=e,this.userCode=`
|
|
void main() {
|
|
float x = getXAtOutCoords();
|
|
float mean = getMeanAtOutCoords();
|
|
float variance = getVarianceAtOutCoords();
|
|
float offset = ${i};
|
|
float scale = ${l};
|
|
float inv = scale * inversesqrt(variance + float(${a}));
|
|
setOutput(dot(vec3(x, -mean, offset), vec3(inv, inv, 1)));
|
|
}
|
|
`}};var vv=class{constructor(e,t,n,o,s,a){this.packedInputs=!0,this.packedOutput=!0,this.variableNames=["x","mean","variance"],C.assertAndGetBroadcastShape(e,t),C.assertAndGetBroadcastShape(e,n);let i="vec4(0.0)";o!=null&&(C.assertAndGetBroadcastShape(e,o),this.variableNames.push("offset"),i="getOffsetAtOutCoords()");let l="vec4(1.0)";s!=null&&(C.assertAndGetBroadcastShape(e,s),this.variableNames.push("scale"),l="getScaleAtOutCoords()"),this.outputShape=e,this.userCode=`
|
|
void main() {
|
|
vec4 offset = ${i};
|
|
vec4 scale = ${l};
|
|
|
|
vec4 x = getXAtOutCoords();
|
|
vec4 mean = getMeanAtOutCoords();
|
|
vec4 variance = getVarianceAtOutCoords();
|
|
|
|
vec4 inv = scale * inversesqrt(variance + vec4(${a}));
|
|
|
|
setOutput((x - mean) * inv + offset);
|
|
}
|
|
`}};var g7=({inputs:r,backend:e,attrs:t})=>{let{x:n,mean:o,variance:s,offset:a,scale:i}=r;y.assert(o.shape.length===s.shape.length,()=>"Batch normalization gradient requires mean and variance to have equal ranks."),y.assert(a==null||o.shape.length===a.shape.length,()=>"Batch normalization gradient requires mean and offset to have equal ranks."),y.assert(i==null||o.shape.length===i.shape.length,()=>"Batch normalization gradient requires mean and scale to have equal ranks.");let{varianceEpsilon:l}=t;l==null&&(l=.001);let u=[n,o,s],c=null;a!=null&&(c=a.shape,u.push(a));let p=null;i!=null&&(p=i.shape,u.push(i));let m=W().getBool("WEBGL_PACK_NORMALIZATION")?new vv(n.shape,o.shape,s.shape,c,p,l):new kv(n.shape,o.shape,s.shape,c,p,l);return e.runWebGLProgram(m,u,u[0].dtype)},R$={kernelName:Oo,backendName:"webgl",kernelFunc:g7};var Cv=class{constructor(e){this.variableNames=["source"],this.outputShape=e,this.rank=e.length;let t=Ve(this.rank),n=`uniform int start[${this.rank}];`,o=x7(this.rank),s,a=e.map((i,l)=>`sourceLoc.${Iv[l]} = start[${l}] + coords.${Iv[l]};`);s=`
|
|
${t} sourceLoc;
|
|
${t} coords = getOutputCoords();
|
|
${a.join(`
|
|
`)}
|
|
`,this.userCode=`
|
|
${n}
|
|
void main() {
|
|
${s}
|
|
setOutput(getSource(${o}));
|
|
}
|
|
`}getCustomSetupFunc(e){if(e.length!==this.rank)throw Error(`The rank (${this.rank}) of the program must match the length of start (${e.length})`);return(t,n)=>{this.startLoc==null&&(this.startLoc=t.getUniformLocationNoThrow(n,"start"),this.startLoc==null)||t.gl.uniform1iv(this.startLoc,e)}}},Iv=["x","y","z","w","u","v"];function x7(r){if(r===1)return"sourceLoc";if(r<=6)return Iv.slice(0,r).map(e=>"sourceLoc."+e).join(",");throw Error(`Slicing for rank ${r} is not yet supported`)}var Nv=class{constructor(e){this.variableNames=["source"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=e,this.rank=e.length;let t=Ve(this.rank),n=jt("coords",this.rank),o=jt("sourceLoc",this.rank),s=this.rank===1?"sourceLoc":`vec2(${o.slice(-2).join()})`,a=`getChannel(getSource(${o.join()}), ${s})`,i=`
|
|
result.x = ${a};
|
|
if (++${n[this.rank-1]} < ${e[this.rank-1]}) {
|
|
++${o[this.rank-1]};
|
|
result.y = ${a};
|
|
--${o[this.rank-1]};
|
|
}
|
|
`,l=this.rank===1?"":`
|
|
--${n[this.rank-1]};
|
|
if (++${n[this.rank-2]} < ${e[this.rank-2]}) {
|
|
++${o[this.rank-2]};
|
|
result.z = ${a};
|
|
if (++${n[this.rank-1]} < ${e[this.rank-1]}) {
|
|
++${o[this.rank-1]};
|
|
result.w = ${a};
|
|
}
|
|
}
|
|
`,u=this.rank<=4?`sourceLoc = coords +
|
|
${t}(${e.map((c,p)=>`start[${p}]`).join()});`:e.map((c,p)=>`${o[p]} = ${n[p]} + start[${p}];`).join(`
|
|
`);this.userCode=`
|
|
uniform int start[${this.rank}];
|
|
void main() {
|
|
${t} coords = getOutputCoords();
|
|
${t} sourceLoc;
|
|
${u}
|
|
vec4 result = vec4(0.);
|
|
${i}
|
|
${l}
|
|
setOutput(result);
|
|
}
|
|
`}getCustomSetupFunc(e){if(e.length!==this.rank)throw Error(`The rank (${this.rank}) of the program must match the length of start (${e.length})`);return(t,n)=>{this.startLoc==null&&(this.startLoc=t.getUniformLocationNoThrow(n,"start"),this.startLoc==null)||t.gl.uniform1iv(this.startLoc,e)}}};function y7(r,e,t,n){let o=n.texData.get(r.dataId),s=n.makeTensorInfo(t,r.dtype),a=n.texData.get(s.dataId);Object.assign(a,o),a.refCount=1,a.shape=t,a.dtype=r.dtype;let i=or.computeFlatOffset(e,y.computeStrides(r.shape));o.slice&&(i+=o.slice.flatOffset),a.slice={flatOffset:i,origDataId:o.slice&&o.slice.origDataId||r.dataId};let l=n.dataRefCount.get(a.slice.origDataId)||1;return n.dataRefCount.set(a.slice.origDataId,l+1),s}function Fa(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{begin:s,size:a}=n,[i,l]=or.parseSliceParams(o,s,a);if(or.assertParamsValid(o,i,l),y.sizeFromShape(l)===0)return t.makeTensorInfo(l,o.dtype,[]);if(t.shouldExecuteOnCPU([o])||o.dtype==="string"){let p=t.texData.get(o.dataId),m=SD(p.values,i,l,o.shape,o.dtype);return t.makeTensorInfo(l,o.dtype,m)}let{isPacked:u}=t.texData.get(o.dataId),c=or.isSliceContinous(o.shape,i,l);if(u||!c){let p=W().getBool("WEBGL_PACK_ARRAY_OPERATIONS")?new Nv(l):new Cv(l),m=p.getCustomSetupFunc(i);return t.runWebGLProgram(p,[o],o.dtype,m)}return t.uploadToGPU(o.dataId),y7(o,i,l,t)}var F$={kernelName:Xs,backendName:"webgl",kernelFunc:Fa};var b7=r=>{let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{blockShape:s,crops:a}=n;y.assert(o.shape.length<=4,()=>"batchToSpaceND for rank > 4 with a WebGL backend not implemented yet");let i=s.reduce((b,w)=>b*w),l=C.getReshaped(o.shape,s,i),u=C.getPermuted(l.length,s.length),c=C.getReshapedPermuted(o.shape,s,i),p=C.getSliceBeginCoords(a,s.length),m=C.getSliceSize(c,a,s.length),f=[],d=ce({inputs:{x:o},backend:t,attrs:{shape:l}}),h=Rt({inputs:{x:d},backend:t,attrs:{perm:u}}),g=ce({inputs:{x:h},backend:t,attrs:{shape:c}}),x=Fa({inputs:{x:g},backend:t,attrs:{begin:p,size:m}});return f.push(d),f.push(h),f.push(g),f.forEach(b=>t.disposeIntermediateTensorInfo(b)),x},O$={kernelName:Ja,backendName:"webgl",kernelFunc:b7};function w7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,weights:s}=e,{size:a}=n,i=t.readSync(o.dataId),l=t.readSync(s.dataId),u=mx(i,l,s.dtype,s.shape,a);return t.makeTensorInfo([a],s.dtype,u)}var P$={kernelName:kc,backendName:"webgl",kernelFunc:w7};var _7="return float(a != b);",Sv=st({opSnippet:_7,dtype:"bool"}),M$={kernelName:Xi,backendName:"webgl",kernelFunc:Sv};function Oa(r){let{inputs:e,backend:t}=r,{input:n}=e,o=t.texData.get(n.dataId);return Ut({inputs:{x:o.complexTensorInfos.real},backend:t})}var L$={kernelName:Gc,backendName:"webgl",kernelFunc:Oa};var k7="return float(int(x));";function z$(r,e){let t=new xn(r.shape,k7),n=e.runWebGLProgram(t,[r],"int32");return{dataId:n.dataId,shape:n.shape,dtype:n.dtype}}function Tv(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{dtype:s}=n;if(s==="complex64"){if(o.dtype==="complex64")return Ut({inputs:{x:o},backend:t});let a=ht(o.shape),i=Tv({inputs:{x:o},backend:t,attrs:{dtype:"float32"}}),l=yn({inputs:{real:i,imag:a},backend:t});return a.dispose(),t.disposeIntermediateTensorInfo(i),l}if(o.dtype==="complex64"){let a=Oa({inputs:{input:o},backend:t}),i=Tv({inputs:{x:a},backend:t,attrs:{dtype:s}});return t.disposeIntermediateTensorInfo(a),i}if(!y.hasEncodingLoss(o.dtype,s)){let a=Ut({inputs:{x:o},backend:t});return{dataId:a.dataId,shape:a.shape,dtype:s}}if(s==="int32")return z$(o,t);if(s==="bool"){let a=t.makeTensorInfo([],"bool",y.getTypedArrayFromDType("bool",1)),l=Sv({inputs:{a:o,b:a},backend:t});return t.disposeIntermediateTensorInfo(a),l}throw new Error(`Error in Cast: failed to cast ${o.dtype} to ${s}`)}var B$={kernelName:qn,backendName:"webgl",kernelFunc:Tv};var V$="return ceil(x);",v7=_e({opSnippet:V$,packedOpSnippet:V$,cpuKernelImpl:uD}),G$={kernelName:Io,backendName:"webgl",kernelFunc:v7};var Av=class{constructor(e){this.variableNames=["A"],this.outputShape=e,this.userCode=`
|
|
uniform float minVal;
|
|
uniform float maxVal;
|
|
|
|
void main() {
|
|
float value = getAAtOutCoords();
|
|
if (isnan(value)) {
|
|
setOutput(value);
|
|
return;
|
|
}
|
|
|
|
setOutput(clamp(value, minVal, maxVal));
|
|
}
|
|
`}getCustomSetupFunc(e,t){return(n,o)=>{this.minLoc==null&&(this.minLoc=n.getUniformLocationNoThrow(o,"minVal"),this.maxLoc=n.getUniformLocationNoThrow(o,"maxVal")),n.gl.uniform1f(this.minLoc,e),n.gl.uniform1f(this.maxLoc,t)}}};var Ev=class{constructor(e){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=e,this.userCode=`
|
|
uniform float minVal;
|
|
uniform float maxVal;
|
|
|
|
void main() {
|
|
vec4 value = getAAtOutCoords();
|
|
|
|
if (any(isnan(value))) {
|
|
setOutput(value);
|
|
return;
|
|
}
|
|
|
|
setOutput(clamp(value, vec4(minVal), vec4(maxVal)));
|
|
}
|
|
`}getCustomSetupFunc(e,t){return(n,o)=>{this.minLoc==null&&(this.minLoc=n.getUniformLocationNoThrow(o,"minVal"),this.maxLoc=n.getUniformLocationNoThrow(o,"maxVal")),n.gl.uniform1f(this.minLoc,e),n.gl.uniform1f(this.maxLoc,t)}}};function C7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{clipValueMin:s,clipValueMax:a}=n,i;W().getBool("WEBGL_PACK_CLIP")?i=new Ev(o.shape):i=new Av(o.shape);let l=i.getCustomSetupFunc(s,a);return t.runWebGLProgram(i,[o],o.dtype,l)}var W$={kernelName:Kn,backendName:"webgl",kernelFunc:C7};var Dv=class{constructor(e){this.variableNames=["real","imag"],this.outputShape=e,this.userCode=`
|
|
void main() {
|
|
float re = abs(getRealAtOutCoords());
|
|
float im = abs(getImagAtOutCoords());
|
|
float mx = max(re, im);
|
|
|
|
// sadly the length function in glsl is not underflow-safe
|
|
// (at least not on Intel GPUs). So the safe solution is
|
|
// to ensure underflow-safety in all cases.
|
|
setOutput(
|
|
mx == 0.0 ? 0.0 : mx * length(vec2(1, min(re, im)/mx))
|
|
);
|
|
}
|
|
`}};function j$(r,e){return{dataId:e.dataId,dtype:e.dtype,shape:r.shape}}function I7(r){let{inputs:e,backend:t}=r,{x:n}=e,o=t.texData.get(n.dataId),s=new Dv(n.shape),a=[j$(n,o.complexTensorInfos.real),j$(n,o.complexTensorInfos.imag)];return t.runWebGLProgram(s,a,a[0].dtype)}var U$={kernelName:Qa,backendName:"webgl",kernelFunc:I7};var $v=class{constructor(e){this.outputShape=[],this.outputShape=C.computeOutShape(e,1),this.variableNames=e.map((a,i)=>`T${i}`);let t=new Array(e.length-1);t[0]=e[0][1];for(let a=1;a<t.length;a++)t[a]=t[a-1]+e[a][1];let n=[`if (yC < ${t[0]}) setOutput(getT0(yR, yC));`];for(let a=1;a<t.length;a++){let i=t[a-1];n.push(`else if (yC < ${t[a]}) setOutput(getT${a}(yR, yC-${i}));`)}let o=t.length,s=t[t.length-1];n.push(`else setOutput(getT${o}(yR, yC-${s}));`),this.userCode=`
|
|
void main() {
|
|
ivec2 coords = getOutputCoords();
|
|
int yR = coords.x;
|
|
int yC = coords.y;
|
|
|
|
${n.join(`
|
|
`)}
|
|
}
|
|
`}};var Rv=class{constructor(e,t){this.packedInputs=!0,this.packedOutput=!0,this.outputShape=[],this.outputShape=C.computeOutShape(e,t);let n=this.outputShape,o=n.length,s=Ve(o),a=jt("coords",o),i=["x","y","z","w","u","v"].slice(0,o);this.variableNames=e.map((h,g)=>`T${g}`);let l=new Array(e.length-1);l[0]=e[0][t];for(let h=1;h<l.length;h++)l[h]=l[h-1]+e[h][t];let u=i[t],c=i.slice(-2),p=i.join(),m=`if (${u} < ${l[0]}) {
|
|
return getChannel(
|
|
getT0(${p}), vec2(${c.join()}));
|
|
}`;for(let h=1;h<l.length;h++){let g=l[h-1];m+=`
|
|
if (${u} < ${l[h]} && ${u} >= ${l[h-1]}) {
|
|
return getChannel(
|
|
getT${h}(${_x(i,u,g)}),
|
|
vec2(${_x(c,u,g)}));
|
|
}`}let f=l.length,d=l[l.length-1];m+=`
|
|
return getChannel(
|
|
getT${f}(${_x(i,u,d)}),
|
|
vec2(${_x(c,u,d)}));`,this.userCode=`
|
|
float getValue(${i.map(h=>"int "+h)}) {
|
|
${m}
|
|
}
|
|
|
|
void main() {
|
|
${s} coords = getOutputCoords();
|
|
vec4 result = vec4(getValue(${a}), 0., 0., 0.);
|
|
|
|
${a[o-1]} = ${a[o-1]} + 1;
|
|
if (${a[o-1]} < ${n[o-1]}) {
|
|
result.g = getValue(${a});
|
|
}
|
|
|
|
${a[o-2]} = ${a[o-2]} + 1;
|
|
if (${a[o-2]} < ${n[o-2]}) {
|
|
result.a = getValue(${a});
|
|
}
|
|
|
|
${a[o-1]} = ${a[o-1]} - 1;
|
|
if (${a[o-2]} < ${n[o-2]} &&
|
|
${a[o-1]} < ${n[o-1]}) {
|
|
result.b = getValue(${a});
|
|
}
|
|
setOutput(result);
|
|
}
|
|
`}};function _x(r,e,t){let n=r.indexOf(e);return r.map((s,a)=>a===n?`${s} - ${t}`:s).join()}function Gu(r){let{inputs:e,backend:t}=r,{input:n}=e,o=t.texData.get(n.dataId);return Ut({inputs:{x:o.complexTensorInfos.imag},backend:t})}var H$={kernelName:Oc,backendName:"webgl",kernelFunc:Gu};function Wu(r,e,t){let n=r[0].dtype;if(n==="complex64"){let c=r.map(h=>Oa({inputs:{input:h},backend:t})),p=r.map(h=>Gu({inputs:{input:h},backend:t})),m=Wu(c,e,t),f=Wu(p,e,t),d=yn({inputs:{real:m,imag:f},backend:t});return c.forEach(h=>t.disposeIntermediateTensorInfo(h)),p.forEach(h=>t.disposeIntermediateTensorInfo(h)),t.disposeIntermediateTensorInfo(m),t.disposeIntermediateTensorInfo(f),d}let o=t.shouldExecuteOnCPU(r);if(n==="string"&&(o=!0),o){let c=r.map(x=>{let b=y.sizeFromShape(x.shape.slice(e));return ce({inputs:{x},backend:t,attrs:{shape:[-1,b]}})}),p=c.map(x=>({vals:t.readSync(x.dataId),shape:x.shape})),m=C.computeOutShape(c.map(x=>x.shape),1),f=c[0].shape[0]===1,d=cD(p,m,n,f),h=C.computeOutShape(r.map(x=>x.shape),e),g=t.makeTensorInfo(h,n,d);return c.forEach(x=>t.disposeIntermediateTensorInfo(x)),g}if(r.length>W().getNumber("WEBGL_MAX_TEXTURES_IN_SHADER")){let c=Math.floor(r.length/2),p=Wu(r.slice(0,c),e,t),m=Wu(r.slice(c),e,t),f=Wu([p,m],e,t);return t.disposeIntermediateTensorInfo(p),t.disposeIntermediateTensorInfo(m),f}if(W().getBool("WEBGL_PACK_ARRAY_OPERATIONS")&&r[0].shape.length>1){let c=new Rv(r.map(p=>p.shape),e);return t.runWebGLProgram(c,r,n)}let{tensors2D:s,outShape:a}=N7(r,e,t),i=new $v(s.map(c=>c.shape)),l=t.runWebGLProgram(i,s,n);s.forEach(c=>t.disposeIntermediateTensorInfo(c));let u=ce({inputs:{x:l},attrs:{shape:a},backend:t});return t.disposeIntermediateTensorInfo(l),u}function N7(r,e,t){let n=C.computeOutShape(r.map(s=>s.shape),e);return{tensors2D:r.map(s=>ce({inputs:{x:s},attrs:{shape:[-1,y.sizeFromShape(s.shape.slice(e))]},backend:t})),outShape:n}}function Fv(r){let{inputs:e,backend:t,attrs:n}=r,{axis:o}=n,s=y.parseAxisParam(o,e[0].shape)[0],a=C.computeOutShape(e.map(u=>u.shape),s);if(y.sizeFromShape(a)===0)return t.makeTensorInfo(a,e[0].dtype,[]);let i=e.filter(u=>y.sizeFromShape(u.shape)>0);if(i.length===1)return Ut({inputs:{x:i[0]},backend:t});let l=i.map(u=>u.shape);return C.assertParamsConsistent(l,s),Wu(i,s,t)}var q$={kernelName:Vs,backendName:"webgl",kernelFunc:Fv};var Bf=class{constructor(e,t=!1,n=null,o=!1,s=!1){this.variableNames=["x","W"],this.outputShape=e.outShape;let a=e.padInfo.top,i=e.padInfo.left,l=e.strideHeight,u=e.strideWidth,c=e.dilationHeight,p=e.dilationWidth,m=e.filterHeight,f=e.filterWidth,d=Math.floor(e.inChannels/4)*4,h=e.inChannels%4,g=e.dataFormat==="channelsLast",x=g?1:2,b=g?2:3,w=g?3:1,_="",I="";n&&(o?_=`float activation(float a) {
|
|
float b = getPreluActivationWeightsAtOutCoords();
|
|
${n}
|
|
}`:s?_=`float activation(float a) {
|
|
float b = getLeakyreluAlphaAtOutCoords();
|
|
${n}
|
|
}`:_=`
|
|
float activation(float x) {
|
|
${n}
|
|
}
|
|
`,I="result = activation(result);");let E=t?"result += getBiasAtOutCoords();":"";t&&this.variableNames.push("bias"),o&&this.variableNames.push("preluActivationWeights"),s&&this.variableNames.push("leakyreluAlpha"),this.userCode=`
|
|
${_}
|
|
|
|
const ivec2 strides = ivec2(${l}, ${u});
|
|
const ivec2 pads = ivec2(${a}, ${i});
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int batch = coords[0];
|
|
int d2 = coords[${w}];
|
|
|
|
ivec2 xRCCorner =
|
|
ivec2(coords[${x}], coords[${b}]) * strides - pads;
|
|
int xRCorner = xRCCorner.x;
|
|
int xCCorner = xRCCorner.y;
|
|
|
|
// Convolve x(?, ?, d1) with w(:, :, d1, d2) to get y(yR, yC, d2).
|
|
// ? = to be determined. : = across all values in that axis.
|
|
float dotProd = 0.0;
|
|
for (int wR = 0; wR < ${m}; wR++) {
|
|
int xR = xRCorner + wR * ${c};
|
|
|
|
if (xR < 0 || xR >= ${e.inHeight}) {
|
|
continue;
|
|
}
|
|
|
|
for (int wC = 0; wC < ${f}; wC++) {
|
|
int xC = xCCorner + wC * ${p};
|
|
|
|
if (xC < 0 || xC >= ${e.inWidth}) {
|
|
continue;
|
|
}
|
|
|
|
for (int d1 = 0; d1 < ${d}; d1 += 4) {
|
|
vec4 wValues = vec4(
|
|
getW(wR, wC, d1, d2),
|
|
getW(wR, wC, d1 + 1, d2),
|
|
getW(wR, wC, d1 + 2, d2),
|
|
getW(wR, wC, d1 + 3, d2)
|
|
);
|
|
|
|
if (${g}) {
|
|
vec4 xValues = vec4(
|
|
getX(batch, xR, xC, d1),
|
|
getX(batch, xR, xC, d1 + 1),
|
|
getX(batch, xR, xC, d1 + 2),
|
|
getX(batch, xR, xC, d1 + 3)
|
|
);
|
|
dotProd += dot(xValues, wValues);
|
|
} else {
|
|
vec4 xValues = vec4(
|
|
getX(batch, d1, xR, xC),
|
|
getX(batch, d1 + 1, xR, xC),
|
|
getX(batch, d1 + 2, xR, xC),
|
|
getX(batch, d1 + 3, xR, xC)
|
|
);
|
|
dotProd += dot(xValues, wValues);
|
|
}
|
|
}
|
|
|
|
if (${h===1}) {
|
|
|
|
if (${g}) {
|
|
dotProd +=
|
|
getX(batch, xR, xC, ${d}) *
|
|
getW(wR, wC, ${d}, d2);
|
|
} else {
|
|
dotProd +=
|
|
getX(batch, ${d}, xR, xC) *
|
|
getW(wR, wC, ${d}, d2);
|
|
}
|
|
|
|
} else if (${h===2}) {
|
|
vec2 wValues = vec2(
|
|
getW(wR, wC, ${d}, d2),
|
|
getW(wR, wC, ${d} + 1, d2)
|
|
);
|
|
|
|
if (${g}) {
|
|
vec2 xValues = vec2(
|
|
getX(batch, xR, xC, ${d}),
|
|
getX(batch, xR, xC, ${d} + 1)
|
|
);
|
|
dotProd += dot(xValues, wValues);
|
|
} else {
|
|
vec2 xValues = vec2(
|
|
getX(batch, ${d}, xR, xC),
|
|
getX(batch, ${d} + 1, xR, xC)
|
|
);
|
|
dotProd += dot(xValues, wValues);
|
|
}
|
|
|
|
} else if (${h===3}) {
|
|
vec3 wValues = vec3(
|
|
getW(wR, wC, ${d}, d2),
|
|
getW(wR, wC, ${d} + 1, d2),
|
|
getW(wR, wC, ${d} + 2, d2)
|
|
);
|
|
|
|
if (${g}) {
|
|
vec3 xValues = vec3(
|
|
getX(batch, xR, xC, ${d}),
|
|
getX(batch, xR, xC, ${d} + 1),
|
|
getX(batch, xR, xC, ${d} + 2)
|
|
);
|
|
dotProd += dot(xValues, wValues);
|
|
} else {
|
|
vec3 xValues = vec3(
|
|
getX(batch, ${d}, xR, xC),
|
|
getX(batch, ${d} + 1, xR, xC),
|
|
getX(batch, ${d} + 2, xR, xC)
|
|
);
|
|
dotProd += dot(xValues, wValues);
|
|
}
|
|
|
|
}
|
|
}
|
|
}
|
|
|
|
float result = dotProd;
|
|
${E}
|
|
${I}
|
|
setOutput(result);
|
|
}
|
|
`}},Ov=class{constructor(e){this.variableNames=["x","W"],this.outputShape=e.outShape;let t=e.padInfo.front,n=e.padInfo.top,o=e.padInfo.left,s=e.strideDepth,a=e.strideHeight,i=e.strideWidth,l=e.dilationDepth,u=e.dilationHeight,c=e.dilationWidth,p=e.filterDepth,m=e.filterHeight,f=e.filterWidth,d=Math.floor(e.inChannels/4)*4,h=e.inChannels%4;this.userCode=`
|
|
const ivec3 strides = ivec3(${s}, ${a}, ${i});
|
|
const ivec3 pads = ivec3(${t}, ${n}, ${o});
|
|
|
|
void main() {
|
|
ivec5 coords = getOutputCoords();
|
|
int batch = coords.x;
|
|
int d2 = coords.u;
|
|
|
|
ivec3 xFRCCorner = ivec3(coords.y, coords.z, coords.w) * strides - pads;
|
|
int xFCorner = xFRCCorner.x;
|
|
int xRCorner = xFRCCorner.y;
|
|
int xCCorner = xFRCCorner.z;
|
|
|
|
// Convolve x(?, ?, ?, d1) with w(:, :, :, d1, d2) to get
|
|
// y(yF, yR, yC, d2). ? = to be determined. : = across all
|
|
// values in that axis.
|
|
float dotProd = 0.0;
|
|
for (int wF = 0; wF < ${p}; wF++) {
|
|
int xF = xFCorner + wF * ${l};
|
|
|
|
if (xF < 0 || xF >= ${e.inDepth}) {
|
|
continue;
|
|
}
|
|
|
|
for (int wR = 0; wR < ${m}; wR++) {
|
|
int xR = xRCorner + wR * ${u};
|
|
|
|
if (xR < 0 || xR >= ${e.inHeight}) {
|
|
continue;
|
|
}
|
|
|
|
for (int wC = 0; wC < ${f}; wC++) {
|
|
int xC = xCCorner + wC * ${c};
|
|
|
|
if (xC < 0 || xC >= ${e.inWidth}) {
|
|
continue;
|
|
}
|
|
|
|
for (int d1 = 0; d1 < ${d}; d1 += 4) {
|
|
vec4 xValues = vec4(
|
|
getX(batch, xF, xR, xC, d1),
|
|
getX(batch, xF, xR, xC, d1 + 1),
|
|
getX(batch, xF, xR, xC, d1 + 2),
|
|
getX(batch, xF, xR, xC, d1 + 3)
|
|
);
|
|
vec4 wValues = vec4(
|
|
getW(wF, wR, wC, d1, d2),
|
|
getW(wF, wR, wC, d1 + 1, d2),
|
|
getW(wF, wR, wC, d1 + 2, d2),
|
|
getW(wF, wR, wC, d1 + 3, d2)
|
|
);
|
|
|
|
dotProd += dot(xValues, wValues);
|
|
}
|
|
|
|
if (${h===1}) {
|
|
dotProd +=
|
|
getX(batch, xF, xR, xC, ${d}) *
|
|
getW(wF, wR, wC, ${d}, d2);
|
|
} else if (${h===2}) {
|
|
vec2 xValues = vec2(
|
|
getX(batch, xF, xR, xC, ${d}),
|
|
getX(batch, xF, xR, xC, ${d} + 1)
|
|
);
|
|
vec2 wValues = vec2(
|
|
getW(wF, wR, wC, ${d}, d2),
|
|
getW(wF, wR, wC, ${d} + 1, d2)
|
|
);
|
|
dotProd += dot(xValues, wValues);
|
|
} else if (${h===3}) {
|
|
vec3 xValues = vec3(
|
|
getX(batch, xF, xR, xC, ${d}),
|
|
getX(batch, xF, xR, xC, ${d} + 1),
|
|
getX(batch, xF, xR, xC, ${d} + 2)
|
|
);
|
|
vec3 wValues = vec3(
|
|
getW(wF, wR, wC, ${d}, d2),
|
|
getW(wF, wR, wC, ${d} + 1, d2),
|
|
getW(wF, wR, wC, ${d} + 2, d2)
|
|
);
|
|
dotProd += dot(xValues, wValues);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}};var Pv=class{constructor(e,t,n){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=e;let{filterWidth:o,inChannels:s,strideWidth:a,strideHeight:i,padInfo:l,outWidth:u,dilationWidth:c,dilationHeight:p,dataFormat:m}=n,{left:f,top:d}=l,h=s*o,g=Pt(),x=m==="channelsLast",b=x?0:1,w=x?1:2,_="";for(let I=0;I<=1;I++)for(let E=0;E<=1;E++)_+=`
|
|
blockIndex = rc.y + ${E};
|
|
pos = rc.x + ${I};
|
|
|
|
if(blockIndex < ${e[1]} && pos < ${e[0]}) {
|
|
offsetY = int(blockIndex / (${u})) * ${i} - ${d};
|
|
d0 = offsetY + ${p} * (pos / ${h});
|
|
|
|
if(d0 < ${t[b]} && d0 >= 0) {
|
|
|
|
offsetX = int(mod(float(blockIndex), ${u}.) * ${a}. - ${f}.);
|
|
d1 = offsetX + ${c} * (int(mod(float(pos), ${h}.) / ${s}.));
|
|
|
|
if(d1 < ${t[w]} && d1 >= 0) {
|
|
|
|
ch = int(mod(float(pos), ${s}.));
|
|
|
|
if (${x}) {
|
|
innerDims = vec2(d1, ch);
|
|
result[${I*2+E}] = getChannel(
|
|
getA(d0, int(innerDims.x),
|
|
int(innerDims.y)), innerDims);
|
|
} else {
|
|
innerDims = vec2(d0, d1);
|
|
result[${I*2+E}] = getChannel(
|
|
getA(ch, int(innerDims.x),
|
|
int(innerDims.y)), innerDims);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
`;this.userCode=`
|
|
void main() {
|
|
ivec2 rc = getOutputCoords();
|
|
|
|
vec4 result = vec4(0);
|
|
|
|
int blockIndex, pos, offsetY, d0, offsetX, d1, ch;
|
|
vec2 innerDims;
|
|
|
|
${_}
|
|
|
|
${g.output} = result;
|
|
}
|
|
`}};function kx({x:r,filter:e,convInfo:t,backend:n,bias:o=null,preluActivationWeights:s=null,leakyreluAlpha:a=0,activation:i=null}){let l=r.shape,u=n.texData.get(r.dataId),c=t.inChannels,p=l[0]*l[1]*l[2],m=t.outChannels,f=t.dataFormat==="channelsLast",d=!1,h=!1,g,x=[],b=(p===1||m===1)&&c>hv,w=l[2]%2!=0&&!!u.isPacked;if(b||!W().getBool("WEBGL_LAZILY_UNPACK")||!W().getBool("WEBGL_PACK_BINARY_OPERATIONS")||!w){let _=f?l[0]*l[1]*l[2]:l[0]*l[2]*l[3],I=ce({inputs:{x:r},backend:n,attrs:{shape:[1,_,t.inChannels]}}),E=ce({inputs:{x:e},backend:n,attrs:{shape:[1,t.inChannels,t.outChannels]}}),D=Bu({a:I,b:E,transposeA:d,transposeB:h,backend:n,bias:o,activation:i,preluActivationWeights:s,leakyreluAlpha:a});g=ce({inputs:{x:D},backend:n,attrs:{shape:t.outShape}}),x.push(I),x.push(E),x.push(D)}else{let _=f?l[0]*l[1]*(l[2]+1):l[0]*l[2]*(l[3]+1),I={dataId:r.dataId,shape:[1,_,t.inChannels],dtype:r.dtype},E=u.shape;u.shape=u.shape.slice(),u.shape[u.shape.length-2]++,y.assert(_l(u.shape,I.shape),()=>`packed reshape ${u.shape} to ${I.shape} isn't free`);let D=ce({inputs:{x:e},backend:n,attrs:{shape:[1,t.inChannels,t.outChannels]}});x.push(D);let $=Bu({a:I,b:D,backend:n,transposeA:d,transposeB:h,bias:o,activation:i,preluActivationWeights:s,leakyreluAlpha:a}),R=n.texData.get($.dataId);y.assert(R.isPacked,()=>"batchMatMul result is expected to be packed"),u.shape=E,R.shape=t.outShape,g=Ut({inputs:{x:$},backend:n}),g.shape=t.outShape,x.push($)}for(let _ of x)n.disposeIntermediateTensorInfo(_);return g}function vx({x:r,filter:e,convInfo:t,backend:n,bias:o=null,preluActivationWeights:s=null,leakyreluAlpha:a=0,activation:i=null}){let{filterWidth:l,filterHeight:u,inChannels:c,outWidth:p,outHeight:m,dataFormat:f}=t,d=f==="channelsLast",h=l*u*c,g=m*p,x=[h,g],b=!0,w=!1,_=[],I=ce({inputs:{x:r},backend:n,attrs:{shape:r.shape.slice(1)}}),E=ce({inputs:{x:e},backend:n,attrs:{shape:[1,h,y.sizeFromShape(e.shape)/h]}});_.push(I),_.push(E);let D=new Pv(x,I.shape,t),$=n.runWebGLProgram(D,[I],"float32"),R=ce({inputs:{x:$},backend:n,attrs:{shape:[1,x[0],x[1]]}});_.push($),_.push(R);let M=o!=null,G=s!=null,j=i==="leakyrelu",U=i?vl(i,!0):null,H=new Lf(R.shape,E.shape,[1,g,t.outChannels],b,w,M,U,G,j),q=[R,E];if(o&&q.push(o),G&&q.push(s),j){let re=n.makeTensorInfo([],"float32",y.createScalarValue(a,"float32"));q.push(re),_.push(re)}let X=n.runWebGLProgram(H,q,"float32"),ne=d?[1,m,p,t.outChannels]:[1,t.outChannels,m,p],Y=ce({inputs:{x:X},backend:n,attrs:{shape:ne}});_.push(X);for(let re of _)n.disposeIntermediateTensorInfo(re);return Y}function S7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s}=e,{strides:a,pad:i,dataFormat:l,dilations:u,dimRoundingMode:c}=n,p=C.convertConv2DDataFormat(l),m=C.computeConv2DInfo(o.shape,s.shape,a,u,i,c,!1,p),f;if(m.filterHeight===1&&m.filterWidth===1&&m.dilationHeight===1&&m.dilationWidth===1&&m.strideHeight===1&&m.strideWidth===1&&(m.padInfo.type==="SAME"||m.padInfo.type==="VALID"))f=kx({x:o,filter:s,convInfo:m,backend:t});else if(W().getBool("WEBGL_CONV_IM2COL")&&o.shape[0]===1)f=vx({x:o,filter:s,convInfo:m,backend:t});else{let h=new Bf(m);f=t.runWebGLProgram(h,[o,s],"float32")}let d=ce({inputs:{x:f},backend:t,attrs:{shape:m.outShape}});return t.disposeIntermediateTensorInfo(f),d}var K$={kernelName:No,backendName:"webgl",kernelFunc:S7};var Mv=class{constructor(e){this.variableNames=["x","dy"],this.outputShape=e.filterShape;let t=e.strideHeight,n=e.strideWidth,o=e.padInfo.top,s=e.padInfo.left,a=e.dataFormat==="channelsLast";this.userCode=`
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int wR = coords.x;
|
|
int wC = coords.y;
|
|
int d1 = coords.z;
|
|
int d2 = coords.w;
|
|
|
|
// Convolve x(?, ?, d1) with dy(:, :, d2) to get dw(wR, wC, d1, d2).
|
|
// ? = to be determined. : = across all values in that axis.
|
|
float dotProd = 0.0;
|
|
|
|
for (int b = 0; b < ${e.batchSize}; b++) {
|
|
for (int yR = 0; yR < ${e.outHeight}; yR++) {
|
|
int xR = wR + yR * ${t} - ${o};
|
|
|
|
if (xR < 0 || xR >= ${e.inHeight}) {
|
|
continue;
|
|
}
|
|
|
|
for (int yC = 0; yC < ${e.outWidth}; yC++) {
|
|
int xC = wC + yC * ${n} - ${s};
|
|
|
|
if (xC < 0 || xC >= ${e.inWidth}) {
|
|
continue;
|
|
}
|
|
|
|
if (${a}) {
|
|
float dyValue = getDy(b, yR, yC, d2);
|
|
float xValue = getX(b, xR, xC, d1);
|
|
dotProd += (xValue * dyValue);
|
|
} else {
|
|
float dyValue = getDy(b, d2, yR, yC);
|
|
float xValue = getX(b, d1, xR, xC);
|
|
dotProd += (xValue * dyValue);
|
|
}
|
|
|
|
}
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}},Lv=class{constructor(e){this.variableNames=["dy","W"],this.outputShape=e.inShape;let t=e.filterHeight,n=e.filterWidth,o=e.strideHeight,s=e.strideWidth,a=e.dataFormat==="channelsLast",i=t-1-e.padInfo.top,l=n-1-e.padInfo.left,u=a?1:2,c=a?2:3,p=a?3:1;this.userCode=`
|
|
const ivec2 pads = ivec2(${i}, ${l});
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int batch = coords[0];
|
|
int d1 = coords[${p}];
|
|
|
|
ivec2 dyCorner = ivec2(coords[${u}], coords[${c}]) - pads;
|
|
int dyRCorner = dyCorner.x;
|
|
int dyCCorner = dyCorner.y;
|
|
|
|
// Convolve dy(?, ?, d2) with w(:, :, d1, d2) to compute dx(xR, xC, d1).
|
|
// ? = to be determined. : = across all values in that axis.
|
|
float dotProd = 0.0;
|
|
for (int wR = 0; wR < ${t}; wR++) {
|
|
float dyR = float(dyRCorner + wR) / ${o}.0;
|
|
|
|
if (dyR < 0.0 || dyR >= ${e.outHeight}.0 || fract(dyR) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyR = int(dyR);
|
|
|
|
int wRPerm = ${t} - 1 - wR;
|
|
|
|
for (int wC = 0; wC < ${n}; wC++) {
|
|
float dyC = float(dyCCorner + wC) / ${s}.0;
|
|
|
|
if (dyC < 0.0 || dyC >= ${e.outWidth}.0 ||
|
|
fract(dyC) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyC = int(dyC);
|
|
|
|
int wCPerm = ${n} - 1 - wC;
|
|
|
|
for (int d2 = 0; d2 < ${e.outChannels}; d2++) {
|
|
|
|
if (${a}) {
|
|
float xValue = getDy(batch, idyR, idyC, d2);
|
|
float wValue = getW(wRPerm, wCPerm, d1, d2);
|
|
dotProd += xValue * wValue;
|
|
} else {
|
|
float xValue = getDy(batch, d2, idyR, idyC);
|
|
float wValue = getW(wRPerm, wCPerm, d1, d2);
|
|
dotProd += xValue * wValue;
|
|
}
|
|
|
|
}
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}},zv=class{constructor(e){this.variableNames=["x","dy"],this.outputShape=e.filterShape;let t=e.strideDepth,n=e.strideHeight,o=e.strideWidth,s=e.padInfo.front,a=e.padInfo.top,i=e.padInfo.left;this.userCode=`
|
|
void main() {
|
|
ivec5 coords = getOutputCoords();
|
|
int wF = coords.x;
|
|
int wR = coords.y;
|
|
int wC = coords.z;
|
|
int d1 = coords.w;
|
|
int d2 = coords.u;
|
|
|
|
float dotProd = 0.0;
|
|
|
|
for (int b = 0; b < ${e.batchSize}; b++) {
|
|
for (int yF = 0; yF < ${e.outDepth}; yF++) {
|
|
int xF = wF + yF * ${t} - ${s};
|
|
|
|
if (xF < 0 || xF >= ${e.inDepth}) {
|
|
continue;
|
|
}
|
|
|
|
for (int yR = 0; yR < ${e.outHeight}; yR++) {
|
|
int xR = wR + yR * ${n} - ${a};
|
|
|
|
if (xR < 0 || xR >= ${e.inHeight}) {
|
|
continue;
|
|
}
|
|
|
|
for (int yC = 0; yC < ${e.outWidth}; yC++) {
|
|
int xC = wC + yC * ${o} - ${i};
|
|
|
|
if (xC < 0 || xC >= ${e.inWidth}) {
|
|
continue;
|
|
}
|
|
|
|
float dyValue = getDy(b, yF, yR, yC, d2);
|
|
float xValue = getX(b, xF, xR, xC, d1);
|
|
dotProd += (xValue * dyValue);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}},Bv=class{constructor(e){this.variableNames=["dy","W"],this.outputShape=e.inShape;let t=e.filterDepth,n=e.filterHeight,o=e.filterWidth,s=e.strideDepth,a=e.strideHeight,i=e.strideWidth,l=t-1-e.padInfo.front,u=n-1-e.padInfo.top,c=o-1-e.padInfo.left;this.userCode=`
|
|
const ivec3 pads = ivec3(${l}, ${u}, ${c});
|
|
|
|
void main() {
|
|
ivec5 coords = getOutputCoords();
|
|
int batch = coords.x;
|
|
int d1 = coords.u;
|
|
|
|
|
|
ivec3 dyCorner = ivec3(coords.y, coords.z, coords.w) - pads;
|
|
int dyFCorner = dyCorner.x;
|
|
int dyRCorner = dyCorner.y;
|
|
int dyCCorner = dyCorner.z;
|
|
|
|
float dotProd = 0.0;
|
|
for (int wF = 0; wF < ${t}; wF++) {
|
|
float dyF = float(dyFCorner + wF) / ${s}.0;
|
|
|
|
if (dyF < 0.0 || dyF >= ${e.outDepth}.0 || fract(dyF) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyF = int(dyF);
|
|
|
|
int wFPerm = ${t} - 1 - wF;
|
|
|
|
for (int wR = 0; wR < ${n}; wR++) {
|
|
float dyR = float(dyRCorner + wR) / ${a}.0;
|
|
|
|
if (dyR < 0.0 || dyR >= ${e.outHeight}.0 ||
|
|
fract(dyR) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyR = int(dyR);
|
|
|
|
int wRPerm = ${n} - 1 - wR;
|
|
|
|
for (int wC = 0; wC < ${o}; wC++) {
|
|
float dyC = float(dyCCorner + wC) / ${i}.0;
|
|
|
|
if (dyC < 0.0 || dyC >= ${e.outWidth}.0 ||
|
|
fract(dyC) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyC = int(dyC);
|
|
|
|
int wCPerm = ${o} - 1 - wC;
|
|
|
|
for (int d2 = 0; d2 < ${e.outChannels}; d2++) {
|
|
float xValue = getDy(batch, idyF, idyR, idyC, d2);
|
|
float wValue = getW(wFPerm, wRPerm, wCPerm, d1, d2);
|
|
dotProd += xValue * wValue;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}};function T7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,dy:s}=e,{strides:a,pad:i,dataFormat:l,dimRoundingMode:u,filterShape:c}=n,p=C.convertConv2DDataFormat(l),m=C.computeConv2DInfo(o.shape,c,a,1,i,u,!1,p),f=new Mv(m);return t.runWebGLProgram(f,[o,s],"float32")}var X$={kernelName:Cc,backendName:"webgl",kernelFunc:T7};function A7(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,filter:s}=e,{inputShape:a,strides:i,pad:l,dataFormat:u,dimRoundingMode:c}=n,p=C.convertConv2DDataFormat(u),m=C.computeConv2DInfo(a,s.shape,i,1,l,c,!1,p),f=new Lv(m);return t.runWebGLProgram(f,[o,s],"float32")}var Y$={kernelName:So,backendName:"webgl",kernelFunc:A7};function E7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s}=e,{strides:a,pad:i,dilations:l}=n,u=C.computeConv3DInfo(o.shape,s.shape,a,l,i),c=new Ov(u);return t.runWebGLProgram(c,[o,s],"float32")}var Z$={kernelName:el,backendName:"webgl",kernelFunc:E7};function D7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,dy:s}=e,{strides:a,pad:i,filterShape:l}=n,u=C.computeConv3DInfo(o.shape,l,a,1,i),c=new zv(u);return t.runWebGLProgram(c,[o,s],"float32")}var J$={kernelName:Ic,backendName:"webgl",kernelFunc:D7};function $7(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,filter:s}=e,{pad:a,strides:i,inputShape:l}=n,u=C.computeConv3DInfo(l,s.shape,i,1,a),c=new Bv(u);return t.runWebGLProgram(c,[o,s],"float32")}var Q$={kernelName:Nc,backendName:"webgl",kernelFunc:$7};var R7=gx+`
|
|
return cos(x);
|
|
`,F7=_e({opSnippet:R7}),eR={kernelName:To,backendName:"webgl",kernelFunc:F7};var O7=`
|
|
float e2x = exp(-x);
|
|
return (e2x + 1.0 / e2x) / 2.0;
|
|
`,P7=_e({opSnippet:O7}),tR={kernelName:Di,backendName:"webgl",kernelFunc:P7};var Vv=class{constructor(e,t,n,o,s){this.variableNames=["Image","Boxes","BoxInd"],this.outputShape=[];let[a,i,l,u]=e,[c]=t,[p,m]=n;this.outputShape=[c,p,m,u];let f=o==="bilinear"?1:0,[d,h]=[`${i-1}.0`,`${l-1}.0`],[g,x,b]=p>1?[`${(i-1)/(p-1)}`,"(y2-y1) * height_ratio",`y1*${d} + float(y)*(height_scale)`]:["0.0","0.0",`0.5 * (y1+y2) * ${d}`],[w,_,I]=m>1?[`${(l-1)/(m-1)}`,"(x2-x1) * width_ratio",`x1*${h} + float(x)*(width_scale)`]:["0.0","0.0",`0.5 * (x1+x2) * ${h}`];this.userCode=`
|
|
const float height_ratio = float(${g});
|
|
const float width_ratio = float(${w});
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int y = coords[1];
|
|
int x = coords[2];
|
|
int d = coords[3];
|
|
|
|
// get box vals
|
|
float y1 = getBoxes(b,0);
|
|
float x1 = getBoxes(b,1);
|
|
float y2 = getBoxes(b,2);
|
|
float x2 = getBoxes(b,3);
|
|
|
|
// get image in batch index
|
|
int bInd = round(getBoxInd(b));
|
|
if(bInd < 0 || bInd >= ${a}) {
|
|
return;
|
|
}
|
|
|
|
float height_scale = ${x};
|
|
float width_scale = ${_};
|
|
|
|
float in_y = ${b};
|
|
if( in_y < 0.0 || in_y > ${d} ) {
|
|
setOutput(float(${s}));
|
|
return;
|
|
}
|
|
float in_x = ${I};
|
|
if( in_x < 0.0 || in_x > ${h} ) {
|
|
setOutput(float(${s}));
|
|
return;
|
|
}
|
|
|
|
vec2 sourceFracIndexCR = vec2(in_x,in_y);
|
|
if(${f} == 1) {
|
|
// Compute the four integer indices.
|
|
ivec2 sourceFloorCR = ivec2(sourceFracIndexCR);
|
|
ivec2 sourceCeilCR = ivec2(ceil(sourceFracIndexCR));
|
|
|
|
float topLeft = getImage(b, sourceFloorCR.y, sourceFloorCR.x, d);
|
|
float bottomLeft = getImage(b, sourceCeilCR.y, sourceFloorCR.x, d);
|
|
float topRight = getImage(b, sourceFloorCR.y, sourceCeilCR.x, d);
|
|
float bottomRight = getImage(b, sourceCeilCR.y, sourceCeilCR.x, d);
|
|
|
|
vec2 fracCR = sourceFracIndexCR - vec2(sourceFloorCR);
|
|
|
|
float top = topLeft + (topRight - topLeft) * fracCR.x;
|
|
float bottom = bottomLeft + (bottomRight - bottomLeft) * fracCR.x;
|
|
float newValue = top + (bottom - top) * fracCR.y;
|
|
setOutput(newValue);
|
|
} else {
|
|
// Compute the coordinators of nearest neighbor point.
|
|
ivec2 sourceNearestCR = ivec2(floor(
|
|
sourceFracIndexCR + vec2(0.5,0.5)));
|
|
float newValue = getImage(b, sourceNearestCR.y, sourceNearestCR.x, d);
|
|
setOutput(newValue);
|
|
}
|
|
}
|
|
`}};var M7=r=>{let{inputs:e,backend:t,attrs:n}=r,{image:o,boxes:s,boxInd:a}=e,{cropSize:i,method:l,extrapolationValue:u}=n,c=new Vv(o.shape,s.shape,i,l,u);return t.runWebGLProgram(c,[o,s,a],"float32")},rR={kernelName:$i,backendName:"webgl",kernelFunc:M7};var Cx=class{constructor(e,t,n){this.variableNames=["x"],this.outputShape=e;let o=e.length,s=t?"0.0":`getX(${nR(o,"coords")})`,a=e[e.length-1],i="",l="";t?(i=n?`end != ${a-1}`:"end != 0",l=n?"end + 1":"end - 1"):(i=n?`end + pow2 < ${a}`:"end >= pow2",l=n?"end + pow2":"end - pow2"),this.userCode=`
|
|
uniform float index;
|
|
void main() {
|
|
${Ve(o)} coords = getOutputCoords();
|
|
int end = ${oR(o,"coords")};
|
|
float val = ${s};
|
|
int pow2 = int(pow(2.0, index));
|
|
if (${i}) {
|
|
int idx = ${l};
|
|
${oR(o,"coords")} = idx;
|
|
val += getX(${nR(o,"coords")});
|
|
}
|
|
setOutput(val);
|
|
}
|
|
`}getCustomSetupFunc(e){return(t,n)=>{this.index==null&&(this.index=t.getUniformLocation(n,"index")),t.gl.uniform1f(this.index,e)}}};function nR(r,e){if(r===1)return`${e}`;if(r===2)return`${e}.x, ${e}.y`;if(r===3)return`${e}.x, ${e}.y, ${e}.z`;if(r===4)return`${e}.x, ${e}.y, ${e}.z, ${e}.w`;throw Error(`Cumulative sum for rank ${r} is not yet supported`)}function oR(r,e){if(r===1)return`${e}`;if(r===2)return`${e}.y`;if(r===3)return`${e}.z`;if(r===4)return`${e}.w`;throw Error(`Cumulative sum for rank ${r} is not yet supported`)}function L7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,exclusive:a,reverse:i}=n,l=o.shape.length,u=C.getAxesPermutation([s],l),c=o;u!=null&&(c=Rt({inputs:{x:o},backend:t,attrs:{perm:u}}));let p=C.getInnerMostAxes(1,l)[0];if(p!==l-1)throw new Error(`WebGL cumsum shader expects an inner-most axis=${o.shape.length-1} but got axis=${s}`);let m=c.shape[p],f=Ut({inputs:{x:c},backend:t});for(let d=0;d<=Math.ceil(Math.log2(m))-1;d++){let h=new Cx(c.shape,!1,i),g=h.getCustomSetupFunc(d),x=f;f=t.runWebGLProgram(h,[f],f.dtype,g),t.disposeIntermediateTensorInfo(x)}if(a){let d=new Cx(c.shape,a,i),h=f;f=t.runWebGLProgram(d,[f],f.dtype),t.disposeIntermediateTensorInfo(h)}if(u!=null){let d=C.getUndoAxesPermutation(u),h=Rt({inputs:{x:f},backend:t,attrs:{perm:d}});return t.disposeIntermediateTensorInfo(f),t.disposeIntermediateTensorInfo(c),h}return f}var sR={kernelName:Ao,backendName:"webgl",kernelFunc:L7};function z7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,weights:s}=e,{size:a,binaryOutput:i}=n;if(o.shape.length===1){let l=t.readSync(o.dataId),u=t.readSync(s.dataId),c=mx(l,u,s.dtype,s.shape,a);return t.makeTensorInfo([a],s.dtype,c)}else if(o.shape.length===2){let l=t.bufferSync(o),u=t.bufferSync(s),c=lD(l,u,a,i);return t.makeTensorInfo(c.shape,s.dtype,c.values)}throw new Error(`Error in denseBincount: input must be at most rank 2, but got rank${o.shape.length}.`)}var iR={kernelName:Sc,backendName:"webgl",kernelFunc:z7};var Gv=class{constructor(e,t,n){this.variableNames=["x"],this.outputShape=[],this.outputShape=e,this.blockSize=t,this.dataFormat=n,this.userCode=`
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int h = ${this.getHeightCoordString()};
|
|
int w = ${this.getWidthCoordString()};
|
|
int d = ${this.getDepthCoordString()};
|
|
|
|
int in_h = h / ${t};
|
|
int offset_h = imod(h, ${t});
|
|
int in_w = w / ${t};
|
|
int offset_w = imod(w, ${t});
|
|
int offset_d = (offset_h * ${t} + offset_w) *
|
|
${this.getOutputDepthSize()};
|
|
int in_d = d + offset_d;
|
|
|
|
float result = ${this.getInputSamplingString()};
|
|
setOutput(result);
|
|
}
|
|
`}getHeightCoordString(){return this.dataFormat==="NHWC"?"coords[1]":"coords[2]"}getWidthCoordString(){return this.dataFormat==="NHWC"?"coords[2]":"coords[3]"}getDepthCoordString(){return this.dataFormat==="NHWC"?"coords[3]":"coords[1]"}getOutputDepthSize(){return this.dataFormat==="NHWC"?this.outputShape[3]:this.outputShape[1]}getInputSamplingString(){return this.dataFormat==="NHWC"?"getX(b, in_h, in_w, in_d)":"getX(b, in_d, in_h, in_w)"}};function B7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{blockSize:s,dataFormat:a}=n;y.assert(s>1,()=>`blockSize should be > 1 for depthToSpace, but was: ${s}`);let i=o.shape[0],l=a==="NHWC"?o.shape[1]:o.shape[2],u=a==="NHWC"?o.shape[2]:o.shape[3],c=a==="NHWC"?o.shape[3]:o.shape[1],p=l*s,m=u*s,f=c/(s*s),d=a==="NHWC"?[i,p,m,f]:[i,f,p,m],h=new Gv(d,s,a);return t.runWebGLProgram(h,[o],o.dtype)}var aR={kernelName:Ri,backendName:"webgl",kernelFunc:B7};var Vf=class{constructor(e,t=!1,n=null,o=!1,s=!1){this.variableNames=["x","W"],this.outputShape=e.outShape;let a=e.inHeight,i=e.inWidth,l=e.padInfo.top,u=e.padInfo.left,c=e.strideHeight,p=e.strideWidth,m=e.dilationHeight,f=e.dilationWidth,d=e.filterHeight,h=e.filterWidth,g=e.outChannels/e.inChannels,x="",b="";n&&(o?x=`float activation(float a) {
|
|
float b = getPreluActivationWeightsAtOutCoords();
|
|
${n}
|
|
}`:s?x=`float activation(float a) {
|
|
float b = getLeakyreluAlphaAtOutCoords();
|
|
${n}
|
|
}`:x=`
|
|
float activation(float x) {
|
|
${n}
|
|
}
|
|
`,b="result = activation(result);");let w=t?"result += getBiasAtOutCoords();":"";t&&this.variableNames.push("bias"),o&&this.variableNames.push("preluActivationWeights"),s&&this.variableNames.push("leakyreluAlpha"),this.userCode=`
|
|
${x}
|
|
|
|
const ivec2 strides = ivec2(${c}, ${p});
|
|
const ivec2 pads = ivec2(${l}, ${u});
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int batch = coords.x;
|
|
ivec2 xRCCorner = coords.yz * strides - pads;
|
|
int d2 = coords.w;
|
|
int d1 = d2 / ${g};
|
|
int q = d2 - d1 * ${g};
|
|
|
|
int xRCorner = xRCCorner.x;
|
|
int xCCorner = xRCCorner.y;
|
|
|
|
// Convolve x(?, ?, d1) with w(:, :, d1, q) to get y(yR, yC, d2).
|
|
// ? = to be determined. : = across all values in that axis.
|
|
float dotProd = 0.0;
|
|
// TO DO(dsmilkov): Flatten the two for loops and vec4 the operations.
|
|
for (int wR = 0; wR < ${d}; wR++) {
|
|
int xR = xRCorner + wR * ${m};
|
|
|
|
if (xR < 0 || xR >= ${a}) {
|
|
continue;
|
|
}
|
|
|
|
for (int wC = 0; wC < ${h}; wC++) {
|
|
int xC = xCCorner + wC * ${f};
|
|
|
|
if (xC < 0 || xC >= ${i}) {
|
|
continue;
|
|
}
|
|
|
|
float xVal = getX(batch, xR, xC, d1);
|
|
float wVal = getW(wR, wC, d1, q);
|
|
dotProd += xVal * wVal;
|
|
}
|
|
}
|
|
|
|
float result = dotProd;
|
|
${w}
|
|
${b}
|
|
setOutput(result);
|
|
}
|
|
`}};var Gf=class{constructor(e,t=!1,n=null,o=!1,s=!1){this.variableNames=["x","W"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=e.outShape;let a=e.outChannels/e.inChannels,i=e.inHeight,l=e.inWidth,u=e.padInfo.top,c=e.padInfo.left,p=e.strideHeight,m=e.strideWidth,f=e.dilationHeight,d=e.dilationWidth,h=e.filterHeight,g=e.filterWidth,x=g,b=`
|
|
int xR; int xC; int xCOffset;
|
|
vec4 wTexel; vec4 previous; vec4 final;`;for(let E=0;E<g;E++)b+=`
|
|
vec4 xTexelC${E*2};
|
|
int xTexelC${E*2}Ready;
|
|
vec4 xC${E};`;for(let E=0;E<h;E++){for(let D=0;D<g;D++)b+=`
|
|
xTexelC${D*2} = vec4(0.0);
|
|
xTexelC${D*2}Ready = 0;
|
|
xC${D} = vec4(0.0);`;b+=`
|
|
xR = xRCorner + ${E*f};
|
|
if (xR >=0 && xR < ${i}) {
|
|
`;for(let D=0;D<(x+1)/2;D++){let $=D*2,R=$*d;if(b+=`
|
|
xC = xCCorner + ${R};
|
|
`,m===1){if($<g&&(c%2==1?(b+=`
|
|
xCOffset = xC + 1;
|
|
if (xCOffset >= 0 && xCOffset < ${l} && xTexelC${R}Ready == 0) {
|
|
xTexelC${R} = getX(batch, xR, xCOffset, d1);
|
|
|
|
// Need to manually clear unused channels in case
|
|
// we're reading from recycled texture.
|
|
if (xCOffset + 1 >= ${l}) {
|
|
xTexelC${R}.zw = vec2(0.0);
|
|
}
|
|
xTexelC${R}Ready = 1;
|
|
}
|
|
`,d===1&&R>0?b+=`
|
|
xC${$} = vec4(xTexelC${R-2}.zw, xTexelC${R}.xy);
|
|
`:b+=`
|
|
xCOffset = xC + 1 - 2;
|
|
|
|
if (xCOffset >= 0 && xCOffset < ${l}) {
|
|
previous = getX(batch, xR, xCOffset, d1);
|
|
|
|
// Need to manually clear unused channels in case
|
|
// we're reading from recycled texture.
|
|
if (xCOffset + 1 >= ${l}) {
|
|
previous.zw = vec2(0.0);
|
|
}
|
|
|
|
xC${$} = vec4(previous.zw, xTexelC${R}.xy);
|
|
} else {
|
|
xC${$} = vec4(0.0, 0.0, xTexelC${R}.xy);
|
|
}
|
|
`):b+=`
|
|
if (xC >= 0 && xC < ${l} && xTexelC${R}Ready == 0) {
|
|
xTexelC${R} = getX(batch, xR, xC, d1);
|
|
if (xC + 1 >= ${l}) {
|
|
xTexelC${R}.zw = vec2(0.0);
|
|
}
|
|
xTexelC${R}Ready = 1;
|
|
}
|
|
|
|
xC${$} = xTexelC${R};
|
|
`,R+1<g)){let M=c%2==0?y.nearestLargerEven(d):d;d%2==0&&c%2==1||d%2!=0&&c%2!=1?(b+=`
|
|
xCOffset = xC + ${c%2} + ${M};
|
|
|
|
if (xCOffset >= 0 && xCOffset < ${l} && xTexelC${R+2}Ready == 0) {
|
|
xTexelC${R+2} = getX(batch, xR, xCOffset, d1);
|
|
|
|
// Need to manually clear unused channels in case
|
|
// we're reading from recycled texture.
|
|
if (xCOffset + 1 >= ${l}) {
|
|
xTexelC${R+2}.zw = vec2(0.0);
|
|
}
|
|
xTexelC${R+2}Ready = 1;
|
|
}
|
|
`,d>1&&(b+=`
|
|
xCOffset -= 2;
|
|
if (xCOffset >= 0 && xCOffset < ${l} && xTexelC${R}Ready == 0) {
|
|
xTexelC${R} = getX(batch, xR, xCOffset, d1);
|
|
xTexelC${R}Ready = 1;
|
|
}
|
|
`),b+=`
|
|
xC${$+1} = vec4(xTexelC${R}.zw, xTexelC${R+2}.xy);
|
|
`):M===1?b+=`
|
|
xC${$+1} = xTexelC${R};
|
|
`:b+=`
|
|
xCOffset = xC + ${M};
|
|
|
|
if (xCOffset >= 0 && xCOffset < ${l} && xTexelC${R+2}Ready == 0) {
|
|
xTexelC${R+2} = getX(batch, xR, xCOffset, d1);
|
|
if (xCOffset + 1 >= ${l}) {
|
|
xTexelC${R+2}.zw = vec2(0.0);
|
|
}
|
|
xTexelC${R+2}Ready = 1;
|
|
}
|
|
|
|
xC${$+1} = xTexelC${R+2};
|
|
`}}else R<g&&(c%2==1?(b+=`
|
|
xCOffset = xC + 1 - ${m};
|
|
if(xCOffset >= 0 && xCOffset < ${l} && xTexelC${R}Ready == 0) {
|
|
xTexelC${R} = getX(batch, xR, xCOffset, d1);
|
|
// Need to manually clear unused channels in case
|
|
// we're reading from recycled texture.
|
|
if (xCOffset + 1 >= ${l}) {
|
|
xTexelC${R}.zw = vec2(0.0);
|
|
}
|
|
xTexelC${R}Ready = 1;
|
|
}
|
|
|
|
if(xC + 1 >= 0 && xC + 1 < ${l} && xTexelC${R+2}Ready == 0) {
|
|
xTexelC${R+2} = getX(batch, xR, xC + 1, d1);
|
|
// Need to manually clear unused channels in case
|
|
// we're reading from recycled texture.
|
|
if (xC + 2 >= ${l}) {
|
|
xTexelC${R+2}.zw = vec2(0.0);
|
|
}
|
|
xTexelC${R+2}Ready = 1;
|
|
}
|
|
|
|
xC${$} = vec4(xTexelC${R}.zw, xTexelC${R+2}.zw);
|
|
`,R+1<g&&(b+=`
|
|
final = vec4(0.0);
|
|
xCOffset = xC + 1 + ${m};
|
|
if(xCOffset >= 0 && xCOffset < ${l}) {
|
|
final = getX(batch, xR, xCOffset, d1);
|
|
}
|
|
xC${$+1} = vec4(xTexelC${R+2}.xy, final.xy);
|
|
`)):(b+=`
|
|
if(xC >= 0 && xC < ${l} && xTexelC${R}Ready == 0) {
|
|
xTexelC${R} = getX(batch, xR, xC, d1);
|
|
if (xC + 1 >= ${l}) {
|
|
xTexelC${R}.zw = vec2(0.0);
|
|
}
|
|
xTexelC${R}Ready = 1;
|
|
}
|
|
|
|
xCOffset = xC + ${m};
|
|
if(xCOffset >= 0 && xCOffset < ${l} && xTexelC${R+2}Ready == 0) {
|
|
xTexelC${R+2} = getX(batch, xR, xCOffset, d1);
|
|
if (xCOffset + 1 >= ${l}) {
|
|
xTexelC${R+2}.zw = vec2(0.);
|
|
}
|
|
xTexelC${R+2}Ready = 1;
|
|
}
|
|
|
|
xC${$} = vec4(
|
|
xTexelC${R}.xy, xTexelC${R+2}.xy);
|
|
`,R+1<g&&(b+=`
|
|
xC${$+1} = vec4(xTexelC${R}.zw, xTexelC${R+2}.zw);
|
|
`)));$<g&&(b+=`
|
|
wTexel = getW(${E}, ${R}, d1, q);
|
|
dotProd += xC${$} * vec4(wTexel.xz, wTexel.xz);
|
|
`,R+1<g&&(b+=`
|
|
wTexel = getW(${E}, ${R+1}, d1, q);
|
|
dotProd += xC${$+1} * vec4(wTexel.xz, wTexel.xz);
|
|
`))}b+=`
|
|
}
|
|
`}let w="",_="";n&&(o?w=`vec4 activation(vec4 a) {
|
|
vec4 b = getPreluActivationWeightsAtOutCoords();
|
|
${n}
|
|
}`:s?w=`vec4 activation(vec4 a) {
|
|
vec4 b = getLeakyreluAlphaAtOutCoords();
|
|
${n}
|
|
}`:w=`vec4 activation(vec4 x) {
|
|
${n}
|
|
}`,_="result = activation(result);");let I=t?"result += getBiasAtOutCoords();":"";t&&this.variableNames.push("bias"),o&&this.variableNames.push("preluActivationWeights"),s&&this.variableNames.push("leakyreluAlpha"),this.userCode=`
|
|
${w}
|
|
|
|
const ivec2 strides = ivec2(${p}, ${m});
|
|
const ivec2 pads = ivec2(${u}, ${c});
|
|
|
|
void main() {
|
|
|
|
ivec4 coords = getOutputCoords();
|
|
int batch = coords.x;
|
|
ivec2 xRCCorner = coords.yz * strides - pads;
|
|
int d2 = coords.w;
|
|
int d1 = d2 / ${a};
|
|
int q = d2 - d1 * ${a};
|
|
int xRCorner = xRCCorner.x;
|
|
int xCCorner = xRCCorner.y;
|
|
|
|
//intialize dotProd with a small epsilon seems to reduce GPU accuracy loss.
|
|
vec4 dotProd = vec4(0.000000000000001);
|
|
|
|
${b}
|
|
|
|
vec4 result = dotProd - vec4(0.000000000000001);
|
|
${I}
|
|
${_}
|
|
setOutput(result);
|
|
}
|
|
`}};function V7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s}=e,{strides:a,pad:i,dilations:l,dimRoundingMode:u}=n,c=l;c==null&&(c=[1,1]),y.assert(C.eitherStridesOrDilationsAreOne(a,c),()=>`Error in depthwiseConv2d: Either strides or dilations must be 1. Got strides ${a} and dilations '${c}'`);let p=C.computeConv2DInfo(o.shape,s.shape,a,c,i,u,!0),m;return W().getBool("WEBGL_PACK_DEPTHWISECONV")&&p.strideWidth<=2&&p.outChannels/p.inChannels==1?m=new Gf(p):m=new Vf(p),t.runWebGLProgram(m,[o,s],"float32")}var lR={kernelName:Eo,backendName:"webgl",kernelFunc:V7};var Wv=class{constructor(e){this.variableNames=["x","dy"],this.outputShape=e.filterShape;let t=e.strideHeight,n=e.strideWidth,o=e.padInfo.top,s=e.padInfo.left,a=e.outChannels/e.inChannels;this.userCode=`
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int wR = coords.x;
|
|
int wC = coords.y;
|
|
int d1 = coords.z;
|
|
int dm = coords.w;
|
|
int d2 = d1 * ${a} + dm;
|
|
|
|
float dotProd = 0.0;
|
|
|
|
// TO DO: Vec4 over the batch size
|
|
for (int b = 0; b < ${e.batchSize}; b++) {
|
|
for (int yR = 0; yR < ${e.outHeight}; yR++) {
|
|
int xR = wR + yR * ${t} - ${o};
|
|
|
|
if (xR < 0 || xR >= ${e.inHeight}) {
|
|
continue;
|
|
}
|
|
|
|
for (int yC = 0; yC < ${e.outWidth}; yC++) {
|
|
int xC = wC + yC * ${n} - ${s};
|
|
|
|
if (xC < 0 || xC >= ${e.inWidth}) {
|
|
continue;
|
|
}
|
|
|
|
float dyValue = getDy(b, yR, yC, d2);
|
|
float xValue = getX(b, xR, xC, d1);
|
|
dotProd += (xValue * dyValue);
|
|
}
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}},jv=class{constructor(e){this.variableNames=["dy","W"],this.outputShape=e.inShape;let t=e.filterHeight,n=e.filterWidth,o=e.strideHeight,s=e.strideWidth,a=t-1-e.padInfo.top,i=n-1-e.padInfo.left,l=e.outChannels/e.inChannels;this.userCode=`
|
|
const ivec2 pads = ivec2(${a}, ${i});
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int batch = coords[0];
|
|
int d1 = coords[3];
|
|
ivec2 dyCorner = coords.yz - pads;
|
|
int dyRCorner = dyCorner.x;
|
|
int dyCCorner = dyCorner.y;
|
|
|
|
float dotProd = 0.0;
|
|
|
|
for (int wR = 0; wR < ${t}; wR++) {
|
|
float dyR = float(dyRCorner + wR) / ${o}.0;
|
|
|
|
if (dyR < 0.0 || dyR >= ${e.outHeight}.0 || fract(dyR) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyR = int(dyR);
|
|
|
|
int wRPerm = ${t} - 1 - wR;
|
|
|
|
for (int wC = 0; wC < ${n}; wC++) {
|
|
float dyC = float(dyCCorner + wC) / ${s}.0;
|
|
|
|
if (dyC < 0.0 || dyC >= ${e.outWidth}.0 ||
|
|
fract(dyC) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyC = int(dyC);
|
|
|
|
int wCPerm = ${n} - 1 - wC;
|
|
|
|
// TO DO: Vec4 over the channelMul
|
|
for (int dm = 0; dm < ${l}; dm++) {
|
|
int d2 = d1 * ${l} + dm;
|
|
float xValue = getDy(batch, idyR, idyC, d2);
|
|
float wValue = getW(wRPerm, wCPerm, d1, dm);
|
|
dotProd += xValue * wValue;
|
|
}
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}};function G7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,dy:s}=e,{strides:a,dilations:i,pad:l,dimRoundingMode:u,filterShape:c}=n,p=C.computeConv2DInfo(o.shape,c,a,i,l,u,!0),m=new Wv(p);return t.runWebGLProgram(m,[o,s],"float32")}var uR={kernelName:Tc,backendName:"webgl",kernelFunc:G7};function W7(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,filter:s}=e,{strides:a,dilations:i,pad:l,dimRoundingMode:u,inputShape:c}=n,p=C.computeConv2DInfo(c,s.shape,a,i,l,u,!0),m=new jv(p);return t.runWebGLProgram(m,[o,s],"float32")}var cR={kernelName:Ac,backendName:"webgl",kernelFunc:W7};var Uv=class{constructor(e){this.variableNames=["X"],this.outputShape=[e,e],this.userCode=`
|
|
void main() {
|
|
ivec2 coords = getOutputCoords();
|
|
float val = coords[0] == coords[1] ? getX(coords[0]) : 0.0;
|
|
setOutput(val);
|
|
}
|
|
`}};function j7(r){let{inputs:e,backend:t}=r,{x:n}=e,o=[...n.shape,...n.shape],s=y.sizeFromShape(n.shape),a=ce({inputs:{x:n},backend:t,attrs:{shape:[s]}}),i=new Uv(s),l=t.runWebGLProgram(i,[a],a.dtype),u=ce({inputs:{x:l},backend:t,attrs:{shape:o}});return t.disposeIntermediateTensorInfo(a),t.disposeIntermediateTensorInfo(l),u}var pR={kernelName:Ec,backendName:"webgl",kernelFunc:j7};var Hv=class{constructor(e){this.variableNames=["x","W"],this.outputShape=e.outShape;let{inHeight:t,inWidth:n,padInfo:o,strideHeight:s,strideWidth:a,filterHeight:i,filterWidth:l,dilationHeight:u,dilationWidth:c}=e,{top:p,left:m}=o;this.userCode=`
|
|
const ivec2 strides = ivec2(${s}, ${a});
|
|
const ivec2 pads = ivec2(${p}, ${m});
|
|
const float neg_infinity = -3.4e38;
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int batch = coords.x;
|
|
int d1 = coords.w;
|
|
ivec2 outTopLeftCorner =
|
|
coords.yz * strides - pads;
|
|
int hBeg = outTopLeftCorner.x;
|
|
int wBeg = outTopLeftCorner.y;
|
|
|
|
float curVal = neg_infinity;
|
|
for (int h = 0; h < ${i}; h++) {
|
|
int hIn = hBeg + h * ${u};
|
|
|
|
if (hIn >= 0 && hIn < ${t}) {
|
|
for (int w = 0; w < ${l}; w++) {
|
|
int wIn = wBeg + w * ${c};
|
|
|
|
if (wIn >= 0 && wIn < ${n}) {
|
|
float xVal = getX(batch, hIn, wIn, d1);
|
|
float wVal = getW(h, w, d1);
|
|
|
|
float val = xVal + wVal;
|
|
if (val > curVal) {
|
|
curVal = val;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
float result = curVal;
|
|
setOutput(result);
|
|
}
|
|
`}};function U7(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s}=e,{strides:a,pad:i,dilations:l}=n,u=C.computeDilation2DInfo(o.shape,s.shape,a,i,"NHWC",l),c,p=new Hv(u);c=t.runWebGLProgram(p,[o,s],"float32");let m=ce({inputs:{x:c},backend:t,attrs:{shape:u.outShape}});return t.disposeIntermediateTensorInfo(c),m}var mR={kernelName:tl,backendName:"webgl",kernelFunc:U7};function H7(r){let{inputs:e,backend:t,attrs:n}=r,{equation:o}=n,s=e,{allDims:a,summedDims:i,idDims:l}=C.decodeEinsumEquation(o,s.length);C.checkEinsumDimSizes(a.length,l,s);let{path:u,steps:c}=C.getEinsumComputePath(i,l),p=c.length,m=null,f=a.length,d=[];for(let h=0;h<p;++h){for(let g of c[h]){let{permutationIndices:x,expandDims:b}=C.getEinsumPermutation(f,l[g]),w;C.isIdentityPermutation(x)?w=s[g]:(w=Rt({inputs:{x:s[g]},backend:t,attrs:{perm:x}}),d.push(w));let _=w.shape.slice();for(let I=0;I<b.length;++I)_.splice(b[I],0,1);y.arraysEqual(w.shape,_)||(w=ce({inputs:{x:w},backend:t,attrs:{shape:_}}),d.push(w)),m===null?m=w:(m=zf({inputs:{a:w,b:m},backend:t}),d.push(m))}h<p-1&&(u[h]>=0&&(m=zu({inputs:{x:m},backend:t,attrs:{axis:u[h]-(a.length-f),keepDims:!1}}),d.push(m)),f--)}for(let h of d)h!==m&&t.disposeIntermediateTensorInfo(h);return m}var fR={kernelName:Dc,backendName:"webgl",kernelFunc:H7};var q7="return (x >= 0.0) ? x : (exp(x) - 1.0);",K7=`
|
|
vec4 result;
|
|
|
|
result.r = (x.r >= 0.0) ? x.r : (exp(x.r) - 1.0);
|
|
result.g = (x.g >= 0.0) ? x.g : (exp(x.g) - 1.0);
|
|
result.b = (x.b >= 0.0) ? x.b : (exp(x.b) - 1.0);
|
|
result.a = (x.a >= 0.0) ? x.a : (exp(x.a) - 1.0);
|
|
|
|
return result;
|
|
`,X7=_e({opSnippet:q7,packedOpSnippet:K7}),dR={kernelName:Fi,backendName:"webgl",kernelFunc:X7};var Y7="return (b >= 1.0) ? a : a * (b + 1.0);",Z7=`
|
|
vec4 bGTEZero = vec4(greaterThanEqual(b, vec4(0.)));
|
|
return (bGTEZero * a) + ((vec4(1.0) - bGTEZero) * (a * (b + vec4(1.0))));
|
|
`,J7=r=>{let{inputs:e,backend:t}=r,{dy:n,y:o}=e,s=W().getBool("WEBGL_PACK_BINARY_OPERATIONS")?new Ss(Z7,n.shape,o.shape):new co(Y7,n.shape,o.shape);return t.runWebGLProgram(s,[n,o],n.dtype)},hR={kernelName:$c,backendName:"webgl",kernelFunc:J7};var Q7=`
|
|
return vec4(equal(a, b));
|
|
`,eY="return float(a == b);",tY=st({opSnippet:eY,packedOpSnippet:Q7,dtype:"bool"}),gR={kernelName:Pi,backendName:"webgl",kernelFunc:tY};var rY=`
|
|
// Error function is calculated approximately with elementary function.
|
|
// See "Handbook of Mathematical Functions with Formulas,
|
|
// Graphs, and Mathematical Tables", Abramowitz and Stegun.
|
|
float p = ${C.ERF_P};
|
|
float a1 = ${C.ERF_A1};
|
|
float a2 = ${C.ERF_A2};
|
|
float a3 = ${C.ERF_A3};
|
|
float a4 = ${C.ERF_A4};
|
|
float a5 = ${C.ERF_A5};
|
|
|
|
float sign = sign(x);
|
|
x = abs(x);
|
|
float t = 1.0 / (1.0 + p * x);
|
|
return sign * (1.0 - (((((a5*t + a4)*t) + a3)*t + a2)*t + a1)*t*exp(-x*x));
|
|
`,nY=_e({opSnippet:rY}),xR={kernelName:Oi,backendName:"webgl",kernelFunc:nY};var yR="return exp(x);",qv=_e({opSnippet:yR,packedOpSnippet:yR,cpuKernelImpl:pD}),bR={kernelName:$o,backendName:"webgl",kernelFunc:qv};function Ix(r){let{inputs:e,attrs:t,backend:n}=r,{dim:o}=t,{input:s}=e,a=s.shape.length,i=s.shape.slice(),l=o;return o<0&&(y.assert(-(a+1)<=o,()=>`Axis must be in the interval [${-(a+1)}, ${a}]`),l=a+o+1),i.splice(l,0,1),ce({inputs:{x:s},backend:n,attrs:{shape:i}})}var wR={kernelName:Gs,backendName:"webgl",kernelFunc:Ix};var _R="return exp(x) - 1.0;",oY=_e({opSnippet:_R,packedOpSnippet:_R,cpuKernelImpl:mD}),kR={kernelName:Mi,backendName:"webgl",kernelFunc:oY};var Nx=class{constructor(e,t,n){this.variableNames=["real","imag"];let o=t[1];this.outputShape=t;let s=n?`2.0 * ${Math.PI}`:`-2.0 * ${Math.PI}`,a=n?`${o}.0`:"1.0",i;if(e==="real")i="return real * expR - imag * expI;";else if(e==="imag")i="return real * expI + imag * expR;";else throw new Error(`FFT component must be either "real" or "imag", got ${e}.`);this.userCode=`
|
|
const float exponentMultiplier = ${s};
|
|
|
|
float unaryOpComplex(float real, float expR, float imag, float expI) {
|
|
${i}
|
|
}
|
|
|
|
float mulMatDFT(int batch, int index) {
|
|
float indexRatio = float(index) / float(${o});
|
|
float exponentMultiplierTimesIndexRatio =
|
|
exponentMultiplier * indexRatio;
|
|
|
|
float result = 0.0;
|
|
|
|
for (int i = 0; i < ${o}; i++) {
|
|
// x = (-2|2 * PI / N) * index * i;
|
|
float x = exponentMultiplierTimesIndexRatio * float(i);
|
|
float expR = cos(x);
|
|
float expI = sin(x);
|
|
float real = getReal(batch, i);
|
|
float imag = getImag(batch, i);
|
|
|
|
result +=
|
|
unaryOpComplex(real, expR, imag, expI) / ${a};
|
|
}
|
|
|
|
return result;
|
|
}
|
|
|
|
void main() {
|
|
ivec2 coords = getOutputCoords();
|
|
setOutput(mulMatDFT(coords[0], coords[1]));
|
|
}
|
|
`}};function Sx(r,e,t){let n=t.texData.get(r.dataId),o=y.sizeFromShape(r.shape),s=r.shape[r.shape.length-1],a=o/s,i=ce({inputs:{x:r},backend:t,attrs:{shape:[a,s]}}),l=i.shape,u=new Nx("real",l,e),c=new Nx("imag",l,e),p=[{dataId:n.complexTensorInfos.real.dataId,dtype:n.complexTensorInfos.real.dtype,shape:l},{dataId:n.complexTensorInfos.imag.dataId,dtype:n.complexTensorInfos.imag.dtype,shape:l}],m=t.runWebGLProgram(u,p,"float32"),f=t.runWebGLProgram(c,p,"float32"),d=yn({inputs:{real:m,imag:f},backend:t});t.disposeIntermediateTensorInfo(m),t.disposeIntermediateTensorInfo(f);let h=ce({inputs:{x:d},backend:t,attrs:{shape:r.shape}});return t.disposeIntermediateTensorInfo(i),t.disposeIntermediateTensorInfo(d),h}function sY(r){let{inputs:e,backend:t}=r,{input:n}=e;return Sx(n,!1,t)}var vR={kernelName:Rc,backendName:"webgl",kernelFunc:sY};var Kv=class{constructor(e,t){this.outputShape=[],this.variableNames=["x"],this.outputShape=e,this.userCode=`
|
|
uniform float value;
|
|
void main() {
|
|
// Input can be obtained from uniform value.
|
|
setOutput(value);
|
|
}
|
|
`}getCustomSetupFunc(e){return(t,n)=>{this.valueLoc==null&&(this.valueLoc=t.getUniformLocationNoThrow(n,"value")),t.gl.uniform1f(this.valueLoc,e)}}};function Wf(r){let{backend:e,attrs:t}=r,{shape:n,value:o}=t,{dtype:s}=t;if(s=s||y.inferDtype(o),s==="string"){let a=y.getArrayFromDType(s,y.sizeFromShape(n));return a.fill(o),e.makeTensorInfo(n,s,a)}else{let a=new Kv(n,o),i=a.getCustomSetupFunc(o);return e.runWebGLProgram(a,[],s,i)}}var CR={kernelName:rl,backendName:"webgl",kernelFunc:Wf};var Xv=class{constructor(e){this.variableNames=["Image"],this.outputShape=[];let t=e[2];this.outputShape=e,this.userCode=`
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int x = coords[2];
|
|
|
|
int coordX = ${t} - x;
|
|
float outputValue;
|
|
if(coordX >= 0 && coordX < ${t}) {
|
|
outputValue = getImage(coords[0], coords[1], coordX, coords[3]);
|
|
} else {
|
|
outputValue = getImage(coords[0], coords[1], coords[2], coords[3]);
|
|
}
|
|
setOutput(outputValue);
|
|
}
|
|
`}};var IR={kernelName:Li,backendName:"webgl",kernelFunc:({inputs:r,backend:e})=>{let{image:t}=r,n=e,o=new Xv(t.shape);return n.runWebGLProgram(o,[t],t.dtype)}};var NR="return floor(x);",iY=_e({opSnippet:NR,packedOpSnippet:NR,cpuKernelImpl:fD}),SR={kernelName:Ro,backendName:"webgl",kernelFunc:iY};var aY=`
|
|
float s = sign(a) * sign(b);
|
|
int ia = round(a);
|
|
int ib = round(b);
|
|
if (ib != 0) {
|
|
// Windows (D3D) wants guaranteed non-zero int division at compile-time.
|
|
return float(idiv(ia, ib, s));
|
|
} else {
|
|
return NAN;
|
|
}
|
|
`,lY=`
|
|
ivec4 ia = round(a);
|
|
ivec4 ib = round(b);
|
|
bvec4 cond = notEqual(ib, ivec4(0));
|
|
ivec4 result = ivec4(0);
|
|
vec4 s = sign(a) * sign(b);
|
|
|
|
// Windows (D3D) wants guaranteed non-zero int division at compile-time.
|
|
if (cond[0]) {
|
|
result[0] = idiv(ia[0], ib[0], s[0]);
|
|
}
|
|
if (cond[1]) {
|
|
result[1] = idiv(ia[1], ib[1], s[1]);
|
|
}
|
|
if (cond[2]) {
|
|
result[2] = idiv(ia[2], ib[2], s[2]);
|
|
}
|
|
if (cond[3]) {
|
|
result[3] = idiv(ia[3], ib[3], s[3]);
|
|
}
|
|
return vec4(result);
|
|
`,uY=st({opSnippet:aY,packedOpSnippet:lY,dtype:"int32"}),TR={kernelName:Fo,backendName:"webgl",kernelFunc:uY};var Yv=class{constructor(e){this.variableNames=["A"];let t=Pt(),[n,o]=e;this.outputShape=e,this.userCode=`
|
|
void main() {
|
|
ivec3 coords = getOutputCoords();
|
|
int texR = coords[0];
|
|
int texC = coords[1];
|
|
int depth = coords[2];
|
|
vec2 uv = (vec2(texC, texR) + halfCR) / vec2(${o}.0, ${n}.0);
|
|
|
|
vec4 values = ${t.texture2D}(A, uv);
|
|
float value;
|
|
if (depth == 0) {
|
|
value = values.r;
|
|
} else if (depth == 1) {
|
|
value = values.g;
|
|
} else if (depth == 2) {
|
|
value = values.b;
|
|
} else if (depth == 3) {
|
|
value = values.a;
|
|
}
|
|
|
|
setOutput(floor(value * 255.0 + 0.5));
|
|
}
|
|
`}};var Zv=class{constructor(e){this.variableNames=["A"],this.packedInputs=!1,this.packedOutput=!0;let t=Pt(),[n,o]=e;this.outputShape=e,this.userCode=`
|
|
void main() {
|
|
ivec3 coords = getOutputCoords();
|
|
int texR = coords[0];
|
|
int texC = coords[1];
|
|
int depth = coords[2];
|
|
|
|
vec4 result = vec4(0.);
|
|
|
|
for(int row=0; row<=1; row++) {
|
|
for(int col=0; col<=1; col++) {
|
|
texC = coords[1] + row;
|
|
depth = coords[2] + col;
|
|
|
|
vec2 uv = (vec2(texC, texR) + halfCR) /
|
|
vec2(${o}.0, ${n}.0);
|
|
vec4 values = ${t.texture2D}(A, uv);
|
|
float value;
|
|
if (depth == 0) {
|
|
value = values.r;
|
|
} else if (depth == 1) {
|
|
value = values.g;
|
|
} else if (depth == 2) {
|
|
value = values.b;
|
|
} else if (depth == 3) {
|
|
value = values.a;
|
|
}
|
|
|
|
result[row * 2 + col] = floor(value * 255.0 + 0.5);
|
|
}
|
|
}
|
|
|
|
${t.output} = result;
|
|
}
|
|
`}};var AR={kernelName:Rm,backendName:"webgl",kernelFunc:cY},Rp;function cY(r){let{inputs:e,backend:t,attrs:n}=r,{pixels:o}=e,{numChannels:s}=n,a=typeof HTMLVideoElement!="undefined"&&o instanceof HTMLVideoElement,i=typeof HTMLImageElement!="undefined"&&o instanceof HTMLImageElement,[l,u]=a?[o.videoWidth,o.videoHeight]:[o.width,o.height],c=[u,l],p=[u,l,s];(i||a)&&(Rp==null&&(Rp=document.createElement("canvas").getContext("2d")),Rp.canvas.width=l,Rp.canvas.height=u,Rp.drawImage(o,0,0,l,u),o=Rp.canvas);let m=t.makeTensorInfo(c,"int32");t.texData.get(m.dataId).usage=Pr.PIXELS,t.gpgpu.uploadPixelDataToTexture(t.getTexture(m.dataId),o);let f=W().getBool("WEBGL_PACK")?new Zv(p):new Yv(p),d=t.runWebGLProgram(f,[m],"int32");return t.disposeData(m.dataId),d}function pY(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s,bias:a,preluActivationWeights:i}=e,{strides:l,pad:u,dataFormat:c,dilations:p,dimRoundingMode:m,activation:f,leakyreluAlpha:d}=n,h=C.convertConv2DDataFormat(c),g=C.computeConv2DInfo(o.shape,s.shape,l,p,u,m,!1,h),x,b=[];if(g.filterHeight===1&&g.filterWidth===1&&g.dilationHeight===1&&g.dilationWidth===1&&g.strideHeight===1&&g.strideWidth===1&&(g.padInfo.type==="SAME"||g.padInfo.type==="VALID"))x=kx({x:o,filter:s,convInfo:g,backend:t,bias:a,activation:f,preluActivationWeights:i,leakyreluAlpha:d});else if(W().getBool("WEBGL_CONV_IM2COL")&&o.shape[0]===1)x=vx({x:o,filter:s,convInfo:g,backend:t,bias:a,activation:f,preluActivationWeights:i,leakyreluAlpha:d});else{let _=a!=null,I=i!=null,E=f==="leakyrelu",D=f?vl(f,!1):null,$=new Bf(g,_,D,I,E),R=[o,s];if(a&&R.push(a),i&&R.push(i),E){let M=t.makeTensorInfo([],"float32",y.createScalarValue(d,"float32"));R.push(M),b.push(M)}x=t.runWebGLProgram($,R,"float32")}let w=ce({inputs:{x},backend:t,attrs:{shape:g.outShape}});return b.push(x),b.forEach(_=>t.disposeIntermediateTensorInfo(_)),w}var ER={kernelName:ei,backendName:"webgl",kernelFunc:pY};function mY(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,filter:s,bias:a,preluActivationWeights:i}=e,{strides:l,pad:u,dilations:c,dimRoundingMode:p,activation:m,leakyreluAlpha:f}=n,d=[],h=c;h==null&&(h=[1,1]),y.assert(C.eitherStridesOrDilationsAreOne(l,h),()=>`Error in depthwiseConv2d: Either strides or dilations must be 1. Got strides ${l} and dilations '${h}'`);let g=C.computeConv2DInfo(o.shape,s.shape,l,h,u,p,!0),x=W().getBool("WEBGL_PACK_DEPTHWISECONV")&&g.strideWidth<=2&&g.outChannels/g.inChannels==1,b=m?vl(m,x):null,w=[o,s],_=a!=null,I=i!=null,E=m==="leakyrelu";if(_&&w.push(a),I&&w.push(i),E){let R=t.makeTensorInfo([],"float32",y.createScalarValue(f,"float32"));w.push(R),d.push(R)}let D;x?D=new Gf(g,_,b,I,E):D=new Vf(g,_,b,I,E);let $=t.runWebGLProgram(D,w,"float32");return d.forEach(R=>t.disposeIntermediateTensorInfo(R)),$}var DR={kernelName:ti,backendName:"webgl",kernelFunc:mY};var Jv=class{constructor(e,t,n){this.sliceDim=e,this.strides=t,this.variableNames=["x","indices"],this.outputShape=n;let o=Ve(t.length),s=Ve(n.length),a=this.sliceDim>1?"strides[j]":"strides";this.userCode=`
|
|
${o} strides = ${o}(${this.strides});
|
|
void main() {
|
|
${s} coords = getOutputCoords();
|
|
int flattenIndex = 0;
|
|
for (int j = 0; j < ${this.sliceDim}; j++) {
|
|
int index = round(getIndices(coords[0], j));
|
|
flattenIndex += index * ${a};
|
|
}
|
|
setOutput(getX(flattenIndex, coords[1]));
|
|
}
|
|
`}};function fY(r){let{inputs:e,backend:t}=r,{params:n,indices:o}=e,s=o.shape,a=s[s.length-1],[i,l,u,c]=C.prepareAndValidate(n,o),p=ce({inputs:{x:o},backend:t,attrs:{shape:[l,a]}}),m=ce({inputs:{x:n},backend:t,attrs:{shape:[y.sizeFromShape(n.shape)/u,u]}}),f=new Jv(a,c,[l,u]),d=t.runWebGLProgram(f,[m,p],m.dtype),h=ce({inputs:{x:d},backend:t,attrs:{shape:i}});return t.disposeIntermediateTensorInfo(p),t.disposeIntermediateTensorInfo(m),t.disposeIntermediateTensorInfo(d),h}var $R={kernelName:zi,backendName:"webgl",kernelFunc:fY};var Qv=class{constructor(e,t){this.variableNames=["A","indices"],this.outputShape=t,this.rank=t.length;let n=Ve(this.rank),o=dY(e,2);this.userCode=`
|
|
void main() {
|
|
${n} resRC = getOutputCoords();
|
|
setOutput(getA(${o}));
|
|
}
|
|
`}};function dY(r,e){let t=["resRC.x","resRC.y","resRC.z","resRC.w"],n=[];for(let o=0;o<r.length;o++)o===2?n.push("int(getIndices(resRC.x, resRC.z))"):n.push(`${t[o]}`);return n.join()}function hY(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,indices:s}=e,{axis:a,batchDims:i}=n,l=y.parseAxisParam(a,o.shape)[0],u=C.segment_util.collectGatherOpShapeInfo(o,s,l,i),c=y.sizeFromShape(s.shape),p=[],m=ce({inputs:{x:o},backend:t,attrs:{shape:[u.batchSize,u.outerSize,u.dimSize,u.sliceSize]}}),f=ce({inputs:{x:s},backend:t,attrs:{shape:[u.batchSize,c/u.batchSize]}});p.push(m),p.push(f);let d=[u.batchSize,u.outerSize,c/u.batchSize,u.sliceSize];if(t.shouldExecuteOnCPU([o,s])||o.dtype==="string"){let b=t.bufferSync(f),w=t.bufferSync(m),_=dD(w,b,d);return p.forEach(I=>t.disposeIntermediateTensorInfo(I)),t.makeTensorInfo(u.outputShape,_.dtype,_.values)}let h=new Qv(m.shape,d),g=t.runWebGLProgram(h,[m,f],m.dtype);p.push(g);let x=ce({inputs:{x:g},backend:t,attrs:{shape:u.outputShape}});return p.forEach(b=>t.disposeIntermediateTensorInfo(b)),x}var RR={kernelName:Ws,backendName:"webgl",kernelFunc:hY};var gY="return float(a > b);",xY=`
|
|
return vec4(greaterThan(a, b));
|
|
`,yY=st({opSnippet:gY,packedOpSnippet:xY,cpuKernelImpl:hD,dtype:"bool"}),FR={kernelName:Bi,backendName:"webgl",kernelFunc:yY};var bY="return float(a >= b);",wY=`
|
|
return vec4(greaterThanEqual(a, b));
|
|
`,_Y=st({opSnippet:bY,packedOpSnippet:wY,dtype:"bool"}),OR={kernelName:Po,backendName:"webgl",kernelFunc:_Y};function kY(r){let{inputs:e,backend:t}=r,{input:n}=e;return Sx(n,!0,t)}var PR={kernelName:Fc,backendName:"webgl",kernelFunc:kY};var vY="return float(!isnan(x) && !isinf(x));",CY=_e({opSnippet:vY,dtype:"bool"}),MR={kernelName:Vi,backendName:"webgl",kernelFunc:CY};var IY="return float(isinf(x));",NY=_e({opSnippet:IY,dtype:"bool"}),LR={kernelName:Gi,backendName:"webgl",kernelFunc:NY};var SY="return float(isnan(x));",TY=_e({opSnippet:SY,dtype:"bool"}),zR={kernelName:Wi,backendName:"webgl",kernelFunc:TY};var AY="return float(a < b);",EY=`
|
|
return vec4(lessThan(a, b));
|
|
`,DY=st({opSnippet:AY,packedOpSnippet:EY,cpuKernelImpl:gD,dtype:"bool"}),BR={kernelName:ji,backendName:"webgl",kernelFunc:DY};var $Y="return float(a <= b);",RY=`
|
|
return vec4(lessThanEqual(a, b));
|
|
`,FY=st({opSnippet:$Y,packedOpSnippet:RY,dtype:"bool"}),VR={kernelName:Ui,backendName:"webgl",kernelFunc:FY};function OY(r){let{backend:e,attrs:t}=r,{start:n,stop:o,num:s}=t,a=xD(n,o,s);return e.makeTensorInfo([a.length],"float32",a)}var GR={kernelName:Pc,backendName:"webgl",kernelFunc:OY};var PY=`if (x < 0.0) return NAN;
|
|
return log(x);`,MY=`
|
|
vec4 result = log(x);
|
|
vec4 isNaN = vec4(lessThan(x, vec4(0.0)));
|
|
result.r = isNaN.r == 1.0 ? NAN : result.r;
|
|
result.g = isNaN.g == 1.0 ? NAN : result.g;
|
|
result.b = isNaN.b == 1.0 ? NAN : result.b;
|
|
result.a = isNaN.a == 1.0 ? NAN : result.a;
|
|
|
|
return result;
|
|
`,LY=_e({opSnippet:PY,packedOpSnippet:MY,cpuKernelImpl:yD}),WR={kernelName:Lo,backendName:"webgl",kernelFunc:LY};var zY="return log(1.0 + x);",BY=_e({opSnippet:zY}),jR={kernelName:Hi,backendName:"webgl",kernelFunc:BY};var VY="return float(a >= 1.0 && b >= 1.0);",GY=`
|
|
return vec4(
|
|
vec4(greaterThanEqual(a, vec4(1.0))) *
|
|
vec4(greaterThanEqual(b, vec4(1.0))));
|
|
`,WY=st({opSnippet:VY,packedOpSnippet:GY,dtype:"bool"}),UR={kernelName:qi,backendName:"webgl",kernelFunc:WY};var jY="return float(!(x >= 1.0));",UY=_e({opSnippet:jY}),HR={kernelName:Hl,backendName:"webgl",kernelFunc:UY};var HY="return float(a >= 1.0 || b >= 1.0);",qY=`
|
|
return min(
|
|
vec4(greaterThanEqual(a, vec4(1.0))) +
|
|
vec4(greaterThanEqual(b, vec4(1.0))),
|
|
vec4(1.0));
|
|
`,KY=st({opSnippet:HY,packedOpSnippet:qY,dtype:"bool"}),qR={kernelName:ql,backendName:"webgl",kernelFunc:KY};var eC=class{constructor(e,t,n,o,s){this.variableNames=["x"],this.outputShape=[];let a=t,i=e[3]-1;this.outputShape=e;let l,u=`float(${n}) + float(${o}) * sum`;s===.5?l=`inversesqrt(${u})`:s===1?l=`1.0/(${u})`:l=`exp(log(${u}) * float(-${s}));`,this.userCode=`
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int r = coords[1];
|
|
int c = coords[2];
|
|
int d = coords[3];
|
|
float x = getX(b, r, c, d);
|
|
float sum = 0.0;
|
|
for (int j = -${a}; j <= ${a}; j++) {
|
|
int idx = d + j;
|
|
if (idx >= 0 && idx <= ${i}) {
|
|
float z = getX(b, r, c, idx);
|
|
sum += z * z;
|
|
}
|
|
}
|
|
float val = x * ${l};
|
|
setOutput(val);
|
|
}
|
|
`}};var tC=class{constructor(e,t,n,o,s){this.variableNames=["x"],this.outputShape=[],this.packedInputs=!0,this.packedOutput=!0;let a=t,i=e[3]-1;this.outputShape=e;let l,u=`float(${n}) + float(${o}) * sum`;s===.5?l=`inversesqrt(${u})`:s===1?l=`1.0/(${u})`:l=`exp(log(${u}) * float(-${s}));`,this.userCode=`
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords.x;
|
|
int r = coords.y;
|
|
int c = coords.z;
|
|
int d = coords.w;
|
|
|
|
bool hasNextCol = d < ${this.outputShape[3]};
|
|
bool hasNextRow = c < ${this.outputShape[2]};
|
|
|
|
vec4 sum = vec4(0.);
|
|
vec4 xFragAtOutputCoords = getX(b, r, c, d);
|
|
|
|
vec4 xAtOutputCoords = vec4(
|
|
getChannel(xFragAtOutputCoords, vec2(c, d)),
|
|
hasNextCol ?
|
|
getChannel(xFragAtOutputCoords, vec2(c, d + 1)) : 0.0,
|
|
hasNextRow ?
|
|
getChannel(xFragAtOutputCoords , vec2(c + 1, d)) : 0.0,
|
|
(hasNextRow && hasNextCol) ?
|
|
getChannel(xFragAtOutputCoords, vec2(c + 1, d + 1)) : 0.0
|
|
);
|
|
|
|
int firstChannel = d - ${a};
|
|
vec2 cache = vec2(0.);
|
|
if(firstChannel >= 0){
|
|
vec4 firstChannelFrag = getX(b, r, c, firstChannel);
|
|
cache.x = getChannel(firstChannelFrag, vec2(c, firstChannel));
|
|
if(hasNextRow){
|
|
cache.y = getChannel(firstChannelFrag, vec2(c + 1, firstChannel));
|
|
}
|
|
}
|
|
|
|
ivec2 depth = ivec2(d, d + 1);
|
|
for (int j = - ${a}; j <= ${a}; j++) {
|
|
ivec2 idx = depth + j;
|
|
bvec2 aboveLowerBound = greaterThanEqual(idx, ivec2(0));
|
|
bvec2 belowUpperBound = lessThanEqual(idx, ivec2(${i}));
|
|
|
|
bool depthInRange = aboveLowerBound.x && belowUpperBound.x;
|
|
bool depthPlusOneInRange = aboveLowerBound.y && belowUpperBound.y;
|
|
|
|
if(depthInRange || depthPlusOneInRange){
|
|
vec4 z = vec4(0.);
|
|
vec4 xFragAtCurrentDepth;
|
|
z.xz = cache.xy;
|
|
if(depthPlusOneInRange && hasNextCol){
|
|
xFragAtCurrentDepth = idx.y != d ?
|
|
getX(b, r, c, idx.y) : xFragAtOutputCoords;
|
|
z.y = getChannel(xFragAtCurrentDepth, vec2(c, idx.y));
|
|
if(hasNextRow){
|
|
z.w = getChannel(xFragAtCurrentDepth, vec2(c + 1, idx.y));
|
|
}
|
|
}
|
|
cache.xy = z.yw;
|
|
sum += z * z;
|
|
}
|
|
}
|
|
vec4 result = xAtOutputCoords * ${l};
|
|
setOutput(result);
|
|
}
|
|
`}};var XY=r=>{let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{depthRadius:s,bias:a,alpha:i,beta:l}=n,u=W().getBool("WEBGL_PACK_NORMALIZATION")?new tC(o.shape,s,a,i,l):new eC(o.shape,s,a,i,l);return t.runWebGLProgram(u,[o],o.dtype)},KR={kernelName:nl,backendName:"webgl",kernelFunc:XY};var rC=class{constructor(e,t,n,o,s){this.variableNames=["inputImage","outputImage","dy"],this.outputShape=[],this.outputShape=e,this.depth=e[3],this.depthRadius=t,this.bias=n,this.alpha=o,this.beta=s,this.userCode=`
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int r = coords[1];
|
|
int c = coords[2];
|
|
|
|
float result = 0.0;
|
|
for (int d = 0; d < ${this.depth}; ++d) {
|
|
int depthBegin = int(max(0.0, float(d - ${t})));
|
|
int depthEnd = int(min(float(${this.depth}),
|
|
float(d + ${t} + 1)));
|
|
|
|
const int MIN_DEPTH_BEGIN = 0;
|
|
const int MAX_DEPTH_END = ${this.depth};
|
|
|
|
float norm = 0.0;
|
|
for (int k = MIN_DEPTH_BEGIN; k < MAX_DEPTH_END; ++k) {
|
|
if (k < depthBegin){
|
|
continue;
|
|
}
|
|
else if (k >= depthBegin && k < depthEnd) {
|
|
norm += getInputImage(b, r, c, k) * getInputImage(b, r, c, k);
|
|
}
|
|
else {
|
|
break;
|
|
}
|
|
}
|
|
|
|
norm = float(${o}) * norm + float(${n});
|
|
|
|
for(int k = MIN_DEPTH_BEGIN; k < MAX_DEPTH_END; ++k){
|
|
if (k < depthBegin){
|
|
continue;
|
|
}
|
|
else if (k >= depthBegin && k < depthEnd){
|
|
float dyi = -2.0 * float(${o})
|
|
* float(${s})
|
|
* getInputImage(b ,r ,c, k) * getOutputImage(b, r, c, d)
|
|
/ norm;
|
|
if (k == d) {
|
|
dyi += pow(norm, -1.0 * ${s});
|
|
}
|
|
if (k == coords[3]) {
|
|
dyi *= getDy(b, r, c, d);
|
|
result += dyi;
|
|
}
|
|
}
|
|
else {
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
setOutput(result);
|
|
}
|
|
`}};var YY=r=>{let{inputs:e,backend:t,attrs:n}=r,{x:o,y:s,dy:a}=e,{depthRadius:i,bias:l,alpha:u,beta:c}=n,p=new rC(o.shape,i,l,u,c);return t.runWebGLProgram(p,[o,s,a],o.dtype)},XR={kernelName:Mc,backendName:"webgl",kernelFunc:YY};function YR(r,e,t,n){let o=y.sizeFromShape(e),a=y.sizeFromShape(r.shape)/o,i=ce({inputs:{x:r},attrs:{shape:[a,o]},backend:n}),l=En(i,r.dtype,"max",n),u=ce({inputs:{x:l},attrs:{shape:t},backend:n});return n.disposeIntermediateTensorInfo(i),n.disposeIntermediateTensorInfo(l),u}function nC(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{reductionIndices:s,keepDims:a}=n,i=o.shape.length,l=y.parseAxisParam(s,o.shape),u=l,c=C.getAxesPermutation(u,i),p=c!=null,m=t.shouldExecuteOnCPU([o]),f=o;if(p){if(m){let w=t.texData.get(f.dataId).values,_=new Array(i);for(let D=0;D<_.length;D++)_[D]=o.shape[c[D]];let I=Mu(w,o.shape,o.dtype,c,_);f=t.makeTensorInfo(_,o.dtype);let E=t.texData.get(f.dataId);E.values=I}else f=Cl(o,c,t);u=C.getInnerMostAxes(u.length,i)}C.assertAxesAreInnerMostDims("max",u,i);let[d,h]=C.computeOutAndReduceShapes(f.shape,u),g=d;a&&(g=C.expandShapeToKeepDim(d,l));let x;if(m){let w=t.texData.get(f.dataId).values,_=bD(w,y.sizeFromShape(h),g,o.dtype);x=t.makeTensorInfo(g,o.dtype);let I=t.texData.get(x.dataId);I.values=_}else x=YR(f,h,g,t);return p&&t.disposeIntermediateTensorInfo(f),x}var ZR={kernelName:zo,backendName:"webgl",kernelFunc:nC};var ZY=hx+`
|
|
return max(a, b);
|
|
`,JY=`
|
|
vec4 result = vec4(max(a, b));
|
|
vec4 isNaN = min(vec4(isnan(a)) + vec4(isnan(b)), vec4(1.0));
|
|
`+kl+`
|
|
return result;
|
|
`,QY=st({opSnippet:ZY,packedOpSnippet:JY,cpuKernelImpl:wD}),JR={kernelName:Bo,backendName:"webgl",kernelFunc:QY};function e9(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e;Cs(o,"maxPool");let{filterSize:s,strides:a,pad:i,dimRoundingMode:l}=n,u=1;y.assert(C.eitherStridesOrDilationsAreOne(a,u),()=>`Error in maxPool: Either strides or dilations must be 1. Got strides ${a} and dilations '${u}'`);let c=C.computePool2DInfo(o.shape,s,a,u,i,l);if(c.filterWidth===1&&c.filterHeight===1&&y.arraysEqual(c.inShape,c.outShape))return Ut({inputs:{x:o},backend:t});let p=new ci(c,"max",!1);return t.runWebGLProgram(p,[o],o.dtype)}var QR={kernelName:Vo,backendName:"webgl",kernelFunc:e9};function t9(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{filterSize:s,strides:a,pad:i,dataFormat:l,dimRoundingMode:u}=n,c=[1,1,1],p=C.computePool3DInfo(o.shape,s,a,c,i,u,l),m=new Vu(p,"max",!1);return t.runWebGLProgram(m,[o],o.dtype)}var eF={kernelName:ol,backendName:"webgl",kernelFunc:t9};var oC=class{constructor(e){this.variableNames=["dy","maxPos"],this.outputShape=e.inShape;let t=e.strideHeight,n=e.strideWidth,o=e.dilationHeight,s=e.effectiveFilterHeight,a=e.effectiveFilterWidth,i=s-1-e.padInfo.top,l=a-1-e.padInfo.left,u=s*a-1;this.userCode=`
|
|
const ivec2 pads = ivec2(${i}, ${l});
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int d = coords[3];
|
|
|
|
ivec2 dyRCCorner = coords.yz - pads;
|
|
int dyRCorner = dyRCCorner.x;
|
|
int dyCCorner = dyRCCorner.y;
|
|
|
|
// Convolve dy(?, ?, d) with pos mask(:, :, d) to get dx(xR, xC, d).
|
|
// ? = to be determined. : = across all values in that axis.
|
|
float dotProd = 0.0;
|
|
for (int wR = 0; wR < ${s};
|
|
wR += ${o}) {
|
|
float dyR = float(dyRCorner + wR) / ${t}.0;
|
|
|
|
if (dyR < 0.0 || dyR >= ${e.outHeight}.0 || fract(dyR) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyR = int(dyR);
|
|
|
|
for (int wC = 0; wC < ${a}; wC++) {
|
|
float dyC = float(dyCCorner + wC) / ${n}.0;
|
|
|
|
if (dyC < 0.0 || dyC >= ${e.outWidth}.0 ||
|
|
fract(dyC) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyC = int(dyC);
|
|
|
|
float dyValue = getDy(b, idyR, idyC, d);
|
|
int maxPosValue = ${u} - int(getMaxPos(b, idyR, idyC, d));
|
|
|
|
// Get the current value, check it against the value from the
|
|
// position matrix.
|
|
int curPosValue = wR * ${a} + wC;
|
|
float mask = float(maxPosValue == curPosValue ? 1.0 : 0.0);
|
|
|
|
dotProd += dyValue * mask;
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}},sC=class{constructor(e){this.variableNames=["dy","maxPos"],this.outputShape=e.inShape;let t=e.strideDepth,n=e.strideHeight,o=e.strideWidth,s=e.dilationDepth,a=e.dilationHeight,i=e.dilationWidth,l=e.effectiveFilterDepth,u=e.effectiveFilterHeight,c=e.effectiveFilterWidth,p=l-1-e.padInfo.front,m=u-1-e.padInfo.top,f=c-1-e.padInfo.left,d=l*u*c-1;this.userCode=`
|
|
const ivec3 pads = ivec3(${p}, ${m}, ${f});
|
|
|
|
void main() {
|
|
ivec5 coords = getOutputCoords();
|
|
int batch = coords.x;
|
|
int ch = coords.u;
|
|
|
|
ivec3 dyCorner = ivec3(coords.y, coords.z, coords.w) - pads;
|
|
int dyDCorner = dyCorner.x;
|
|
int dyRCorner = dyCorner.y;
|
|
int dyCCorner = dyCorner.z;
|
|
|
|
// Convolve dy(?, ?, ?, ch) with pos mask(:, :, :, d) to get
|
|
// dx(xD, xR, xC, ch).
|
|
// ? = to be determined. : = across all values in that axis.
|
|
float dotProd = 0.0;
|
|
|
|
for (int wD = 0; wD < ${l};
|
|
wD += ${s}) {
|
|
float dyD = float(dyDCorner + wD) / ${t}.0;
|
|
|
|
if (dyD < 0.0 || dyD >= ${e.outDepth}.0 || fract(dyD) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyD = int(dyD);
|
|
|
|
for (int wR = 0; wR < ${u};
|
|
wR += ${a}) {
|
|
float dyR = float(dyRCorner + wR) / ${n}.0;
|
|
|
|
if (dyR < 0.0 || dyR >= ${e.outHeight}.0 ||
|
|
fract(dyR) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyR = int(dyR);
|
|
|
|
for (int wC = 0; wC < ${c};
|
|
wC += ${i}) {
|
|
float dyC = float(dyCCorner + wC) / ${o}.0;
|
|
|
|
if (dyC < 0.0 || dyC >= ${e.outWidth}.0 ||
|
|
fract(dyC) > 0.0) {
|
|
continue;
|
|
}
|
|
int idyC = int(dyC);
|
|
|
|
float dyValue = getDy(batch, idyD, idyR, idyC, ch);
|
|
int maxPosValue = ${d} -
|
|
int(getMaxPos(batch, idyD, idyR, idyC, ch));
|
|
|
|
// Get the current value, check it against the value from the
|
|
// position matrix.
|
|
int curPosValue =
|
|
wD * ${u} * ${c} +
|
|
wR * ${c} + wC;
|
|
float mask = float(maxPosValue == curPosValue ? 1.0 : 0.0);
|
|
|
|
dotProd += dyValue * mask;
|
|
}
|
|
}
|
|
}
|
|
setOutput(dotProd);
|
|
}
|
|
`}};function r9(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,input:s}=e,a=s,{filterSize:i,strides:l,pad:u,dimRoundingMode:c}=n,p=[1,1,1],m=C.computePool3DInfo(a.shape,i,l,p,u,c),f=new Vu(m,"max",!0),d=t.runWebGLProgram(f,[a],a.dtype),h=new sC(m),g=t.runWebGLProgram(h,[o,d],a.dtype);return t.disposeIntermediateTensorInfo(d),g}var tF={kernelName:zc,backendName:"webgl",kernelFunc:r9};function n9(r){let{inputs:e,backend:t,attrs:n}=r,{dy:o,input:s,output:a}=e,i=s;Cs([s,a],"maxPoolGrad");let{filterSize:l,strides:u,pad:c,dimRoundingMode:p}=n,m=C.computePool2DInfo(i.shape,l,u,1,c,p),f=!0,d=new ci(m,"max",f),h=t.runWebGLProgram(d,[i],i.dtype),g=new oC(m),x=t.runWebGLProgram(g,[o,h],i.dtype);return t.disposeIntermediateTensorInfo(h),x}var rF={kernelName:Lc,backendName:"webgl",kernelFunc:n9};function nF(r,e,t,n){let o=new ci(t,"max",!1),s=n.runWebGLProgram(o,[r],"float32");o=new ci(t,"max",!0,!0,e);let a=n.runWebGLProgram(o,[r],"float32");return[s,a]}var oF={kernelName:Bc,backendName:"webgl",kernelFunc:({inputs:r,attrs:e,backend:t})=>{let{x:n}=r,{filterSize:o,strides:s,pad:a,includeBatchInIndex:i}=e,l=t;y.assert(n.shape.length===4,()=>`Error in maxPool: input must be rank 4 but got rank ${n.shape.length}.`);let u=[1,1];y.assert(C.eitherStridesOrDilationsAreOne(s,u),()=>`Error in maxPool: Either strides or dilations must be 1. Got strides ${s} and dilations '${u}'`);let c=C.computePool2DInfo(n.shape,o,s,u,a),[p,m]=nF(n,i,c,l);return[p,m]}};function sF(r,e,t,n){let o=y.sizeFromShape(e),a=y.sizeFromShape(r.shape)/o,i=ce({inputs:{x:r},attrs:{shape:[a,o]},backend:n}),l=En(i,"float32","mean",n),u=ce({inputs:{x:l},attrs:{shape:t},backend:n});return n.disposeIntermediateTensorInfo(i),n.disposeIntermediateTensorInfo(l),u}var iF={kernelName:Go,backendName:"webgl",kernelFunc:({inputs:r,attrs:e,backend:t})=>{let{x:n}=r,{keepDims:o,axis:s}=e,a=t,i=n.shape.length,l=y.parseAxisParam(s,n.shape),u=l,c=C.getAxesPermutation(u,i),p=c!=null,m=a.shouldExecuteOnCPU([n]),f=[],d=n;if(p){if(m){let _=a.texData.get(d.dataId).values,I=new Array(i);for(let $=0;$<I.length;$++)I[$]=n.shape[c[$]];let E=Mu(_,n.shape,n.dtype,c,I);d=a.makeTensorInfo(I,n.dtype);let D=a.texData.get(d.dataId);D.values=E}else d=Cl(n,c,a);f.push(d),u=C.getInnerMostAxes(u.length,i)}C.assertAxesAreInnerMostDims("sum",u,i);let[h,g]=C.computeOutAndReduceShapes(d.shape,u),x=h;o&&(x=C.expandShapeToKeepDim(h,l));let b=sF(d,g,x,a);for(let w of f)a.disposeIntermediateTensorInfo(w);return b}};function o9(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n,i=o.shape.length,l=y.parseAxisParam(s,o.shape),u=l,c=C.getAxesPermutation(u,i),p=o;c!=null&&(p=Rt({inputs:{x:o},backend:t,attrs:{perm:c}}),u=C.getInnerMostAxes(u.length,o.shape.length)),C.assertAxesAreInnerMostDims("min",u,i);let[m,f]=C.computeOutAndReduceShapes(p.shape,u),d=y.sizeFromShape(f),h=ce({inputs:{x:p},backend:t,attrs:{shape:[-1,d]}}),g=En(h,h.dtype,"min",t),x;if(a){let b=C.expandShapeToKeepDim(m,l);x=ce({inputs:{x:g},backend:t,attrs:{shape:b}})}else x=ce({inputs:{x:g},backend:t,attrs:{shape:m}});return t.disposeIntermediateTensorInfo(h),t.disposeIntermediateTensorInfo(g),c!=null&&t.disposeIntermediateTensorInfo(p),x}var aF={kernelName:Wo,backendName:"webgl",kernelFunc:o9};var s9=hx+`
|
|
return min(a, b);
|
|
`,i9=`
|
|
vec4 result = vec4(min(a, b));
|
|
vec4 isNaN = min(vec4(isnan(a)) + vec4(isnan(b)), vec4(1.0));
|
|
`+kl+`
|
|
return result;
|
|
`,a9=st({opSnippet:s9,packedOpSnippet:i9,cpuKernelImpl:_D}),lF={kernelName:jo,backendName:"webgl",kernelFunc:a9};var iC=class{constructor(e,t,n){this.variableNames=["x"],this.outputShape=t.map((c,p)=>c[0]+e[p]+c[1]);let o=e.length,s=Ve(o),a=t.map(c=>c[0]).join(","),i=t.map((c,p)=>c[0]+e[p]).join(","),l=["coords[0]","coords[1]","coords[2]","coords[3]"].slice(0,o),u=n==="reflect"?0:1;if(o===1){this.userCode=`
|
|
int start = ${a};
|
|
int end = ${i};
|
|
|
|
void main() {
|
|
int outC = getOutputCoords();
|
|
if (outC < start) {
|
|
outC = start * 2 - outC - ${u};
|
|
} else if(outC >= end) {
|
|
outC = (end - 1) * 2 - outC + ${u};
|
|
}
|
|
setOutput(getX(outC - start));
|
|
}
|
|
`;return}this.userCode=`
|
|
${s} start = ${s}(${a});
|
|
${s} end = ${s}(${i});
|
|
|
|
void main() {
|
|
${s} outC = getOutputCoords();
|
|
for (int i = 0; i < ${o}; i++) {
|
|
if (outC[i] < start[i]) {
|
|
outC[i] = start[i] * 2 - outC[i] - ${u};
|
|
} else if(outC[i] >= end[i]) {
|
|
outC[i] = (end[i] - 1) * 2 - outC[i] + ${u};
|
|
}
|
|
}
|
|
${s} coords = outC - start;
|
|
setOutput(getX(${l}));
|
|
}
|
|
`}};var aC=class{constructor(e,t,n){this.variableNames=["x"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=t.map((d,h)=>d[0]+e[h]+d[1]);let o=e.length,s=Ve(o),a=t.map(d=>d[0]).join(","),i=t.map((d,h)=>d[0]+e[h]).join(","),l=jt("rc",o),u=jt("source",o),c=`${l[o-1]} < ${this.outputShape[o-1]}`,p=o===1?"source":`vec2(${u.slice(-2).join()})`,m=n==="reflect"?0:1,f="";if(o===1){let d=`
|
|
${s} source = rc;
|
|
if (source < start) {
|
|
source = start * 2 - source - ${m};
|
|
} else if (source >= end) {
|
|
source = (end - 1) * 2 - source + ${m};
|
|
}
|
|
source -= start;
|
|
`;f=`
|
|
${s} rc = outputLoc;
|
|
${d}
|
|
result[0] = getChannel(getX(${u.join()}), ${p});
|
|
${l[o-1]} += 1;
|
|
if(${c}) {
|
|
${d}
|
|
result[1] = getChannel(getX(${u.join()}), ${p});
|
|
}
|
|
`}else{let d=`
|
|
${s} source = rc;
|
|
${s} lt = ${s}(lessThan(source, start));
|
|
${s} gte = ${s}(greaterThanEqual(source, end));
|
|
${s} orig = 1 - (lt + gte);
|
|
source = orig * source +
|
|
lt * (start * 2 - source - ${m}) +
|
|
gte * ((end - 1) * 2 - source + ${m});
|
|
source -= start;
|
|
`;f=`
|
|
${s} rc = outputLoc;
|
|
${d}
|
|
result[0] = getChannel(getX(${u.join()}), ${p});
|
|
${l[o-1]} += 1;
|
|
if(${c}) {
|
|
${d}
|
|
result[1] = getChannel(getX(${u.join()}), ${p});
|
|
}
|
|
rc = outputLoc;
|
|
${l[o-2]} += 1;
|
|
if(${l[o-2]} < ${this.outputShape[o-2]}) {
|
|
${d}
|
|
result[2] = getChannel(getX(${u.join()}), ${p});
|
|
${l[o-1]} += 1;
|
|
if(${c}) {
|
|
${d}
|
|
result[3] = getChannel(getX(${u.join()}), ${p});
|
|
}
|
|
}
|
|
`}this.userCode=`
|
|
const ${s} start = ${s}(${a});
|
|
const ${s} end = ${s}(${i});
|
|
|
|
void main() {
|
|
${s} outputLoc = getOutputCoords();
|
|
vec4 result = vec4(0.);
|
|
${f}
|
|
setOutput(result);
|
|
}
|
|
`}};var l9=({inputs:r,backend:e,attrs:t})=>{let{x:n}=r,{paddings:o,mode:s}=t,a=W().getBool("WEBGL_PACK_ARRAY_OPERATIONS")?new aC(n.shape,o,s):new iC(n.shape,o,s);return e.runWebGLProgram(a,[n],n.dtype)},uF={kernelName:Uo,backendName:"webgl",kernelFunc:l9};var u9=`if (b == 0.0) return NAN;
|
|
return mod(a, b);`,c9=`
|
|
vec4 result = mod(a, b);
|
|
vec4 isNaN = vec4(equal(b, vec4(0.0)));
|
|
`+kl+`
|
|
return result;
|
|
`,p9=st({opSnippet:u9,packedOpSnippet:c9}),cF={kernelName:Ki,backendName:"webgl",kernelFunc:p9};var lC=class{constructor(e,t,n){this.variableNames=["probs"],this.outputShape=[e,n],this.userCode=`
|
|
uniform float seed;
|
|
|
|
void main() {
|
|
ivec2 coords = getOutputCoords();
|
|
int batch = coords[0];
|
|
|
|
float r = random(seed);
|
|
float cdf = 0.0;
|
|
|
|
for (int i = 0; i < ${t-1}; i++) {
|
|
cdf += getProbs(batch, i);
|
|
|
|
if (r < cdf) {
|
|
setOutput(float(i));
|
|
return;
|
|
}
|
|
}
|
|
|
|
// If no other event happened, last event happened.
|
|
setOutput(float(${t-1}));
|
|
}
|
|
`}getCustomSetupFunc(e){return(t,n)=>{this.seedLoc==null&&(this.seedLoc=t.getUniformLocation(n,"seed")),t.gl.uniform1f(this.seedLoc,e)}}};var m9=`
|
|
if (a == b) {
|
|
return 1.0;
|
|
};
|
|
return a / b;`,f9=`
|
|
// vec4 one = vec4(equal(a, b));
|
|
// return one + (vec4(1.0) - one) * a / b;
|
|
vec4 result = a / b;
|
|
if(a.x == b.x) {
|
|
result.x = 1.;
|
|
}
|
|
if(a.y == b.y) {
|
|
result.y = 1.;
|
|
}
|
|
if(a.z == b.z) {
|
|
result.z = 1.;
|
|
}
|
|
if(a.w == b.w) {
|
|
result.w = 1.;
|
|
}
|
|
|
|
return result;
|
|
`,uC=st({opSnippet:m9,packedOpSnippet:f9,checkOutOfBounds:!0}),pF={kernelName:Do,backendName:"webgl",kernelFunc:uC};var mF="return a - b;",cC=st({opSnippet:mF,packedOpSnippet:mF,supportsComplex:!0,cpuKernelImpl:DD}),fF={kernelName:us,backendName:"webgl",kernelFunc:cC};function pC(r){let{inputs:e,backend:t,attrs:n}=r,{logits:o}=e,{dim:s}=n,a=y.parseAxisParam([s],o.shape),i=nC({inputs:{x:o},backend:t,attrs:{reductionIndices:a,keepDims:!1}}),l=C.expandShapeToKeepDim(i.shape,a),u=ce({inputs:{x:i},backend:t,attrs:{shape:l}}),c=cC({inputs:{a:o,b:u},backend:t}),p=qv({inputs:{x:c},backend:t}),m=zu({inputs:{x:p},backend:t,attrs:{axis:a,keepDims:!1}}),f=ce({inputs:{x:m},backend:t,attrs:{shape:l}}),d=uC({inputs:{a:p,b:f},backend:t});return t.disposeIntermediateTensorInfo(i),t.disposeIntermediateTensorInfo(u),t.disposeIntermediateTensorInfo(c),t.disposeIntermediateTensorInfo(p),t.disposeIntermediateTensorInfo(m),t.disposeIntermediateTensorInfo(f),d}var dF={kernelName:as,backendName:"webgl",kernelFunc:pC};function d9(r){let{inputs:e,backend:t,attrs:n}=r,{logits:o}=e,{numSamples:s,seed:a,normalized:i}=n,l=i?o:pC({inputs:{logits:o},backend:t,attrs:{dim:o.shape.length-1}}),u=l.shape[0],c=l.shape[1],p=new lC(u,c,s),m=p.getCustomSetupFunc(a),f=t.runWebGLProgram(p,[l],"int32",m);return i||t.disposeIntermediateTensorInfo(l),f}var hF={kernelName:Vc,backendName:"webgl",kernelFunc:d9};var gF="return -x;";function h9(r){let{inputs:e,backend:t}=r,{x:n}=e;if(t.shouldExecuteOnCPU([n])){let s=t.texData.get(n.dataId),[a,i]=vD(s.values,n.shape,n.dtype);return t.makeTensorInfo(i,n.dtype,a)}let o;return W().getBool("WEBGL_PACK_UNARY_OPERATIONS")?o=new Ns(n.shape,gF):o=new xn(n.shape,gF),t.runWebGLProgram(o,[n],n.dtype)}var xF={kernelName:js,backendName:"webgl",kernelFunc:h9};var g9=Fr.nonMaxSuppressionV3Impl;function x9(r){C.warn("tf.nonMaxSuppression() in webgl locks the UI thread. Call tf.nonMaxSuppressionAsync() instead");let{inputs:e,backend:t,attrs:n}=r,{boxes:o,scores:s}=e,{maxOutputSize:a,iouThreshold:i,scoreThreshold:l}=n,u=t.readSync(o.dataId),c=t.readSync(s.dataId),{selectedIndices:p}=g9(u,c,a,i,l);return t.makeTensorInfo([p.length],"int32",new Int32Array(p))}var yF={kernelName:Yi,backendName:"webgl",kernelFunc:x9};var y9=Fr.nonMaxSuppressionV4Impl;function b9(r){C.warn("tf.nonMaxSuppression() in webgl locks the UI thread. Call tf.nonMaxSuppressionAsync() instead");let{inputs:e,backend:t,attrs:n}=r,{boxes:o,scores:s}=e,{maxOutputSize:a,iouThreshold:i,scoreThreshold:l,padToMaxOutputSize:u}=n,c=t.readSync(o.dataId),p=t.readSync(s.dataId),{selectedIndices:m,validOutputs:f}=y9(c,p,a,i,l,u);return[t.makeTensorInfo([m.length],"int32",new Int32Array(m)),t.makeTensorInfo([],"int32",new Int32Array([f]))]}var bF={kernelName:Zi,backendName:"webgl",kernelFunc:b9};var w9=Fr.nonMaxSuppressionV5Impl;function _9(r){C.warn("tf.nonMaxSuppression() in webgl locks the UI thread. Call tf.nonMaxSuppressionAsync() instead");let{inputs:e,backend:t,attrs:n}=r,{boxes:o,scores:s}=e,{maxOutputSize:a,iouThreshold:i,scoreThreshold:l,softNmsSigma:u}=n,c=t.readSync(o.dataId),p=t.readSync(s.dataId),m=a,f=i,d=l,h=u,{selectedIndices:g,selectedScores:x}=w9(c,p,m,f,d,h);return[t.makeTensorInfo([g.length],"int32",new Int32Array(g)),t.makeTensorInfo([x.length],"float32",new Float32Array(x))]}var wF={kernelName:Ji,backendName:"webgl",kernelFunc:_9};var mC=class{constructor(e,t,n,o){this.variableNames=["indices"],this.outputShape=[e,t],this.userCode=`
|
|
void main() {
|
|
ivec2 coords = getOutputCoords();
|
|
int index = round(getIndices(coords.x));
|
|
setOutput(mix(float(${o}), float(${n}),
|
|
float(index == coords.y)));
|
|
}
|
|
`}};var k9=r=>{let{inputs:e,backend:t,attrs:n}=r,{indices:o}=e,{depth:s,onValue:a,offValue:i}=n,l=y.sizeFromShape(o.shape),u=new mC(l,s,a,i),c=ce({inputs:{x:o},backend:t,attrs:{shape:[l]}}),p=t.runWebGLProgram(u,[c],o.dtype);t.disposeIntermediateTensorInfo(c);let m=[...o.shape,s],f=ce({inputs:{x:p},backend:t,attrs:{shape:m}});return t.disposeIntermediateTensorInfo(p),f},_F={kernelName:qo,backendName:"webgl",kernelFunc:k9};function jf(r){let{inputs:e,backend:t}=r,{x:n}=e;if(n.dtype==="complex64"){let o=Oa({inputs:{input:n},backend:t}),s=jf({inputs:{x:o},backend:t}),a=Gu({inputs:{input:n},backend:t}),i=jf({inputs:{x:a},backend:t}),l=yn({inputs:{real:s,imag:i},backend:t});return t.disposeIntermediateTensorInfo(o),t.disposeIntermediateTensorInfo(s),t.disposeIntermediateTensorInfo(a),t.disposeIntermediateTensorInfo(i),l}else return Wf({attrs:{shape:n.shape,dtype:n.dtype,value:n.dtype==="string"?"":0},backend:t})}var kF={kernelName:Js,backendName:"webgl",kernelFunc:jf};function vF(r){let{inputs:e,backend:t}=r,{x:n}=e;if(n.dtype==="string")throw new Error("onesLike is not supported under string dtype");if(n.dtype==="complex64"){let o=Oa({inputs:{input:n},backend:t}),s=vF({inputs:{x:o},backend:t}),a=Gu({inputs:{input:n},backend:t}),i=jf({inputs:{x:a},backend:t}),l=yn({inputs:{real:s,imag:i},backend:t});return t.disposeIntermediateTensorInfo(o),t.disposeIntermediateTensorInfo(s),t.disposeIntermediateTensorInfo(a),t.disposeIntermediateTensorInfo(i),l}else return Wf({attrs:{shape:n.shape,dtype:n.dtype,value:1},backend:t})}var CF={kernelName:Us,backendName:"webgl",kernelFunc:vF};function v9(r){let{inputs:e,backend:t,attrs:n}=r,{axis:o}=n;if(e.length===1)return Ix({inputs:{input:e[0]},backend:t,attrs:{dim:o}});let s=e[0].shape,a=e[0].dtype;e.forEach(c=>{y.assertShapesMatch(s,c.shape,"All tensors passed to stack must have matching shapes"),y.assert(a===c.dtype,()=>"All tensors passed to stack must have matching dtypes")});let i=[],l=e.map(c=>{let p=Ix({inputs:{input:c},backend:t,attrs:{dim:o}});return i.push(p),p}),u=Fv({inputs:l,backend:t,attrs:{axis:o}});return i.forEach(c=>t.disposeIntermediateTensorInfo(c)),u}var IF={kernelName:Hs,backendName:"webgl",kernelFunc:v9};var fC=class{constructor(e,t,n){this.variableNames=["x"],this.outputShape=t.map((u,c)=>u[0]+e[c]+u[1]);let o=e.length,s=Ve(o),a=t.map(u=>u[0]).join(","),i=t.map((u,c)=>u[0]+e[c]).join(","),l=["coords[0]","coords[1]","coords[2]","coords[3]"].slice(0,o);if(o===1){this.userCode=`
|
|
int start = ${a};
|
|
int end = ${i};
|
|
uniform float value;
|
|
|
|
void main() {
|
|
int outC = getOutputCoords();
|
|
if (outC < start || outC >= end) {
|
|
setOutput(value);
|
|
} else {
|
|
setOutput(getX(outC - start));
|
|
}
|
|
}
|
|
`;return}this.userCode=`
|
|
${s} start = ${s}(${a});
|
|
${s} end = ${s}(${i});
|
|
uniform float value;
|
|
|
|
void main() {
|
|
${s} outC = getOutputCoords();
|
|
if (any(lessThan(outC, start)) || any(greaterThanEqual(outC, end))) {
|
|
setOutput(value);
|
|
} else {
|
|
${s} coords = outC - start;
|
|
setOutput(getX(${l}));
|
|
}
|
|
}
|
|
`}getCustomSetupFunc(e){return(t,n)=>{this.valueLoc==null&&(this.valueLoc=t.getUniformLocationNoThrow(n,"value")),t.gl.uniform1f(this.valueLoc,e)}}};var dC=class{constructor(e,t,n){this.variableNames=["x"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=t.map((h,g)=>h[0]+e[g]+h[1]);let o=e.length,s=Ve(o),a=t.map(h=>h[0]).join(","),i=t.map((h,g)=>h[0]+e[g]).join(","),l=jt("rc",o),u=jt("source",o),c=`${l[o-1]} < ${this.outputShape[o-1]}`,p=o===1?"source":`vec2(${u.slice(-2).join()})`,m=[`${s} rc = outputLoc;`,`${l[o-1]} += 1;
|
|
if(${c}) {
|
|
`,o===1?"":`}
|
|
rc = outputLoc;
|
|
${l[o-2]} += 1;
|
|
if(${l[o-2]} < ${this.outputShape[o-2]}) {`,o===1?"":` ${l[o-1]} += 1;
|
|
if(${c}) {`],f=o===1?"rc < start || rc >= end":"any(lessThan(rc, start)) || any(greaterThanEqual(rc, end))",d="";for(let h=0,g=o===1?2:4;h<g;h++)d+=`
|
|
${m[h]}
|
|
if (${f}) {
|
|
result[${h}] = float(value);
|
|
} else {
|
|
${s} source = rc - start;
|
|
result[${h}] = getChannel(getX(${u.join()}), ${p});
|
|
}
|
|
`;d+=o===1?"} ":"}}",this.userCode=`
|
|
const ${s} start = ${s}(${a});
|
|
const ${s} end = ${s}(${i});
|
|
uniform float value;
|
|
|
|
void main() {
|
|
${s} outputLoc = getOutputCoords();
|
|
vec4 result = vec4(0.);
|
|
${d}
|
|
setOutput(result);
|
|
}
|
|
`}getCustomSetupFunc(e){return(t,n)=>{this.valueLoc==null&&(this.valueLoc=t.getUniformLocationNoThrow(n,"value")),t.gl.uniform1f(this.valueLoc,e)}}};var hC=r=>{let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{paddings:s,constantValue:a}=n,i=W().getBool("WEBGL_PACK_ARRAY_OPERATIONS")?new dC(o.shape,s,a):new fC(o.shape,s,a),l=i.getCustomSetupFunc(a);return t.runWebGLProgram(i,[o],o.dtype,l)},NF={kernelName:Ko,backendName:"webgl",kernelFunc:hC};var C9=`
|
|
if(a < 0.0 && floor(b) < b){
|
|
return NAN;
|
|
}
|
|
if (b == 0.0) {
|
|
return 1.0;
|
|
}
|
|
return (round(mod(b, 2.0)) != 1) ?
|
|
pow(abs(a), b) : sign(a) * pow(abs(a), b);
|
|
`,I9=`
|
|
// isModRound1 has 1 for components with round(mod(b, 2.0)) == 1, 0 otherwise.
|
|
vec4 isModRound1 = vec4(equal(round(mod(b, 2.0)), ivec4(1)));
|
|
vec4 multiplier = sign(a) * isModRound1 + (vec4(1.0) - isModRound1);
|
|
vec4 result = multiplier * pow(abs(a), b);
|
|
|
|
// Ensure that a^0 = 1, including 0^0 = 1 as this correspond to TF and JS
|
|
bvec4 isExpZero = equal(b, vec4(0.0));
|
|
result.r = isExpZero.r ? 1.0 : result.r;
|
|
result.g = isExpZero.g ? 1.0 : result.g;
|
|
result.b = isExpZero.b ? 1.0 : result.b;
|
|
result.a = isExpZero.a ? 1.0 : result.a;
|
|
|
|
vec4 isNaN = vec4(lessThan(a, vec4(0.0))) * vec4(lessThan(floor(b), b));
|
|
`+kl+`
|
|
return result;
|
|
`,N9=st({opSnippet:C9,packedOpSnippet:I9}),SF={kernelName:Xo,backendName:"webgl",kernelFunc:N9};function S9(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,keepDims:a}=n,i=o.shape.length,l=[],u=y.parseAxisParam(s,o.shape),c=u,p=C.getAxesPermutation(c,i),m=o;p!=null&&(m=Rt({inputs:{x:o},backend:t,attrs:{perm:p}}),c=C.getInnerMostAxes(c.length,i),l.push(m)),C.assertAxesAreInnerMostDims("prod",c,i);let f;if(t.shouldExecuteOnCPU([m])){let d=t.texData.get(m.dataId).values,{outVals:h,outShape:g,outDtype:x}=CD(m.shape,m.dtype,d,c);f=t.makeTensorInfo(g,x,h)}else{let[d,h]=C.computeOutAndReduceShapes(m.shape,c),g=y.sizeFromShape(h),x=ce({inputs:{x:m},backend:t,attrs:{shape:[-1,g]}}),b=Yl(o.dtype),w=En(x,b,"prod",t);f=ce({inputs:{x:w},backend:t,attrs:{shape:d}}),l.push(x),l.push(w)}if(a){l.push(f);let d=C.expandShapeToKeepDim(f.shape,u);f=ce({inputs:{x:f},backend:t,attrs:{shape:d}})}return l.forEach(d=>t.disposeIntermediateTensorInfo(d)),f}var TF={kernelName:Qi,backendName:"webgl",kernelFunc:S9};var gC=r=>{let{backend:e,attrs:t}=r,{start:n,stop:o,step:s,dtype:a}=t,i=ID(n,o,s,a);return e.makeTensorInfo([i.length],a,i)},AF={kernelName:sl,backendName:"webgl",kernelFunc:gC};var T9="return 1.0 / x;",A9=_e({opSnippet:T9}),EF={kernelName:ea,backendName:"webgl",kernelFunc:A9};var E9=br+`
|
|
return (x < 0.0) ? 0.0 : x;
|
|
`,D9=`
|
|
vec4 result = x * vec4(greaterThanEqual(x, vec4(0.0)));
|
|
bvec4 isNaN = isnan(x);
|
|
|
|
result.r = isNaN.r ? x.r : result.r;
|
|
result.g = isNaN.g ? x.g : result.g;
|
|
result.b = isNaN.b ? x.b : result.b;
|
|
result.a = isNaN.a ? x.a : result.a;
|
|
|
|
return result;
|
|
`,$9=_e({opSnippet:E9,packedOpSnippet:D9}),DF={kernelName:Zo,backendName:"webgl",kernelFunc:$9};var R9=br+`
|
|
return (x < 0.0) ? 0.0 : min(6.0, x);
|
|
`,F9=`
|
|
vec4 result = min(x, vec4(6.)) * vec4(greaterThanEqual(x, vec4(0.0)));
|
|
bvec4 isNaN = isnan(x);
|
|
|
|
result.r = isNaN.r ? x.r : result.r;
|
|
result.g = isNaN.g ? x.g : result.g;
|
|
result.b = isNaN.b ? x.b : result.b;
|
|
result.a = isNaN.a ? x.a : result.a;
|
|
|
|
return result;
|
|
`,O9=_e({opSnippet:R9,packedOpSnippet:F9}),$F={kernelName:Qo,backendName:"webgl",kernelFunc:O9};var xC=class{constructor(e,t,n,o,s){this.variableNames=["A"],this.outputShape=[];let[a,i,l,u]=e;this.outputShape=[a,t,n,u];let c=[o&&t>1?i-1:i,o&&n>1?l-1:l],p=[o&&t>1?t-1:t,o&&n>1?n-1:n],m;s?m="(vec2(yRC) + vec2(0.5)) * effectiveInputOverOutputRatioRC - vec2(0.5)":m="vec2(yRC) * effectiveInputOverOutputRatioRC",this.userCode=`
|
|
const vec2 effectiveInputOverOutputRatioRC = vec2(
|
|
${c[0]/p[0]},
|
|
${c[1]/p[1]});
|
|
const vec2 inputShapeRC = vec2(${i}.0, ${l}.0);
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int d = coords[3];
|
|
ivec2 yRC = coords.yz;
|
|
|
|
// Fractional source index.
|
|
vec2 sourceFracIndexRC = ${m};
|
|
|
|
// Compute the four integer indices.
|
|
ivec2 sourceFloorRC = ivec2(max(sourceFracIndexRC, vec2(0.0)));
|
|
ivec2 sourceCeilRC = ivec2(
|
|
min(inputShapeRC - 1.0, ceil(sourceFracIndexRC)));
|
|
|
|
float topLeft = getA(b, sourceFloorRC.x, sourceFloorRC.y, d);
|
|
float bottomLeft = getA(b, sourceCeilRC.x, sourceFloorRC.y, d);
|
|
float topRight = getA(b, sourceFloorRC.x, sourceCeilRC.y, d);
|
|
float bottomRight = getA(b, sourceCeilRC.x, sourceCeilRC.y, d);
|
|
|
|
vec2 fracRC = sourceFracIndexRC - vec2(sourceFloorRC);
|
|
|
|
float top = topLeft + (topRight - topLeft) * fracRC.y;
|
|
float bottom = bottomLeft + (bottomRight - bottomLeft) * fracRC.y;
|
|
float newValue = top + (bottom - top) * fracRC.x;
|
|
|
|
setOutput(newValue);
|
|
}
|
|
`}};var yC=class{constructor(e,t,n,o,s){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=[];let[a,i,l,u]=e;this.outputShape=[a,t,n,u];let c=[o&&t>1?i-1:i,o&&n>1?l-1:l],p=[o&&t>1?t-1:t,o&&n>1?n-1:n],m;s?m="(vec3(yRC) + vec3(0.5)) * effectiveInputOverOutputRatioRC - vec3(0.5)":m="vec3(yRC) * effectiveInputOverOutputRatioRC",this.userCode=`
|
|
const vec3 effectiveInputOverOutputRatioRC = vec3(
|
|
${c[0]/p[0]},
|
|
${c[1]/p[1]},
|
|
${c[1]/p[1]});
|
|
const vec3 inputShapeRC = vec3(${i}.0, ${l}.0,
|
|
${l}.0);
|
|
|
|
float getAValue(int b, int r, int c, int d) {
|
|
return getChannel(getA(b, r, c, d), vec2(c, d));
|
|
}
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int d = coords[3];
|
|
// Calculate values for next column in yRC.z.
|
|
ivec3 yRC = coords.yzz + ivec3(0, 0, 1);
|
|
|
|
// Fractional source index.
|
|
vec3 sourceFracIndexRC = ${m};
|
|
|
|
// Compute the four integer indices.
|
|
ivec3 sourceFloorRC = ivec3(max(sourceFracIndexRC, vec3(0.0)));
|
|
ivec3 sourceCeilRC = ivec3(
|
|
min(inputShapeRC - 1.0, ceil(sourceFracIndexRC)));
|
|
|
|
// Should we calculate next column and row elements in 2x2 packed cell.
|
|
bool hasNextCol = d < ${u-1};
|
|
bool hasNextRow = coords.z < ${n-1};
|
|
|
|
// In parallel, construct four corners for all four components in
|
|
// packed 2x2 cell.
|
|
vec4 topLeft = vec4(
|
|
getAValue(b, sourceFloorRC.x, sourceFloorRC.y, d),
|
|
hasNextCol ? getAValue(b, sourceFloorRC.x, sourceFloorRC.y, d + 1)
|
|
: 0.0,
|
|
hasNextRow ? getAValue(b, sourceFloorRC.x, sourceFloorRC.z, d)
|
|
: 0.0,
|
|
(hasNextRow && hasNextCol) ?
|
|
getAValue(b, sourceFloorRC.x, sourceFloorRC.z, d + 1) : 0.0);
|
|
|
|
vec4 bottomLeft = vec4(
|
|
getAValue(b, sourceCeilRC.x, sourceFloorRC.y, d),
|
|
hasNextCol ? getAValue(b, sourceCeilRC.x, sourceFloorRC.y, d + 1)
|
|
: 0.0,
|
|
hasNextRow ? getAValue(b, sourceCeilRC.x, sourceFloorRC.z, d)
|
|
: 0.0,
|
|
(hasNextRow && hasNextCol) ?
|
|
getAValue(b, sourceCeilRC.x, sourceFloorRC.z, d + 1) : 0.0);
|
|
|
|
vec4 topRight = vec4(
|
|
getAValue(b, sourceFloorRC.x, sourceCeilRC.y, d),
|
|
hasNextCol ? getAValue(b, sourceFloorRC.x, sourceCeilRC.y, d + 1)
|
|
: 0.0,
|
|
hasNextRow ? getAValue(b, sourceFloorRC.x, sourceCeilRC.z, d)
|
|
: 0.0,
|
|
(hasNextRow && hasNextCol) ?
|
|
getAValue(b, sourceFloorRC.x, sourceCeilRC.z, d + 1) : 0.0);
|
|
|
|
vec4 bottomRight = vec4(
|
|
getAValue(b, sourceCeilRC.x, sourceCeilRC.y, d),
|
|
hasNextCol ? getAValue(b, sourceCeilRC.x, sourceCeilRC.y, d + 1)
|
|
: 0.0,
|
|
hasNextRow ? getAValue(b, sourceCeilRC.x, sourceCeilRC.z, d)
|
|
: 0.0,
|
|
(hasNextRow && hasNextCol) ?
|
|
getAValue(b, sourceCeilRC.x, sourceCeilRC.z, d + 1) : 0.0);
|
|
|
|
vec3 fracRC = sourceFracIndexRC - vec3(sourceFloorRC);
|
|
|
|
vec4 top = mix(topLeft, topRight, fracRC.yyzz);
|
|
vec4 bottom = mix(bottomLeft, bottomRight, fracRC.yyzz);
|
|
vec4 newValue = mix(top, bottom, fracRC.x);
|
|
|
|
setOutput(newValue);
|
|
}
|
|
`}};function P9(r){let{inputs:e,backend:t,attrs:n}=r,{images:o}=e,{alignCorners:s,halfPixelCenters:a,size:i}=n,[l,u]=i,c=W().getBool("WEBGL_PACK_IMAGE_OPERATIONS")?new yC(o.shape,l,u,s,a):new xC(o.shape,l,u,s,a);return t.runWebGLProgram(c,[o],"float32")}var RF={kernelName:Jo,backendName:"webgl",kernelFunc:P9};var bC=class{constructor(e,t,n){this.variableNames=["dy"],this.outputShape=[],this.outputShape=t;let[,o,s]=t,[,a,i]=e,l=[n&&a>1?o-1:o,n&&i>1?s-1:s],u=[n&&a>1?a-1:a,n&&i>1?i-1:i],c=l[0]/u[0],p=l[1]/u[1],m=1/c,f=1/p,d=Math.ceil(m)*2+2,h=Math.ceil(f)*2+2;this.userCode=`
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int d = coords[3];
|
|
int r = coords[1];
|
|
int c = coords[2];
|
|
|
|
float accumulator = 0.0;
|
|
|
|
const float heightScale = float(${c});
|
|
const float widthScale = float(${p});
|
|
|
|
const float invHeightScale = float(${m});
|
|
const float invWidthScale = float(${f});
|
|
|
|
const int winHeight = int(${d});
|
|
const int winWidth = int(${h});
|
|
|
|
// Compute bounds for where in dy we will look
|
|
float startRLerp = floor(float(r) * invHeightScale);
|
|
int startDyR = int(startRLerp - float(winHeight / 2));
|
|
|
|
float startCLerp = floor(float(c) * invWidthScale);
|
|
int startDyC = int(startCLerp - float(winWidth / 2));
|
|
|
|
// Loop over dy
|
|
for (int dyROffset = 0; dyROffset < winHeight; dyROffset++) {
|
|
int dyR = dyROffset + startDyR;
|
|
|
|
// Guard against the window exceeding the bounds of dy
|
|
if (dyR < 0 || dyR >= ${a}) {
|
|
continue;
|
|
}
|
|
|
|
for (int dyCOffset = 0; dyCOffset < winWidth; dyCOffset++) {
|
|
int dyC = dyCOffset + startDyC;
|
|
|
|
// Guard against the window exceeding the bounds of dy
|
|
if (dyC < 0 || dyC >= ${i}) {
|
|
continue;
|
|
}
|
|
|
|
float dxR = float(dyR) * heightScale;
|
|
int topDxRIndex = int(floor(dxR));
|
|
int bottomDxRIndex = int(min(ceil(dxR), ${o-1}.0));
|
|
float dxRLerp = dxR - float(topDxRIndex);
|
|
float inverseDxRLerp = 1.0 - dxRLerp;
|
|
|
|
float dxC = float(dyC) * widthScale;
|
|
int leftDxCIndex = int(floor(dxC));
|
|
int rightDxCIndex = int(min(ceil(dxC), ${s-1}.0));
|
|
float dxCLerp = dxC - float(leftDxCIndex);
|
|
float inverseDxCLerp = 1.0 - dxCLerp;
|
|
|
|
if (r == topDxRIndex && c == leftDxCIndex) {
|
|
// topLeft
|
|
accumulator +=
|
|
getDy(b, dyR, dyC, d) * inverseDxRLerp * inverseDxCLerp;
|
|
}
|
|
|
|
if (r == topDxRIndex && c == rightDxCIndex) {
|
|
// topRight
|
|
accumulator += getDy(b, dyR, dyC, d) * inverseDxRLerp * dxCLerp;
|
|
}
|
|
|
|
if (r == bottomDxRIndex && c == leftDxCIndex) {
|
|
// bottomLeft
|
|
accumulator += getDy(b, dyR, dyC, d) * dxRLerp * inverseDxCLerp;
|
|
}
|
|
|
|
if (r == bottomDxRIndex && c == rightDxCIndex) {
|
|
// bottomRight
|
|
accumulator += getDy(b, dyR, dyC, d) * dxRLerp * dxCLerp;
|
|
}
|
|
}
|
|
}
|
|
// End loop over dy
|
|
|
|
setOutput(accumulator);
|
|
}
|
|
`}};function M9(r){let{inputs:e,backend:t,attrs:n}=r,{images:o,dy:s}=e,{alignCorners:a}=n,i=new bC(s.shape,o.shape,a);return t.runWebGLProgram(i,[s],s.dtype)}var FF={kernelName:jc,backendName:"webgl",kernelFunc:M9};var wC=class{constructor(e,t,n,o,s){this.variableNames=["A"],this.outputShape=[];let[a,i,l,u]=e;this.outputShape=[a,t,n,u];let c=[o&&t>1?i-1:i,o&&n>1?l-1:l],p=[o&&t>1?t-1:t,o&&n>1?n-1:n],m=o?"0.5":"0.0",f;s?f="max((vec2(yRC) + vec2(0.5)) * effectiveInputOverOutputRatioRC, vec2(0.0))":f="vec2(yRC) * effectiveInputOverOutputRatioRC",this.userCode=`
|
|
const vec2 effectiveInputOverOutputRatioRC = vec2(
|
|
${c[0]/p[0]},
|
|
${c[1]/p[1]});
|
|
const vec2 inputShapeRC = vec2(${i}.0, ${l}.0);
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int d = coords[3];
|
|
ivec2 yRC = coords.yz;
|
|
|
|
// Fractional source index.
|
|
vec2 sourceFracIndexRC = ${f};
|
|
|
|
// Compute the coordinators of nearest neighbor point.
|
|
ivec2 sourceNearestRC = ivec2(
|
|
min(inputShapeRC - 1.0, floor(sourceFracIndexRC + ${m})));
|
|
float newValue = getA(b, sourceNearestRC.x, sourceNearestRC.y, d);
|
|
|
|
setOutput(newValue);
|
|
}
|
|
`}};var _C=class{constructor(e,t,n,o,s){this.variableNames=["A"],this.packedInputs=!0,this.packedOutput=!0,this.outputShape=[];let[a,i,l,u]=e;this.outputShape=[a,t,n,u];let c=[o&&t>1?i-1:i,o&&n>1?l-1:l],p=[o&&t>1?t-1:t,o&&n>1?n-1:n],m=o?"0.5":"0.0",f;s?f="max((vec3(yRC) + vec3(0.5)) * effectiveInputOverOutputRatioRC, vec3(0.0))":f="vec3(yRC) * effectiveInputOverOutputRatioRC",this.userCode=`
|
|
const vec3 effectiveInputOverOutputRatioRC = vec3(
|
|
${c[0]/p[0]},
|
|
${c[1]/p[1]},
|
|
${c[1]/p[1]});
|
|
const vec3 inputShapeRC = vec3(${i}.0, ${l}.0,
|
|
${l}.0);
|
|
|
|
float getAValue(int b, int r, int c, int d) {
|
|
return getChannel(getA(b, r, c, d), vec2(c, d));
|
|
}
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int d = coords[3];
|
|
// Calculate values for next column in yRC.z.
|
|
ivec3 yRC = coords.yzz + ivec3(0, 0, 1);
|
|
|
|
// Fractional source index.
|
|
vec3 sourceFracIndexRC = ${f};
|
|
|
|
// Compute the coordinators of nearest neighbor point.
|
|
ivec3 sourceNearestRC = ivec3(
|
|
min(inputShapeRC - 1.0, floor(sourceFracIndexRC + ${m})));
|
|
|
|
// Should we calculate next column and row elements in 2x2 packed cell.
|
|
bool hasNextCol = d < ${u-1};
|
|
bool hasNextRow = coords.z < ${n-1};
|
|
|
|
vec4 newValue = vec4(
|
|
getAValue(b, sourceNearestRC.x, sourceNearestRC.y, d),
|
|
hasNextCol ? getAValue(b, sourceNearestRC.x, sourceNearestRC.y, d + 1)
|
|
: 0.0,
|
|
hasNextRow ? getAValue(b, sourceNearestRC.x, sourceNearestRC.z, d)
|
|
: 0.0,
|
|
(hasNextRow && hasNextCol) ?
|
|
getAValue(b, sourceNearestRC.x, sourceNearestRC.z, d + 1) : 0.0);
|
|
|
|
setOutput(newValue);
|
|
}
|
|
`}};function L9(r){let{inputs:e,backend:t,attrs:n}=r,{images:o}=e,{alignCorners:s,halfPixelCenters:a,size:i}=n,[l,u]=i,c=W().getBool("WEBGL_PACK_IMAGE_OPERATIONS")?new _C(o.shape,l,u,s,a):new wC(o.shape,l,u,s,a);return t.runWebGLProgram(c,[o],o.dtype)}var OF={kernelName:il,backendName:"webgl",kernelFunc:L9};var kC=class{constructor(e,t,n){this.variableNames=["dy"],this.outputShape=[],this.outputShape=t;let[,o,s]=t,[,a,i]=e,l=[n&&a>1?o-1:o,n&&i>1?s-1:s],u=[n&&a>1?a-1:a,n&&i>1?i-1:i],c=l[0]/u[0],p=l[1]/u[1],m=1/c,f=1/p,d=Math.ceil(m)*2+2,h=Math.ceil(f)*2+2;this.userCode=`
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int b = coords[0];
|
|
int d = coords[3];
|
|
int r = coords[1];
|
|
int c = coords[2];
|
|
|
|
float accumulator = 0.0;
|
|
|
|
const float heightScale = float(${c});
|
|
const float widthScale = float(${p});
|
|
|
|
const float invHeightScale = float(${m});
|
|
const float invWidthScale = float(${f});
|
|
|
|
const int winHeight = int(${d});
|
|
const int winWidth = int(${h});
|
|
|
|
// Compute bounds for where in dy we will look
|
|
float startRLerp = floor(float(r) * invHeightScale);
|
|
int startDyR = int(floor(startRLerp - float(winHeight / 2)));
|
|
|
|
float startCLerp = floor(float(c) * invWidthScale);
|
|
int startDyC = int(floor(startCLerp - float(winWidth / 2)));
|
|
|
|
// Loop over dy
|
|
for (int dyROffset = 0; dyROffset < winHeight; dyROffset++) {
|
|
int dyR = dyROffset + startDyR;
|
|
|
|
// Guard against the window exceeding the bounds of dy
|
|
if (dyR < 0 || dyR >= ${a}) {
|
|
continue;
|
|
}
|
|
|
|
for (int dyCOffset = 0; dyCOffset < winWidth; dyCOffset++) {
|
|
int dyC = dyCOffset + startDyC;
|
|
|
|
// Guard against the window exceeding the bounds of dy
|
|
if (dyC < 0 || dyC >= ${i}) {
|
|
continue;
|
|
}
|
|
|
|
float sourceFracRow =
|
|
float(${l[0]}) *
|
|
(float(dyR) / float(${u[0]}));
|
|
|
|
float sourceFracCol =
|
|
float(${l[1]}) *
|
|
(float(dyC) / float(${u[1]}));
|
|
|
|
int sourceNearestRow = int(min(
|
|
float(int(${o}) - 1),
|
|
${n} ? float(round(sourceFracRow)) :
|
|
float(floor(sourceFracRow))));
|
|
|
|
int sourceNearestCol = int(min(
|
|
float(int(${s}) - 1),
|
|
${n} ? float(round(sourceFracCol)) :
|
|
float(floor(sourceFracCol))));
|
|
|
|
if (r == sourceNearestRow && c == sourceNearestCol) {
|
|
accumulator += getDy(b, dyR, dyC, d);
|
|
}
|
|
}
|
|
}
|
|
// End loop over dy
|
|
|
|
setOutput(accumulator);
|
|
}
|
|
`}};function z9(r){let{inputs:e,backend:t,attrs:n}=r,{images:o,dy:s}=e,{alignCorners:a}=n,i=new kC(s.shape,o.shape,a);return t.runWebGLProgram(i,[s],s.dtype)}var PF={kernelName:Wc,backendName:"webgl",kernelFunc:z9};var vC=class{constructor(e,t){this.variableNames=["x"];let n=e.length;if(n>4)throw new Error(`WebGL backend: Reverse of rank-${n} tensor is not yet supported`);if(this.outputShape=e,n===1){this.userCode=`
|
|
void main() {
|
|
int coord = getOutputCoords();
|
|
setOutput(getX(${e[0]} - coord - 1));
|
|
}
|
|
`;return}let o=i=>t.indexOf(i)!==-1&&e[i]!==1?`${e[i]} - coords[${i}] - 1`:`coords[${i}]`,s=e.map((i,l)=>o(l)).join(","),a=Ve(n);this.userCode=`
|
|
void main() {
|
|
${a} coords = getOutputCoords();
|
|
setOutput(getX(${s}));
|
|
}
|
|
`}};var CC=class{constructor(e,t){this.variableNames=["x"],this.packedInputs=!0,this.packedOutput=!0;let n=e.length;if(n>4)throw new Error(`WebGL backend: Reverse of rank-${n} tensor is not yet supported`);this.outputShape=e;let o=jt("rc",n),s=`${o[n-1]} + 1 < ${this.outputShape[n-1]}`,a=`${o[n-2]} + 1 < ${this.outputShape[n-2]}`,i=Ve(n);n===1?this.userCode=`
|
|
void main(){
|
|
int rc = getOutputCoords();
|
|
vec4 result = vec4(0.);
|
|
result.r = getChannel(getX(${e[0]} - rc - 1),
|
|
${e[0]} - rc - 1);
|
|
if(${s}){
|
|
result.g = getChannel(getX(${e[0]} - (rc + 1) - 1),
|
|
${e[0]} - (rc + 1) - 1);
|
|
}
|
|
setOutput(result);
|
|
}
|
|
`:this.userCode=`
|
|
void main() {
|
|
${i} rc = getOutputCoords();
|
|
vec4 result = vec4(0.);
|
|
result.r = ${l(o.slice())};
|
|
if(${s}){
|
|
result.g = ${u(o.slice())};
|
|
}
|
|
if(${a}) {
|
|
result.b = ${c(o.slice())};
|
|
if(${s}) {
|
|
result.a = ${p(o.slice())};
|
|
}
|
|
}
|
|
setOutput(result);
|
|
}
|
|
`;function l(d){return m(d)}function u(d){return d[n-1]="("+d[n-1]+" + 1)",m(d)}function c(d){return d[n-2]="("+d[n-2]+" + 1)",m(d)}function p(d){return d[n-1]="("+d[n-1]+" + 1)",d[n-2]="("+d[n-2]+" + 1)",m(d)}function m(d){let h=e.map((b,w)=>f(w,d)),g=h.join(","),x=h.slice(-2).join(",");return`getChannel(getX(${g}), vec2(${x}))`}function f(d,h){return t.indexOf(d)!==-1&&e[d]!==1?`${e[d]} - ${h[d]} - 1`:`${h[d]}`}}};function B9(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{dims:s}=n,a=o.shape.length,i=y.parseAxisParam(s,o.shape);if(a===0)return Ut({inputs:{x:o},backend:t});let l=W().getBool("WEBGL_PACK_ARRAY_OPERATIONS")?new CC(o.shape,i):new vC(o.shape,i);return t.runWebGLProgram(l,[o],o.dtype)}var MF={kernelName:es,backendName:"webgl",kernelFunc:B9};var IC=class{constructor(e,t){this.variableNames=["Image"],this.outputShape=[];let n=e[1],o=e[2];this.outputShape=e;let s="";typeof t=="number"?s=`float outputValue = ${t.toFixed(2)};`:s=`
|
|
vec3 fill = vec3(${t.join(",")});
|
|
float outputValue = fill[coords[3]];`,this.userCode=`
|
|
uniform vec4 params;
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
int x = coords[2];
|
|
int y = coords[1];
|
|
float coordXFloat = (float(x) - params[0]) * params[3] -
|
|
(float(y) - params[1]) * params[2];
|
|
float coordYFloat = (float(x) - params[0]) * params[2] +
|
|
(float(y) - params[1]) * params[3];
|
|
int coordX = int(round(coordXFloat + params[0]));
|
|
int coordY = int(round(coordYFloat + params[1]));
|
|
${s}
|
|
if(coordX >= 0 && coordX < ${o} && coordY >= 0 && coordY < ${n}) {
|
|
outputValue = getImage(coords[0], coordY, coordX, coords[3]);
|
|
}
|
|
setOutput(outputValue);
|
|
}
|
|
`}getCustomSetupFunc(e,t,n,o){return(s,a)=>{this.paramsLoc==null&&(this.paramsLoc=s.getUniformLocationNoThrow(a,"params")),s.gl.uniform4f(this.paramsLoc,e,t,n,o)}}};var LF={kernelName:ua,backendName:"webgl",kernelFunc:({inputs:r,attrs:e,backend:t})=>{let{image:n}=r,{radians:o,fillValue:s,center:a}=e,i=t,l=new IC(n.shape,s),[u,c]=C.getImageCenter(a,n.shape[1],n.shape[2]),p=l.getCustomSetupFunc(u,c,Math.sin(o),Math.cos(o));return i.runWebGLProgram(l,[n],n.dtype,p)}};var V9=`
|
|
// OpenGL ES does not support round function.
|
|
// The algorithm is based on banker's rounding.
|
|
float base = floor(x);
|
|
if ((x - base) < 0.5) {
|
|
return floor(x);
|
|
} else if ((x - base) > 0.5) {
|
|
return ceil(x);
|
|
} else {
|
|
if (mod(base, 2.0) == 0.0) {
|
|
return base;
|
|
} else {
|
|
return base + 1.0;
|
|
}
|
|
}
|
|
`,G9=_e({opSnippet:V9}),zF={kernelName:ts,backendName:"webgl",kernelFunc:G9};var W9="return inversesqrt(x);",j9=_e({opSnippet:W9,cpuKernelImpl:ND}),BF={kernelName:rs,backendName:"webgl",kernelFunc:j9};var Uf=class{constructor(e,t,n,o,s,a,i=!0){this.variableNames=["updates","indices","defaultValue"],this.outputShape=a;let l=Ve(s.length),u=Ve(a.length),c="";n===1?c="i":n===2&&(c="i, j");let p=`getIndices(${c})`,m="";o===1?m="i":o===2&&(m="i, coords[1]");let f=`getUpdates(${m})`,d=t>1?"strides[j]":"strides";this.userCode=`
|
|
${l} strides = ${l}(${s});
|
|
|
|
void main() {
|
|
${u} coords = getOutputCoords();
|
|
float sum = 0.0;
|
|
bool found = false;
|
|
for (int i = 0; i < ${e}; i++) {
|
|
int flattenedIndex = 0;
|
|
for (int j = 0; j < ${t}; j++) {
|
|
int index = round(${p});
|
|
flattenedIndex += index * ${d};
|
|
}
|
|
if (flattenedIndex == coords[0]) {
|
|
sum += ${f};
|
|
found = true;
|
|
}
|
|
}
|
|
setOutput(mix(getDefaultValue(), sum, float(found)));
|
|
}
|
|
`}};function U9(r){let{inputs:e,backend:t,attrs:n}=r,{indices:o,updates:s}=e,{shape:a}=n,{sliceRank:i,numUpdates:l,sliceSize:u,strides:c,outputSize:p}=C.calculateShapes(s,o,a),m=[p/u,u];if(p===0)return t.makeTensorInfo(a,o.dtype);let f=ce({inputs:{x:o},backend:t,attrs:{shape:[l,i]}}),d=ce({inputs:{x:s},backend:t,attrs:{shape:[l,u]}}),h=t.makeTensorInfo([],"float32",new Float32Array([0])),g=new Uf(l,i,f.shape.length,d.shape.length,c,m),x=t.runWebGLProgram(g,[d,f,h],d.dtype),b=ce({inputs:{x},backend:t,attrs:{shape:a}});return t.disposeIntermediateTensorInfo(f),t.disposeIntermediateTensorInfo(d),t.disposeIntermediateTensorInfo(x),t.disposeIntermediateTensorInfo(h),b}var VF={kernelName:ta,backendName:"webgl",kernelFunc:U9};var NC=class{constructor(e,t,n){this.variableNames=["c","a","b"],this.outputShape=t;let o,s;if(n>4)throw Error(`Where for rank ${n} is not yet supported`);if(n===1)s="resRC",o="resRC";else{let i=["resRC.x","resRC.y","resRC.z","resRC.w"],l=[],u=[];for(let c=0;c<t.length;c++)u.push(`${i[c]}`),c<e&&l.push(`${i[c]}`);o=l.join(),s=u.join()}let a=Ve(n);this.userCode=`
|
|
void main() {
|
|
${a} resRC = getOutputCoords();
|
|
float cVal = getC(${o});
|
|
if (cVal >= 1.0) {
|
|
setOutput(getA(${s}));
|
|
} else {
|
|
setOutput(getB(${s}));
|
|
}
|
|
}
|
|
`}};function H9(r){let{inputs:e,backend:t}=r,{condition:n,t:o,e:s}=e,a=new NC(n.shape.length,o.shape,o.shape.length);return t.runWebGLProgram(a,[n,o,s],lr(o.dtype,s.dtype))}var GF={kernelName:Ks,backendName:"webgl",kernelFunc:H9};var q9=`
|
|
// Stable and Attracting Fixed Point (0, 1) for Normalized Weights.
|
|
// see: https://arxiv.org/abs/1706.02515
|
|
float scaleAlpha = ${C.SELU_SCALEALPHA};
|
|
float scale = ${C.SELU_SCALE};
|
|
return (x >= 0.0) ? scale * x : scaleAlpha * (exp(x) - 1.0);
|
|
`,K9=_e({opSnippet:q9}),WF={kernelName:ra,backendName:"webgl",kernelFunc:K9};var X9="return 1.0 / (1.0 + exp(-1.0 * x));",Y9=_e({opSnippet:X9}),jF={kernelName:os,backendName:"webgl",kernelFunc:Y9};var Z9=`
|
|
if (isnan(x)) { return 0.0; }
|
|
return sign(x);
|
|
`,J9=_e({opSnippet:Z9}),UF={kernelName:oa,backendName:"webgl",kernelFunc:J9};var Q9=gx+`
|
|
return sin(x);
|
|
`,eZ=_e({opSnippet:Q9}),HF={kernelName:ns,backendName:"webgl",kernelFunc:eZ};var tZ=`
|
|
float e2x = exp(x);
|
|
return (e2x - 1.0 / e2x) / 2.0;
|
|
`,rZ=_e({opSnippet:tZ}),qF={kernelName:na,backendName:"webgl",kernelFunc:rZ};var nZ=`
|
|
float epsilon = 1.1920928955078125e-7;
|
|
float threshold = log(epsilon) + 2.0;
|
|
|
|
bool too_large = x > -threshold;
|
|
bool too_small = x < threshold;
|
|
|
|
float result;
|
|
float exp_x = exp(x);
|
|
|
|
if (too_large){
|
|
result = x;
|
|
}
|
|
else if (too_small){
|
|
result = exp_x;
|
|
}
|
|
else{
|
|
result = log(exp_x + 1.0);
|
|
}
|
|
return result;
|
|
`,oZ=_e({opSnippet:nZ}),KF={kernelName:sa,backendName:"webgl",kernelFunc:oZ};var sZ=r=>{let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{blockShape:s,paddings:a}=n;y.assert(o.shape.length<=4,()=>"spaceToBatchND for rank > 4 with a WebGL backend not implemented yet");let i=s.reduce((x,b)=>x*b),l=[[0,0]];l.push(...a);for(let x=1+s.length;x<o.shape.length;++x)l.push([0,0]);let u=[],c=hC({inputs:{x:o},backend:t,attrs:{paddings:l,constantValue:0}}),p=C.getReshaped(c.shape,s,i,!1),m=C.getPermuted(p.length,s.length,!1),f=C.getReshapedPermuted(c.shape,s,i,!1),d=ce({inputs:{x:c},backend:t,attrs:{shape:p}}),h=Rt({inputs:{x:d},backend:t,attrs:{perm:m}}),g=ce({inputs:{x:h},backend:t,attrs:{shape:f}});return u.push(c),u.push(d),u.push(h),u.forEach(x=>t.disposeIntermediateTensorInfo(x)),g},XF={kernelName:al,backendName:"webgl",kernelFunc:sZ};function iZ(r){let{inputs:e,backend:t}=r,{indices:n,values:o,denseShape:s,defaultValue:a}=e;if(s.shape.length!==1)throw new Error(`Dense shape must be a vector, saw:
|
|
${s.shape}`);if(n.shape.length!==2)throw new Error(`Indices must be a matrix, saw:
|
|
${n.shape}`);if(o.shape.length!==1)throw new Error(`Values must be a vector, saw:
|
|
${o.shape}`);if(a.shape.length!==0)throw new Error(`Default value must be a scalar, saw:
|
|
${a.shape}`);let i=t.readSync(n.dataId),l=t.readSync(o.dataId),u=t.readSync(s.dataId),c=t.readSync(a.dataId)[0],[p,m,f,d,h]=TD(i,n.shape,n.dtype,l,o.dtype,u,c);return[t.makeTensorInfo(m,n.dtype,p),t.makeTensorInfo([m[0]],o.dtype,f),t.makeTensorInfo([d.length],"bool",new Uint8Array(d.map(g=>Number(g)))),t.makeTensorInfo([h.length],n.dtype,new Int32Array(h))]}var YF={kernelName:Uc,backendName:"webgl",kernelFunc:iZ};function aZ(r){let{inputs:e,backend:t}=r,{inputIndices:n,inputShape:o,newShape:s}=e;if(n.shape.length!==2)throw new Error(`Input indices should be a matrix but received shape ${n.shape}`);if(o.shape.length!==1)throw new Error(`Input shape should be a vector but received shape ${o.shape}`);if(s.shape.length!==1)throw new Error(`Target shape should be a vector but received shape ${s.shape}`);let a=Array.from(t.readSync(o.dataId)),i=t.readSync(n.dataId),l=Array.from(t.readSync(s.dataId)),[u,c,p]=AD(i,n.shape,n.dtype,a,l);return[t.makeTensorInfo(c,n.dtype,u),t.makeTensorInfo([p.length],s.dtype,new Int32Array(p))]}var ZF={kernelName:Hc,backendName:"webgl",kernelFunc:aZ};function lZ(r){let{inputs:e,backend:t,attrs:n}=r,{sparseIndices:o,sparseValues:s,defaultValue:a}=e,{outputShape:i}=n,{sliceRank:l,numUpdates:u,strides:c,outputSize:p}=C.calculateShapes(s,o,i),m=!1,f=new Uf(u,l,o.shape.length,s.shape.length,c,[p,1],m),d=t.runWebGLProgram(f,[s,o,a],s.dtype),h=ce({inputs:{x:d},backend:t,attrs:{shape:i}});return t.disposeIntermediateTensorInfo(d),h}var JF={kernelName:qc,backendName:"webgl",kernelFunc:lZ};function uZ(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{numOrSizeSplits:s,axis:a}=n,i=y.parseAxisParam(a,o.shape)[0],l=C.prepareSplitSize(o,s,i),u=o.shape.length,c=new Array(u).fill(0),p=o.shape.slice();return l.map(m=>{let f=[...p];f[i]=m;let d=Fa({inputs:{x:o},backend:t,attrs:{begin:c,size:f}});return c[i]+=m,d})}var QF={kernelName:Ys,backendName:"webgl",kernelFunc:uZ};var cZ="return sqrt(x);",pZ=_e({opSnippet:cZ}),eO={kernelName:ss,backendName:"webgl",kernelFunc:pZ};var mZ="return x * x;",fZ=_e({opSnippet:mZ}),tO={kernelName:ll,backendName:"webgl",kernelFunc:fZ};var rO="return (a - b) * (a - b);",dZ=st({opSnippet:rO,packedOpSnippet:rO}),nO={kernelName:ls,backendName:"webgl",kernelFunc:dZ};function hZ({inputs:r,attrs:e,backend:t}){let{x:n}=r,o=br+`
|
|
return x > 0.0 ? 1.0 : float(${e.alpha});
|
|
`,s=new xn(n.shape,o);return t.runWebGLProgram(s,[n],n.dtype)}var oO={kernelName:Yn,backendName:"webgl",kernelFunc:hZ};var SC=class{constructor(e,t,n){this.variableNames=["x"],this.outputShape=n;let o=n.length,s=Ve(n.length),a=Ve(n.length),i="";if(o===1)i="coords * strides + begin";else{let l=0;i=n.map((u,c)=>(l++,n.length===1?`coords * strides[${c}] + begin[${c}]`:`coords[${l-1}] * strides[${c}] + begin[${c}]`)).join(",")}this.userCode=`
|
|
${s} begin = ${s}(${e});
|
|
${s} strides = ${s}(${t});
|
|
|
|
void main() {
|
|
${a} coords = getOutputCoords();
|
|
setOutput(getX(${i}));
|
|
}
|
|
`}};function gZ(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{begin:s,end:a,strides:i,beginMask:l,endMask:u,ellipsisMask:c,newAxisMask:p,shrinkAxisMask:m}=n,{nonStrided:f,$begin:d,$strides:h,size:g,newShape:x,outShape:b}=or.sliceInfo(o.shape,s,a,i,l,u,c,p,m),w=ce({inputs:{x:o},backend:t,attrs:{shape:x}}),_;if(f){let E=Fa({inputs:{x:w},backend:t,attrs:{begin:d,size:g}});_=ce({inputs:{x:E},backend:t,attrs:{shape:b}}),t.disposeIntermediateTensorInfo(E)}else if(b.some(E=>E===0))_=t.makeTensorInfo(b,o.dtype,[]);else if(t.shouldExecuteOnCPU([w])){let $=t.texData.get(w.dataId).values,R=Ce(w.shape,w.dtype,$),M=ED(b,R,h,d);_=t.makeTensorInfo(b,w.dtype,M.values)}else{let D=new SC(d,h,b);_=t.runWebGLProgram(D,[w],w.dtype)}let I=ce({inputs:{x:_},backend:t,attrs:{shape:b}});return t.disposeIntermediateTensorInfo(w),t.disposeIntermediateTensorInfo(_),I}var sO={kernelName:ia,backendName:"webgl",kernelFunc:gZ};var xZ="return tan(x);",yZ=_e({opSnippet:xZ}),iO={kernelName:cs,backendName:"webgl",kernelFunc:yZ};var bZ=`
|
|
float e2x = exp(-2.0 * abs(x));
|
|
return sign(x) * (1.0 - e2x) / (1.0 + e2x);
|
|
`,wZ=_e({opSnippet:bZ}),aO={kernelName:ps,backendName:"webgl",kernelFunc:wZ};var TC=class{constructor(e,t){this.variableNames=["A"];let n=new Array(e.length);for(let a=0;a<n.length;a++)n[a]=e[a]*t[a];this.outputShape=n,this.rank=n.length;let o=Ve(this.rank),s=_Z(e);this.userCode=`
|
|
void main() {
|
|
${o} resRC = getOutputCoords();
|
|
setOutput(getA(${s}));
|
|
}
|
|
`}};function _Z(r){let e=r.length;if(e>5)throw Error(`Tile for rank ${e} is not yet supported`);if(e===1)return`imod(resRC, ${r[0]})`;let t=["resRC.x","resRC.y","resRC.z","resRC.w","resRC.u"],n=[];for(let o=0;o<r.length;o++)n.push(`imod(${t[o]}, ${r[o]})`);return n.join()}function AC(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{reps:s}=n;if(o.dtype==="string"||o.shape.length>5){let l=t.readSync(o.dataId),u=o.dtype==="string"?l.map(m=>y.decodeString(m)):l,c=Ce(o.shape,o.dtype,u),p=$D(c,s);return t.makeTensorInfo(p.shape,p.dtype,p.values)}let a=new TC(o.shape,s);return t.runWebGLProgram(a,[o],o.dtype)}var lO={kernelName:Mn,backendName:"webgl",kernelFunc:AC};function kZ(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{k:s,sorted:a}=n,i=t.readSync(o.dataId),[l,u]=RD(i,o.shape,o.dtype,s,a);return[t.makeTensorInfo(l.shape,l.dtype,l.values),t.makeTensorInfo(u.shape,u.dtype,u.values)]}var uO={kernelName:aa,backendName:"webgl",kernelFunc:kZ};var EC=class{constructor(e,t,n,o,s,a){this.variableNames=["Image","Transforms"],this.outputShape=a;let i=n==="nearest"?1:2,l;switch(o){case"constant":l=1;break;case"reflect":l=2;break;case"wrap":l=3;break;case"nearest":l=4;break;default:l=1;break}this.userCode=`
|
|
float mapCoord(float outCoord, float len) {
|
|
float inCoord = outCoord;
|
|
if(${l} == 2) {
|
|
if (inCoord < 0.0) {
|
|
if (len <= 1.0) {
|
|
inCoord = 0.0;
|
|
} else {
|
|
float sz2 = 2.0 * len;
|
|
if (inCoord < sz2) {
|
|
inCoord = sz2 * float(int(float(-inCoord / sz2))) +
|
|
inCoord;
|
|
}
|
|
inCoord = inCoord < -len ? inCoord + sz2 : -inCoord - 1.0;
|
|
}
|
|
} else if (inCoord > len - 1.0) {
|
|
if (len <= 1.0) {
|
|
inCoord = 0.0;
|
|
} else {
|
|
float sz2 = 2.0 * len;
|
|
inCoord -= sz2 * float(int(float(inCoord / sz2)));
|
|
if (inCoord >= len) {
|
|
inCoord = sz2 - inCoord - 1.0;
|
|
}
|
|
}
|
|
}
|
|
return clamp(inCoord, 0.0, len - 1.0);
|
|
} else if (${l} == 3) {
|
|
if (inCoord < 0.0) {
|
|
if (len <= 1.0) {
|
|
inCoord = 0.0;
|
|
} else {
|
|
float sz = len - 1.0;
|
|
inCoord += len * (float(int(float(-inCoord / sz))) + 1.0);
|
|
}
|
|
} else if (inCoord > len - 1.0) {
|
|
if (len <= 1.0) {
|
|
inCoord = 0.0;
|
|
} else {
|
|
float sz = len - 1.0;
|
|
inCoord -= len * float(int(float(inCoord / sz)));
|
|
}
|
|
}
|
|
return clamp(inCoord, 0.0, len - 1.0);
|
|
} else if (${l} == 4) {
|
|
return clamp(outCoord, 0.0, len - 1.0);
|
|
} else {
|
|
return outCoord;
|
|
}
|
|
}
|
|
|
|
float readWithFillValue(int batch, int coordY, int coordX,
|
|
int channel) {
|
|
float outputValue;
|
|
if (0 <= coordY && coordY < ${e} && 0 <= coordX && coordX < ${t}) {
|
|
outputValue = getImage(batch, coordY, coordX, channel);
|
|
} else {
|
|
outputValue = float(${s});
|
|
}
|
|
return outputValue;
|
|
}
|
|
|
|
void main() {
|
|
ivec4 coords = getOutputCoords();
|
|
float outputValue;
|
|
int batch = coords[0];
|
|
int x = coords[2];
|
|
int y = coords[1];
|
|
int channel = coords[3];
|
|
float xf = float(x);
|
|
float yf = float(y);
|
|
float a1 = getTransforms(batch, 0);
|
|
float a2 = getTransforms(batch, 1);
|
|
float a3 = getTransforms(batch, 2);
|
|
float b1 = getTransforms(batch, 3);
|
|
float b2 = getTransforms(batch, 4);
|
|
float b3 = getTransforms(batch, 5);
|
|
float c1 = getTransforms(batch, 6);
|
|
float c2 = getTransforms(batch, 7);
|
|
float projection = c1 * xf + c2 * yf + 1.0;
|
|
if (projection == 0.0) {
|
|
outputValue = float(${s});
|
|
} else {
|
|
float inX = (a1 * xf + a2 * yf + a3) / projection;
|
|
float inY = (b1 * xf + b2 * yf + b3) / projection;
|
|
float mapX = mapCoord(inX, float(${t}));
|
|
float mapY = mapCoord(inY, float(${e}));
|
|
|
|
if (${i} == 1) {
|
|
int coordY = int(round(mapY));
|
|
int coordX = int(round(mapX));
|
|
outputValue = readWithFillValue(batch, coordY, coordX,
|
|
channel);
|
|
} else {
|
|
float yFloor = floor(mapY);
|
|
float xFloor = floor(mapX);
|
|
float yCeil = yFloor + 1.0;
|
|
float xCeil = xFloor + 1.0;
|
|
float valueYFloor = (xCeil - mapX) *
|
|
readWithFillValue(batch, int(yFloor), int(xFloor), channel) +
|
|
(mapX - xFloor) *
|
|
readWithFillValue(batch, int(yFloor), int(xCeil), channel);
|
|
float valueYCeil = (xCeil - mapX) *
|
|
readWithFillValue(batch, int(yCeil), int(xFloor), channel) +
|
|
(mapX - xFloor) *
|
|
readWithFillValue(batch, int(yCeil), int(xCeil), channel);
|
|
outputValue = (yCeil - mapY) * valueYFloor +
|
|
(mapY - yFloor) * valueYCeil;
|
|
}
|
|
}
|
|
setOutput(outputValue);
|
|
}
|
|
`}};function vZ(r){let{inputs:e,backend:t,attrs:n}=r,{image:o,transforms:s}=e,{interpolation:a,fillMode:i,fillValue:l,outputShape:u}=n,[c,p,m,f]=o.shape,[d,h]=u!=null?u:[p,m],g=[c,d,h,f],x=new EC(p,m,a,i,l,g);return t.runWebGLProgram(x,[o,s],"float32")}var cO={kernelName:la,backendName:"webgl",kernelFunc:vZ};function CZ(r){let{inputs:e,attrs:t,backend:n}=r,{axis:o}=t,{x:s}=e;Cs(s,"unique"),console.warn("WARNING: ","UI might be locked temporarily as data is being downloaded");let a=n.readSync(s.dataId),{outputValues:i,outputShape:l,indices:u}=FD(a,o,s.shape,s.dtype);return[n.makeTensorInfo(l,s.dtype,i),n.makeTensorInfo([u.length],"int32",u)]}var pO={kernelName:Kc,backendName:"webgl",kernelFunc:CZ};function IZ(r){let{inputs:e,backend:t,attrs:n}=r,{value:o}=e,{axis:s}=n;s<0&&(s+=o.shape.length);let a=o,i=a.shape.length,l=o.shape[s],u=new Array(i-1),c=0;for(let h=0;h<i;h++)h!==s&&(u[c++]=a.shape[h]);let p=[],m=new Array(i).fill(0),f=a.shape.slice();f[s]=1;let d=new Array(l);for(let h=0;h<d.length;h++){m[s]=h;let g=Fa({inputs:{x:a},backend:t,attrs:{begin:m,size:f}}),x=ce({inputs:{x:g},backend:t,attrs:{shape:u}});d[h]=x,p.push(g)}return p.forEach(h=>t.disposeIntermediateTensorInfo(h)),d}var mO={kernelName:Zs,backendName:"webgl",kernelFunc:IZ};var DC=class{constructor(e,t){this.variableNames=["x","segmentIds"];let n=e.windowSize,o=e.batchSize,s=e.inSize,a=e.numSegments,i=a*Math.ceil(s/n);this.outputShape=[o,i];let l="0.0",u="sumValue",c=Math.floor(n/4)*4,p=n%4,m=`
|
|
sumValue += dot(values, segFilter);
|
|
`,f="";s%n>0&&(f=`
|
|
if (inIdx < 0 || inIdx >= ${s}) {
|
|
return initializationValue;
|
|
}
|
|
`);let d="";s%n>0&&(d=`
|
|
if (inIdx < 0 || inIdx >= ${s}) {
|
|
return -1.0;
|
|
}
|
|
`),this.userCode=`
|
|
const float initializationValue = ${l};
|
|
|
|
float getValue(int batch, int inIdx) {
|
|
${f}
|
|
return getX(batch, inIdx);
|
|
}
|
|
|
|
float getSegmentIdAtIndex(int inIdx) {
|
|
${d}
|
|
return getSegmentIds(inIdx);
|
|
}
|
|
|
|
void main() {
|
|
ivec2 coords = getOutputCoords();
|
|
int batch = coords[0];
|
|
int outIdx = coords[1];
|
|
int inOffset = int(floor(float(outIdx) / float(
|
|
${a})) * float(${n}));
|
|
int currentSeg = int(mod(float(outIdx), float(${a})));
|
|
|
|
float sumValue = 0.0;
|
|
|
|
for (int i = 0; i < ${c}; i += 4) {
|
|
int inIdx = inOffset + i;
|
|
vec4 values = vec4(
|
|
getValue(batch, inIdx),
|
|
getValue(batch, inIdx + 1),
|
|
getValue(batch, inIdx + 2),
|
|
getValue(batch, inIdx + 3)
|
|
);
|
|
|
|
vec4 segFilter = vec4(
|
|
int(getSegmentIdAtIndex(inIdx)) == currentSeg ? 1 : 0,
|
|
int(getSegmentIdAtIndex(inIdx + 1)) == currentSeg ? 1 : 0,
|
|
int(getSegmentIdAtIndex(inIdx + 2)) == currentSeg ? 1 : 0,
|
|
int(getSegmentIdAtIndex(inIdx + 3)) == currentSeg ? 1 : 0
|
|
);
|
|
|
|
${m}
|
|
}
|
|
|
|
int inIdx = inOffset + ${c};
|
|
if (${p===1}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, inIdx),
|
|
initializationValue,
|
|
initializationValue,
|
|
initializationValue
|
|
);
|
|
|
|
int inIdxSeg = int(getSegmentIdAtIndex(inIdx));
|
|
|
|
vec4 segFilter = vec4(
|
|
int(getSegmentIdAtIndex(inIdx)) == currentSeg ? 1 : 0,
|
|
0,
|
|
0,
|
|
0
|
|
);
|
|
|
|
${m}
|
|
} else if (${p===2}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, inIdx),
|
|
getValue(batch, inIdx + 1),
|
|
initializationValue,
|
|
initializationValue
|
|
);
|
|
|
|
vec4 segFilter = vec4(
|
|
int(getSegmentIdAtIndex(inIdx)) == currentSeg ? 1 : 0,
|
|
int(getSegmentIdAtIndex(inIdx + 1)) == currentSeg ? 1 : 0,
|
|
0,
|
|
0
|
|
);
|
|
|
|
${m}
|
|
} else if (${p===3}) {
|
|
vec4 values = vec4(
|
|
getValue(batch, inIdx),
|
|
getValue(batch, inIdx + 1),
|
|
getValue(batch, inIdx + 2),
|
|
initializationValue
|
|
);
|
|
|
|
vec4 segFilter = vec4(
|
|
int(getSegmentIdAtIndex(inIdx)) == currentSeg ? 1 : 0,
|
|
int(getSegmentIdAtIndex(inIdx + 1)) == currentSeg ? 1 : 0,
|
|
int(getSegmentIdAtIndex(inIdx + 2)) == currentSeg ? 1 : 0,
|
|
0
|
|
);
|
|
|
|
${m}
|
|
}
|
|
setOutput(${u});
|
|
}
|
|
`}};function NZ(r){let{inputs:e,backend:t,attrs:n}=r,{x:o,segmentIds:s}=e,{numSegments:a}=n,i=o.shape.length,l=[],u=0,c=C.getAxesPermutation([u],i),p=o;c!=null&&(p=Rt({inputs:{x:o},backend:t,attrs:{perm:c}}),l.push(p),u=C.getInnerMostAxes(1,i)[0]);let m=C.segment_util.computeOutShape(p.shape,u,a),f=y.sizeFromShape([p.shape[u]]),d=ce({inputs:{x:p},backend:t,attrs:{shape:[-1,f]}});l.push(d);let h=Yl(o.dtype),g=(_,I,E,D,$)=>{let R=_.shape[0],M=_.shape[1],G=C.segment_util.segOpComputeOptimalWindowSize(M,$),j={windowSize:G,inSize:M,batchSize:R,numSegments:$},U=new DC(j,I),H=t.compileAndRun(U,[_,E],D);if(l.push(H),H.shape[1]===$)return H;let q=gC({backend:t,attrs:{start:0,stop:$,step:1,dtype:"float32"}}),X=AC({inputs:{x:q},backend:t,attrs:{reps:[M/G]}});return l.push(q),l.push(X),g(H,I,X,D,$)},x=g(d,"unsortedSegmentSum",s,h,a),b=ce({inputs:{x},backend:t,attrs:{shape:m}}),w=b;if(c!=null){l.push(b);let _=C.getUndoAxesPermutation(c);w=Rt({inputs:{x:w},backend:t,attrs:{perm:_}})}return l.forEach(_=>t.disposeIntermediateTensorInfo(_)),w}var fO={kernelName:ul,backendName:"webgl",kernelFunc:NZ};var SZ=[KR,XR,u$,p$,m$,f$,h$,g$,x$,y$,_$,k$,v$,C$,N$,I$,S$,A$,T$,E$,D$,$$,R$,O$,P$,B$,G$,W$,U$,ZD,q$,X$,Y$,K$,J$,Q$,Z$,eR,tR,rR,sR,iR,aR,uR,cR,lR,pR,mR,fR,dR,hR,gR,xR,bR,wR,kR,vR,CR,IR,SR,TR,AR,ER,DR,$R,RR,FR,OR,YD,PR,H$,MR,LR,zR,JD,BR,VR,GR,jR,WR,UR,HR,qR,ZR,eF,QR,tF,rF,oF,JR,iF,aF,lF,uF,cF,hF,n$,xF,yF,bF,wF,M$,_F,CF,IF,NF,SF,QD,TF,AF,L$,pF,EF,$F,DF,s$,RF,FF,OF,PF,MF,LF,zF,BF,VF,GF,WF,jF,UF,HF,qF,F$,dF,KF,XF,YF,ZF,JF,QF,eO,tO,nO,oO,sO,fF,a$,iO,aO,lO,uO,cO,l$,pO,mO,fO,kF];for(let r of SZ)Kl(r);var Ft;(function(r){r[r.float32=0]="float32",r[r.int32=1]="int32",r[r.bool=2]="bool",r[r.string=3]="string",r[r.complex64=4]="complex64"})(Ft||(Ft={}));var Il;(function(r){r[r.linear=0]="linear",r[r.relu=1]="relu",r[r.relu6=2]="relu6",r[r.prelu=3]="prelu",r[r.leakyrelu=4]="leakyrelu",r[r.sigmoid=5]="sigmoid"})(Il||(Il={}));var dO;function TZ(r){dO=r.wasm.cwrap(Qs,null,["number","array","number","number","array","number","number","number","number","number","number","number","number"])}function AZ(r){let{inputs:e,backend:t,attrs:n}=r,{a:o,b:s,bias:a,preluActivationWeights:i}=e;if(o.dtype!=="float32"||s.dtype!=="float32")throw new Error("_FusedMatMul for non non-float32 tensors not yet supported.");let{transposeA:l,transposeB:u,activation:c,leakyreluAlpha:p}=n,m=t.dataIdMap.get(o.dataId).id,f=t.dataIdMap.get(s.dataId).id,d=0;if(a!=null){let $=t.dataIdMap.get(a.dataId);if($.shape.length!==1)throw new Error(`_FusedMatMul only supports rank-1 bias but got rank ${$.shape.length}.`);d=$.id}let h=i==null?0:t.dataIdMap.get(i.dataId).id,g=Il[c];if(g==null)throw new Error(`${c} activation not yet supported for FusedConv2D in the wasm backend.`);let x=l?o.shape[2]:o.shape[1],b=u?s.shape[1]:s.shape[2],w=o.shape[0],_=t.makeOutput([w,x,b],o.dtype),I=t.dataIdMap.get(_.dataId).id,E=new Uint8Array(new Int32Array(o.shape).buffer),D=new Uint8Array(new Int32Array(s.shape).buffer);return dO(m,E,o.shape.length,f,D,s.shape.length,l,u,g,d,h,p||0,I),_}var hO={kernelName:Qs,backendName:"wasm",setupFunc:TZ,kernelFunc:AZ};function xt(r){let e;function t(o){e=o.wasm.cwrap(r,null,["number","number"])}function n(o){let{backend:s,inputs:{x:a}}=o,i=s.dataIdMap.get(a.dataId).id,l=s.makeOutput(a.shape,a.dtype),u=s.dataIdMap.get(l.dataId).id;return y.sizeFromShape(l.shape)===0||e(i,u),l}return{kernelName:r,backendName:"wasm",setupFunc:t,kernelFunc:n}}var gO=xt(Bs);function yt(r,e,t){let n;function o(a){n=a.wasm.cwrap(r,null,["number","array","number","number","array","number","number","number"])}function s(a){let{backend:i,inputs:l}=a,{a:u,b:c}=l,p=i.dataIdMap.get(u.dataId).id,m=i.dataIdMap.get(c.dataId).id,f=t!=null?t:u.dtype,d=C.assertAndGetBroadcastShape(u.shape,c.shape),h=i.makeOutput(d,f);if(y.sizeFromShape(d)===0)return h;let g=new Uint8Array(new Int32Array(u.shape).buffer),x=new Uint8Array(new Int32Array(c.shape).buffer),b=i.dataIdMap.get(h.dataId).id,w=()=>n(p,g,u.shape.length,m,x,c.shape.length,Ft[u.dtype],b);if(e&&u.dtype==="float32")return w(),h;let _=C.getBroadcastDims(u.shape,d),I=C.getBroadcastDims(c.shape,d),E=_.every(($,R)=>$===R),D=I.every(($,R)=>$===R);if(E&&D)return w(),h;throw new Error(`Broadcasting along outer dims is not yet supported for ${u.dtype} ${r}.`)}return{kernelName:r,backendName:"wasm",setupFunc:o,kernelFunc:s}}var EZ=!0,xO=yt(Pn,EZ);var yO;function DZ(r){yO=r.wasm.cwrap(_o,null,["array","number","number","number"])}function $Z(r){let{inputs:e,backend:t}=r,n=t.makeOutput(e[0].shape,e[0].dtype);if(y.sizeFromShape(n.shape)===0)return n;let o=e.map(i=>t.dataIdMap.get(i.dataId).id),s=new Uint8Array(new Int32Array(o).buffer),a=t.dataIdMap.get(n.dataId).id;return yO(s,o.length,Ft[n.dtype],a),n}var bO={kernelName:_o,backendName:"wasm",setupFunc:DZ,kernelFunc:$Z};function ju(r){let{inputs:{x:e},backend:t}=r,n=t.makeOutput(e.shape,e.dtype),o=t.typedArrayFromHeap(e);return t.typedArrayFromHeap(n).set(o),n}var wO={kernelName:Xn,backendName:"wasm",kernelFunc:ju};var _O;function RZ(r){_O=r.wasm.cwrap(ms,null,["number","array","number","number","number","array","number"])}function Fp(r){let{inputs:e,backend:t,attrs:n}=r,[o,s]=OZ(e.x.shape,n.perm),a=!0;for(let d=0;d<s.length;d++)s[d]!==d&&(a=!1);let i=FZ(e.x.shape,n.perm),l={dataId:e.x.dataId,shape:o,dtype:e.x.dtype};if(a){let d=ju({inputs:e,backend:t});return d.shape=i,d}let u=t.makeOutput(i,l.dtype),c=t.dataIdMap.get(l.dataId).id,p=t.dataIdMap.get(u.dataId).id,m=new Uint8Array(new Int32Array(s).buffer),f=new Uint8Array(new Int32Array(l.shape).buffer);return _O(c,f,l.shape.length,Ft[l.dtype],p,m,s.length),u}function FZ(r,e){let t=new Array(r.length);for(let n=0;n<t.length;n++)t[n]=r[e[n]];return t}function OZ(r,e){let t=[],n=[];for(let o=0;o<r.length;++o)r[o]!==1&&t.push(r[o]),r[e[o]]!==1&&n.push(e[o]);for(let o=0;o<n.length;++o){let s=-1;for(let a=0;a<n.length;++a)n[a]>=o&&(s===-1||n[s]>n[a])&&(s=a);n[s]=o}return[t,n]}var kO={kernelName:ms,backendName:"wasm",kernelFunc:Fp,setupFunc:RZ};function tn(r,e,t){let n=r.shape,o=r.shape.length,s=y.parseAxisParam(e,n),a=s,i=C.getAxesPermutation(a,o),l=null,u=!1;if(i!=null){let c=new Array(o);for(let f=0;f<c.length;f++)c[f]=n[i[f]];a=C.getInnerMostAxes(a.length,o),l=Fp({inputs:{x:r},attrs:{perm:i},backend:t});let p=t.dataIdMap.get(r.dataId).id;t.dataIdMap.get(l.dataId).id!==p&&(u=!0)}return{transposed:l,originalAxes:s,axes:a,inputWasTransposed:u}}var vO;function PZ(r){vO=r.wasm.cwrap(Ci,null,["number, number, number"])}function MZ(r){let{backend:e,inputs:t,attrs:n}=r,{axis:o,keepDims:s}=n,{x:a}=t,l=e.dataIdMap.get(a.dataId).id,u=a,{transposed:c,axes:p,originalAxes:m,inputWasTransposed:f}=tn(a,o,e);if(f){let w=e.dataIdMap.get(c.dataId).id;u=c,l=w}let d=u.shape.length;C.assertAxesAreInnerMostDims("all",p,d);let[h,g]=C.computeOutAndReduceShapes(u.shape,p),x=y.sizeFromShape(g),b=e.makeOutput(h,a.dtype);if(y.sizeFromShape(u.shape)!==0){let w=e.dataIdMap.get(b.dataId).id;vO(l,x,w)}if(f&&e.disposeData(c.dataId),s){let w=C.expandShapeToKeepDim(b.shape,m);b.shape=w}return b}var CO={kernelName:Ci,backendName:"wasm",setupFunc:PZ,kernelFunc:MZ};var IO;function LZ(r){IO=r.wasm.cwrap(Ii,null,["number, number, number"])}function zZ(r){let{backend:e,inputs:t,attrs:n}=r,{axis:o,keepDims:s}=n,{x:a}=t,l=e.dataIdMap.get(a.dataId).id,u=a,{transposed:c,axes:p,originalAxes:m,inputWasTransposed:f}=tn(a,o,e);if(f){let w=e.dataIdMap.get(c.dataId).id;u=c,l=w}let d=u.shape.length;C.assertAxesAreInnerMostDims("any",p,d);let[h,g]=C.computeOutAndReduceShapes(u.shape,p),x=y.sizeFromShape(g),b=e.makeOutput(h,a.dtype);if(y.sizeFromShape(u.shape)!==0){let w=e.dataIdMap.get(b.dataId).id;IO(l,x,w)}if(f&&e.disposeData(c.dataId),s){let w=C.expandShapeToKeepDim(b.shape,m);b.shape=w}return b}var NO={kernelName:Ii,backendName:"wasm",setupFunc:LZ,kernelFunc:zZ};var SO;function BZ(r){SO=r.wasm.cwrap(ko,null,["number","number","number","number","number"])}function VZ(r){let{backend:e,inputs:t,attrs:n}=r,{axis:o}=n,{x:s}=t,a=e.dataIdMap.get(s.dataId).id,i=a,l=s,{transposed:u,axes:c,inputWasTransposed:p}=tn(s,o,e);if(p){let x=e.dataIdMap.get(u.dataId).id;x!==a&&(l=u,i=x)}let m=l.shape.slice(0,-1),f=e.makeOutput(m,"int32"),d=e.dataIdMap.get(f.dataId).id,h=y.sizeFromShape(f.shape),g=l.shape[c[0]];return SO(i,Ft[l.dtype],h,g,d),p&&e.disposeData(u.dataId),f}var TO={kernelName:ko,backendName:"wasm",kernelFunc:VZ,setupFunc:BZ};var AO;function GZ(r){AO=r.wasm.cwrap(vo,null,["number","number","number","number","number","number","number","number","number","number","number","number","number","number"])}function WZ(r){let{inputs:e,attrs:t,backend:n}=r,o=e.x,s=n.dataIdMap.get(o.dataId).id,{filterSize:a,strides:i,pad:l,dimRoundingMode:u}=t,c=C.computePool2DInfo(o.shape,a,i,1,l,u),p=c.filterHeight,m=c.filterWidth,f=c.padInfo.top,d=c.padInfo.right,h=c.padInfo.bottom,g=c.padInfo.left,x=c.strideHeight,b=c.strideWidth,w=c.inChannels;if(c.dataFormat!=="channelsLast")throw new Error(`wasm backend does not support dataFormat:'${c.dataFormat}'. Please use 'channelsLast'.`);if(c.dilationWidth!==1||c.dilationHeight!==1)throw new Error(`was backend only supports average pooling with dilation = [1, 1], got [${c.dilationHeight}, ${c.dilationWidth}].`);let _=n.makeOutput(c.outShape,"float32"),I=n.dataIdMap.get(_.dataId).id;return AO(s,o.shape[0],o.shape[1],o.shape[2],p,m,f,d,h,g,x,b,w,I),_}var EO={kernelName:vo,backendName:"wasm",setupFunc:GZ,kernelFunc:WZ};function Vr(r){let{inputs:e,attrs:t}=r,{x:n}=e,{shape:o}=t,s=y.sizeFromShape(n.shape),a=y.inferFromImplicitShape(o,s);return y.assert(s===y.sizeFromShape(a),()=>`new shape: ${a}, old shape: ${n.shape}. New shape and old shape must have the same number of elements.`),r.backend.incRef(n.dataId),{dataId:n.dataId,shape:a,dtype:n.dtype}}var DO={kernelName:qs,backendName:"wasm",kernelFunc:Vr};var $O;function jZ(r){$O=r.wasm.cwrap(Co,null,["number","array","number","number","array","number","number","number","number"])}function UZ(r){let{inputs:e,backend:t,attrs:n}=r,{a:o,b:s}=e,{transposeA:a,transposeB:i}=n;if(o.dtype!=="float32"||s.dtype!=="float32")throw new Error("BatchMatMul for non non-float32 tensors not yet supported.");let l=o.shape.length,u=s.shape.length,c=a?o.shape[l-2]:o.shape[l-1],p=i?s.shape[u-1]:s.shape[u-2],m=a?o.shape[l-1]:o.shape[l-2],f=i?s.shape[u-2]:s.shape[u-1],d=o.shape.slice(0,-2),h=s.shape.slice(0,-2),g=y.sizeFromShape(d),x=y.sizeFromShape(h),b=g===x||g===1||x===1;y.assert(l>=2&&u>=2&&b,()=>`Error in matMul: the input batch dimensions must either be the same or at least one input batch dimension must be 1. Got input batch dimensions of (${d}) and (${h}).`);let _=(g>x?o.shape.slice(0,-2):s.shape.slice(0,-2)).concat([m,f]);y.assert(c===p,()=>`Error in matMul: inner shapes (${c}) and (${p}) of Tensors with shapes ${o.shape} and ${s.shape} and transposeA=${a} and transposeB=${i} must match.`);let I=a?[g,c,m]:[g,m,c],E=i?[x,f,p]:[x,p,f],D=Vr({inputs:{x:o},backend:t,attrs:{shape:I}}),$=Vr({inputs:{x:s},backend:t,attrs:{shape:E}}),R=t.dataIdMap.get(D.dataId).id,M=t.dataIdMap.get($.dataId).id,G=a?D.shape[2]:D.shape[1],j=i?$.shape[1]:$.shape[2],U=Math.max(g,x),H=t.makeOutput([U,G,j],D.dtype),q=t.dataIdMap.get(H.dataId).id,X=new Uint8Array(new Int32Array(D.shape).buffer),ne=new Uint8Array(new Int32Array($.shape).buffer);return $O(R,X,D.shape.length,M,ne,$.shape.length,a,i,q),t.disposeData(D.dataId),t.disposeData($.dataId),H.shape=_,H}var RO={kernelName:Co,backendName:"wasm",setupFunc:jZ,kernelFunc:UZ};function Uu(r){let{inputs:{x:e},attrs:{dtype:t},backend:n}=r,o=n.makeOutput(e.shape,t),s=n.typedArrayFromHeap(e);return n.typedArrayFromHeap(o).set(s),o}var FO={kernelName:qn,backendName:"wasm",kernelFunc:Uu};var OO=xt(Io);var PO;function HZ(r){PO=r.wasm.cwrap(Kn,null,["number","number","number","number"])}function qZ(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{clipValueMin:s,clipValueMax:a}=n,i=t.dataIdMap.get(o.dataId).id,l=t.makeOutput(o.shape,o.dtype),u=t.dataIdMap.get(l.dataId).id;return PO(i,s,a,u),l}var MO={kernelName:Kn,backendName:"wasm",setupFunc:HZ,kernelFunc:qZ};function $C(r){let{inputs:e,backend:t}=r,n=y.parseAxisParam(r.attrs.axis,e[0].shape)[0],o=C.computeOutShape(e.map(f=>f.shape),n),s=e.filter(f=>y.sizeFromShape(f.shape)>0);if(s.length===1)return ju({inputs:{x:s[0]},backend:t});let a=t.makeOutput(o,e[0].dtype);if(y.sizeFromShape(o)===0)return a;let i=s.map(f=>f.shape);if(C.assertParamsConsistent(i,n),s[0].dtype==="string"){let f=s.map(w=>{let _=y.sizeFromShape(w.shape.slice(n));return Vr({inputs:{x:w},backend:t,attrs:{shape:[-1,_]}})}),d=f.map(w=>({vals:t.readSync(w.dataId),shape:w.shape}));o=C.computeOutShape(f.map(w=>w.shape),1);let h=f[0].shape[0]===1,g=Eu(d,o,e[0].dtype,h),x=C.computeOutShape(s.map(w=>w.shape),n);a.shape=x;let b=t.dataIdMap.get(a.dataId);return b.stringBytes=C.fromStringArrayToUint8(g),f.forEach(w=>t.disposeData(w.dataId)),a}let l=y.sizeFromShape(s[0].shape.slice(0,n)),u=0,c=s.map(f=>{let d=y.sizeFromShape(f.shape.slice(n));return u+=d,d}),p=s.map(f=>t.typedArrayFromHeap(f)),m=t.typedArrayFromHeap(a);for(let f=0;f<l;f++){let d=f*u;for(let h=0;h<p.length;h++){let g=c[h],x=f*g,b=p[h].subarray(x,x+g);m.set(b,d),d+=g}}return a}var LO={kernelName:Vs,backendName:"wasm",kernelFunc:$C};var zO;function KZ(r){zO=r.wasm.cwrap(No,null,["number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number"])}function XZ(r){let{inputs:e,attrs:t,backend:n}=r,{x:o,filter:s}=e,a=n.dataIdMap.get(o.dataId).id,i=n.dataIdMap.get(s.dataId).id,{strides:l,dilations:u,pad:c,dimRoundingMode:p,dataFormat:m}=t,f=C.convertConv2DDataFormat(m),d=C.computeConv2DInfo(o.shape,s.shape,l,u,c,p,!1,f),h=d.filterHeight,g=d.filterWidth,x=d.padInfo.top,b=d.padInfo.right,w=d.padInfo.bottom,_=d.padInfo.left,I=d.dilationHeight,E=d.dilationWidth,D=d.strideHeight,$=d.strideWidth,R=d.inChannels,M=d.outChannels,G=d.padInfo.type==="SAME"?1:0;if(d.dataFormat!=="channelsLast")throw new Error(`wasm backend Conv2D does not support dataFormat:'${d.dataFormat}'. Please use 'channelsLast'.`);let j=n.makeOutput(d.outShape,"float32"),U=n.dataIdMap.get(j.dataId).id;return zO(a,o.shape[0],o.shape[1],o.shape[2],i,h,g,x,b,w,_,G,I,E,D,$,R,M,U),j}var BO={kernelName:No,backendName:"wasm",setupFunc:KZ,kernelFunc:XZ};var VO;function YZ(r){VO=r.wasm.cwrap(So,null,["number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number"])}function ZZ(r){let{backend:e,inputs:t,attrs:n}=r,{dy:o,filter:s}=t,{strides:a,pad:i,dataFormat:l,dimRoundingMode:u,inputShape:c}=n,p=1,m=C.convertConv2DDataFormat(l),f=C.computeConv2DInfo(c,s.shape,a,p,i,u,!1,m),{batchSize:d,filterHeight:h,filterWidth:g,inChannels:x,inHeight:b,inWidth:w,outChannels:_,outHeight:I,outWidth:E,strideHeight:D,strideWidth:$}=f,R=h-1-f.padInfo.top,M=g-1-f.padInfo.left,G=f.dataFormat==="channelsLast",j=y.computeStrides(f.inShape),U=y.computeStrides(o.shape),[H,q,X]=y.computeStrides(s.shape),ne=j[0],Y=G?j[1]:j[2],re=G?j[2]:1,ee=G?1:j[1],ie=U[0],me=G?U[1]:U[2],ae=G?U[2]:1,de=G?1:U[1],he=e.makeOutput(f.inShape,"float32"),xe=e.dataIdMap.get(he.dataId).id,we=e.dataIdMap.get(o.dataId).id,De=e.dataIdMap.get(s.dataId).id;return VO(we,De,d,h,g,b,w,x,I,E,_,D,$,R,M,H,q,X,ne,Y,re,ee,ie,me,ae,de,xe),he}var GO={kernelName:So,backendName:"wasm",setupFunc:YZ,kernelFunc:ZZ};var WO=xt(To);var RC;(function(r){r[r.bilinear=0]="bilinear",r[r.nearest=1]="nearest"})(RC||(RC={}));var jO;function JZ(r){jO=r.wasm.cwrap($i,null,["number","number","number","number","array","number","number","number","number","number"])}function QZ(r){let{backend:e,inputs:t,attrs:n}=r,{method:o,extrapolationValue:s,cropSize:a}=n,{image:i,boxes:l,boxInd:u}=t,c=l.shape[0],[p,m]=a,f=[c,p,m,i.shape[3]],d=e.dataIdMap.get(i.dataId),h;i.dtype!=="float32"&&(h=Uu({backend:e,inputs:{x:i},attrs:{dtype:"float32"}}),d=e.dataIdMap.get(h.dataId));let g=d.id,x=e.dataIdMap.get(l.dataId).id,b=e.dataIdMap.get(u.dataId).id,w=e.makeOutput(f,"float32"),_=e.dataIdMap.get(w.dataId).id,I=new Uint8Array(new Int32Array(i.shape).buffer);return jO(g,x,b,c,I,p,m,RC[o],s,_),h!=null&&e.disposeData(h.dataId),w}var UO={kernelName:$i,backendName:"wasm",setupFunc:JZ,kernelFunc:QZ};var HO;function eJ(r){HO=r.wasm.cwrap(Ao,null,["number","number","number","number","number","number"])}function tJ(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{axis:s,exclusive:a,reverse:i}=n,l=o.shape.length;y.assert(o.dtype==="float32"||o.dtype==="int32",()=>`cumsum does not support ${o.dtype} tensors in the WASM backend`);let u=C.getAxesPermutation([s],l),c=o;u!==null&&(c=Fp({inputs:{x:o},attrs:{perm:u},backend:t}));let p=C.getInnerMostAxes(1,l)[0];C.assertAxesAreInnerMostDims("cumsum",[p],l);let m=t.makeOutput(c.shape,c.dtype),f=c.shape[p],d=t.dataIdMap.get(c.dataId).id,h=t.dataIdMap.get(m.dataId).id;HO(d,a?1:0,i?1:0,f,h,Ft[o.dtype]);let g=m;if(u!==null){let x=C.getUndoAxesPermutation(u);g=Fp({inputs:{x:m},attrs:{perm:x},backend:t}),t.disposeData(c.dataId),t.disposeData(m.dataId)}return g}var qO={kernelName:Ao,backendName:"wasm",setupFunc:eJ,kernelFunc:tJ};var KO;function rJ(r){KO=r.wasm.cwrap(Ri,null,["number","number","number","array","number","array","array","number","number"])}function nJ(r){let{backend:e,inputs:t,attrs:n}=r,{x:o}=t,{blockSize:s,dataFormat:a}=n;y.assert(s>1,()=>`blockSize should be > 1 for depthToSpace, but was: ${s}`);let i=o.shape[0],l=a==="NHWC"?o.shape[1]:o.shape[2],u=a==="NHWC"?o.shape[2]:o.shape[3],c=a==="NHWC"?o.shape[3]:o.shape[1],p=l*s,m=u*s,f=c/(s*s),d=a==="NHWC"?[i,p,m,f]:[i,f,p,m],h=e.makeOutput(d,"float32"),x=e.dataIdMap.get(o.dataId).id,b=new Uint8Array(new Int32Array(y.computeStrides(o.shape)).buffer),w=new Uint8Array(new Int32Array(d).buffer),_=new Uint8Array(new Int32Array(y.computeStrides(d)).buffer),I=e.dataIdMap.get(h.dataId).id;return KO(x,s,a==="NHWC"?1:0,b,o.shape.length-1,w,_,d.length,I),h}var XO={kernelName:Ri,backendName:"wasm",setupFunc:rJ,kernelFunc:nJ};var YO;function oJ(r){YO=r.wasm.cwrap(Eo,null,["number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number"])}function sJ(r){let{inputs:e,attrs:t,backend:n}=r,{x:o,filter:s}=e,a=n.dataIdMap.get(o.dataId).id,i=n.dataIdMap.get(s.dataId).id,{strides:l,dilations:u,pad:c,dimRoundingMode:p}=t,m=u==null?[1,1]:u,f=C.computeConv2DInfo(o.shape,s.shape,l,m,c,p,!0),d=f.filterHeight,h=f.filterWidth,g=f.padInfo.top,x=f.padInfo.right,b=f.padInfo.bottom,w=f.padInfo.left,_=f.dilationHeight,I=f.dilationWidth,E=f.strideHeight,D=f.strideWidth,$=f.inChannels,R=f.outChannels,M=f.padInfo.type==="SAME"?1:0;if(f.dataFormat!=="channelsLast")throw new Error(`wasm backend DepthwiseConv2dNative does not support dataFormat:'${f.dataFormat}'. Please use 'channelsLast'.`);let G=n.makeOutput(f.outShape,"float32"),j=n.dataIdMap.get(G.dataId).id;return YO(a,o.shape[0],o.shape[1],o.shape[2],i,d,h,g,x,b,w,M,_,I,E,D,$,R,j),G}var ZO={kernelName:Eo,backendName:"wasm",setupFunc:oJ,kernelFunc:sJ};var iJ=!1,JO=yt(Pi,iJ,"bool");var QO=xt($o);function Tx(r){let{inputs:e,attrs:t,backend:n}=r,{input:o}=e,{dim:s}=t,a=o.shape.length,i=o.shape.slice(),l=s;return s<0&&(y.assert(-(a+1)<=s,()=>`Axis must be in the interval [${-(a+1)}, ${a}]`),l=a+s+1),i.splice(l,0,1),Vr({inputs:{x:o},backend:n,attrs:{shape:i}})}var eP={kernelName:Gs,backendName:"wasm",kernelFunc:Tx};function aJ(r){let{attrs:{shape:e,value:t,dtype:n},backend:o}=r,s=o.makeOutput(e,n);return o.typedArrayFromHeap(s).fill(t),s}var tP={kernelName:rl,backendName:"wasm",kernelFunc:aJ};var rP;function lJ(r){rP=r.wasm.cwrap(Li,null,["number","number","number","number","number","number"])}function uJ(r){let{inputs:e,backend:t}=r,{image:n}=e,o=t.makeOutput(n.shape,n.dtype),s=t.dataIdMap.get(n.dataId).id,a=t.dataIdMap.get(o.dataId).id,[i,l,u,c]=n.shape;return rP(s,i,l,u,c,a),o}var nP={kernelName:Li,backendName:"wasm",kernelFunc:uJ,setupFunc:lJ};var oP=xt(Ro);var cJ=!1,sP=yt(Fo,cJ);var iP;function pJ(r){iP=r.wasm.cwrap(Oo,null,["number","number","number","number","number","number","number"])}function mJ(r){let{backend:e,inputs:t,attrs:n}=r,{varianceEpsilon:o}=n,{x:s,mean:a,variance:i,offset:l,scale:u}=t,c=e.dataIdMap.get(s.dataId).id,p=e.dataIdMap.get(a.dataId).id,m=e.dataIdMap.get(i.dataId).id,f=l!=null?e.dataIdMap.get(l.dataId).id:0,d=u!=null?e.dataIdMap.get(u.dataId).id:0,h=e.makeOutput(s.shape,s.dtype);if(y.sizeFromShape(s.shape)===0)return h;let g=e.dataIdMap.get(h.dataId).id;return iP(c,p,m,f,d,o,g),h}var aP={kernelName:Oo,backendName:"wasm",setupFunc:pJ,kernelFunc:mJ};var lP;function fJ(r){lP=r.wasm.cwrap(ei,null,["number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number"])}function dJ(r){let{inputs:e,attrs:t,backend:n}=r,{x:o,filter:s,bias:a,preluActivationWeights:i}=e,{strides:l,pad:u,dilations:c,dataFormat:p,dimRoundingMode:m,activation:f,leakyreluAlpha:d}=t,h=C.computeConv2DInfo(o.shape,s.shape,l,c,u,m),g=Il[f];if(g==null)throw new Error(`${f} activation not yet supported for FusedConv2D in the wasm backend.`);let x=n.dataIdMap.get(o.dataId).id,b=n.dataIdMap.get(s.dataId).id,w=h.outChannels,_=0;if(a!=null){let ae=n.dataIdMap.get(a.dataId);if(ae.shape.length!==1)throw new Error(`FusedConv2D only supports rank-1 bias but got rank ${ae.shape.length}.`);if(ae.shape[0]!==w)throw new Error(`FusedConv2D bias shape (${ae.shape}) does not match the number of output channels (${w})`);_=ae.id}let I=h.filterHeight,E=h.filterWidth,D=h.padInfo.top,$=h.padInfo.right,R=h.padInfo.bottom,M=h.padInfo.left,G=h.dilationHeight,j=h.dilationWidth,U=h.strideHeight,H=h.strideWidth,q=h.inChannels,X=h.padInfo.type==="SAME"?1:0,ne=h.batchSize,Y=h.inHeight,re=h.inWidth;if(p!=="NHWC")throw new Error(`wasm backend FusedConv2D does not support dataFormat:'${p}'. Please use 'NHWC'.`);let ee=n.makeOutput(h.outShape,"float32"),ie=n.dataIdMap.get(ee.dataId).id,me=i==null?0:n.dataIdMap.get(i.dataId).id;return lP(x,ne,Y,re,b,I,E,_,D,$,R,M,X,G,j,U,H,q,w,g,me,d||0,ie),ee}var uP={kernelName:ei,backendName:"wasm",setupFunc:fJ,kernelFunc:dJ};var cP;function hJ(r){cP=r.wasm.cwrap(ti,null,["number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number"])}function gJ(r){let{inputs:e,attrs:t,backend:n}=r,{x:o,filter:s,bias:a,preluActivationWeights:i}=e,{strides:l,pad:u,dilations:c,dataFormat:p,dimRoundingMode:m,activation:f,leakyreluAlpha:d}=t,h=C.computeConv2DInfo(o.shape,s.shape,l,c,u,m,!0),g=Il[f];if(g==null)throw new Error(`${f} activation not yet supported for FusedDepthwiseConv2D in the wasm backend.`);let x=n.dataIdMap.get(o.dataId).id,b=n.dataIdMap.get(s.dataId).id,w=h.outChannels,_=0;if(a!=null){let ae=n.dataIdMap.get(a.dataId);if(ae.shape.length!==1)throw new Error(`FusedDepthwiseConv2D only supports rank-1 bias but got rank ${ae.shape.length}.`);if(ae.shape[0]!==w)throw new Error(`FusedDepthwiseConv2D bias shape (${ae.shape}) does not match the number of output channels (${w})`);_=ae.id}let I=h.filterHeight,E=h.filterWidth,D=h.padInfo.top,$=h.padInfo.right,R=h.padInfo.bottom,M=h.padInfo.left,G=h.dilationHeight,j=h.dilationWidth,U=h.strideHeight,H=h.strideWidth,q=h.inChannels,X=h.padInfo.type==="SAME"?1:0,ne=h.batchSize,Y=h.inHeight,re=h.inWidth;if(p!=="NHWC")throw new Error(`wasm backend FusedDepthwiseConv2D does not support dataFormat:'${p}'. Please use 'NHWC'.`);let ee=n.makeOutput(h.outShape,"float32"),ie=n.dataIdMap.get(ee.dataId).id,me=i==null?0:n.dataIdMap.get(i.dataId).id;return cP(x,ne,Y,re,b,I,E,_,D,$,R,M,X,G,j,U,H,q,w,g,me,d||0,ie),ee}var pP={kernelName:ti,backendName:"wasm",setupFunc:hJ,kernelFunc:gJ};var mP;function xJ(r){mP=r.wasm.cwrap(zi,null,["number","number","number","number","number","number","array","number"])}function yJ(r){let{backend:e,inputs:t}=r,{params:n,indices:o}=t,[s,a,i,l]=pg.prepareAndValidate(n,o),u=e.makeOutput(s,n.dtype);if(a===0)return u;let c=o.shape,p=c[c.length-1],f=e.dataIdMap.get(n.dataId).id,h=e.dataIdMap.get(o.dataId).id,g=new Uint8Array(new Int32Array(l).buffer),x=e.dataIdMap.get(u.dataId).id;return mP(f,Ft[n.dtype],h,a,p,i,g,x),u}var fP={kernelName:zi,backendName:"wasm",setupFunc:xJ,kernelFunc:yJ};var dP;function bJ(r){dP=r.wasm.cwrap("Gather",null,["number","number","array","number","number","number","array","number"])}function wJ(r){let{backend:e,inputs:t,attrs:n}=r,{x:o,indices:s}=t,{axis:a,batchDims:i}=n,l=y.parseAxisParam(a,o.shape)[0],u=C.segment_util.collectGatherOpShapeInfo(o,s,l,i),c=Vr({inputs:{x:o},attrs:{shape:[u.batchSize,u.outerSize,u.dimSize,u.sliceSize]},backend:e}),p=y.sizeFromShape(s.shape),m=Vr({inputs:{x:s},attrs:{shape:[u.batchSize,p/u.batchSize]},backend:e}),f=[u.batchSize,u.outerSize,p/u.batchSize,u.sliceSize],d=e.makeOutput(f,o.dtype);if(y.sizeFromShape(o.shape)===0)return d;let h=c.shape.length-1,x=e.dataIdMap.get(c.dataId).id,w=e.dataIdMap.get(m.dataId).id,_=e.dataIdMap.get(d.dataId).id,I=new Uint8Array(new Int32Array(y.computeStrides(c.shape)).buffer),E=new Uint8Array(new Int32Array(y.computeStrides(f)).buffer);return dP(x,Ft[o.dtype],I,h,w,u.batchSize,E,_),e.disposeData(c.dataId),e.disposeData(m.dataId),d.shape=u.outputShape,d}var hP={kernelName:Ws,backendName:"wasm",setupFunc:bJ,kernelFunc:wJ};var _J=!1,gP=yt(Bi,_J,"bool");var kJ=!1,xP=yt(Po,kJ,"bool");var yP;function vJ(r){yP=r.wasm.cwrap(Mo,null,["number","number","number"])}function CJ(r){let{inputs:{x:e},attrs:{alpha:t},backend:n}=r,o=n.dataIdMap.get(e.dataId).id,s=n.makeOutput(e.shape,e.dtype);if(y.sizeFromShape(e.shape)!==0){let a=n.dataIdMap.get(s.dataId).id;yP(o,t,a)}return s}var bP={kernelName:Mo,backendName:"wasm",setupFunc:vJ,kernelFunc:CJ};var IJ=!1,wP=yt(ji,IJ,"bool");var NJ=!1,_P=yt(Ui,NJ,"bool");var kP=xt(Lo);var SJ=!1,vP=yt(qi,SJ,"bool");var CP;function TJ(r){CP=r.wasm.cwrap(zo,null,["number, number, number"])}function AJ(r){let{backend:e,inputs:t,attrs:n}=r,{reductionIndices:o,keepDims:s}=n,{x:a}=t,l=e.dataIdMap.get(a.dataId).id,u=a,{transposed:c,axes:p,originalAxes:m,inputWasTransposed:f}=tn(a,o,e);if(f){let w=e.dataIdMap.get(c.dataId).id;u=c,l=w}let d=u.shape.length;C.assertAxesAreInnerMostDims("max",p,d);let[h,g]=C.computeOutAndReduceShapes(u.shape,p),x=y.sizeFromShape(g),b=e.makeOutput(h,a.dtype);if(y.sizeFromShape(u.shape)!==0){let w=e.dataIdMap.get(b.dataId).id;CP(l,x,w)}if(f&&e.disposeData(c.dataId),s){let w=C.expandShapeToKeepDim(b.shape,m);b.shape=w}return b}var IP={kernelName:zo,backendName:"wasm",setupFunc:TJ,kernelFunc:AJ};var EJ=!1,NP=yt(Bo,EJ);var SP;function DJ(r){SP=r.wasm.cwrap(Vo,null,["number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number","number"])}function $J(r){let{inputs:e,attrs:t,backend:n}=r,o=e.x,s=n.dataIdMap.get(o.dataId).id,{filterSize:a,strides:i,pad:l,dimRoundingMode:u}=t,c=C.computePool2DInfo(o.shape,a,i,1,l,u),p=c.filterHeight,m=c.filterWidth,f=c.padInfo.top,d=c.padInfo.right,h=c.padInfo.bottom,g=c.padInfo.left,x=c.dilationHeight,b=c.dilationWidth,w=c.strideHeight,_=c.strideWidth,I=c.inChannels,E=c.outChannels;if(c.dataFormat!=="channelsLast")throw new Error(`wasm backend does not support dataFormat:'${c.dataFormat}'. Please use 'channelsLast'.`);let D=n.makeOutput(c.outShape,"float32"),$=n.dataIdMap.get(D.dataId).id;return SP(s,o.shape[0],o.shape[1],o.shape[2],p,m,f,d,h,g,x,b,w,_,I,E,$),D}var TP={kernelName:Vo,backendName:"wasm",setupFunc:DJ,kernelFunc:$J};var AP;function RJ(r){AP=r.wasm.cwrap(Go,null,["number, number, number"])}function FJ(r){let{backend:e,inputs:t,attrs:n}=r,{axis:o,keepDims:s}=n,{x:a}=t,i=e.dataIdMap.get(a.dataId).id,l=i,u=a,{transposed:c,axes:p,originalAxes:m,inputWasTransposed:f}=tn(a,o,e),d=p;if(f){let _=e.dataIdMap.get(c.dataId).id;_!==i&&(u=c,l=_,d=C.getInnerMostAxes(d.length,u.shape.length))}C.assertAxesAreInnerMostDims("mean",d,u.shape.length);let[h,g]=C.computeOutAndReduceShapes(u.shape,d),x=y.sizeFromShape(g),b=u;u.dtype!=="float32"&&(b=Uu({backend:e,inputs:{x:u},attrs:{dtype:"float32"}}),l=e.dataIdMap.get(b.dataId).id);let w=e.makeOutput(h,"float32");if(y.sizeFromShape(u.shape)!==0){let _=e.dataIdMap.get(w.dataId).id;AP(l,x,_)}if(f&&e.disposeData(c.dataId),s){let _=C.expandShapeToKeepDim(w.shape,m);w.shape=_}return u.dtype!=="float32"&&e.disposeData(b.dataId),w}var EP={kernelName:Go,backendName:"wasm",setupFunc:RJ,kernelFunc:FJ};var DP;function OJ(r){DP=r.wasm.cwrap(Wo,null,["number, number, number"])}function PJ(r){let{backend:e,inputs:t,attrs:n}=r,{axis:o,keepDims:s}=n,{x:a}=t,i=e.dataIdMap.get(a.dataId).id,l=i,u=a,{transposed:c,axes:p,originalAxes:m,inputWasTransposed:f}=tn(a,o,e);if(f){let w=e.dataIdMap.get(c.dataId).id;w!==i&&(u=c,l=w)}let d=u.shape.length;C.assertAxesAreInnerMostDims("min",p,d);let[h,g]=C.computeOutAndReduceShapes(u.shape,p),x=y.sizeFromShape(g),b=e.makeOutput(h,u.dtype);if(y.sizeFromShape(u.shape)!==0){let w=e.dataIdMap.get(b.dataId).id;DP(l,x,w)}if(f&&e.disposeData(c.dataId),s){let w=C.expandShapeToKeepDim(b.shape,m);b.shape=w}return b}var $P={kernelName:Wo,backendName:"wasm",setupFunc:OJ,kernelFunc:PJ};var MJ=!1,RP=yt(jo,MJ);var FC;(function(r){r[r.reflect=0]="reflect",r[r.symmetric=1]="symmetric"})(FC||(FC={}));var FP;function LJ(r){FP=r.wasm.cwrap(Uo,null,["number","array","number","number","array","array","number","number"])}function zJ(r){let{inputs:{x:e},backend:t,attrs:{paddings:n,mode:o}}=r,s=n.map((d,h)=>d[0]+e.shape[h]+d[1]),a=t.dataIdMap.get(e.dataId).id,i=t.makeOutput(s,e.dtype),l=t.dataIdMap.get(i.dataId).id,u=new Uint8Array(new Int32Array(e.shape).buffer),c=n.map(d=>d[0]),p=n.map(d=>d[1]),m=new Uint8Array(new Int32Array(c).buffer),f=new Uint8Array(new Int32Array(p).buffer);return FP(a,u,e.shape.length,Ft[e.dtype],m,f,FC[o],l),i}var OP={kernelName:Uo,backendName:"wasm",kernelFunc:zJ,setupFunc:LJ};var BJ=!0,PP=yt(Ho,BJ);var MP=xt(js);function Op(r,e){let t=new Int32Array(r.wasm.HEAPU8.buffer,e,4),n=t[0],o=t[1],s=t[2],a=t[3];return r.wasm._free(e),{pSelectedIndices:n,selectedSize:o,pSelectedScores:s,pValidOutputs:a}}var LP;function VJ(r){LP=r.wasm.cwrap(Yi,"number",["number","number","number","number","number"])}function GJ(r){let{backend:e,inputs:t,attrs:n}=r,{iouThreshold:o,maxOutputSize:s,scoreThreshold:a}=n,{boxes:i,scores:l}=t,u=e.dataIdMap.get(i.dataId).id,c=e.dataIdMap.get(l.dataId).id,p=LP(u,c,s,o,a),{pSelectedIndices:m,selectedSize:f,pSelectedScores:d,pValidOutputs:h}=Op(e,p);return e.wasm._free(d),e.wasm._free(h),e.makeOutput([f],"int32",m)}var zP={kernelName:Yi,backendName:"wasm",setupFunc:VJ,kernelFunc:GJ};var BP;function WJ(r){BP=r.wasm.cwrap(Zi,"number",["number","number","number","number","number","bool"])}function jJ(r){let{backend:e,inputs:t,attrs:n}=r,{iouThreshold:o,maxOutputSize:s,scoreThreshold:a,padToMaxOutputSize:i}=n,{boxes:l,scores:u}=t,c=e.dataIdMap.get(l.dataId).id,p=e.dataIdMap.get(u.dataId).id,m=BP(c,p,s,o,a,i),{pSelectedIndices:f,selectedSize:d,pSelectedScores:h,pValidOutputs:g}=Op(e,m);e.wasm._free(h);let x=e.makeOutput([d],"int32",f),b=e.makeOutput([],"int32",g);return[x,b]}var VP={kernelName:Zi,backendName:"wasm",setupFunc:WJ,kernelFunc:jJ};var GP;function UJ(r){GP=r.wasm.cwrap(Ji,"number",["number","number","number","number","number","number"])}function HJ(r){let{backend:e,inputs:t,attrs:n}=r,{iouThreshold:o,maxOutputSize:s,scoreThreshold:a,softNmsSigma:i}=n,{boxes:l,scores:u}=t,c=e.dataIdMap.get(l.dataId).id,p=e.dataIdMap.get(u.dataId).id,m=GP(c,p,s,o,a,i),{pSelectedIndices:f,selectedSize:d,pSelectedScores:h,pValidOutputs:g}=Op(e,m);e.wasm._free(g);let x=e.makeOutput([d],"int32",f),b=e.makeOutput([d],"float32",h);return[x,b]}var WP={kernelName:Ji,backendName:"wasm",setupFunc:UJ,kernelFunc:HJ};var qJ=!1,jP=yt(Xi,qJ,"bool");var UP;function KJ(r){UP=r.wasm.cwrap(qo,null,["number","number","number","number","number"])}function XJ(r){let{inputs:e,backend:t,attrs:n}=r,{indices:o}=e,{depth:s,onValue:a,offValue:i}=n,l=t.makeOutput([...o.shape,s],"int32"),u=t.dataIdMap.get(l.dataId).id,p=t.dataIdMap.get(o.dataId).id;return UP(p,s,a,i,u),l}var HP={kernelName:qo,backendName:"wasm",setupFunc:KJ,kernelFunc:XJ};function YJ(r){let{inputs:{x:e},backend:t}=r,n=t.makeOutput(e.shape,e.dtype);return t.typedArrayFromHeap(n).fill(1),n}var qP={kernelName:Us,backendName:"wasm",kernelFunc:YJ};function ZJ(r){let{inputs:e,backend:t,attrs:n}=r,{axis:o}=n;if(e.length===1)return Tx({inputs:{input:e[0]},backend:t,attrs:{dim:o}});let s=e[0].shape,a=e[0].dtype;e.forEach(c=>{y.assertShapesMatch(s,c.shape,"All tensors passed to stack must have matching shapes"),y.assert(a===c.dtype,()=>"All tensors passed to stack must have matching dtypes")});let i=[],l=e.map(c=>{let p=Tx({inputs:{input:c},backend:t,attrs:{dim:o}});return i.push(p),p}),u=$C({inputs:l,backend:t,attrs:{axis:o}});return i.forEach(c=>t.disposeData(c.dataId)),u}var KP={kernelName:Hs,backendName:"wasm",kernelFunc:ZJ};var XP;function JJ(r){XP=r.wasm.cwrap(Ko,null,["number","array","number","number","array","array","number","number"])}function QJ(r){let{inputs:{x:e},backend:t,attrs:{paddings:n,constantValue:o}}=r,s=n.map((d,h)=>d[0]+e.shape[h]+d[1]),a=t.dataIdMap.get(e.dataId).id,i=t.makeOutput(s,e.dtype),l=t.dataIdMap.get(i.dataId).id,u=new Uint8Array(new Int32Array(e.shape).buffer),c=n.map(d=>d[0]),p=n.map(d=>d[1]),m=new Uint8Array(new Int32Array(c).buffer),f=new Uint8Array(new Int32Array(p).buffer);return XP(a,u,e.shape.length,Ft[e.dtype],m,f,o,l),i}var YP={kernelName:Ko,backendName:"wasm",kernelFunc:QJ,setupFunc:JJ};var eQ=!1,ZP=yt(Xo,eQ);var JP;function tQ(r){JP=r.wasm.cwrap(Yo,null,["number","number","number"])}function rQ(r){let{inputs:e,backend:t}=r,{x:n,alpha:o}=e,s=t.dataIdMap.get(n.dataId).id,a=t.dataIdMap.get(o.dataId).id,i=t.makeOutput(n.shape,"float32"),l=t.dataIdMap.get(i.dataId).id;return JP(s,a,l),i}var QP={kernelName:Yo,backendName:"wasm",setupFunc:tQ,kernelFunc:rQ};var eM;function nQ(r){eM=r.wasm.cwrap(Qi,null,["number","number","number","number"])}function oQ(r){let{backend:e,inputs:t,attrs:n}=r,{axis:o,keepDims:s}=n,{x:a}=t,i=e.dataIdMap.get(a.dataId).id,l=i,u=a,{transposed:c,axes:p,originalAxes:m,inputWasTransposed:f}=tn(a,o,e),d=p;if(f){let w=e.dataIdMap.get(c.dataId).id;w!==i&&(u=c,l=w,d=C.getInnerMostAxes(d.length,u.shape.length))}C.assertAxesAreInnerMostDims("prod",d,u.shape.length);let[h,g]=C.computeOutAndReduceShapes(u.shape,d),x=y.sizeFromShape(g),b=e.makeOutput(h,u.dtype);if(y.sizeFromShape(u.shape)!==0){let w=e.dataIdMap.get(b.dataId).id;eM(l,x,Ft[b.dtype],w)}if(f&&e.disposeData(c.dataId),s){let w=C.expandShapeToKeepDim(b.shape,m);b.shape=w}return b}var tM={kernelName:Qi,backendName:"wasm",setupFunc:nQ,kernelFunc:oQ};var sQ=r=>{let{backend:e,attrs:t}=r,{start:n,stop:o,step:s,dtype:a}=t,i=$u(n,o,s,a),l=e.makeOutput([i.length],a);return e.typedArrayFromHeap(l).set(i),l},rM={kernelName:sl,backendName:"wasm",kernelFunc:sQ};var iQ=!0,nM=yt(Do,iQ);var oM=xt(Zo);var sM=xt(Qo);var iM;function aQ(r){iM=r.wasm.cwrap(Jo,null,["number","number","number","number","number","number","number","number","number","number"])}function lQ(r){let{backend:e,inputs:t,attrs:n}=r,{images:o}=t,{alignCorners:s,halfPixelCenters:a,size:i}=n,[l,u]=i,[c,p,m,f]=o.shape,d=[c,l,u,f],h=e.dataIdMap.get(o.dataId),g;h.dtype!=="float32"&&(g=Uu({backend:e,inputs:{x:o},attrs:{dtype:"float32"}}),h=e.dataIdMap.get(g.dataId));let x=h.id,b=e.makeOutput(d,"float32");if(y.sizeFromShape(o.shape)===0)return b;let w=e.dataIdMap.get(b.dataId).id;return iM(x,c,p,m,f,l,u,s?1:0,a?1:0,w),g!=null&&e.disposeData(g.dataId),b}var aM={kernelName:Jo,backendName:"wasm",setupFunc:aQ,kernelFunc:lQ};var lM;function uQ(r){lM=r.wasm.cwrap(es,null,["number","array","number","array","number","number"])}function cQ(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,{dims:s}=n,a=y.parseAxisParam(s,o.shape);if(o.shape.length===0)return ju({inputs:{x:o},backend:t});let i=t.makeOutput(o.shape,o.dtype),l=t.dataIdMap.get(o.dataId).id,u=t.dataIdMap.get(i.dataId).id,c=new Uint8Array(new Int32Array(a).buffer),p=new Uint8Array(new Int32Array(o.shape).buffer);lM(l,c,a.length,p,o.shape.length,u);let m=Vr({inputs:{x:i},attrs:{shape:o.shape},backend:t});return t.disposeData(i.dataId),m}var uM={kernelName:es,backendName:"wasm",kernelFunc:cQ,setupFunc:uQ};var cM;function pQ(r){cM=r.wasm.cwrap(ua,null,["number","number","number","number","number","number","number","number","array","number","number"])}function mQ(r){let{inputs:e,backend:t,attrs:n}=r,{image:o}=e,{radians:s,fillValue:a,center:i}=n,l=t.makeOutput(o.shape,o.dtype),u=t.dataIdMap.get(o.dataId).id,c=t.dataIdMap.get(l.dataId).id,[p,m,f,d]=o.shape,[h,g]=C.getImageCenter(i,m,f),x=a===0,b=255,w=typeof a=="number"?[a,a,a,x?0:b]:[...a,b],_=new Uint8Array(new Int32Array(w).buffer);return cM(u,p,m,f,d,s,h,g,_,w.length,c),l}var pM={kernelName:ua,backendName:"wasm",kernelFunc:mQ,setupFunc:pQ};var mM=xt(ts);var fM=xt(rs);var dM;function fQ(r){dM=r.wasm.cwrap(ta,null,["number","number","number","number","number","number","array","number","number"])}function dQ(r){let{backend:e,inputs:t,attrs:n}=r,{indices:o,updates:s}=t,{shape:a}=n,i=e.makeOutput(a,s.dtype);if(y.sizeFromShape(a)===0)return i;let{sliceRank:l,numUpdates:u,sliceSize:c,strides:p,outputSize:m}=fg.calculateShapes(s,o,a),d=e.dataIdMap.get(o.dataId).id,g=e.dataIdMap.get(s.dataId).id,x=new Uint8Array(new Int32Array(p).buffer),b=e.dataIdMap.get(i.dataId).id;return dM(d,g,Ft[s.dtype],l,u,c,x,m,b),i}var hM={kernelName:ta,backendName:"wasm",setupFunc:fQ,kernelFunc:dQ};var gM;function hQ(r){gM=r.wasm.cwrap("SelectV2",null,["number","number","number","number","number"])}function gQ(r){let{inputs:e,backend:t}=r,{condition:n,t:o,e:s}=e,a=t.dataIdMap.get(n.dataId).id,i=t.dataIdMap.get(o.dataId).id,l=t.dataIdMap.get(s.dataId).id,u=t.makeOutput(o.shape,o.dtype),c=t.dataIdMap.get(u.dataId).id,p=n.shape.length,m=o.shape.length,f=p===0||p>1||m===1?1:y.sizeFromShape(o.shape.slice(1));return gM(a,i,l,f,c),u}var xM={kernelName:Ks,backendName:"wasm",kernelFunc:gQ,setupFunc:hQ};var yM;function xQ(r){yM=r.wasm.cwrap(os,null,["number","number"])}function yQ(r){let{backend:e,inputs:{x:t}}=r,n=e.dataIdMap.get(t.dataId).id,o=e.makeOutput(t.shape,t.dtype),s=e.dataIdMap.get(o.dataId).id;return y.sizeFromShape(o.shape)===0||yM(n,s),o}var bM={kernelName:"Sigmoid",backendName:"wasm",setupFunc:xQ,kernelFunc:yQ};var wM=xt(ns);function Hu(r){let{inputs:{x:e},attrs:{begin:t,size:n},backend:o}=r,[s,a]=or.parseSliceParams(e,t,n),i=or.isSliceContinous(e.shape,s,a),l=o.readSync(e.dataId),u=o.makeOutput(a,e.dtype),c=y.computeStrides(e.shape),p=o.dataIdMap.get(u.dataId);if(i){let d=or.computeFlatOffset(s,c);return e.dtype==="string"?p.stringBytes=l.slice(d,d+y.sizeFromShape(a)):o.typedArrayFromHeap(u).set(l.subarray(d,d+y.sizeFromShape(a))),u}if(e.dtype==="string"){let d=Ru(l,s,a,e.shape,e.dtype);return p.stringBytes=d,u}let m=o.typedArrayFromHeap(u),f=e.shape.length;if(f===2)bQ(l,c[0],m,s,a);else if(f===3)wQ(l,c[0],c[1],m,s,a);else if(f===4)_Q(l,c[0],c[1],c[2],m,s,a);else{let d=Ru(l,s,a,e.shape,e.dtype);m.set(d)}return u}function bQ(r,e,t,n,o){let s=0,a=n[0],i=n[1],l=a+o[0];for(let u=a;u<l;u++){let c=u*e+i;t.set(r.subarray(c,c+o[1]),s),s+=o[1]}}function wQ(r,e,t,n,o,s){let a=0,i=o[0],l=o[1],u=o[2],c=i+s[0],p=l+s[1];for(let m=i;m<c;m++)for(let f=l;f<p;f++){let d=m*e+f*t+u;n.set(r.subarray(d,d+s[2]),a),a+=s[2]}}function _Q(r,e,t,n,o,s,a){let i=0,l=s[0],u=s[1],c=s[2],p=l+a[0],m=u+a[1],f=c+a[2],d=s[3];for(let h=l;h<p;h++)for(let g=u;g<m;g++)for(let x=c;x<f;x++){let b=h*e+g*t+x*n+d;o.set(r.subarray(b,b+a[3]),i),i+=a[3]}}var _M={kernelName:Xs,backendName:"wasm",kernelFunc:Hu};var kM;function kQ(r){kM=r.wasm.cwrap(as,null,["number","number","number","number"])}function vQ(r){let{backend:e,inputs:{logits:t},attrs:{dim:n}}=r,o=e.dataIdMap.get(t.dataId).id,s=e.makeOutput(t.shape,t.dtype),a=e.dataIdMap.get(s.dataId).id,i=t.shape[n],l=y.sizeFromShape(t.shape)/i;return y.sizeFromShape(s.shape)===0||kM(o,a,i,l),s}var vM={kernelName:as,backendName:"wasm",setupFunc:kQ,kernelFunc:vQ};function CQ(r){let{inputs:e,attrs:t,backend:n}=r,{x:o}=e,{numOrSizeSplits:s,axis:a}=t,i=y.parseAxisParam(a,o.shape)[0],l=C.prepareSplitSize(o,s,i),u=new Array(o.shape.length).fill(0),c=o.shape.slice();return l.map(p=>{let m=[...c];m[i]=p;let f=Hu({inputs:{x:o},attrs:{begin:u,size:m},backend:n});return u[i]+=p,f})}var CM={kernelName:Ys,backendName:"wasm",kernelFunc:CQ};var IM=xt(ss);var NM=xt(ll);var IQ=!0,SM=yt(ls,IQ);var TM;function NQ(r){TM=r.wasm.cwrap(Yn,null,["number","number","number"])}function SQ(r){let{backend:e,inputs:t,attrs:n}=r,{alpha:o}=n,{x:s}=t,a=e.dataIdMap.get(s.dataId).id,i=e.makeOutput(s.shape,s.dtype),l=e.dataIdMap.get(i.dataId).id;return TM(a,o,l),i}var AM={kernelName:Yn,backendName:"wasm",setupFunc:NQ,kernelFunc:SQ};var EM;function TQ(r){EM=r.wasm.cwrap(ia,null,["number","array","number","array","array","array","array","array","number","number"])}function AQ(r){let{backend:e,inputs:t,attrs:n}=r,{x:o}=t,{begin:s,end:a,strides:i}=n;i==null&&(i=new Array(s.length));let{beginMask:l,endMask:u,ellipsisMask:c,newAxisMask:p,shrinkAxisMask:m}=n,f=C.slice_util.maskToAxes(c);if(f.length>1)throw new Error("Multiple ellipses in slice is not allowed.");if(c!==0&&p!==0)throw new Error("Using both ellipsisMask and newAxisMask is not yet supported.");if(c!==0&&m!==0)throw new Error("Using both ellipsisMask and shrinkAxisMask is not yet supported.");let d=o.shape.length-s.length,h=C.slice_util.maskToAxes(p),g=o.shape.slice();h.forEach(G=>{s[G]=0,a[G]=1,g.splice(G,0,1)});let x=Vr({inputs:{x:o},attrs:{shape:g},backend:e}),{begin:b,end:w,strides:_}=C.slice_util.getNormalizedAxes(x.shape,f,d,s,a,i,l,u,c);s=b,a=w,i=_;let I=C.slice_util.maskToAxes(m);I.forEach(G=>{a[G]=s[G]+1,i[G]=1});let E=C.slice_util.computeOutShape(s,a,i),D=E.filter((G,j)=>I.indexOf(j)===-1);if(i.every(G=>G===1)){let G=Hu({inputs:{x},attrs:{begin:s,size:E},backend:e});e.disposeData(x.dataId);let j=Vr({inputs:{x:G},attrs:{shape:D},backend:e});return e.disposeData(G.dataId),j}let R=e.makeOutput(D,"float32");if(!D.some(G=>G===0)){let G=e.dataIdMap.get(x.dataId).id,j=new Uint8Array(new Int32Array(y.computeStrides(x.shape)).buffer),U=new Uint8Array(new Int32Array(s).buffer),H=new Uint8Array(new Int32Array(a).buffer),q=new Uint8Array(new Int32Array(i).buffer),X=new Uint8Array(new Int32Array(D).buffer),ne=new Uint8Array(new Int32Array(y.computeStrides(D)).buffer),Y=e.dataIdMap.get(R.dataId).id;EM(G,j,x.shape.length,U,H,q,X,ne,D.length,Y)}e.disposeData(x.dataId);let M=Vr({inputs:{x:R},attrs:{shape:D},backend:e});return e.disposeData(R.dataId),M}var DM={kernelName:ia,backendName:"wasm",setupFunc:TQ,kernelFunc:AQ};var EQ=!0,$M=yt(us,EQ);var RM;function DQ(r){RM=r.wasm.cwrap(is,null,["number, number, number"])}function $Q(r){let{backend:e,inputs:t,attrs:n}=r,{axis:o,keepDims:s}=n,{x:a}=t,i=e.dataIdMap.get(a.dataId).id,l=i,u=a,{transposed:c,axes:p,originalAxes:m,inputWasTransposed:f}=tn(a,o,e),d=p;if(f){let w=e.dataIdMap.get(c.dataId).id;w!==i&&(u=c,l=w,d=C.getInnerMostAxes(d.length,u.shape.length))}C.assertAxesAreInnerMostDims("sum",d,u.shape.length);let[h,g]=C.computeOutAndReduceShapes(u.shape,d),x=y.sizeFromShape(g),b=e.makeOutput(h,u.dtype);if(y.sizeFromShape(u.shape)!==0){let w=e.dataIdMap.get(b.dataId).id;RM(l,x,w)}if(f&&e.disposeData(c.dataId),s){let w=C.expandShapeToKeepDim(b.shape,m);b.shape=w}return b}var FM={kernelName:is,backendName:"wasm",setupFunc:DQ,kernelFunc:$Q};var OM=xt(cs);var PM=xt(ps);var MM;function RQ(r){MM=r.wasm.cwrap(Mn,null,["number","array","number","array","number","number"])}function FQ(r){let{inputs:e,backend:t,attrs:n}=r,{x:o}=e,s=t.dataIdMap.get(o.dataId).id,{reps:a}=n,i=new Array(o.shape.length);for(let m=0;m<i.length;m++)i[m]=o.shape[m]*a[m];let l=new Uint8Array(new Int32Array(o.shape).buffer),u=new Uint8Array(new Int32Array(i).buffer),c=t.makeOutput(i,o.dtype),p=t.dataIdMap.get(c.dataId).id;return MM(s,l,o.shape.length,u,i.length,Ft[c.dtype],p),c}var LM={kernelName:Mn,backendName:"wasm",setupFunc:RQ,kernelFunc:FQ};var zM;function OQ(r){zM=r.wasm.cwrap(aa,null,["number","array","number","number","number","bool","number","number"])}var PQ=({inputs:r,backend:e,attrs:t})=>{let{x:n}=r,{k:o,sorted:s}=t,a=e.dataIdMap.get(n.dataId).id,i=new Uint8Array(new Int32Array(n.shape).buffer),l=n.shape.slice();l[l.length-1]=o;let u=e.makeOutput(l,n.dtype),c=e.dataIdMap.get(u.dataId).id,p=e.makeOutput(l,"int32"),m=e.dataIdMap.get(p.dataId).id;return zM(a,i,n.shape.length,Ft[n.dtype],o,s,c,m),[u,p]},BM={kernelName:aa,backendName:"wasm",setupFunc:OQ,kernelFunc:PQ};var VM;function MQ(r){VM=r.wasm.cwrap(la,null,["number","number","bool","number","number","number","number","number","number","array","number","number","number","number","number"])}function LQ(r){let{backend:e,inputs:t,attrs:n}=r,{image:o,transforms:s}=t,{interpolation:a,fillMode:i,fillValue:l,outputShape:u}=n,[c,p,m,f]=o.shape,[d,h]=u!=null?u:[p,m],g=[c,d,h,f],x=new Uint8Array(new Int32Array(y.computeStrides(o.shape)).buffer),b=e.makeOutput(g,o.dtype),w=e.dataIdMap.get(b.dataId).id,I=e.dataIdMap.get(o.dataId).id,D=e.dataIdMap.get(s.dataId).id,$=a==="nearest"?1:2,R;switch(i){case"constant":R=1;break;case"reflect":R=2;break;case"wrap":R=3;break;case"nearest":R=4;break;default:R=1;break}return VM(I,D,s.shape[0]>1,c,d,h,f,m,p,x,o.shape.length-1,$,R,l,w),b}var GM={kernelName:la,backendName:"wasm",setupFunc:MQ,kernelFunc:LQ};function zQ(r){let{inputs:e,backend:t,attrs:n}=r,{value:o}=e,{axis:s}=n;s<0&&(s+=o.shape.length);let a=o.shape[s],i=o.shape.length,l=new Array(i-1),u=0;for(let f=0;f<i;f++)f!==s&&(l[u++]=o.shape[f]);let c=new Array(a),p=new Array(i).fill(0),m=o.shape.slice();m[s]=1;for(let f=0;f<c.length;f++)p[s]=f,c[f]=Hu({inputs:{x:o},attrs:{begin:p,size:m},backend:t});return c.map(({dataId:f,dtype:d})=>({dataId:f,dtype:d,shape:l}))}var WM={kernelName:Zs,backendName:"wasm",kernelFunc:zQ};function BQ(r){let{inputs:{x:e},backend:t}=r,n=t.makeOutput(e.shape,e.dtype);return t.typedArrayFromHeap(n).fill(0),n}var jM={kernelName:Js,backendName:"wasm",kernelFunc:BQ};var VQ=[gO,xO,bO,CO,NO,TO,EO,RO,FO,OO,MO,LO,BO,GO,WO,UO,qO,XO,ZO,JO,QO,eP,tP,nP,oP,sP,hO,aP,uP,pP,fP,hP,gP,xP,wO,bP,wP,_P,kP,vP,IP,NP,TP,EP,$P,RP,OP,PP,MP,zP,VP,WP,jP,HP,qP,KP,YP,ZP,QP,tM,rM,nM,oM,sM,DO,aM,uM,pM,fM,mM,hM,xM,bM,wM,_M,vM,CM,IM,NM,SM,AM,DM,$M,FM,OM,PM,LM,BM,GM,kO,WM,jM];for(let r of VQ)Kl(r);var OC=W();OC.registerFlag("WASM_HAS_SIMD_SUPPORT",async()=>WebAssembly.validate(new Uint8Array([0,97,115,109,1,0,0,0,1,4,1,96,0,0,3,2,1,0,10,9,1,7,0,65,0,253,15,26,11])));OC.registerFlag("WASM_HAS_MULTITHREAD_SUPPORT",async()=>{if(OC.get("IS_NODE"))return!1;try{return new MessageChannel().port1.postMessage(new SharedArrayBuffer(1)),WebAssembly.validate(new Uint8Array([0,97,115,109,1,0,0,0,1,4,1,96,0,0,3,2,1,0,5,4,1,3,1,1,10,11,1,9,0,65,0,254,16,2,0,26,11]))}catch(r){return!1}});var BC=hc(qM());var KM='var Module={};function threadPrintErr(){var text=Array.prototype.slice.call(arguments).join(" ");console.error(text)}function threadAlert(){var text=Array.prototype.slice.call(arguments).join(" ");postMessage({cmd:"alert",text:text,threadId:Module["_pthread_self"]()})}var err=threadPrintErr;this.alert=threadAlert;Module["instantiateWasm"]=function(info,receiveInstance){var instance=new WebAssembly.Instance(Module["wasmModule"],info);Module["wasmModule"]=null;receiveInstance(instance);return instance.exports};function moduleLoaded(){}this.onmessage=function(e){try{if(e.data.cmd==="load"){Module["wasmModule"]=e.data.wasmModule;Module["wasmMemory"]=e.data.wasmMemory;Module["buffer"]=Module["wasmMemory"].buffer;Module["ENVIRONMENT_IS_PTHREAD"]=true;if(typeof e.data.urlOrBlob==="string"){importScripts(e.data.urlOrBlob)}else{var objectUrl=URL.createObjectURL(e.data.urlOrBlob);importScripts(objectUrl);URL.revokeObjectURL(objectUrl)}WasmBackendModuleThreadedSimd(Module).then(function(instance){Module=instance;moduleLoaded()})}else if(e.data.cmd==="objectTransfer"){Module["PThread"].receiveObjectTransfer(e.data)}else if(e.data.cmd==="run"){Module["__performance_now_clock_drift"]=performance.now()-e.data.time;Module["__emscripten_thread_init"](e.data.threadInfoStruct,0,0);var max=e.data.stackBase;var top=e.data.stackBase+e.data.stackSize;Module["establishStackSpace"](top,max);Module["_emscripten_tls_init"]();Module["PThread"].receiveObjectTransfer(e.data);Module["PThread"].setThreadStatus(Module["_pthread_self"](),1);try{var result=Module["invokeEntryPoint"](e.data.start_routine,e.data.arg);if(!Module["getNoExitRuntime"]())Module["PThread"].threadExit(result)}catch(ex){if(ex==="Canceled!"){Module["PThread"].threadCancel()}else if(ex!="unwind"){if(ex instanceof Module["ExitStatus"]){if(Module["getNoExitRuntime"]()){}else{Module["PThread"].threadExit(ex.status)}}else{Module["PThread"].threadExit(-2);throw ex}}}}else if(e.data.cmd==="cancel"){if(Module["_pthread_self"]()){Module["PThread"].threadCancel()}}else if(e.data.target==="setimmediate"){}else if(e.data.cmd==="processThreadQueue"){if(Module["_pthread_self"]()){Module["_emscripten_current_thread_process_queued_calls"]()}}else{err("worker.js received unknown command "+e.data.cmd);err(e.data)}}catch(ex){err("worker.js onmessage() captured an uncaught exception: "+ex);if(ex&&ex.stack)err(ex.stack);throw ex}};if(typeof process==="object"&&typeof process.versions==="object"&&typeof process.versions.node==="string"){self={location:{href:__filename}};var onmessage=this.onmessage;var nodeWorkerThreads=require("worker_threads");global.Worker=nodeWorkerThreads.Worker;var parentPort=nodeWorkerThreads.parentPort;parentPort.on("message",function(data){onmessage({data:data})});var nodeFS=require("fs");var nodeRead=function(filename){return nodeFS.readFileSync(filename,"utf8")};function globalEval(x){global.require=require;global.Module=Module;eval.call(null,x)}importScripts=function(f){globalEval(nodeRead(f))};postMessage=function(msg){parentPort.postMessage(msg)};if(typeof performance==="undefined"){performance={now:function(){return Date.now()}}}}';var YM=hc(XM());var Dx=class extends Ms{constructor(e){super();this.wasm=e,this.dataIdNextNumber=1,this.wasm.tfjs.init(),this.dataIdMap=new Xa(this,hs())}write(e,t,n){let o={id:this.dataIdNextNumber++};return this.move(o,e,t,n,1),o}numDataIds(){return this.dataIdMap.numDataIds()}async time(e){let t=y.now();return e(),{kernelMs:y.now()-t}}move(e,t,n,o,s){let a=this.dataIdNextNumber++;if(o==="string"){let c=t;this.dataIdMap.set(e,{id:a,stringBytes:c,shape:n,dtype:o,memoryOffset:null,refCount:s});return}let i=y.sizeFromShape(n),l=i*y.bytesPerElement(o),u=this.wasm._malloc(l);this.dataIdMap.set(e,{id:a,memoryOffset:u,shape:n,dtype:o,refCount:s}),this.wasm.tfjs.registerTensor(a,i,u),t!=null&&this.wasm.HEAPU8.set(new Uint8Array(t.buffer,t.byteOffset,l),u)}async read(e){return this.readSync(e)}readSync(e){let{memoryOffset:t,dtype:n,shape:o,stringBytes:s}=this.dataIdMap.get(e);if(n==="string")return s;let a=this.wasm.HEAPU8.slice(t,t+y.sizeFromShape(o)*y.bytesPerElement(n));return WQ(a.buffer,n)}disposeData(e,t=!1){if(this.dataIdMap.has(e)){let n=this.dataIdMap.get(e);if(n.refCount--,!t&&n.refCount>0)return!1;this.wasm._free(n.memoryOffset),this.wasm.tfjs.disposeData(n.id),this.dataIdMap.delete(e)}return!0}refCount(e){return this.dataIdMap.has(e)?this.dataIdMap.get(e).refCount:0}incRef(e){let t=this.dataIdMap.get(e);t!=null&&t.refCount++}floatPrecision(){return 32}getMemoryOffset(e){return this.dataIdMap.get(e).memoryOffset}dispose(){this.wasm.tfjs.dispose(),"PThread"in this.wasm&&this.wasm.PThread.terminateAllThreads(),this.wasm=null}memory(){return{unreliable:!1}}makeOutput(e,t,n){let o;if(n==null)o=this.write(null,e,t);else{let s=this.dataIdNextNumber++;o={id:s},this.dataIdMap.set(o,{id:s,memoryOffset:n,shape:e,dtype:t,refCount:1});let a=y.sizeFromShape(e);this.wasm.tfjs.registerTensor(s,a,n)}return{dataId:o,shape:e,dtype:t}}typedArrayFromHeap({shape:e,dtype:t,dataId:n}){let o=this.wasm.HEAPU8.buffer,{memoryOffset:s}=this.dataIdMap.get(n),a=y.sizeFromShape(e);switch(t){case"float32":return new Float32Array(o,s,a);case"int32":return new Int32Array(o,s,a);case"bool":return new Uint8Array(o,s,a);default:throw new Error(`Unknown dtype ${t}`)}}};function GQ(r){return(e,t)=>(y.fetch(r,{credentials:"same-origin"}).then(n=>{n.ok||e.env.a(`failed to load wasm binary file at '${r}'`),n.arrayBuffer().then(o=>{WebAssembly.instantiate(o,e).then(s=>{t(s.instance,s.module)})})}),{})}function ZM(r,e,t){if($x!=null)return $x;let n="tfjs-backend-wasm.wasm";return r&&e?n="tfjs-backend-wasm-threaded-simd.wasm":r&&(n="tfjs-backend-wasm-simd.wasm"),qf!=null&&qf[n]!=null?qf[n]:t+n}async function JM(){let[r,e]=await Promise.all([W().getAsync("WASM_HAS_SIMD_SUPPORT"),W().getAsync("WASM_HAS_MULTITHREAD_SUPPORT")]);return new Promise((t,n)=>{let o={};o.locateFile=(i,l)=>{if(i.endsWith(".worker.js")){let u=KM,c=new Blob([u],{type:"application/javascript"});return URL.createObjectURL(c)}return i.endsWith(".wasm")?ZM(r,e,Hf!=null?Hf:l):l+i},VC&&(o.instantiateWasm=GQ(ZM(r,e,Hf!=null?Hf:"")));let s=!1;o.onAbort=()=>{if(s||Kf)return;Kf=!0,n({message:"Make sure the server can serve the `.wasm` file relative to the bundled js file. For more details see https://github.com/tensorflow/tfjs/blob/master/tfjs-backend-wasm/README.md#using-bundlers"})};let a;e&&r&&$x==null?(o.mainScriptUrlOrBlob=new Blob(["var WasmBackendModuleThreadedSimd = "+BC.default.toString()],{type:"text/javascript"}),a=(0,BC.default)(o)):a=(0,YM.default)(o),a.then(i=>{s=!0,Kf=!1;let l=null;i.tfjs={init:i.cwrap("init",null,[]),registerTensor:i.cwrap("register_tensor",null,["number","number","number"]),disposeData:i.cwrap("dispose_data",l,["number"]),dispose:i.cwrap("dispose",l,[])},t({wasm:i})})})}function WQ(r,e){switch(e){case"float32":return new Float32Array(r);case"int32":return new Int32Array(r);case"bool":return new Uint8Array(r);default:throw new Error(`Unknown dtype ${e}`)}}var jQ=["tfjs-backend-wasm.wasm","tfjs-backend-wasm-simd.wasm","tfjs-backend-wasm-threaded-simd.wasm"],$x=null,Hf=null,qf={},Kf=!1,VC=!1;function UQ(r,e=!1){if(Vw("setWasmPath has been deprecated in favor of setWasmPaths and will be removed in a future release."),Kf)throw new Error("The WASM backend was already initialized. Make sure you call `setWasmPath()` before you call `tf.setBackend()` or `tf.ready()`");$x=r,VC=e}function HQ(r,e=!1){if(Kf)throw new Error("The WASM backend was already initialized. Make sure you call `setWasmPaths()` before you call `tf.setBackend()` or `tf.ready()`");if(typeof r=="string")Hf=r;else{qf=r;let t=jQ.filter(n=>qf[n]==null);if(t.length>0)throw new Error(`There were no entries found for the following binaries: ${t.join(",")}. Please either call setWasmPaths with a map providing a path for each binary, or with a string indicating the directory where all the binaries can be found.`)}VC=e}var GC="3.6.0";var qQ=2;np("wasm",async()=>{let{wasm:r}=await JM();return new Dx(r)},qQ);P().prototype.abs=function(){return this.throwIfDisposed(),Tt(this)};P().prototype.acos=function(){return this.throwIfDisposed(),Gm(this)};P().prototype.acosh=function(){return this.throwIfDisposed(),Wm(this)};P().prototype.add=function(r){return this.throwIfDisposed(),J(this,r)};P().prototype.all=function(r,e){return this.throwIfDisposed(),tu(this,r,e)};P().prototype.any=function(r,e){return this.throwIfDisposed(),dl(this,r,e)};P().prototype.argMax=function(r){return this.throwIfDisposed(),hl(this,r)};P().prototype.argMin=function(r){return this.throwIfDisposed(),jm(this,r)};P().prototype.asScalar=function(){return this.throwIfDisposed(),T(this.size===1,()=>"The array must have only 1 element."),L(this,[])};P().prototype.asType=function(r){return this.throwIfDisposed(),oe(this,r)};P().prototype.as1D=function(){return this.throwIfDisposed(),L(this,[this.size])};P().prototype.as2D=function(r,e){return this.throwIfDisposed(),L(this,[r,e])};P().prototype.as3D=function(r,e,t){return this.throwIfDisposed(),L(this,[r,e,t])};P().prototype.as4D=function(r,e,t,n){return this.throwIfDisposed(),L(this,[r,e,t,n])};P().prototype.as5D=function(r,e,t,n,o){return this.throwIfDisposed(),L(this,[r,e,t,n,o])};P().prototype.asin=function(){return this.throwIfDisposed(),Um(this)};P().prototype.asinh=function(){return this.throwIfDisposed(),Hm(this)};P().prototype.atan=function(){return this.throwIfDisposed(),qm(this)};P().prototype.atan2=function(r){return this.throwIfDisposed(),Km(this,r)};P().prototype.atanh=function(){return this.throwIfDisposed(),Xm(this)};P().prototype.avgPool=function(r,e,t,n){return this.throwIfDisposed(),ha(this,r,e,t,n)};P().prototype.batchToSpaceND=function(r,e){return this.throwIfDisposed(),ga(this,r,e)};P().prototype.batchNorm=function(r,e,t,n,o){return this.throwIfDisposed(),Jn(this,r,e,t,n,o)};P().prototype.broadcastTo=function(r){return this.throwIfDisposed(),xa(this,r)};P().prototype.cast=function(r){return this.throwIfDisposed(),oe(this,r)};P().prototype.ceil=function(){return this.throwIfDisposed(),Jm(this)};P().prototype.clipByValue=function(r,e){return this.throwIfDisposed(),ur(this,r,e)};P().prototype.concat=function(r,e){return this.throwIfDisposed(),r instanceof Pe&&(r=[r]),Qe([this,...r],e)};P().prototype.conv1d=function(r,e,t,n,o,s){return this.throwIfDisposed(),ou(this,r,e,t,n,o,s)};P().prototype.conv2dTranspose=function(r,e,t,n,o){return this.throwIfDisposed(),su(this,r,e,t,n,o)};P().prototype.conv2d=function(r,e,t,n,o,s){return this.throwIfDisposed(),Jr(this,r,e,t,n,o,s)};P().prototype.cos=function(){return this.throwIfDisposed(),ya(this)};P().prototype.cosh=function(){return this.throwIfDisposed(),iu(this)};P().prototype.cumsum=function(r,e,t){return this.throwIfDisposed(),au(this,r,e,t)};P().prototype.depthToSpace=function(r,e){return this.throwIfDisposed(),ef(this,r,e)};P().prototype.depthwiseConv2d=function(r,e,t,n,o,s){return this.throwIfDisposed(),xs(this,r,e,t,n,o,s)};P().prototype.dilation2d=function(r,e,t,n,o){return this.throwIfDisposed(),tf(this,r,e,t,n,o)};P().prototype.divNoNan=function(r){return this.throwIfDisposed(),rf(this,r)};P().prototype.div=function(r){return this.throwIfDisposed(),pe(this,r)};P().prototype.dot=function(r){return this.throwIfDisposed(),t_(this,r)};P().prototype.elu=function(){return this.throwIfDisposed(),ys(this)};P().prototype.equal=function(r){return this.throwIfDisposed(),Sn(this,r)};P().prototype.erf=function(){return this.throwIfDisposed(),nf(this)};P().prototype.exp=function(){return this.throwIfDisposed(),Qt(this)};P().prototype.expandDims=function(r){return this.throwIfDisposed(),cr(this,r)};P().prototype.expm1=function(){return this.throwIfDisposed(),of(this)};P().prototype.fft=function(){return this.throwIfDisposed(),Na(this)};P().prototype.flatten=function(){return this.throwIfDisposed(),L(this,[this.size])};P().prototype.floor=function(){return this.throwIfDisposed(),ws(this)};P().prototype.floorDiv=function(r){return this.throwIfDisposed(),eu(this,r)};P().prototype.gather=function(r,e){return this.throwIfDisposed(),Qn(this,r,e)};P().prototype.greaterEqual=function(r){return this.throwIfDisposed(),hn(this,r)};P().prototype.greater=function(r){return this.throwIfDisposed(),Vt(this,r)};P().prototype.ifft=function(){return this.throwIfDisposed(),si(this)};P().prototype.irfft=function(){return this.throwIfDisposed(),_u(this)};P().prototype.isFinite=function(){return this.throwIfDisposed(),n_(this)};P().prototype.isInf=function(){return this.throwIfDisposed(),o_(this)};P().prototype.isNaN=function(){return this.throwIfDisposed(),sf(this)};P().prototype.leakyRelu=function(r){return this.throwIfDisposed(),ba(this,r)};P().prototype.lessEqual=function(r){return this.throwIfDisposed(),gn(this,r)};P().prototype.less=function(r){return this.throwIfDisposed(),uu(this,r)};P().prototype.localResponseNormalization=function(r,e,t,n){return this.throwIfDisposed(),af(this,r,e,t,n)};P().prototype.logSigmoid=function(){return this.throwIfDisposed(),i_(this)};P().prototype.logSoftmax=function(r){return this.throwIfDisposed(),pu(this,r)};P().prototype.logSumExp=function(r,e){return this.throwIfDisposed(),uf(this,r,e)};P().prototype.log=function(){return this.throwIfDisposed(),pr(this)};P().prototype.log1p=function(){return this.throwIfDisposed(),cu(this)};P().prototype.logicalAnd=function(r){return this.throwIfDisposed(),yr(this,r)};P().prototype.logicalNot=function(){return this.throwIfDisposed(),wa(this)};P().prototype.logicalOr=function(r){return this.throwIfDisposed(),mu(this,r)};P().prototype.logicalXor=function(r){return this.throwIfDisposed(),c_(this,r)};P().prototype.matMul=function(r,e,t){return this.throwIfDisposed(),ze(this,r,e,t)};P().prototype.maxPool=function(r,e,t,n){return this.throwIfDisposed(),_a(this,r,e,t,n)};P().prototype.max=function(r,e){return this.throwIfDisposed(),mr(this,r,e)};P().prototype.maximum=function(r){return this.throwIfDisposed(),en(this,r)};P().prototype.mean=function(r,e){return this.throwIfDisposed(),dt(this,r,e)};P().prototype.min=function(r,e){return this.throwIfDisposed(),oi(this,r,e)};P().prototype.minimum=function(r){return this.throwIfDisposed(),_s(this,r)};P().prototype.mirrorPad=function(r,e){return this.throwIfDisposed(),pf(this,r,e)};P().prototype.mod=function(r){return this.throwIfDisposed(),mf(this,r)};P().prototype.mul=function(r){return this.throwIfDisposed(),O(this,r)};P().prototype.neg=function(){return this.throwIfDisposed(),qe(this)};P().prototype.norm=function(r,e,t){return this.throwIfDisposed(),pp(this,r,e,t)};P().prototype.notEqual=function(r){return this.throwIfDisposed(),ro(this,r)};P().prototype.oneHot=function(r,e=1,t=0){return this.throwIfDisposed(),ds(this,r,e,t)};P().prototype.onesLike=function(){return this.throwIfDisposed(),sr(this)};P().prototype.pad=function(r,e){return this.throwIfDisposed(),Lr(this,r,e)};P().prototype.pool=function(r,e,t,n,o){return this.throwIfDisposed(),f_(this,r,e,t,n,o)};P().prototype.pow=function(r){return this.throwIfDisposed(),zr(this,r)};P().prototype.prelu=function(r){return this.throwIfDisposed(),va(this,r)};P().prototype.prod=function(r,e){return this.throwIfDisposed(),fu(this,r,e)};P().prototype.reciprocal=function(){return this.throwIfDisposed(),ff(this)};P().prototype.relu=function(){return this.throwIfDisposed(),$r(this)};P().prototype.relu6=function(){return this.throwIfDisposed(),hu(this)};P().prototype.reshapeAs=function(r){return this.throwIfDisposed(),L(this,r.shape)};P().prototype.reshape=function(r){return this.throwIfDisposed(),L(this,r)};P().prototype.resizeBilinear=function(r,e,t){return this.throwIfDisposed(),Fg(this,r,e,t)};P().prototype.resizeNearestNeighbor=function(r,e,t){return this.throwIfDisposed(),Og(this,r,e,t)};P().prototype.reverse=function(r){return this.throwIfDisposed(),Kt(this,r)};P().prototype.rfft=function(){return this.throwIfDisposed(),Sa(this)};P().prototype.round=function(){return this.throwIfDisposed(),gu(this)};P().prototype.rsqrt=function(){return this.throwIfDisposed(),xu(this)};P().prototype.selu=function(){return this.throwIfDisposed(),yu(this)};P().prototype.separableConv2d=function(r,e,t,n,o,s){return this.throwIfDisposed(),df(this,r,e,t,n,o,s)};P().prototype.sigmoid=function(){return this.throwIfDisposed(),Dr(this)};P().prototype.sign=function(){return this.throwIfDisposed(),hf(this)};P().prototype.sin=function(){return this.throwIfDisposed(),bu(this)};P().prototype.sinh=function(){return this.throwIfDisposed(),wu(this)};P().prototype.slice=function(r,e){return this.throwIfDisposed(),Re(this,r,e)};P().prototype.softmax=function(r){return this.throwIfDisposed(),Ia(this,r)};P().prototype.softplus=function(){return this.throwIfDisposed(),eo(this)};P().prototype.spaceToBatchND=function(r,e){return this.throwIfDisposed(),ka(this,r,e)};P().prototype.split=function(r,e){return this.throwIfDisposed(),tr(this,r,e)};P().prototype.sqrt=function(){return this.throwIfDisposed(),gt(this)};P().prototype.square=function(){return this.throwIfDisposed(),Me(this)};P().prototype.squaredDifference=function(r){return this.throwIfDisposed(),ku(this,r)};P().prototype.squeeze=function(r){return this.throwIfDisposed(),Tn(this,r)};P().prototype.stack=function(r,e){this.throwIfDisposed();let t=r instanceof Pe?[this,r]:[this,...r];return Gt(t,e)};P().prototype.step=function(r){return this.throwIfDisposed(),vs(this,r)};P().prototype.stridedSlice=function(r,e,t,n,o,s,a,i){return this.throwIfDisposed(),yf(this,r,e,t,n,o,s,a,i)};P().prototype.sub=function(r){return this.throwIfDisposed(),le(this,r)};P().prototype.sum=function(r,e){return this.throwIfDisposed(),fe(this,r,e)};P().prototype.tan=function(){return this.throwIfDisposed(),bf(this)};P().prototype.tanh=function(){return this.throwIfDisposed(),gs(this)};P().prototype.tile=function(r){return this.throwIfDisposed(),Bn(this,r)};P().prototype.toBool=function(){return this.throwIfDisposed(),oe(this,"bool")};P().prototype.toFloat=function(){return this.throwIfDisposed(),oe(this,"float32")};P().prototype.toInt=function(){return this.throwIfDisposed(),oe(this,"int32")};P().prototype.topk=function(r,e){return this.throwIfDisposed(),wf(this,r,e)};P().prototype.transpose=function(r){return this.throwIfDisposed(),Ue(this,r)};P().prototype.unique=function(r){return this.throwIfDisposed(),cp(this,r)};P().prototype.unsortedSegmentSum=function(r,e){return this.throwIfDisposed(),_f(this,r,e)};P().prototype.unstack=function(r){return this.throwIfDisposed(),fr(this,r)};P().prototype.where=function(r,e){return this.throwIfDisposed(),vt(r,this,e)};P().prototype.zerosLike=function(){return this.throwIfDisposed(),Ie(this)};var Rx={kernelName:Bs,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(r,vs(oe(t,"float32"),-1))}}};var QM={kernelName:ki,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>{let n=Me(oe(t,"float32")),o=gt(le(ue(1),n));return qe(pe(r,o))}}}};var eL={kernelName:vi,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>{let n=gt(le(Me(oe(t,"float32")),1));return pe(r,n)}}}};var tL={kernelName:Pn,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e,o=Be(t.shape,n.shape);return{a:()=>{let i=r,l=kt(t.shape,o);return l.length>0&&(i=fe(i,l)),L(i,t.shape)},b:()=>{let i=r,l=kt(n.shape,o);return l.length>0&&(i=fe(i,l)),L(i,n.shape)}}}};var rL={kernelName:_o,saveAllInputs:!0,gradFunc:(r,e)=>{let t={};return e.forEach((n,o)=>{t[o]=()=>r.clone()}),t}};var nL={kernelName:ko,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>Ie(t)}}};var oL={kernelName:Ya,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>Ie(t)}}};var sL={kernelName:Ni,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>pe(r,gt(le(ue(1),Me(oe(t,"float32")))))}}};var iL={kernelName:Si,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>{let n=gt(J(ue(1),Me(oe(t,"float32"))));return pe(r,n)}}}};var aL={kernelName:Ei,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e,o=Be(t.shape,n.shape);return{a:()=>{let i=J(Me(t),Me(n)),l=O(r,pe(n,i)),u=kt(t.shape,o);return u.length>0&&(l=fe(l,u)),L(l,t.shape)},b:()=>{let i=J(Me(t),Me(n)),l=qe(O(r,pe(t,i))),u=kt(n.shape,o);return u.length>0&&(l=fe(l,u)),L(l,n.shape)}}}};var lL={kernelName:Ti,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>pe(r,J(Me(oe(t,"float32")),1))}}};var uL={kernelName:Ai,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>pe(r,le(ue(1),Me(oe(t,"float32"))))}}};function KQ(r,e,t,n,o,s){let a=k(r,"dy","avgPool3dGrad"),i=k(e,"input","avgPool3dGrad"),l=a,u=i,c=!1;i.rank===4&&(c=!0,l=L(a,[1,a.shape[0],a.shape[1],a.shape[2],a.shape[3]]),u=L(i,[1,i.shape[0],i.shape[1],i.shape[2],i.shape[3]])),T(l.rank===5,()=>`Error in avgPool3dGrad: dy must be rank 5 but got rank ${l.rank}.`),T(u.rank===5,()=>`Error in avgPool3dGrad: input must be rank 5 but got rank ${u.rank}.`),s!=null&&T(ot(o),()=>`Error in avgPool3dGrad: pad must be an integer when using, dimRoundingMode ${s} but got pad ${o}.`);let p={dy:l,input:u},m={filterSize:t,strides:n,pad:o,dimRoundingMode:s},f=A.runKernel(_c,p,m);return c?L(f,[f.shape[1],f.shape[2],f.shape[3],f.shape[4]]):f}var cL=S({avgPool3dGrad_:KQ});var pL={kernelName:Za,inputsToSave:["x"],gradFunc:(r,e,t)=>{let[n]=e,{filterSize:o,strides:s,pad:a,dimRoundingMode:i}=t;return{x:()=>cL(r,n,o,s,a,i)}}};function XQ(r,e,t,n,o){let s=k(r,"dy","avgPoolGrad"),a=k(e,"input","avgPoolGrad");T(a.rank===s.rank,()=>`Rank of input (${a.rank}) does not match rank of dy (${s.rank})`);let i=a,l=s,u=!1;a.rank===3&&(u=!0,i=L(a,[1,a.shape[0],a.shape[1],a.shape[2]]),l=L(s,[1,s.shape[0],s.shape[1],s.shape[2]])),T(l.rank===4,()=>`Error in avgPoolGrad: dy must be rank 4 but got rank ${l.rank}.`),T(i.rank===4,()=>`Error in avgPoolGrad: input must be rank 4 but got rank ${i.rank}.`);let c={dy:l,input:i},p={filterSize:t,strides:n,pad:o},m=A.runKernel(wc,c,p);return u?L(m,[m.shape[1],m.shape[2],m.shape[3]]):m}var mL=S({avgPoolGrad_:XQ});var fL={kernelName:vo,inputsToSave:["x"],gradFunc:(r,e,t)=>{let[n]=e,{filterSize:o,strides:s,pad:a}=t;return{x:()=>mL(r,n,o,s,a)}}};var dL={kernelName:Co,inputsToSave:["a","b"],gradFunc:(r,e,t)=>{let[n,o]=e,{transposeA:s,transposeB:a}=t;return!s&&!a?{a:()=>ze(r,o,!1,!0),b:()=>ze(n,r,!0,!1)}:!s&&a?{a:()=>ze(r,o,!1,!1),b:()=>ze(r,n,!0,!1)}:s&&!a?{a:()=>ze(o,r,!1,!0),b:()=>ze(n,r,!1,!1)}:{a:()=>ze(o,r,!0,!0),b:()=>ze(r,n,!0,!0)}}};var hL={kernelName:Ja,gradFunc:(r,e,t)=>{let{blockShape:n,crops:o}=t;return{x:()=>ka(r,n,o)}}};var gL={kernelName:TS,gradFunc:(r,e,t)=>{let n=t,o=n.inputShape,s=n.shape,a=Array.from(s);for(let l=o.length-1;l>=0;l--)if(o[l]===s[l])a[l]=1;else if(o[l]!==1)throw new Error(`broadcastTo(): [${o}] cannot be broadcast to [${s}].`);let i=[];for(let l=0;l<a.length;l++)a[l]>1&&i.push(l);return{x:()=>fe(r,i,!0)}}};var xL={kernelName:qn,gradFunc:r=>({x:()=>r.clone()})};var yL={kernelName:Io,gradFunc:r=>({x:()=>Ie(r)})};var bL={kernelName:Kn,inputsToSave:["x"],gradFunc:(r,e,t)=>{let[n]=e,{clipValueMin:o,clipValueMax:s}=t;return{x:()=>vt(yr(hn(n,o),gn(n,s)),r,Ie(r))}}};var wL={kernelName:Qa,inputsToSave:["x"],gradFunc:Rx.gradFunc};var _L={kernelName:Vs,saveAllInputs:!0,gradFunc:(r,e,t)=>{let n=e.map(l=>l.shape),{axis:o}=t,s=nr(o,e[0].shape)[0],a=n.map(l=>l[s]);return tr(r,a,s).map(l=>()=>l)}};var kL={kernelName:No,inputsToSave:["x","filter"],gradFunc:(r,e,t)=>{let[n,o]=e,{dilations:s,strides:a,pad:i,dataFormat:l}=t;return T(zn(s),()=>`Error in gradient of conv2D: dilation rates greater than 1 are not yet supported in gradients. Got dilations '${s}'`),{x:()=>sp(n.shape,r,o,a,i,l),filter:()=>mp(n,r,o.shape,a,i,l)}}};var vL={kernelName:So,inputsToSave:["dy","filter"],gradFunc:(r,e,t)=>{let[n,o]=e,{strides:s,pad:a,dataFormat:i,dimRoundingMode:l}=t;return{dy:()=>Jr(r,o,s,a,i,1,l),filter:()=>mp(r,n,o.shape,s,a,i,l)}}};function YQ(r,e,t,n,o){let s=r;r.rank===4&&(s=L(r,[1,r.shape[0],r.shape[1],r.shape[2],r.shape[3]]));let a=e;a.rank===4&&(a=L(e,[1,e.shape[0],e.shape[1],e.shape[2],e.shape[3]])),T(s.rank===5,()=>`Error in conv3dDerFilter: input must be rank 5, but got shape ${s.shape}.`),T(a.rank===5,()=>`Error in conv3dDerFilter: dy must be rank 5, but got shape ${a.shape}.`),T(t.length===5,()=>`Error in conv3dDerFilter: filterShape must be length 5, but got ${t}.`),T(s.shape[4]===t[3],()=>`Error in conv3dDerFilter: depth of input ${s.shape[4]}) must match input depth in filter (${t[3]}.`),T(a.shape[4]===t[4],()=>`Error in conv3dDerFilter: depth of dy (${a.shape[4]}) must match output depth for filter (${t[4]}).`);let i={x:s,dy:a},l={strides:n,pad:o,filterShape:t};return A.runKernel(Ic,i,l)}var CL=S({conv3DBackpropFilter_:YQ});var IL={kernelName:el,inputsToSave:["x","filter"],gradFunc:(r,e,t)=>{let{dilations:n,strides:o,pad:s}=t;T(zn(n),()=>`Error in gradient of conv3D: dilation rates greater than 1 are not yet supported in gradients. Got dilations '${n}'`);let[a,i]=e;return{x:()=>xg(a.shape,r,i,o,s),filter:()=>CL(a,r,i.shape,o,s)}}};var NL={kernelName:To,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(qe(bu(oe(t,"float32"))),r)}}};var SL={kernelName:Di,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(wu(oe(t,"float32")),r)}}};var TL={kernelName:Ao,inputsToSave:["x"],gradFunc:(r,e,t)=>{let[n]=e,{axis:o,exclusive:s,reverse:a}=t;return{x:()=>{let i=u_([o],n.rank),l=au(r,o,s,!a);return i!=null&&(l=Ue(l,i)),l}}}};var AL={kernelName:Eo,inputsToSave:["x","filter"],gradFunc:(r,e,t)=>{let{dilations:n,strides:o,pad:s,dimRoundingMode:a}=t,i=n==null?[1,1]:n;T(zn(i),()=>`Error in gradient of depthwiseConv2dNative: dilation rates greater than 1 are not yet supported. Got dilations '${i}'`);let[l,u]=e;return T(l.rank===4,()=>`Error in gradient of depthwiseConv2dNative: input must be rank 4, but got rank ${l.rank}.`),T(u.rank===4,()=>`Error in gradient of depthwiseConv2dNative: filter must be rank 4, but got rank ${u.rank}.`),T(l.shape[3]===u.shape[2],()=>`Error in gradient of depthwiseConv2d: number of input channels (${l.shape[3]}) must match the inChannels dimension in filter ${u.shape[2]}.`),T(Cr(o,i),()=>`Error in gradient of depthwiseConv2d: Either strides or dilations must be 1. Got strides ${o} and dilations '${i}'.`),a!=null&&T(ot(s),()=>`Error in depthwiseConv2d: pad must be an integer when using, dimRoundingMode ${a} but got pad ${s}.`),{x:()=>Tg(l.shape,r,u,o,s,n,a),filter:()=>Sg(l,r,u.shape,o,s,n,a)}}};var EL={kernelName:tl,inputsToSave:["x","filter"],gradFunc:(r,e,t)=>{let[n,o]=e,s={x:n,filter:o,dy:r},a={x:n,filter:o,dy:r};return{x:()=>A.runKernel(Dm,s,t),filter:()=>A.runKernel($m,a,t)}}};var DL={kernelName:Fi,outputsToSave:[!0],gradFunc:(r,e)=>{let[t]=e,n={dy:r,y:t};return{x:()=>A.runKernel($c,n)}}};var $L={kernelName:Oi,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e,n=O(Qt(qe(Me(t))),2/Math.sqrt(Math.PI));return{x:()=>O(r,n)}}};var RL={kernelName:$o,outputsToSave:[!0],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(r,t)}}};var FL={kernelName:Gs,inputsToSave:["input"],gradFunc:(r,e)=>{let[t]=e;return{input:()=>L(r,t.shape)}}};var OL={kernelName:Mi,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(r,Qt(t))}}};var PL={kernelName:Ro,gradFunc:r=>({x:()=>Ie(r)})};var ML={kernelName:Fo,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e,o=Be(t.shape,n.shape);return{a:()=>{let i=pe(r,oe(n,"float32")),l=kt(t.shape,o);return l.length>0?L(fe(i,l),t.shape):i},b:()=>{let i=O(r,oe(t,"float32")),l=kt(n.shape,o);l.length>0&&(i=L(fe(i,l),n.shape));let u=Me(n);return qe(pe(i,oe(u,"float32")))}}}};var LL={kernelName:Oo,inputsToSave:["x","mean","variance","scale"],gradFunc:(r,e,t)=>{let{varianceEpsilon:n}=t,[o,s,a,i]=e,l=i==null?ue(1):i,u=kt(s.shape,o.shape),c=[];if(s.rank===1){for(let _=0;_<o.shape.length-1;++_)c.push(o.shape[_]);c.push(1)}let p=le(o,s),m=O(r,l),f=xu(J(a,ue(n))),d=O(O(O(f,f),f),ue(-.5));return{x:()=>s.rank===1?L(O(O(r,Bn(L(f,[1,1,1,s.shape[0]]),c)),l),o.shape):L(O(O(r,f),l),o.shape),mean:()=>{let _=O(O(f,ue(-1)),m);return s.rank===1&&(_=fe(_,u)),L(_,s.shape)},variance:()=>{let _=O(O(d,p),m);return s.rank===1&&(_=fe(_,u)),L(_,s.shape)},scale:()=>{let _=O(p,f),I=O(r,_);return s.rank===1&&(I=fe(I,u)),L(I,s.shape)},offset:()=>{let _=r;return s.rank===1&&(_=fe(_,u)),L(_,s.shape)}}}};var zL={kernelName:Ws,inputsToSave:["x","indices"],gradFunc:(r,e,t)=>{let[n,o]=e,{axis:s}=t,a=nr(s,n.shape)[0];return{x:()=>{let l=n.shape,u=o.size,c=l.slice(0,a),p=c.length,m=l.slice(s,l.length).slice(1),f=m.length,d=BL(0,p),h=BL(p+1,p+1+f),g=VL([c,[u],m]),x=L(r,g),b=L(o,[u]),w=VL([[p],d,h]),_=Ue(x,w),I=_f(_,b,n.shape[a]),E=lf(w);return I=Ue(I,E),I},indices:()=>o}}};function BL(r,e){let t=[];for(let n=r;n<e;++n)t.push(n);return t}function VL(r){let e=[];for(let t=0;t<r.length;++t)for(let n=0;n<r[t].length;++n)e.push(r[t][n]);return e}var GL={kernelName:Po,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e;return{a:()=>Ie(t),b:()=>Ie(n)}}};var WL={kernelName:Xn,gradFunc:r=>({x:()=>oe(r,"float32")})};var jL={kernelName:Vi,gradFunc:r=>({x:()=>Ie(r)})};var UL={kernelName:Gi,gradFunc:r=>({x:()=>Ie(r)})};var HL={kernelName:Wi,gradFunc:r=>({x:()=>Ie(r)})};var qL={kernelName:Mo,inputsToSave:["x"],gradFunc:(r,e,t)=>{let[n]=e,{alpha:o}=t,s=Vt(n,0);return{x:()=>vt(s,r,O(r,o))}}};var KL={kernelName:Hi,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>pe(r,J(t,1))}}};var XL={kernelName:Lo,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>pe(r,oe(t,"float32"))}}};var YL={kernelName:AS,inputsToSave:[],outputsToSave:[!0],gradFunc:(r,e,t)=>{let[n]=e,{axis:o}=t;return{logits:()=>{let s=!0,a=Qt(n);return le(r,O(fe(r,o,s),a))}}}};function ZQ(r,e,t,n=5,o=1,s=1,a=.5){let i={x:r,y:e,dy:t},l={depthRadius:n,bias:o,alpha:s,beta:a};return A.runKernel(Mc,i,l)}var ZL=S({localResponseNormalizationBackprop_:ZQ});var JL={kernelName:nl,inputsToSave:["x"],outputsToSave:[!0],gradFunc:(r,e,t)=>{let[n,o]=e,{depthRadius:s,bias:a,alpha:i,beta:l}=t;return{x:()=>ZL(n,o,r,s,a,i,l)}}};function Fx(r,e,t,n){return e.rank<t.rank&&(e=L(e,to(e.shape,n))),r.rank<t.rank&&(r=L(r,to(r.shape,n))),{x:()=>O(r,oe(Sn(t,e),r.dtype))}}var WC={kernelName:zo,inputsToSave:["x"],outputsToSave:[!0],gradFunc:(r,e,t)=>{let n=t,{reductionIndices:o}=n,s=e[0],a=e[1],i=nr(o,s.shape),l=Fx(r,a,s,i);return{x:()=>l.x()}}};var QL={kernelName:Bo,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e;return{a:()=>O(r,oe(hn(t,n),"float32")),b:()=>O(r,oe(uu(t,n),"float32"))}}};function JQ(r,e,t,n,o,s,a){let i=k(r,"dy","maxPool3dGrad"),l=k(e,"input","maxPool3dGrad"),u=k(t,"output","maxPool3dGrad"),c=i,p=l,m=u,f=!1;l.rank===4&&(f=!0,c=L(i,[1,i.shape[0],i.shape[1],i.shape[2],i.shape[3]]),p=L(l,[1,l.shape[0],l.shape[1],l.shape[2],l.shape[3]]),m=L(u,[1,u.shape[0],u.shape[1],u.shape[2],u.shape[3]])),T(c.rank===5,()=>`Error in maxPool3dGrad: dy must be rank 5 but got rank ${c.rank}.`),T(p.rank===5,()=>`Error in maxPool3dGrad: input must be rank 5 but got rank ${p.rank}.`),T(m.rank===5,()=>`Error in maxPool3dGrad: output must be rank 5 but got rank ${m.rank}.`),a!=null&&T(ot(s),()=>`Error in maxPool3dGrad: pad must be an integer when using, dimRoundingMode ${a} but got pad ${s}.`);let d={dy:c,input:p,output:m},h={filterSize:n,strides:o,pad:s,dimRoundingMode:a},g=A.runKernel(zc,d,h);return f?L(g,[g.shape[1],g.shape[2],g.shape[3],g.shape[4]]):g}var ez=S({maxPool3dGrad_:JQ});var tz={kernelName:ol,inputsToSave:["x"],outputsToSave:[!0],gradFunc:(r,e,t)=>{let[n,o]=e,{filterSize:s,strides:a,pad:i,dimRoundingMode:l}=t;return{x:()=>ez(r,n,o,s,a,i,l)}}};function QQ(r,e,t,n,o,s,a){let i=k(r,"dy","maxPoolGrad"),l=k(e,"input","maxPoolGrad"),u=k(t,"output","maxPoolGrad");T(l.rank===i.rank,()=>`Rank of input (${l.rank}) does not match rank of dy (${i.rank})`),T(i.rank===4,()=>`Error in maxPoolGrad: dy must be rank 4 but got rank ${i.rank}.`),T(l.rank===4,()=>`Error in maxPoolGrad: input must be rank 4 but got rank ${l.rank}.`),a!=null&&T(ot(s),()=>`Error in maxPoolGrad: pad must be an integer when using, dimRoundingMode ${a} but got pad ${s}.`);let c={dy:i,input:l,output:u},p={filterSize:n,strides:o,pad:s,dimRoundingMode:a};return A.runKernel(Lc,c,p)}var rz=S({maxPoolGrad_:QQ});var nz={kernelName:Vo,inputsToSave:["x"],outputsToSave:[!0],gradFunc:(r,e,t)=>{let[n,o]=e,{filterSize:s,strides:a,pad:i}=t;return{x:()=>rz(r,n,o,s,a,i)}}};var oz={kernelName:Go,inputsToSave:["x"],gradFunc:(r,e,t)=>{let[n]=e,{axis:o}=t,s=nr(o,n.shape),i=l_(n.shape,s)[1],l=nt(i);return{x:()=>{let c=n.shape.slice();s.forEach(f=>{c[f]=1});let p=L(r,c);return pe(O(p,er(n.shape,"float32")),l)}}}};var sz={kernelName:Wo,inputsToSave:["x"],outputsToSave:[!0],gradFunc:(r,e,t)=>{let n=t,{axis:o}=n,[s,a]=e,i=nr(o,s.shape),l=Fx(r,a,s,i);return{x:()=>l.x()}}};var iz={kernelName:jo,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e;return{a:()=>O(r,oe(gn(t,n),"float32")),b:()=>O(r,oe(Vt(t,n),"float32"))}}};var az={kernelName:Uo,inputsToSave:["x"],gradFunc:(r,e,t)=>{let n=e[0],{paddings:o}=t,s=o.map(a=>a[0]);return{x:()=>Re(r,s,n.shape)}}};var lz={kernelName:Ki,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e,o=Be(t.shape,n.shape);return{a:()=>{let i=kt(t.shape,o);return i.length>0?L(fe(r,i),t.shape):r},b:()=>{let i=O(r,qe(ws(pe(t,n)))),l=kt(n.shape,o);return l.length>0?L(fe(i,l),n.shape):i}}}};var uz={kernelName:Ho,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e,o=Be(t.shape,n.shape);return{a:()=>{let i=O(r,oe(n,"float32")),l=kt(t.shape,o);return l.length>0?L(fe(i,l),t.shape):i},b:()=>{let i=O(r,oe(t,"float32")),l=kt(n.shape,o);return l.length>0?L(fe(i,l),n.shape):i}}}};var cz={kernelName:js,gradFunc:r=>({x:()=>qe(r)})};var pz={kernelName:qo,inputsToSave:["indices"],gradFunc:(r,e)=>{let t=e[0];return{indices:()=>ht(t.shape,"float32")}}};var mz={kernelName:Us,gradFunc:r=>({x:()=>Ie(r)})};var fz={kernelName:Hs,saveAllInputs:!0,gradFunc:(r,e,t)=>{let{axis:n}=t;return fr(r,n).map(s=>()=>s)}};var jC={kernelName:Ko,inputsToSave:["x"],gradFunc:(r,e,t)=>{let n=e[0],{paddings:o}=t,s=o.map(a=>a[0]);return{x:()=>Re(r,s,n.shape)}}};var dz={kernelName:Xo,inputsToSave:["a","b"],outputsToSave:[!0],gradFunc:(r,e)=>{let[t,n,o]=e,s=t,a=n,i=Be(s.shape,a.shape);return{a:()=>{let c=oe(a,"float32"),p=O(r,O(c,zr(s,le(c,ue(1))))),m=kt(s.shape,i);return m.length>0&&(p=fe(p,m)),L(p,s.shape)},b:()=>{let c=Vt(s,0),p=vt(c,pr(s),Ie(s)),m=O(r,O(o,p)),f=kt(a.shape,i);return f.length>0&&(m=fe(m,f)),L(m,a.shape)}}}};var hz={kernelName:Yo,inputsToSave:["x","alpha"],gradFunc:(r,e)=>{let[t,n]=e,o=Vt(t,0);return{x:()=>vt(o,r,O(r,n)),alpha:()=>{let s=vt(o,Ie(r),O(r,t)),a=kt(n.shape,r.shape);return a.length>0&&(s=fe(s,a)),L(s,n.shape)}}}};var gz={kernelName:Do,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e,o=Be(t.shape,n.shape);return{a:()=>{let i=pe(r,oe(n,"float32")),l=kt(t.shape,o);return l.length>0?L(fe(i,l),t.shape):i},b:()=>{let i=O(r,oe(t,"float32")),l=kt(n.shape,o);l.length>0&&(i=L(fe(i,l),n.shape));let u=Me(n);return qe(pe(i,oe(u,"float32")))}}}};var xz={kernelName:ea,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>pe(r,qe(Me(t)))}}};var yz={kernelName:Qo,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e,n=O(gn(t,6),vs(t));return{x:()=>O(r,oe(n,"float32"))}}};var bz={kernelName:Zo,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(r,oe(vs(t),"float32"))}}};var wz={kernelName:qs,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>L(r,t.shape)}}};var _z={kernelName:Jo,inputsToSave:["images"],gradFunc:(r,e,t)=>{let[n]=e,o={dy:r,images:n};return{images:()=>A.runKernel(jc,o,t)}}};var kz={kernelName:il,inputsToSave:["images"],gradFunc:(r,e,t)=>{let[n]=e,o={dy:r,images:n};return{images:()=>A.runKernel(Wc,o,t)}}};var vz={kernelName:es,gradFunc:(r,e,t)=>{let{dims:n}=t,o=nr(n,r.shape);return{x:()=>Kt(r,o)}}};var Cz={kernelName:ts,gradFunc:r=>({x:()=>Ie(r)})};var Iz={kernelName:rs,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>qe(pe(r,O(zr(t,1.5),2)))}}};var Nz={kernelName:Ks,inputsToSave:["condition"],gradFunc:(r,e)=>{let[t]=e;return{condition:()=>oe(Ie(t),"float32"),t:()=>O(r,oe(t,r.dtype)),e:()=>O(r,oe(wa(t),r.dtype))}}};var Sz={kernelName:ra,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>{let n=Vt(t,ue(0)),o=ue(S_),s=ue(T_),a=O(r,s),i=O(O(r,o),Qt(oe(t,"float32")));return vt(n,a,i)}}}};var Tz={kernelName:os,outputsToSave:[!0],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(r,O(t,le(ue(1),t)))}}};var Az={kernelName:oa,gradFunc:r=>({x:()=>Ie(r)})};var Ez={kernelName:ns,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(ya(oe(t,"float32")),r)}}};var Dz={kernelName:na,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(iu(oe(t,"float32")),r)}}};var $z={kernelName:Xs,inputsToSave:["x"],gradFunc:(r,e,t)=>{let[n]=e,{begin:o,size:s}=t,a=n.shape,[i,l]=Mw(n,o,s),u=[];for(let c=0;c<r.rank;c++)u.push([i[c],a[c]-i[c]-l[c]]);return{x:()=>Lr(r,u)}}};var Rz={kernelName:as,outputsToSave:[!0],gradFunc:(r,e,t)=>{let[n]=e,{dim:o}=t,s=!0,a=O(r,n);return{logits:()=>le(a,O(fe(a,[o],s),n))}}};var Fz={kernelName:sa,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(r,Dr(t))}}};var UC={kernelName:al,gradFunc:(r,e,t)=>{let{blockShape:n,paddings:o}=t;return{x:()=>ga(r,n,o)}}};var HC={kernelName:Ys,gradFunc:(r,e,t)=>{let{axis:n}=t;return{x:()=>Qe(r,n)}}};var Oz={kernelName:ss,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>pe(r,O(gt(oe(t,"float32")),2))}}};var Pz={kernelName:ll,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(r,O(oe(t,"float32"),2))}}};var Mz={kernelName:ls,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e,o=ue(2);return{a:()=>O(r,O(o,le(t,n))),b:()=>O(r,O(o,le(n,t)))}}};var Lz={kernelName:Yn,gradFunc:r=>({x:()=>Ie(r)})};var zz={kernelName:us,inputsToSave:["a","b"],gradFunc:(r,e)=>{let[t,n]=e,o=Be(t.shape,n.shape);return{a:()=>{let i=r,l=kt(t.shape,o);return l.length>0&&(i=fe(i,l)),L(i,t.shape)},b:()=>{let i=r,l=kt(n.shape,o);return l.length>0&&(i=fe(i,l)),L(qe(i),n.shape)}}}};var Bz={kernelName:is,inputsToSave:["x"],gradFunc:(r,e,t)=>{let[n]=e,o=n.shape.slice(),{axis:s}=t;nr(s,n.shape).forEach(u=>{o[u]=1});let i=L(r,o),l=O(i,er(n.shape,"float32"));return{x:()=>l}}};var Vz={kernelName:cs,inputsToSave:["x"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>pe(r,Me(ya(t)))}}};var Gz={kernelName:ps,outputsToSave:[!0],gradFunc:(r,e)=>{let[t]=e;return{x:()=>O(le(ue(1),Me(t)),r)}}};var Wz={kernelName:Mn,inputsToSave:["x"],gradFunc:(r,e,t)=>{let[n]=e,{reps:o}=t;return{x:()=>{let a=Ie(n);if(n.rank===1)for(let i=0;i<o[0];++i)a=J(a,Re(r,[i*n.shape[0]],[n.shape[0]]));else if(n.rank===2)for(let i=0;i<o[0];++i)for(let l=0;l<o[1];++l)a=J(a,Re(r,[i*n.shape[0],l*n.shape[1]],[n.shape[0],n.shape[1]]));else if(n.rank===3)for(let i=0;i<o[0];++i)for(let l=0;l<o[1];++l)for(let u=0;u<o[2];++u)a=J(a,Re(r,[i*n.shape[0],l*n.shape[1],u*n.shape[2]],[n.shape[0],n.shape[1],n.shape[2]]));else if(n.rank===4)for(let i=0;i<o[0];++i)for(let l=0;l<o[1];++l)for(let u=0;u<o[2];++u)for(let c=0;c<o[3];++c)a=J(a,Re(r,[i*n.shape[0],l*n.shape[1],u*n.shape[2],c*n.shape[3]],[n.shape[0],n.shape[1],n.shape[2],n.shape[3]]));else throw new Error(`Gradient for tile operation is not implemented for rank-${n.rank} tensors yet.`);return a}}}};var jz={kernelName:ms,gradFunc:(r,e,t)=>{let n=t,{perm:o}=n,s=lf(o);return{x:()=>Ue(r,s)}}};var Uz={kernelName:Zs,gradFunc:(r,e,t)=>{let n=t,{axis:o}=n;return{value:()=>Gt(r,o)}}};var Hz={kernelName:ul,inputsToSave:["segmentIds"],gradFunc:(r,e)=>{let[t]=e;return{x:()=>eee(r,t)}}};function eee(r,e){let t=en(e,Ie(e)),n=Qn(r,t),o=hn(e,ue(0,"int32")),s=n.rank-o.rank;for(let i=0;i<s;++i)o=cr(o,i+1);o=yr(o,er(n.shape,"bool"));let a=Ie(n);return vt(o,n,a)}var qz={kernelName:Js,gradFunc:r=>({x:()=>Ie(r)})};var tee=[Rx,QM,eL,tL,rL,nL,oL,sL,iL,aL,lL,uL,pL,fL,dL,hL,gL,xL,yL,bL,wL,_L,vL,kL,IL,NL,SL,TL,AL,EL,gz,DL,$L,RL,FL,OL,ML,PL,LL,zL,GL,WL,jL,UL,HL,qL,KL,XL,YL,JL,WC,WC,QL,tz,nz,oz,sz,iz,az,lz,uz,cz,pz,mz,fz,jC,jC,dz,hz,xz,yz,bz,wz,_z,kz,vz,Cz,Iz,Nz,Sz,Tz,Az,Ez,Dz,$z,Rz,Fz,UC,UC,HC,HC,Oz,Mz,Pz,Lz,zz,Bz,Vz,Gz,Wz,jz,Uz,Hz,qz];for(let r of tee)ES(r);var Qz={};We(Qz,{maxNorm:()=>nee,minMaxNorm:()=>iee,nonNeg:()=>see,unitNorm:()=>oee});var qC;function rr(){return qC==null&&(qC=LT().epsilon()),qC}function rn(){return"channelsLast"}var Dn=class extends Error{constructor(e){super(e);Object.setPrototypeOf(this,Dn.prototype)}},Gr=class extends Error{constructor(e){super(e);Object.setPrototypeOf(this,Gr.prototype)}},z=class extends Error{constructor(e){super(e);Object.setPrototypeOf(this,z.prototype)}},Se=class extends Error{constructor(e){super(e);Object.setPrototypeOf(this,Se.prototype)}},Xf=class extends Error{constructor(e){super(e);Object.setPrototypeOf(this,Xf.prototype)}};function po(r,e){if(Array.isArray(r)){let t=[];for(let n=0;n<e;n++)t=t.concat(r);return t}else{let t=new Array(e);return t.fill(r),t}}function Gn(r,e){if(!r)throw new Xf(e)}function KC(r,e){let t=0;for(let n of r)n===e&&t++;return t}function wr(r){return r.length===1?r[0]:r}function bt(r){return Array.isArray(r)?r:[r]}function mo(r){let t=r.replace(/(.)([A-Z][a-z0-9]+)/g,"$1_$2").replace(/([a-z])([A-Z])/g,"$1_$2").toLowerCase();return t[0]!=="_"?t:"private"+t}function Pa(r){return r.length<=1||r.indexOf("_")===-1?r:r.replace(/[_]+(\w|$)/g,(e,t)=>t.toUpperCase())}var fo={};function Mp(r){if(r==null)return null;let e={};return e.className=r.getClassName(),e.config=r.getConfig(),e}function XC(r){if(!(r==null||typeof r!="object"))if(Array.isArray(r))r.forEach(e=>XC(e));else{let e=Object.keys(r);for(let t of e){let n=r[t];n!=null&&typeof n=="object"&&(!Array.isArray(n)&&n.type==="ndarray"&&typeof n.value=="number"?r[t]=n.value:XC(n))}}}function pi(r,e={},t={},n="object",o=!1){if(typeof r=="string"){let s=r,a;if(s in t)a=t[s];else if(s in fo)a=fo[s];else if(a=e[s],a==null)throw new z(`Unknown ${n}: ${r}. This may be due to one of the following reasons:
|
|
1. The ${n} is defined in Python, in which case it needs to be ported to TensorFlow.js or your JavaScript code.
|
|
2. The custom ${n} is defined in JavaScript, but is not registered properly with tf.serialization.registerClass().`);return a}else{let s=r;if(s.className==null||s.config==null)throw new z(`${n}: Improper config format: ${JSON.stringify(s)}.
|
|
'className' and 'config' must set.`);let a=s.className,i,l;if(a in t?[i,l]=t[a]:a in fo?[i,l]=fo.className:a in e&&([i,l]=e[a]),i==null)throw new z(`Unknown ${n}: ${a}. This may be due to one of the following reasons:
|
|
1. The ${n} is defined in Python, in which case it needs to be ported to TensorFlow.js or your JavaScript code.
|
|
2. The custom ${n} is defined in JavaScript, but is not registered properly with tf.serialization.registerClass().`);if(l!=null){let u={};for(let f of Object.keys(fo))u[f]=fo[f];for(let f of Object.keys(t))u[f]=t[f];let c=s.config;c.customObjects=u;let p=Object.assign({},fo);for(let f of Object.keys(t))fo[f]=t[f];XC(s.config);let m=l(i,s.config,t,o);return fo=Object.assign({},p),m}else{let u=Object.assign({},fo);for(let p of Object.keys(t))fo[p]=t[p];let c=new i(s.config);return fo=Object.assign({},u),c}}}function ree(r,e){return r<e?-1:r>e?1:0}function Yf(r,e){return-1*ree(r,e)}function ho(r){if(r==null)return r;let e=[];for(let t of r)e.indexOf(t)===-1&&e.push(t);return e}function Kz(r){if(r==null)throw new z(`Invalid value in obj: ${JSON.stringify(r)}`);for(let e in r)if(r.hasOwnProperty(e))return!1;return!0}function mi(r,e,t){if(t!=null&&r.indexOf(t)<0)throw new z(`${t} is not a valid ${e}. Valid values are ${r} or null/undefined.`)}function Ox(r,e,t=0,n=Infinity){return Gn(t>=0),Gn(n>=t),Array.isArray(r)&&r.length>=t&&r.length<=n&&r.every(o=>typeof o===e)}function Ht(r,e){Array.isArray(r)?(y.assert(r.length>0,()=>`${e} is unexpectedly an empty array.`),r.forEach((t,n)=>Ht(t,`element ${n+1} of ${e}`))):y.assert(Number.isInteger(r)&&r>0,()=>`Expected ${e} to be a positive integer, but got ${Xz(r)}.`)}function Xz(r){return r===null?"null":Array.isArray(r)?"["+r.map(e=>Xz(e)).join(",")+"]":typeof r=="string"?`"${r}"`:`${r}`}function Yz(r,e){let t=y.now(),n;return(...s)=>{let a=y.now();return a-t<e||(t=a,n=r(...s)),n}}function Px(r){return r==="relu"?"relu":r==="linear"?"linear":r==="elu"?"elu":null}function YC(r,e){return V(()=>gt(fe(O(r,r),e,!0)))}var Lp=class extends Q.Serializable{getConfig(){return{}}},Zf=class extends Lp{constructor(e){super();this.defaultMaxValue=2,this.defaultAxis=0,this.maxValue=e.maxValue!=null?e.maxValue:this.defaultMaxValue,this.axis=e.axis!=null?e.axis:this.defaultAxis}apply(e){return V(()=>{let t=YC(e,this.axis),n=ur(t,0,this.maxValue);return O(e,pe(n,J(rr(),t)))})}getConfig(){return{maxValue:this.maxValue,axis:this.axis}}};Zf.className="MaxNorm";Q.registerClass(Zf);var Jf=class extends Lp{constructor(e){super();this.defaultAxis=0,this.axis=e.axis!=null?e.axis:this.defaultAxis}apply(e){return V(()=>pe(e,J(rr(),YC(e,this.axis))))}getConfig(){return{axis:this.axis}}};Jf.className="UnitNorm";Q.registerClass(Jf);var Qf=class extends Lp{apply(e){return $r(e)}};Qf.className="NonNeg";Q.registerClass(Qf);var ed=class extends Lp{constructor(e){super();this.defaultMinValue=0,this.defaultMaxValue=1,this.defaultRate=1,this.defaultAxis=0,this.minValue=e.minValue!=null?e.minValue:this.defaultMinValue,this.maxValue=e.maxValue!=null?e.maxValue:this.defaultMaxValue,this.rate=e.rate!=null?e.rate:this.defaultRate,this.axis=e.axis!=null?e.axis:this.defaultAxis}apply(e){return V(()=>{let t=YC(e,this.axis),n=J(O(this.rate,ur(t,this.minValue,this.maxValue)),O(1-this.rate,t));return O(e,pe(n,J(rr(),t)))})}getConfig(){return{minValue:this.minValue,maxValue:this.maxValue,rate:this.rate,axis:this.axis}}};ed.className="MinMaxNorm";Q.registerClass(ed);var Zz={maxNorm:"MaxNorm",minMaxNorm:"MinMaxNorm",nonNeg:"NonNeg",unitNorm:"UnitNorm"};function Mt(r){return Mp(r)}function Jz(r,e={}){return pi(r,Q.SerializationMap.getMap().classNameMap,e,"constraint")}function Lt(r){if(r==null)return null;if(typeof r=="string"){let t={className:r in Zz?Zz[r]:r,config:{}};return Jz(t)}else return r instanceof Lp?r:Jz(r)}function nee(r){return new Zf(r)}function oee(r){return new Jf(r)}function see(){return new Qf}function iee(r){return new ed(r)}var w3={};We(w3,{constant:()=>dee,glorotNormal:()=>_ee,glorotUniform:()=>wee,heNormal:()=>kee,heUniform:()=>vee,identity:()=>yee,leCunNormal:()=>Cee,leCunUniform:()=>Iee,ones:()=>fee,orthogonal:()=>Nee,randomNormal:()=>gee,randomUniform:()=>hee,truncatedNormal:()=>xee,varianceScaling:()=>bee,zeros:()=>mee});var e3=["channelsFirst","channelsLast"],t3=["nearest","bilinear"],r3=["valid","same","causal"],n3=["max","avg"],o3=["sum","mul","concat","ave"];var zp=new Map;function Ot(r){mi(e3,"DataFormat",r)}function s3(r){mi(t3,"InterpolationFormat",r)}function nn(r){mi(r3,"PaddingMode",r)}function ZC(r){mi(n3,"PoolMode",r)}var td=[],i3="/";function Ts(r,e){td.push(r);try{let t=e();return td.pop(),t}catch(t){throw td.pop(),t}}function aee(){return td.length===0?"":td.join(i3)+i3}function Mx(r){if(!a3(r))throw new Error("Not a valid tensor name: '"+r+"'");return aee()+r}function Lx(r){if(!a3(r))throw new Error("Not a valid tensor name: '"+r+"'");zp.has(r)||zp.set(r,0);let e=zp.get(r);if(zp.set(r,zp.get(r)+1),e>0){let t=`${r}_${e}`;return zp.set(t,1),t}else return r}var lee=new RegExp(/^[A-Za-z0-9][-A-Za-z0-9\._\/]*$/);function a3(r){return!!r.match(lee)}function l3(r){return r===parseInt(r.toString(),10)}function go(r,e,t){e==null&&(e=0),t==null&&(t=r.length);let n=1;for(let o=e;o<t;++o)n*=r[o];return n}function u3(r){return r=Array.isArray(r)?new Float32Array(r):r,Ct(r)}function qu(r){return oi(u3(r)).dataSync()[0]}function As(r){return mr(u3(r)).dataSync()[0]}function Wr(r,e){if(e<r)throw new z(`end (${e}) < begin (${r}) is forbidden.`);let t=[];for(let n=r;n<e;++n)t.push(n);return t}function Ma(r,e){return r.asType(e)}function La(r,e=-1){let t=r.shape.slice();return e<0&&(e=t.length+e+1),t.splice(e,0,1),r.reshape(t)}function c3(r,e){return V(()=>{if(r.shape.length!==2)throw new z(`repeat() expects a rank-2 tensor, but received a rank-${r.shape.length} tensor.`);let t=La(r,1);return zx(t,[1,e,1])})}function p3(r){let e=[go(r.shape)];return r.reshape(e)}function m3(r){if(r.rank<=1)throw new z(`batchFlatten requires a minimum rank of 2. Got rank: ${r.rank}.`);let e=[r.shape[0],go(r.shape,1)];return r.reshape(e)}function za(r,e,t){return V(()=>{switch(r.rank){case 1:return gf(r,e,t);case 2:return vg(r,[e,0],[t,r.shape[1]]);case 3:return xf(r,[e,0,0],[t,r.shape[1],r.shape[2]]);case 4:return up(r,[e,0,0,0],[t,r.shape[1],r.shape[2],r.shape[3]]);case 5:return Re(r,[e,0,0,0,0],[t,r.shape[1],r.shape[2],r.shape[3],r.shape[4]]);case 6:return Re(r,[e,0,0,0,0,0],[t,r.shape[1],r.shape[2],r.shape[3],r.shape[4],r.shape[5]]);default:throw new z(`sliceAlongFirstAxis() received an unsupported tensor rank: ${r.rank}`)}})}function JC(r,e,t){return V(()=>{switch(r.rank){case 1:return gf(r,e,t);case 2:return vg(r,[0,e],[r.shape[0],t]);case 3:return xf(r,[0,0,e],[r.shape[0],r.shape[1],t]);case 4:return up(r,[0,0,0,e],[r.shape[0],r.shape[1],r.shape[2],t]);default:throw new z(`sliceAlongLastAxis() received an unsupported tensor rank: ${r.rank}`)}})}function rd(r,e,t,n){return V(()=>{switch(r.rank){case 1:return gf(r,e,t);case 2:switch(n){case 1:return za(r,e,t);case 2:return JC(r,e,t);default:throw new z(`The axis is not within the rank of the tensor ${n}`)}case 3:switch(n){case 1:return za(r,e,t);case 2:return xf(r,[0,e,0],[r.shape[0],t,r.shape[2]]);case 3:return JC(r,e,t);default:throw new z(`The axis is not within the rank of the tensor ${n}`)}case 4:switch(n){case 1:return za(r,e,t);case 2:return up(r,[0,e,0,0],[r.shape[0],t,r.shape[2],r.shape[3]]);case 3:return up(r,[0,0,e,0],[r.shape[0],r.shape[1],t,r.shape[3]]);case 4:return JC(r,e,t);default:throw new z(`The axis is not within the rank of the tensor ${n}`)}default:throw new z(`sliceAlongLastAxis() received an unsupported tensor rank: ${r.rank}`)}})}function Bp(r,e=-1){let t;return e<0&&(t=r[0].rank,t!==0?e=t:e=0),e===r[0].rank&&(e=-1),Qe(r,e)}function QC(r,e){switch(r.rank){case 1:return Xw([r,e]);case 2:return Yw([r,e],0);case 3:return Zw([r,e],0);case 4:return Jw([r,e],0);default:throw new z(`concatAlongFirstAxis() received an unsupported tensor rank: ${r.rank}`)}}function zx(r,e){if(Array.isArray(e)||(e=[e]),r.rank!==e.length)throw new z(`The length of input n (${e.length}) does not match the number of dimensions in input x (${r.rank})`);return Bn(r,e)}function Vp(r,e=0,t=1,n,o){return kg(r,e,t,n,o)}function xo(r,e,t,n){if(r.rank<2||e.rank<2)throw new Se(`dot requires both inputs to be rank >= 2 but got x shape = ${r.shape} and y shape = ${e.shape}`);if(e.rank>=3){let o=r.shape.slice(-1)[0],s=e.shape.slice(-2)[0];if(o!==s)throw new Se(`If rank y >= 3, then the second last dim of y must equal the last dim of x but got x shape = ${r.shape} and y shape = ${e.shape}`)}if(r.rank===2&&e.rank===2){let o=!1,s=!1;return no.matMul({a:r,b:e,transposeA:o,transposeB:s,bias:n?e0(r.rank,n,rn()):null,activation:t})}else{let o=r.shape.slice(),s=o.pop();r=r.reshape([-1,s]);let a=e.shape.slice(),i=a.pop(),l=a.pop(),u=[...a,i],c=Array.from({length:e.rank},(d,h)=>h===0?e.rank-2:h<=e.rank-2?h-1:h);e=e.transpose(c).reshape([l,-1]);let p=[...o,...u],m=!1,f=!1;return no.matMul({a:r,b:e,transposeA:m,transposeB:f,bias:n?e0(r.rank,n,rn()):null,activation:t}).reshape(p)}}function Bx(r,e,t){return V(()=>(Array.isArray(e)?e=Ct(e,"int32"):e=e.toInt(),Qn(r,e,t)))}function Ku(r){return O(r,r)}function e0(r,e,t){let n=e.shape;if(e.rank!==1&&e.rank!==r)throw new z(`Unexpected bias dimensions: ${e.rank}; expected it to be 1 or ${r}`);if(r===5){if(t==="channelsFirst")return n.length===1?e.reshape([1,n[0],1,1,1]):e.reshape([1,n[3],n[0],n[1],n[2]]);if(t==="channelsLast")return n.length===1?e.reshape([1,1,1,1,n[0]]):e.reshape([1].concat(n))}else if(r===4){if(t==="channelsFirst")return n.length===1?e.reshape([1,n[0],1,1]):e.reshape([1,n[2],n[0],n[1]]);if(t==="channelsLast")return n.length===1?e.reshape([1,1,1,n[0]]):e.reshape([1].concat(n))}else if(r===3){if(t==="channelsFirst")return n.length===1?e.reshape([1,n[0],1]):e.reshape([1,n[1],n[0]]);if(t==="channelsLast")return n.length===1?e.reshape([1,1,n[0]]):e.reshape([1].concat(n))}else if(r<3)return e;throw new z(`Unsupported input rank by biasAdd: ${e.rank}`)}function on(r,e,t){return V(()=>(t==null&&(t=rn()),Ot(t),r.add(e0(r.rank,e,t))))}function f3(r,e=1){if(e!==1)throw new Se(`Support for alpha values other than 1 (${e}) is not implemented yet.`);return ys(r)}function d3(r){return V(()=>pe(r,Tt(r).add(1)))}function Vx(r,e,t,n){return V(()=>l1(r,e,t,n))}function h3(r){return V(()=>{let e=J(.5,O(.2,r));return ur(e,0,1)})}function Nl(r,e,t=!1){return t?r():e()}var g3=["fanIn","fanOut","fanAvg"],x3=["normal","uniform","truncatedNormal"];function uee(r){mi(g3,"FanMode",r)}function cee(r){mi(x3,"Distribution",r)}var bn=class extends Q.Serializable{fromConfigUsesCustomObjects(){return!1}getConfig(){return{}}},nd=class extends bn{apply(e,t){return ht(e,t)}};nd.className="Zeros";Q.registerClass(nd);var Xu=class extends bn{apply(e,t){return er(e,t)}};Xu.className="Ones";Q.registerClass(Xu);var od=class extends bn{constructor(e){super();if(typeof e!="object")throw new z(`Expected argument of type ConstantConfig but got ${e}`);if(e.value===void 0)throw new z(`config must have value set but got ${e}`);this.value=e.value}apply(e,t){return V(()=>O(ue(this.value),er(e,t)))}getConfig(){return{value:this.value}}};od.className="Constant";Q.registerClass(od);var sd=class extends bn{constructor(e){super();this.DEFAULT_MINVAL=-.05,this.DEFAULT_MAXVAL=.05,this.minval=e.minval||this.DEFAULT_MINVAL,this.maxval=e.maxval||this.DEFAULT_MAXVAL,this.seed=e.seed}apply(e,t){return ks(e,this.minval,this.maxval,t)}getConfig(){return{minval:this.minval,maxval:this.maxval,seed:this.seed}}};sd.className="RandomUniform";Q.registerClass(sd);var id=class extends bn{constructor(e){super();this.DEFAULT_MEAN=0,this.DEFAULT_STDDEV=.05,this.mean=e.mean||this.DEFAULT_MEAN,this.stddev=e.stddev||this.DEFAULT_STDDEV,this.seed=e.seed}apply(e,t){if(t=t||"float32",t!=="float32"&&t!=="int32")throw new Se(`randomNormal does not support dType ${t}.`);return Vp(e,this.mean,this.stddev,t,this.seed)}getConfig(){return{mean:this.mean,stddev:this.stddev,seed:this.seed}}};id.className="RandomNormal";Q.registerClass(id);var ad=class extends bn{constructor(e){super();this.DEFAULT_MEAN=0,this.DEFAULT_STDDEV=.05,this.mean=e.mean||this.DEFAULT_MEAN,this.stddev=e.stddev||this.DEFAULT_STDDEV,this.seed=e.seed}apply(e,t){if(t=t||"float32",t!=="float32"&&t!=="int32")throw new Se(`truncatedNormal does not support dType ${t}.`);return vu(e,this.mean,this.stddev,t,this.seed)}getConfig(){return{mean:this.mean,stddev:this.stddev,seed:this.seed}}};ad.className="TruncatedNormal";Q.registerClass(ad);var ld=class extends bn{constructor(e){super();this.gain=e.gain!=null?e.gain:1}apply(e,t){return V(()=>{if(e.length!==2||e[0]!==e[1])throw new z("Identity matrix initializer can only be used for 2D square matrices.");return O(this.gain,ip(e[0]))})}getConfig(){return{gain:this.gain}}};ld.className="Identity";Q.registerClass(ld);function pee(r,e="channelsLast"){let t,n;if(Ot(e),r.length===2)t=r[0],n=r[1];else if([3,4,5].indexOf(r.length)!==-1){if(e==="channelsFirst"){let o=go(r,2);t=r[1]*o,n=r[0]*o}else if(e==="channelsLast"){let o=go(r,0,r.length-2);t=r[r.length-2]*o,n=r[r.length-1]*o}}else{let o=go(r);t=Math.sqrt(o),n=Math.sqrt(o)}return[t,n]}var jr=class extends bn{constructor(e){super();if(e.scale<0)throw new z(`scale must be a positive float. Got: ${e.scale}`);this.scale=e.scale==null?1:e.scale,this.mode=e.mode==null?"fanIn":e.mode,uee(this.mode),this.distribution=e.distribution==null?"normal":e.distribution,cee(this.distribution),this.seed=e.seed}apply(e,t){let n=pee(e),o=n[0],s=n[1],a=this.scale;if(this.mode==="fanIn"?a/=Math.max(1,o):this.mode==="fanOut"?a/=Math.max(1,s):a/=Math.max(1,(o+s)/2),this.distribution==="normal"){let i=Math.sqrt(a);if(t=t||"float32",t!=="float32"&&t!=="int32")throw new Se(`${this.getClassName()} does not support dType ${t}.`);return vu(e,0,i,t,this.seed)}else{let i=Math.sqrt(3*a);return ks(e,-i,i,t)}}getConfig(){return{scale:this.scale,mode:this.mode,distribution:this.distribution,seed:this.seed}}};jr.className="VarianceScaling";Q.registerClass(jr);var Gp=class extends jr{constructor(e){super({scale:1,mode:"fanAvg",distribution:"uniform",seed:e==null?null:e.seed})}getClassName(){return jr.className}};Gp.className="GlorotUniform";Q.registerClass(Gp);var Wp=class extends jr{constructor(e){super({scale:1,mode:"fanAvg",distribution:"normal",seed:e==null?null:e.seed})}getClassName(){return jr.className}};Wp.className="GlorotNormal";Q.registerClass(Wp);var jp=class extends jr{constructor(e){super({scale:2,mode:"fanIn",distribution:"normal",seed:e==null?null:e.seed})}getClassName(){return jr.className}};jp.className="HeNormal";Q.registerClass(jp);var Up=class extends jr{constructor(e){super({scale:2,mode:"fanIn",distribution:"uniform",seed:e==null?null:e.seed})}getClassName(){return jr.className}};Up.className="HeUniform";Q.registerClass(Up);var Hp=class extends jr{constructor(e){super({scale:1,mode:"fanIn",distribution:"normal",seed:e==null?null:e.seed})}getClassName(){return jr.className}};Hp.className="LeCunNormal";Q.registerClass(Hp);var qp=class extends jr{constructor(e){super({scale:1,mode:"fanIn",distribution:"uniform",seed:e==null?null:e.seed})}getClassName(){return jr.className}};qp.className="LeCunNormal";Q.registerClass(qp);var ud=class extends bn{constructor(e){super();if(this.DEFAULT_GAIN=1,this.gain=e.gain==null?this.DEFAULT_GAIN:e.gain,this.seed=e.seed,this.seed!=null)throw new Se("Random seed is not implemented for Orthogonal Initializer yet.")}apply(e,t){return V(()=>{if(e.length<2)throw new Se("Shape must be at least 2D.");e[0]*e[1]>2e3&&console.warn(`Orthogonal initializer is being called on a matrix with more than 2000 (${e[0]*e[1]}) elements: Slowness may result.`);let n=e[0]>e[1]?[e[1],e[0]]:e,o=Vp(n,0,1,"float32"),s=G1.gramSchmidt(o);return e[0]>e[1]&&(s=s.transpose()),O(this.gain,s)})}getConfig(){return{gain:this.gain,seed:this.seed}}};ud.className="Orthogonal";Q.registerClass(ud);var y3={constant:"Constant",glorotNormal:"GlorotNormal",glorotUniform:"GlorotUniform",heNormal:"HeNormal",heUniform:"HeUniform",identity:"Identity",leCunNormal:"LeCunNormal",leCunUniform:"LeCunUniform",ones:"Ones",orthogonal:"Orthogonal",randomNormal:"RandomNormal",randomUniform:"RandomUniform",truncatedNormal:"TruncatedNormal",varianceScaling:"VarianceScaling",zeros:"Zeros"};function b3(r,e={}){return pi(r,Q.SerializationMap.getMap().classNameMap,e,"initializer")}function It(r){return Mp(r)}function pt(r){if(typeof r=="string"){let e=r in y3?y3[r]:r;if(e==="GlorotNormal")return new Wp;if(e==="GlorotUniform")return new Gp;if(e==="HeNormal")return new jp;if(e==="HeUniform")return new Up;if(e==="LeCunNormal")return new Hp;if(e==="LeCunUniform")return new qp;{let t={};return t.className=e,t.config={},b3(t)}}else return r instanceof bn?r:b3(r)}function mee(){return new nd}function fee(){return new Xu}function dee(r){return new od(r)}function hee(r){return new sd(r)}function gee(r){return new id(r)}function xee(r){return new ad(r)}function yee(r){return new ld(r)}function bee(r){return new jr(r)}function wee(r){return new Gp(r)}function _ee(r){return new Wp(r)}function kee(r){return new jp(r)}function vee(r){return new Up(r)}function Cee(r){return new Hp(r)}function Iee(r){return new qp(r)}function Nee(r){return new ud(r)}var tB={};We(tB,{Layer:()=>Le,RNN:()=>$n,RNNCell:()=>Ol,activation:()=>Xte,add:()=>ore,alphaDropout:()=>Vre,average:()=>sre,averagePooling1d:()=>V0,averagePooling2d:()=>G0,averagePooling3d:()=>W0,avgPool1d:()=>dre,avgPool2d:()=>gre,avgPool3d:()=>yre,avgPooling1d:()=>hre,avgPooling2d:()=>xre,avgPooling3d:()=>bre,batchNormalization:()=>pre,bidirectional:()=>Rre,concatenate:()=>ire,conv1d:()=>Bte,conv2d:()=>Vte,conv2dTranspose:()=>Gte,conv3d:()=>Wte,conv3dTranspose:()=>jte,convLstm2d:()=>Are,convLstm2dCell:()=>Ere,cropping2D:()=>Hte,dense:()=>Yte,depthwiseConv2d:()=>Kte,dot:()=>cre,dropout:()=>Zte,elu:()=>Fte,embedding:()=>nre,flatten:()=>Qte,gaussianDropout:()=>Bre,gaussianNoise:()=>zre,globalAveragePooling1d:()=>wre,globalAveragePooling2d:()=>_re,globalMaxPool1d:()=>Ore,globalMaxPool2d:()=>Pre,globalMaxPooling1d:()=>Z3,globalMaxPooling2d:()=>J3,gru:()=>vre,gruCell:()=>Cre,input:()=>f0,inputLayer:()=>Rte,layerNormalization:()=>mre,leakyReLU:()=>Pte,lstm:()=>Ire,lstmCell:()=>Nre,masking:()=>Gre,maxPool1d:()=>Mre,maxPool2d:()=>Lre,maxPooling1d:()=>Q3,maxPooling2d:()=>eB,maxPooling3d:()=>kre,maximum:()=>are,minimum:()=>lre,multiply:()=>ure,permute:()=>rre,prelu:()=>Mte,reLU:()=>Ote,repeatVector:()=>ere,reshape:()=>tre,rnn:()=>Dre,separableConv2d:()=>Ute,simpleRNN:()=>Sre,simpleRNNCell:()=>Tre,softmax:()=>Lte,spatialDropout1d:()=>Jte,stackedRNNCells:()=>$re,thresholdedReLU:()=>zte,timeDistributed:()=>Fre,upSampling2d:()=>qte,zeroPadding2d:()=>fre});var See=0;function Gx(){return See++}var Wx={};function Sl(r=""){return r in Wx||(Wx[r]=0),Wx[r]+=1,r+Wx[r].toString()}function jx(r){return Array.isArray(r)&&Array.isArray(r[0])}function Kp(r){return r.length===0?[]:Array.isArray(r[0])?r:[r]}function Fe(r){let e;if(Array.isArray(r)){if(r.length!==1)throw new z(`Expected Tensor length to be 1; got ${r.length}`);e=r[0]}else e=r;return e}function Xe(r){if(Array.isArray(r)&&Array.isArray(r[0])){if(r.length===1)return r=r,r[0];throw new z(`Expected exactly 1 Shape; got ${r.length}`)}else return r}function Xp(r){let e=0;for(let t of r)t.shape.length===0?e+=1:e+=t.shape.reduce((n,o)=>n*o);return e}var _3="Variable",Ux=class{constructor(e,t="float32",n=_3,o=!0,s=null){this.dtype=t==null?"float32":t,this.shape=e.shape,this.id=Gx(),n=n==null?_3:n,this.originalName=Mx(n),this.name=Lx(this.originalName),this.trainable_=o,this.constraint=s,this.val=I_(e,this.trainable_,this.name,this.dtype)}read(){return this.assertNotDisposed(),this.val}write(e){return this.assertNotDisposed(),Tee(this.val,e),this.val.id!==e.id&&(this.val.assign(e),this.constraint!=null&&this.val.assign(this.constraint.apply(this.val))),this}dispose(){this.assertNotDisposed(),this.val.dispose()}assertNotDisposed(){if(this.val.isDisposed)throw new Error(`LayersVariable ${this.name} is already disposed.`)}get trainable(){return this.trainable_}set trainable(e){this.trainable_=e,this.val.trainable=e}};function Tee(r,e){if(r.shape.toString()!==e.shape.toString())throw new Error("Shape mismatch: "+JSON.stringify(r.shape)+" vs. "+JSON.stringify(e.shape))}function cd(r){return r.map(e=>e.read())}function Yp(r){r.forEach(e=>{e[0].write(e[1])})}var Nt=class{constructor(e){this.dtype=e.dtype,this.shape=e.shape,e.shape!=null?this.ndim=e.shape.length:this.ndim=e.ndim,this.maxNDim=e.maxNDim,this.minNDim=e.minNDim,this.axes=e.axes||{}}},sn=class{constructor(e,t,n,o,s,a,i){this.dtype=e,this.shape=t,this.sourceLayer=n,this.inputs=o,this.callArgs=s,this.outputTensorIndex=i,this.id=Gx(),a!=null&&(this.originalName=Mx(a),this.name=Lx(this.originalName)),this.rank=t.length}},Aee=0,Tl=class{constructor(e,t){this.callArgs=t,this.id=Aee++,this.outboundLayer=e.outboundLayer,this.inboundLayers=e.inboundLayers,this.nodeIndices=e.nodeIndices,this.tensorIndices=e.tensorIndices,this.inputTensors=e.inputTensors,this.outputTensors=e.outputTensors,this.inputMasks=e.inputMasks,this.outputMasks=e.outputMasks,this.inputShapes=e.inputShapes,this.outputShapes=e.outputShapes;for(let n of e.inboundLayers)n!=null&&n.outboundNodes.push(this);e.outboundLayer.inboundNodes.push(this)}getConfig(){let e=[];for(let t of this.inboundLayers)t!=null?e.push(t.name):e.push(null);return{outboundLayer:this.outboundLayer?this.outboundLayer.name:null,inboundLayers:e,nodeIndices:this.nodeIndices,tensorIndices:this.tensorIndices}}},Eee=0,Le=class extends Q.Serializable{constructor(e={}){super();this._callHook=null,this._addedWeightNames=[],this._stateful=!1,this.id=Eee++,this.activityRegularizer=null,this.inputSpec=null,this.supportsMasking=!1,this._trainableWeights=[],this._nonTrainableWeights=[],this._losses=[],this._updates=[],this._built=!1,this.inboundNodes=[],this.outboundNodes=[];let t=e.name;if(!t){let n=this.getClassName();t=mo(n)+"_"+Sl(n)}if(this.name=t,this.trainable_=e.trainable==null?!0:e.trainable,e.inputShape!=null||e.batchInputShape!=null){let n;if(e.batchInputShape!=null)n=e.batchInputShape;else if(e.inputShape!=null){let s=null;e.batchSize!=null&&(s=e.batchSize),n=[s].concat(e.inputShape)}this.batchInputShape=n;let o=e.dtype;o==null&&(o=e.inputDType),o==null&&(o="float32"),this.dtype=o}e.weights!=null?this.initialWeights=e.weights:this.initialWeights=null,this._refCount=null,this.fastWeightInitDuringBuild=!1}static nodeKey(e,t){return e.name+"_ib-"+t.toString()}getNodeAtIndex(e,t){if(this.inboundNodes.length===0)throw new Gr(`The layer has never been called and thus has no defined ${t}.`);if(this.inboundNodes.length<=e)throw new z(`Asked to get ${t} at node ${e}, but the layer has only ${this.inboundNodes.length} inbound nodes.`);return this.inboundNodes[e]}getInputAt(e){return wr(this.getNodeAtIndex(e,"input").inputTensors)}getOutputAt(e){return wr(this.getNodeAtIndex(e,"output").outputTensors)}get input(){if(this.inboundNodes.length>1)throw new Dn(`Layer ${this.name} has multiple inbound nodes, hence the notion of "layer input" is ill-defined. Use \`getInputAt(nodeIndex)\` instead.`);if(this.inboundNodes.length===0)throw new Dn(`Layer ${this.name} is not connected, no input to return.`);return wr(this.getNodeAtIndex(0,"input").inputTensors)}get output(){if(this.inboundNodes.length===0)throw new Dn(`Layer ${this.name} has no inbound nodes.`);if(this.inboundNodes.length>1)throw new Dn(`Layer ${this.name} has multiple inbound nodes, hence the notion of "layer output" is ill-defined. Use \`getOutputAt(nodeIndex)\` instead.`);return wr(this.getNodeAtIndex(0,"output").outputTensors)}get losses(){return this._losses}calculateLosses(){return this.losses.map(e=>e())}get updates(){return this._updates}get built(){return this._built}set built(e){this._built=e}get trainable(){return this.trainable_}set trainable(e){this._trainableWeights.forEach(t=>t.trainable=e),this.trainable_=e}get trainableWeights(){return this.trainable_?this._trainableWeights.filter(e=>e.trainable):[]}set trainableWeights(e){this._trainableWeights=e}get nonTrainableWeights(){return this.trainable?this._trainableWeights.filter(e=>!e.trainable).concat(this._nonTrainableWeights):this._trainableWeights.concat(this._nonTrainableWeights)}set nonTrainableWeights(e){this._nonTrainableWeights=e}get weights(){return this.trainableWeights.concat(this.nonTrainableWeights)}get stateful(){return this._stateful}resetStates(){if(!this.stateful)throw new Error("Cannot call the resetStates() method of a non-stateful Layer object.")}assertInputCompatibility(e){if(e=bt(e),this.inputSpec==null||this.inputSpec.length===0)return;let t=bt(this.inputSpec);if(e.length!==t.length)throw new z(`Layer ${this.name} expects ${t.length} inputs, but it received ${e.length} input tensors. Input received: ${e}`);for(let n=0;n<e.length;n++){let o=e[n],s=t[n];if(s==null)continue;let a=o.rank;if(s.ndim!=null&&a!==s.ndim)throw new z(`Input ${n} is incompatible with layer ${this.name}: expected ndim=${s.ndim}, found ndim=${a}`);if(s.maxNDim!=null&&a>s.maxNDim)throw new z(`Input ${n} is incompatible with layer ${this.name}: expected max_ndim=${s.maxNDim}, found ndim=${a}`);if(s.minNDim!=null&&a<s.minNDim)throw new z(`Input ${n} is incompatible with layer ${this.name}: expected min_ndim=${s.minNDim}, found ndim=${a}.`);if(s.dtype!=null&&o.dtype!==s.dtype)throw new z(`Input ${n} is incompatible with layer ${this.name} : expected dtype=${s.dtype}, found dtype=${o.dtype}.`);if(s.axes){let i=o.shape;for(let l in s.axes){let u=Number(l),c=s.axes[l],p=u>=0?i[u]:i[i.length+u];if(c!=null&&[c,null].indexOf(p)===-1)throw new z(`Input ${n} is incompatible with layer ${this.name}: expected axis ${u} of input shape to have value ${c} but got shape ${i}.`)}}if(s.shape!=null)for(let i=0;i<s.shape.length;++i){let l=s.shape[i],u=o.shape[i];if(l!=null&&u!=null&&l!==u)throw new z(`Input ${n} is incompatible with layer ${this.name}: expected shape=${s.shape}, found shape=${o.shape}.`)}}}call(e,t){return e}invokeCallHook(e,t){this._callHook!=null&&this._callHook(e,t)}setCallHook(e){this._callHook=e}clearCallHook(){this._callHook=null}apply(e,t){t=t||{},this.assertNotDisposed();let n=bt(e),o=!0;for(let a of n)if(!(a instanceof sn)){o=!1;break}let s=!0;for(let a of n)if(a instanceof sn){s=!1;break}if(o===s)throw new z("Arguments to apply() must be all SymbolicTensors or all Tensors");return Ts(this.name,()=>{if(!this.built){this.assertInputCompatibility(e);let a=[];for(let i of bt(e))a.push(i.shape);this.build(wr(a)),this.built=!0,this.initialWeights&&this.setWeights(this.initialWeights),this._refCount===null&&s&&(this._refCount=1)}if(this.assertInputCompatibility(e),s){let a=this.call(e,t),i=bt(a),l=[];for(let u of i)n.indexOf(u)!==-1&&(u=u.clone()),l.push(u);if(a=wr(l),this.activityRegularizer!=null)throw new Se("Layer invocation in the presence of activity regularizer(s) is not supported yet.");return a}else{let a=Dee(e),i=this.computeOutputShape(a),l,u=$ee(e);if(this.warnOnIncompatibleInputShape(Array.isArray(e)?a[0]:a),i!=null&&i.length>0&&Array.isArray(i[0])?l=i.map((c,p)=>new sn(u,c,this,bt(e),t,this.name,p)):l=new sn(u,i,this,bt(e),t,this.name),this.addInboundNode(e,l,null,null,a,i,t),this._refCount++,this.activityRegularizer!=null)throw new Se("Layer invocation in the presence of activity regularizer(s) is not supported yet.");return l}})}warnOnIncompatibleInputShape(e){if(this.batchInputShape!=null)if(e.length!==this.batchInputShape.length)console.warn(`The rank of the input tensor provided (shape: ${JSON.stringify(e)}) does not match that of the batchInputShape (${JSON.stringify(this.batchInputShape)}) of the layer ${this.name}`);else{let t=!1;this.batchInputShape.forEach((n,o)=>{n!=null&&e[o]!=null&&e[o]!==n&&(t=!0)}),t&&console.warn(`The shape of the input tensor (${JSON.stringify(e)}) does not match the expectation of layer ${this.name}: ${JSON.stringify(this.batchInputShape)}`)}}get outputShape(){if(this.inboundNodes==null||this.inboundNodes.length===0)throw new Dn(`The layer ${this.name} has never been called and thus has no defined output shape.`);let e=[];for(let t of this.inboundNodes){let n=JSON.stringify(t.outputShapes);e.indexOf(n)===-1&&e.push(n)}if(e.length===1){let t=this.inboundNodes[0].outputShapes;return Array.isArray(t)&&Array.isArray(t[0])&&t.length===1?t[0]:t}else throw new Dn(`The layer ${this.name} has multiple inbound nodes with different output shapes. Hence the notion of "output shape" is ill-defined for the layer.`)}countParams(){if(!this.built)throw new Gr(`You tried to call countParams() on ${this.name}, but the layer is not built yet. Build it first by calling build(batchInputShape).`);return Xp(this.weights)}build(e){this.built=!0}getWeights(e=!1){return cd(e?this.trainableWeights:this.weights)}setWeights(e){V(()=>{let t=this.weights;if(t.length!==e.length)throw new z(`You called setWeights(weights) on layer "${this.name}" with a weight list of length ${e.length}, but the layer was expecting ${t.length} weights. Provided weights: ${e}...`);if(t.length===0)return;let n=[],o=cd(t);for(let s=0;s<o.length;++s){let a=o[s],i=t[s],l=e[s];if(!y.arraysEqual(a.shape,l.shape))throw new z(`Layer weight shape ${a.shape} not compatible with provided weight shape ${l.shape}`);n.push([i,l])}Yp(n)})}addWeight(e,t,n,o,s,a,i){if(this._addedWeightNames.indexOf(e)!==-1)throw new z(`Duplicate weight name ${e} for layer ${this.name}`);this._addedWeightNames.push(e),n==null&&(n="float32"),this.fastWeightInitDuringBuild&&(o=pt("zeros"));let l=o.apply(t,n),u=new Ux(l,n,e,a,i);return l.dispose(),s!=null&&this.addLoss(()=>s.apply(u.read())),a==null&&(a=!0),a?this._trainableWeights.push(u):this._nonTrainableWeights.push(u),u}setFastWeightInitDuringBuild(e){this.fastWeightInitDuringBuild=e}addLoss(e){e==null||Array.isArray(e)&&e.length===0||(e=bt(e),this._losses!==void 0&&this._losses!==null&&this.losses.push(...e))}computeOutputShape(e){return e}computeMask(e,t){if(!this.supportsMasking){if(t!=null)if(Array.isArray(t))t.forEach(n=>{if(n!=null)throw new TypeError(`Layer ${this.name} does not support masking, but was passed an inputMask.`)});else throw new TypeError(`Layer ${this.name} does not support masking, but was passed an inputMask.`);return null}return t}addInboundNode(e,t,n,o,s,a,i=null){let l=bt(e);t=bt(t),n=bt(n),o=bt(o),s=Kp(s),a=Kp(a);let u=[],c=[],p=[];for(let m of l)u.push(m.sourceLayer),c.push(m.nodeIndex),p.push(m.tensorIndex);new Tl({outboundLayer:this,inboundLayers:u,nodeIndices:c,tensorIndices:p,inputTensors:l,outputTensors:t,inputMasks:n,outputMasks:o,inputShapes:s,outputShapes:a},i);for(let m=0;m<t.length;m++)t[m].sourceLayer=this,t[m].nodeIndex=this.inboundNodes.length-1,t[m].tensorIndex=m}getConfig(){let e={name:this.name,trainable:this.trainable};return this.batchInputShape!=null&&(e.batchInputShape=this.batchInputShape),this.dtype!=null&&(e.dtype=this.dtype),e}disposeWeights(){return this.weights.forEach(e=>e.dispose()),this.weights.length}assertNotDisposed(){if(this._refCount===0)throw new Error(`Layer '${this.name}' is already disposed.`)}dispose(){if(!this.built)throw new Error(`Cannot dispose Layer ${this.name} because it has not been built yet.`);if(this._refCount===null)throw new Error(`Cannot dispose Layer ${this.name} because it has not been used yet.`);this.assertNotDisposed();let e=0;return--this._refCount==0&&(e=this.disposeWeights()),{refCountAfterDispose:this._refCount,numDisposedVariables:e}}};function Dee(r){r=bt(r);let e=[];for(let t of r)e.push(t.shape);return wr(e)}function $ee(r){return"float32"}function t0(r,e,t){if((e==null||t!=null&&t>0)&&(e=r.sourceLayer,t=r.nodeIndex),e.inboundNodes.length===0)return[r];{let n=e.inboundNodes[t];if(n.inboundLayers.length===0)return n.inputTensors;{let o=[];for(let s=0;s<n.inboundLayers.length;s++){let a=n.inputTensors[s],i=n.inboundLayers[s],l=n.nodeIndices[s],u=t0(a,i,l);for(let c of u)o.indexOf(c)===-1&&o.push(c)}return o}}}var fi=class extends Le{constructor(e){super({dtype:e.dtype,name:e.name!=null?e.name:Sl("input").toString()});if(e.batchSize==null&&(e.batchSize=null),e.sparse==null&&(e.sparse=!1),this.trainable=!1,this.built=!0,this.sparse=e.sparse,e.inputShape!=null&&e.batchInputShape!=null)throw new z("Only provide the inputShape OR batchInputShape argument to inputLayer, not both at the same time.");let t=e.batchInputShape;if(t==null){if(e.inputShape==null)throw new z("An InputLayer should be passed either a `batchInputShape` or an `inputShape`.");t=[e.batchSize].concat(e.inputShape)}else if(e.batchSize!=null)throw new z("Cannot specify batchSize if batchInputShape is specified when creating an InputLayer.");let n=e.dtype||"float32";this.batchInputShape=t,this.dtype=n,this.inputSpec=[{shape:t}];let o=new sn(this.dtype,this.batchInputShape,this,[],{},this.name);o.nodeIndex=0,o.tensorIndex=0,new Tl({outboundLayer:this,inboundLayers:[],nodeIndices:[],tensorIndices:[],inputTensors:[o],outputTensors:[o],inputMasks:[null],outputMasks:[null],inputShapes:[t],outputShapes:[t]})}apply(e,t){throw new z(`Cannot pass any input to an InputLayer's apply() method. InputLayer name: ${this.name}`)}dispose(){return{refCountAfterDispose:this._refCount,numDisposedVariables:0}}getConfig(){return{batchInputShape:this.batchInputShape,dtype:this.dtype,sparse:this.sparse,name:this.name}}};fi.className="InputLayer";Q.registerClass(fi);function Hx(r){if(r.batchShape==null&&r.shape==null)throw new Error("Please provide to Input either a `shape` or a `batchShape` argument. Note that `shape` does not include the batch dimension.");if(r.batchShape!=null&&r.shape!=null)throw new z("Please provide either a `shape` or `batchShape` argument to Input, but not both.");let e=r.batchShape;r.shape!=null&&e==null&&(e=[null].concat(r.shape));let t=r.dtype;return t==null&&(t="float32"),new fi({batchInputShape:e,name:r.name,dtype:t,sparse:r.sparse}).inboundNodes[0].outputTensors[0]}async function di(r){if(r==null)return;let e=[],t=[],n=[];for(let o in r){let s=r[o];if(typeof s!="number"){let a=s;e.push(a.data()),t.push(o),n.push(a)}}if(e.length>0){let o=await Promise.all(e);for(let s=0;s<o.length;++s)r[t[s]]=o[s][0];Ee(n)}}function qx(r){if(r!=null)for(let e in r){let t=r[e];typeof t!="number"&&t.dispose()}}var k3;(function(r){r[r.SILENT=0]="SILENT",r[r.VERBOSE=1]="VERBOSE"})(k3||(k3={}));var Ree=125,Al=class{constructor(){this.validationData=null}setParams(e){this.params=e}async onEpochBegin(e,t){}async onEpochEnd(e,t){}async onBatchBegin(e,t){}async onBatchEnd(e,t){}async onTrainBegin(e){}async onTrainEnd(e){}setModel(e){}},r0=class{constructor(e,t=10){e==null&&(e=[]),this.callbacks=e,this.queueLength=t}append(e){this.callbacks.push(e)}setParams(e){for(let t of this.callbacks)t.setParams(e)}setModel(e){for(let t of this.callbacks)t.setModel(e)}async onEpochBegin(e,t){t==null&&(t={});for(let n of this.callbacks)await n.onEpochBegin(e,t)}async onEpochEnd(e,t){t==null&&(t={});for(let n of this.callbacks)await n.onEpochEnd(e,t)}async onBatchBegin(e,t){t==null&&(t={});for(let n of this.callbacks)await n.onBatchBegin(e,t)}async onBatchEnd(e,t){t==null&&(t={});for(let n of this.callbacks)await n.onBatchEnd(e,t)}async onTrainBegin(e){e==null&&(e={});for(let t of this.callbacks)await t.onTrainBegin(e)}async onTrainEnd(e){e==null&&(e={});for(let t of this.callbacks)await t.onTrainEnd(e)}},v3=class extends Al{constructor(){super()}async onEpochBegin(e){this.seen=0,this.totals={}}async onBatchEnd(e,t){t==null&&(t={});let n=t.size==null?0:t.size;this.seen+=n;for(let o in t){let s=t[o];if(typeof s=="number")this.totals.hasOwnProperty(o)||(this.totals[o]=0),this.totals[o]=this.totals[o]+s*n;else{let a;o in this.totals?a=this.totals[o]:this.totals[o]=0;let i=V(()=>J(this.totals[o],O(s,n)));this.totals[o]=i,a!=null&&a.dispose()}}}async onEpochEnd(e,t){if(t!=null)for(let n of this.params.metrics)this.totals[n]!=null&&(typeof this.totals[n]=="number"?t[n]=this.totals[n]/this.seen:V(()=>{let o=O(pe(1,this.seen),this.totals[n]);t[n]=o,this.totals[n].dispose(),$t(t[n])}))}},n0=class extends Al{async onTrainBegin(e){this.epoch=[],this.history={}}async onEpochEnd(e,t){t==null&&(t={}),this.epoch.push(e);for(let n in t)this.history[n]==null&&(this.history[n]=[]),this.history[n].push(t[n])}async syncData(){let e=[],t=[],n=[];for(let s in this.history){let a=this.history[s];for(let i=0;i<a.length;++i)if(typeof a[i]!="number"){let l=a[i];e.push(l.data()),t.push(s),n.push(i)}}let o=await Promise.all(e);for(let s=0;s<o.length;++s)this.history[t[s]][n[s]].dispose(),this.history[t[s]][n[s]]=o[s][0]}},o0=class extends Al{constructor(e,t){super();if(this.currentEpoch=0,this.yieldEvery=t||"auto",this.yieldEvery==="auto"&&(this.yieldEvery=Ree),this.yieldEvery==="never"&&e.onYield!=null)throw new Error("yieldEvery is `never` but you provided an `onYield` callback. Either change `yieldEvery` or remove the callback");y.isNumber(this.yieldEvery)&&(this.maybeWait=Yz(this.maybeWait.bind(this),this.yieldEvery)),this.trainBegin=e.onTrainBegin,this.trainEnd=e.onTrainEnd,this.epochBegin=e.onEpochBegin,this.epochEnd=e.onEpochEnd,this.batchBegin=e.onBatchBegin,this.batchEnd=e.onBatchEnd,this.yield=e.onYield}async maybeWait(e,t,n){let o=[];this.yield!=null&&(await di(n),o.push(this.yield(e,t,n))),o.push(vf()),await Promise.all(o)}async onEpochBegin(e,t){this.currentEpoch=e,this.epochBegin!=null&&(await di(t),await this.epochBegin(e,t))}async onEpochEnd(e,t){let n=[];this.epochEnd!=null&&(await di(t),n.push(this.epochEnd(e,t))),this.yieldEvery==="epoch"&&n.push(vf()),await Promise.all(n)}async onBatchBegin(e,t){this.batchBegin!=null&&(await di(t),await this.batchBegin(e,t))}async onBatchEnd(e,t){let n=[];this.batchEnd!=null&&(await di(t),n.push(this.batchEnd(e,t))),this.yieldEvery==="batch"?n.push(vf()):y.isNumber(this.yieldEvery)&&n.push(this.maybeWait(this.currentEpoch,e,t)),await Promise.all(n)}async onTrainBegin(e){this.trainBegin!=null&&(await di(e),await this.trainBegin(e))}async onTrainEnd(e){this.trainEnd!=null&&(await di(e),await this.trainEnd(e))}};function Kx(r,e){return r==null&&(r={}),r instanceof Al?[r]:Array.isArray(r)&&r[0]instanceof Al?r:bt(r).map(n=>new o0(n,e))}var wn=class{constructor(){}static registerCallbackConstructor(e,t){y.assert(e>=0&&Number.isInteger(e),()=>`Verbosity level is expected to be an integer >= 0, but got ${e}`),wn.checkForDuplicate(t),wn.constructors[e]==null&&(wn.constructors[e]=[]),wn.constructors[e].push(t)}static checkForDuplicate(e){for(let t in wn.constructors)wn.constructors[+t].forEach(o=>{if(o===e)throw new z("Duplicate callback constructor.")})}static clear(){wn.constructors={}}static createCallbacks(e){let t=[];for(let n in wn.constructors){let o=+n;e>=o&&t.push(...wn.constructors[o])}return t.map(n=>new n)}};wn.constructors={};function Xx(r,e,t,n,o,s,a,i,l){let u=new n0,c=[new v3,...wn.createCallbacks(e)];r!=null&&c.push(...r),c.push(u);let p=new r0(c);return p.setParams({epochs:t,initialEpoch:n,samples:o,steps:s,batchSize:a,verbose:e,doValidation:i,metrics:l}),{callbackList:p,history:u}}function an(r,e={},t=!1){return pi(r,Q.SerializationMap.getMap().classNameMap,e,"layer",t)}function pd(r,e){return V(()=>{r.dtype!=="float32"&&(r=r.asType("float32"));let t=fe(Ku(r),e,!0),n=bs(t.shape,rr()),o=gt(en(t,n));return pe(r,o)})}function hi(r,e){return V(()=>dt(Ku(le(e,r)),-1))}function Zp(r,e){return V(()=>dt(Tt(le(e,r)),-1))}function El(r,e){return V(()=>{let t=le(r,e),n=ur(Tt(r),rr(),Number.MAX_VALUE),o=Tt(pe(t,n));return O(100,dt(o,-1))})}function Fee(r,e){return V(()=>{let t=ur(e,rr(),Number.MAX_VALUE),n=pr(J(1,t)),o=ur(r,rr(),Number.MAX_VALUE),s=pr(J(1,o));return dt(Ku(le(n,s)),-1)})}function Oee(r,e){return V(()=>{let t=en(0,le(1,O(r,e)));return dt(Ku(t),-1)})}function Pee(r,e){return V(()=>{let t=en(0,le(1,O(r,e)));return dt(t,-1)})}function Mee(r,e){return V(()=>{let t=fe(O(r,e),-1),n=mr(O(le(1,r),e),-1);return en(0,J(1,le(n,t)))})}function Lee(r,e){return V(()=>{let t=Math.log(2),n=le(e,r),o=le(J(n,eo(O(-2,n))),t);return dt(o,-1)})}function Yu(r,e,t=!1){return V(()=>{if(t)e=Ia(e);else{let n=fe(e,e.shape.length-1,!0);e=pe(e,n)}return e=ur(e,rr(),1-rr()),qe(fe(O(r.toFloat(),pr(e)),e.shape.length-1))})}function Jp(r,e,t=!1){return V(()=>{let n=ws(p3(r)).toInt();e=ur(e,rr(),1-rr());let o=e.shape,s=ds(n,o[o.length-1]).reshape(o);return Yu(s,e,t)})}function zee(r,e){if(!y.arraysEqual(r.shape,e.shape))throw new z(`logits and labels must have the same shape, but got shapes ${JSON.stringify(r.shape)} and ${JSON.stringify(e.shape)}`);return V(()=>{let t=e.relu(),n=e.abs().neg();return t.sub(e.mul(r)).add(n.exp().log1p())})}function Qp(r,e){return V(()=>{let t;return t=ur(e,rr(),1-rr()),t=pr(pe(t,le(1,t))),dt(zee(r,t),-1)})}function Bee(r,e){return V(()=>{let t=ur(r,rr(),1),n=ur(e,rr(),1);return fe(O(r,pr(pe(t,n))),-1)})}function Vee(r,e){return V(()=>{let t=pr(J(rr(),e));return dt(le(e,O(r,t)),-1)})}function md(r,e){return V(()=>{let t=pd(r,-1),n=pd(e,-1),o=O(t,n);return qe(fe(o,-1))})}var fd={meanSquaredError:hi,meanAbsoluteError:Zp,meanAbsolutePercentageError:El,meanSquaredLogarithmicError:Fee,squaredHinge:Oee,hinge:Pee,categoricalHinge:Mee,logcosh:Lee,categoricalCrossentropy:Yu,sparseCategoricalCrossentropy:Jp,binaryCrossentropy:Qp,kullbackLeiblerDivergence:Bee,poisson:Vee,cosineProximity:md};function Yx(r){if(typeof r=="string"){if(r in fd)return fd[r];let e=`Unknown loss ${r}`;throw r.toLowerCase().includes("softmaxcrossentropy")&&(e=`Unknown loss ${r}. Use "categoricalCrossentropy" as the string name for tf.losses.softmaxCrossEntropy`),new z(e)}else return r}function dd(r,e){return V(()=>{let t=O(.5,sr(e)),n=Ma(Vt(e,t),r.dtype);return dt(Sn(r,n),-1)})}function hd(r,e){return V(()=>Ma(Sn(hl(r,-1),hl(e,-1)),"float32"))}function C3(r,e){return V(()=>yr(r.equal(1),e.equal(1)).sum().cast("float32"))}function Gee(r,e){return V(()=>yr(r.equal(1),e.equal(0)).sum().cast("float32"))}function Wee(r,e){return V(()=>yr(r.equal(0),e.equal(1)).sum().cast("float32"))}function s0(r,e){return V(()=>{let t=C3(r,e),n=Wee(r,e),o=t.add(n);return vt(Vt(o,0),t.div(o),0).cast("float32")})}function I3(r,e){return V(()=>{let t=C3(r,e),n=Gee(r,e),o=t.add(n);return vt(Vt(o,0),t.div(o),0).cast("float32")})}function Zx(r,e){return Qp(r,e)}function Jx(r,e){return r.rank===e.rank&&(r=r.squeeze([r.rank-1])),e=e.argMax(-1),e.dtype!==r.dtype&&(e=e.asType(r.dtype)),Sn(r,e).asType("float32")}var jee=hi,Uee=hi,Hee=Zp,qee=Zp,Kee=El,Xee=El,gd=Yu,Yee=md,i0=Jp,Qx={binaryAccuracy:dd,categoricalAccuracy:hd,precision:s0,categoricalCrossentropy:gd,sparseCategoricalCrossentropy:i0,mse:jee,MSE:Uee,mae:Hee,MAE:qee,mape:Kee,MAPE:Xee,cosine:Yee};function N3(r){if(typeof r=="string"&&r in Qx)return Qx[r];if(typeof r!="string"&&r!=null)return r;throw new z(`Unknown metric ${r}`)}function xd(r){if(Gn(r!==null,`Unknown LossOrMetricFn ${r}`),typeof r=="string")return r;{let e;for(let t of Object.keys(fd))if(fd[t]===r){e=t;break}if(e!==void 0)return e;for(let t of Object.keys(Qx))if(Qx[t]===r){e=t;break}return e!==void 0?e:r.name}}function S3(r){let e={Adagrad:()=>Tu.adagrad(.01),Adadelta:()=>Tu.adadelta(1,.95,rr()),Adam:()=>Tu.adam(.001,.9,.999,rr()),Adamax:()=>Tu.adamax(.002,.9,.999,rr(),0),RMSProp:()=>Tu.rmsprop(.001,.9,0,rr()),SGD:()=>Tu.sgd(.01)};if(e.adagrad=e.Adagrad,e.adadelta=e.Adadelta,e.adam=e.Adam,e.adamax=e.Adamax,e.rmsprop=e.RMSProp,e.sgd=e.SGD,r in e)return e[r]();throw new z(`Unknown Optimizer ${r}`)}var T3=1*1024*1024;function a0(r,e,t=!1){if(r==null||typeof r!="object"||Object.getPrototypeOf(r)!==Object.prototype||!l0(r))throw new Error("User-defined metadata is expected to be a JSON object, but is not.");if(t){let n=JSON.stringify(r);n.length>T3&&console.warn(`User-defined metadata of model "${e}" is too large in size (length=${n.length} when serialized). It is not recommended to store such large objects in user-defined metadata. Please make sure its serialized length is <= ${T3}.`)}}function l0(r){if(r===null)return!0;if(typeof r=="object")if(Object.getPrototypeOf(r)===Object.prototype){let e=Object.keys(r);for(let t of e)if(typeof t!="string"||!l0(r[t]))return!1;return!0}else if(Array.isArray(r)){for(let e of r)if(!l0(e))return!1;return!0}else return!1;else{let e=typeof r;return e==="string"||e==="number"||e==="boolean"}}function A3(r,e,t,n=console.log){let o=Jee(r),s=["Layer (type)","Output shape","Param #"];o?(e=e||65,t=t||[.45,.85,1]):(e=e||98,t=t||[.33,.55,.67,1]),t[t.length-1]<=1&&(t=t.map(c=>Math.floor(e*c)));let a;if(!o){s.push("Receives inputs"),a=[];for(let c in r.nodesByDepth)a.push(...r.nodesByDepth[c])}n("_".repeat(e)),ey(s,t,n),n("=".repeat(e));let i=r.layers;for(let c=0;c<i.length;++c)o?Qee(i[c],t,n):ete(i[c],t,a,n),n((c===i.length-1?"=":"_").repeat(e));r.checkTrainableWeightsConsistency();let l=Zee(r),u=Xp(r.nonTrainableWeights);n(`Total params: ${l+u}`),n(`Trainable params: ${l}`),n(`Non-trainable params: ${u}`),n("_".repeat(e))}function Zee(r){let e;return r.collectedTrainableWeights!=null?e=Xp(r.collectedTrainableWeights):e=Xp(r.trainableWeights),e}function Jee(r){let e=!0,t=[],n=[];for(let o in r.nodesByDepth)t.push(r.nodesByDepth[o]);for(let o of t){if(o.length>1||o.length===1&&o[0].inboundLayers.length>1){e=!1;break}n.push(...o)}if(e)for(let o of r.layers){let s=!1;for(let a of o.inboundNodes)if(n.indexOf(a)!==-1)if(s){e=!1;break}else s=!0;if(!e)break}return e}function ey(r,e,t=console.log){let n="";for(let o=0;o<r.length;++o)o>0&&(n=n.slice(0,n.length-1)+" "),n+=r[o],n=n.slice(0,e[o]),n+=" ".repeat(e[o]-n.length);t(n)}function Qee(r,e,t){let n;try{n=JSON.stringify(r.outputShape)}catch(i){n="multiple"}let o=r.name,s=r.getClassName(),a=[`${o} (${s})`,n,r.countParams().toString()];ey(a,e,t)}function ete(r,e,t,n){let o;try{o=JSON.stringify(r.outputShape)}catch(c){o="multiple"}let s=[];for(let c of r.inboundNodes)if(!(t!=null&&t.length>0&&t.indexOf(c)===-1))for(let p=0;p<c.inboundLayers.length;++p){let m=c.inboundLayers[p].name,f=c.nodeIndices[p],d=c.tensorIndices[p];s.push(`${m}[${f}][${d}]`)}let a=r.name,i=r.getClassName(),l=s.length===0?"":s[0],u=[`${a} (${i})`,o,r.countParams().toString(),l];ey(u,e,n);for(let c=1;c<s.length;++c)ey(["","","",s[c]],e,n)}function E3(r,e,t){return(r==="inboundNodes"||r==="outputLayers"||r==="inputLayers")&&e===0&&typeof t=="string"}function Zu(r,e){if(r===null)return null;if(typeof r=="string")return Pa(r);if(typeof r=="number"||typeof r=="boolean")return r;if(r instanceof Array){let t=[],n=r.length;for(let o=0;o<n;++o){let s=r[o];E3(e,o,s)?t.push(s):t.push(Zu(s,e))}return t}else{let t={};for(let n of Object.keys(r)){let o=r[n];if(n==="name"&&typeof o=="string")t[n]=o;else{let s=Pa(n);t[s]=Zu(o,s)}}return t}}function ty(r,e){if(r==null)return null;if(typeof r=="string")return mo(r);if(typeof r=="number"||typeof r=="boolean")return r;if(r instanceof Array){let t=[],n=r.length;for(let o=0;o<n;++o){let s=r[o];E3(e,o,s)?t.push(s):t.push(ty(s,e))}return t}else{let t={};for(let n of Object.keys(r)){let o=r[n],s=mo(n);(n==="name"||n==="className")&&typeof o=="string"?t[s]=o:t[s]=ty(o,n)}return t}}var yd="3.6.0";function tte(r,e){if(r.dtype==null||r.dtype===e.dtype)return e;try{return oe(e,r.dtype)}catch(t){throw new z(`The dtype of the feed (${e.dtype}) can not be cast to the dtype of the key '${r.name}' (${r.dtype}).`)}}var Es=class{constructor(e){if(this.id2Value={},this.id2Mask={},this.name2Id={},e instanceof Es)for(let t in e.id2Value)this.id2Value[t]=e.id2Value[t],t in e.id2Mask&&(this.id2Mask[t]=e.id2Mask[t]);else{if(e==null)return;for(let t of e)this.add(t.key,t.value)}}add(e,t,n){if(this.id2Value[e.id]==null)this.id2Value[e.id]=tte(e,t),this.name2Id[e.name]=e.id,n!=null&&(this.id2Mask[e.id]=n);else throw new z(`Duplicate key: name=${e.name}, id=${e.id}`);return this}addFeed(e){this.add(e.key,e.value)}hasKey(e){return this.id2Value[e.id]!=null}names(){return Object.keys(this.name2Id)}getValue(e){if(e instanceof sn){if(this.id2Value[e.id]==null)throw new z(`Nonexistent key: ${e.name}`);return this.id2Value[e.id]}else{let t=this.name2Id[e];if(t==null)throw new z(`Feed dict has no SymbolicTensor name: ${e}`);return this.id2Value[t]}}getMask(e){if(e instanceof sn){if(this.id2Value[e.id]==null)throw new z(`Nonexistent key: ${e.name}`);return this.id2Mask[e.id]}else{let t=this.name2Id[e];if(t==null)throw new z(`Feed dict has no SymbolicTensor name: ${e}`);return this.id2Mask[t]}}disposeMasks(){this.id2Mask!=null&&Ee(this.id2Mask)}},u0={},D3={};function Ju(r,e,t,n){let o=t==null?!1:t.training,s=Array.isArray(r),a=s?r:[r],i=a.map(d=>d.name),l=[],u=e.names();for(let d of i)u.indexOf(d)!==-1?l.push(e.getValue(d)):l.push(null);n!=null&&(n.maxNumTensors=-Infinity,n.minNumTensors=Infinity);let c=i.join(",")+"|"+e.names().join(","),p,m;if(u0[c]==null){let d=rte(a,e);p=d.sorted,m=d.recipientCounts,u0[c]=p,D3[c]=m}p=u0[c],m={},o||Object.assign(m,D3[c]);let f=new Es(e);for(let d=0;d<p.length;++d){if(n!=null){let R=Vm().numTensors;R>n.maxNumTensors&&(n.maxNumTensors=R),R<n.minNumTensors&&(n.minNumTensors=R)}let h=p[d],g=h.sourceLayer;if(g instanceof fi)continue;let x=[],b=[],w=[],_=!1;for(let R of h.inputs){let M=f.getValue(R),G=f.getMask(R);x.push(M),b.push(G),G!=null&&(_=!0),o||(m[R.name]--,m[R.name]===0&&!e.hasKey(R)&&i.indexOf(R.name)===-1&&!M.isDisposed&&R.sourceLayer.stateful!==!0&&w.push(M))}_&&(t=t||{},t.mask=b[0]);let I=bt(g.apply(x,t)),E=null;g.supportsMasking&&(E=g.computeMask(x,b));let D=ote(h),$=Array.isArray(D)?D:[D];for(let R=0;R<$.length;++R){f.hasKey($[R])||f.add($[R],I[R],Array.isArray(E)?E[0]:E);let M=i.indexOf($[R].name);M!==-1&&(l[M]=I[R])}o||Ee(w)}return f.disposeMasks(),s?l:l[0]}function rte(r,e){y.assert(r!=null&&r.length>0,()=>"Expected at least one fetch, got none");let t=[],n={};if(r.length===1){let o=$3(r[0],e);t=o.sorted,n=o.recipientMap}else{let o=new Set;for(let s of r){let{sorted:a,recipientMap:i}=$3(s,e);for(let l of a)o.has(l.name)||(t.push(l),o.add(l.name));for(let l in i)n[l]==null&&(n[l]=new Set),i[l].forEach(u=>n[l].add(u))}}return{sorted:t,recipientCounts:nte(n)}}function nte(r){let e={};for(let t in r)e[t]=r[t].size;return e}function $3(r,e){let t=new Set,n=[],o={};for(let i of e.names())t.add(i);let s=[],a=[];for(s.push(r);s.length>0;){let i=s[s.length-1];if(t.has(i.name)){s.pop();continue}let l=a[a.length-1]===s.length-1;if(i.inputs.length===0||l)s.pop(),n.push(i),t.add(i.name),l&&a.pop();else{a.push(s.length-1);for(let u of i.inputs)o[u.name]==null&&(o[u.name]=new Set),o[u.name].add(i.name),!t.has(u.name)&&s.push(u)}}return{sorted:n,recipientMap:o}}function ote(r){let e;if(r.sourceLayer.inboundNodes.length===1)e=r.sourceLayer.output;else{let t=null;for(let n=0;n<r.sourceLayer.inboundNodes.length;++n)for(let o of r.sourceLayer.inboundNodes[n].outputTensors)if(o.id===r.id){t=n;break}e=r.sourceLayer.getOutputAt(t)}return e}var Wn=class extends Le{constructor(e){super({});if(this.containerNodes=new Set,this.name=e.name,this.name==null){let b=this.getClassName().toLowerCase();this.name=Sl(b)}if(this.supportsMasking=!1,this.trainable_=!0,Array.isArray(e.inputs)?this.inputs=e.inputs.slice():this.inputs=[e.inputs],Array.isArray(e.outputs)?this.outputs=e.outputs.slice():this.outputs=[e.outputs],ho(this.inputs).length!==this.inputs.length)throw new z(`The list of inputs passed to the model is redundant. All inputs should only appear once. Found: ${this.inputs.map(b=>b.name)}`);ho(this.outputs).length!==this.outputs.length&&console.warn(`The list of outputs passed to the model is redundant. All outputs should only appear once. Found: ${this.outputs.map(b=>b.name)}`),this.inputLayers=[],this.inputLayersNodeIndices=[],this.inputLayersTensorIndices=[],this.outputLayers=[],this.outputLayersNodeIndices=[],this.outputLayersTensorIndices=[],this.layers=[],this.internalContainerRefs=[];for(let b of this.outputs){let w=b.sourceLayer,_=b.nodeIndex,I=b.tensorIndex;this.outputLayers.push(w),this.outputLayersNodeIndices.push(_),this.outputLayersTensorIndices.push(I)}for(let b of this.inputs){let w=b.sourceLayer,_=b.nodeIndex,I=b.tensorIndex;Gn(_===0,"input layer has >1 nodes"),Gn(I===0,"input layer has >1 tensors"),this.inputLayers.push(w),this.inputLayersNodeIndices.push(_),this.inputLayersTensorIndices.push(I)}this.inputNames=[],this.outputNames=[],this.feedInputShapes=[],this.feedInputNames=[],this.feedOutputNames=[];for(let b=0;b<this.inputLayers.length;b++){let w=this.inputLayers[b];if(!(w instanceof fi))throw new TypeError(`Input layers to a LayersModel must be InputLayer objects. Received inputs: ${e.inputs}. Input ${b} (0-based) originates from layer type ${w.getClassName()}.`);this.inputNames.push(w.name),this.feedInputShapes.push(w.batchInputShape),this.feedInputNames.push(w.name)}for(let b of this.outputLayers)this.outputNames.push(b.name);this.internalInputShapes=this.inputs.map(b=>b.shape),this.internalOutputShapes=this.outputs.map(b=>b.shape);let t={},n={},o={},s={},a={},i=[],l=(b,w,_,I,E,D)=>{(I==null||E==null||D==null)&&(I=b.sourceLayer,E=b.nodeIndex,D=b.tensorIndex);let $=I.inboundNodes[E];if(_.indexOf($)!==-1)throw new Gr(`The tensor ${b.name} at layer "${I.name}" is part of a cycle.`);if(w.indexOf($)!==-1)return;this.containerNodes.add(Wn.nodeKey(I,E)),I.id in a||(a[I.id]=Object.keys(a).length),_.indexOf($)===-1&&_.push($);let R=$.inboundLayers.length;for(let M=0;M<R;M++){let G=$.inputTensors[M],j=$.inboundLayers[M],U=$.nodeIndices[M],H=$.tensorIndices[M];l(G,w,_,j,U,H)}for(w.push($);_.indexOf($)>=0;)_.splice(_.indexOf($),1);i.push($)},u=[],c=[];for(let b of this.outputs)l(b,u,c);let p=i.slice().reverse();for(let b of p){n[b.id]=b,b.id in t||(t[b.id]=0);let w=t[b.id],_=o[b.outboundLayer.id]==null?0:o[b.outboundLayer.id];w=Math.max(w,_),o[b.outboundLayer.id]=w,s[b.outboundLayer.id]=b.outboundLayer,t[b.id]=w;for(let I=0;I<b.inboundLayers.length;I++){let E=b.inboundLayers[I],D=b.nodeIndices[I],$=E.inboundNodes[D],R=t[$.id]==null?0:t[$.id];t[$.id]=Math.max(w+1,R),n[$.id]=$}}let m={};for(let b in t){let w=t[b];w in m||(m[w]=[]),m[w].push(n[b])}let f={};for(let b in o){let w=o[b];w in f||(f[w]=[]),f[w].push(s[b])}let d=Object.keys(f).map(b=>parseInt(b,10)).sort(Yf);this.layers=[];for(let b of d){let w=f[b];w.sort((_,I)=>{let E=a[_.id],D=a[I.id];return E<D?-1:E>D?1:0});for(let _ of w)_ instanceof Wn&&this.internalContainerRefs.push(_),this.layers.push(_)}this.layersByDepth=f,d=Object.keys(m).map(b=>parseInt(b,10)).sort(Yf);let h=this.inputs.slice(),g=[];for(let b of d)for(let w of m[b]){let _=w.outboundLayer;if(_!=null){for(let I of w.inputTensors)if(h.indexOf(I)===-1)throw new Gr(`Graph disconnected: cannot obtain value for tensor ${I} at layer "${_.name}". The following previous layers were accessed without issue: ${g}`);for(let I of w.outputTensors)h.push(I);g.push(_.name)}}this.nodesByDepth=m;let x=this.layers.map(b=>b.name);for(let b of x){let w=x.filter(_=>_===b).length;if(w!==1)throw new Gr(`The name "${b}" is used ${w} times in the model. All layer names should be unique. Layer names: `+JSON.stringify(x))}this.outboundNodes=[],this.inboundNodes=[],new Tl({outboundLayer:this,inboundLayers:[],nodeIndices:[],tensorIndices:[],inputTensors:this.inputs,outputTensors:this.outputs,inputMasks:this.inputs.map(b=>null),outputMasks:this.outputs.map(b=>null),inputShapes:this.inputs.map(b=>b.shape),outputShapes:this.outputs.map(b=>b.shape)}),this.built=!0,this._refCount=1}assertNotDisposed(){if(this._refCount===0)throw new Error(`Container '${this.name}' is already disposed.`)}dispose(){this.assertNotDisposed();let e={refCountAfterDispose:null,numDisposedVariables:0};if(--this._refCount==0){for(let t of this.layers)e.numDisposedVariables+=t.dispose().numDisposedVariables;for(let t of this.internalContainerRefs)e.numDisposedVariables+=t.dispose().numDisposedVariables}return e.refCountAfterDispose=this._refCount,e}get trainable(){return this.trainable_}set trainable(e){this.layers.forEach(t=>{t._trainableWeights.forEach(n=>n.trainable=e)}),this.trainable_=e}get trainableWeights(){if(this._trainableWeights.length>0)throw new z("Container instance unexpectedly contains _trainableWeights.The trainable weights of a Container are a union of the trainable weights of its consituent Layers. Its own _trainableWeights must remain an empty Array.");if(!this.trainable)return[];let e=[];for(let t of this.layers)e=e.concat(t.trainableWeights);return e}get nonTrainableWeights(){let e=[];for(let t of this.layers)e.push(...t.nonTrainableWeights);if(!this.trainable){let t=[];for(let n of this.layers)t.push(...n.trainableWeights);return t.concat(e)}return e}get weights(){return this.trainableWeights.concat(this.nonTrainableWeights)}loadWeights(e,t=!0){let n={},o=0;for(let a of this.layers)for(let i of a.weights){if(n[i.originalName]!=null)throw new z(`Duplicate weight name: ${i.originalName}`);n[i.originalName]=i,o++}let s=[];for(let a in e){let i=a;if(n[a]==null){let l=a.split("/");i=l.slice(0,-2).concat([l[l.length-1]]).join("/")}if(n[i]!=null)s.push([n[i],e[a]]);else if(t)throw new z(`Provided weight data has no target variable: ${a}`);delete n[i]}if(t){let a=[];for(let i in n)a.push(i);if(a.length>0)throw new z(`${a.length} of ${o} weights are not set: ${a}`)}Yp(s)}updatedConfig(){let e=this.getConfig(),t={};return t.className=this.getClassName(),t.config=e,t.kerasVersion=`tfjs-layers ${yd}`,t.backend="TensorFlow.js",t}toJSON(e,t=!0){let n=ty(this.updatedConfig());return t?JSON.stringify(n):n}call(e,t){return V(()=>{e=bt(e);let n=new Es;for(let o=0;o<this.inputs.length;++o)n.add(this.inputs[o],e[o]);return Ju(this.outputs,n,t)})}computeMask(e,t){return V(()=>{e=bt(e);let n;return t==null?n=po(null,e.length):n=bt(t),this.runInternalGraph(e,n)[1]})}computeOutputShape(e){let t=Kp(e);if(t.length!==this.inputLayers.length)throw new z(`Invalid inputShape argument ${e}: model has ${this.inputLayers.length} tensor inputs.`);let n={};for(let i=0;i<t.length;i++){let l=this.inputLayers[i],u=t[i],c=l.name+"_0_0";n[c]=u}let o=Object.keys(this.nodesByDepth).map(i=>parseInt(i,10)).sort(Yf);if(o.length>1)for(let i of o){let l=this.nodesByDepth[i];for(let u of l){let c=u.outboundLayer;if(this.inputLayers.map(h=>h.id).indexOf(c.id)!==-1)continue;let p=[];for(let h=0;h<u.inboundLayers.length;h++){let g=u.inboundLayers[h],x=u.nodeIndices[h],b=u.tensorIndices[h],w=`${g.name}_${x}_${b}`,_=n[w];p.push(_)}let m=c.computeOutputShape(wr(p)),f=Kp(m),d=c.inboundNodes.indexOf(u);for(let h=0;h<f.length;h++){let g=`${c.name}_${d}_${h}`;n[g]=f[h]}}}let s=[],a=[];for(let i=0;i<this.outputLayers.length;i++){let l=this.outputLayers[i],u=this.outputLayersNodeIndices[i],c=this.outputLayersTensorIndices[i],p=`${l.name}_${u}_${c}`;a.push(p)}for(let i=0;i<a.length;i++){let l=a[i];Gn(l in n),s.push(n[l])}return wr(s)}runInternalGraph(e,t){t==null&&(t=po(null,e.length));let n={};for(let l=0;l<this.inputs.length;++l){let u=this.inputs[l],c=e[l],p=t[l];n[u.id]=[c,p]}let o=Object.keys(this.nodesByDepth).map(l=>parseInt(l,10)).sort(Yf);for(let l of o){let u=this.nodesByDepth[l];for(let c of u){let p=c.outboundLayer,m=c.inputTensors,f=c.outputTensors,d=new Array;for(let h of m)h.id in n&&d.push(n[h.id]);if(d.length===m.length){let h={},g,x,b,w;if(c.callArgs!=null&&(h=c.callArgs),d.length===1){let[_,I]=d[0];h.mask==null&&(h.mask=I),b=bt(p.call(_,h)),w=bt(p.computeMask(_,I)),g=[_],x=[I]}else g=d.map(_=>_[0]),x=d.map(_=>_[1]),h.mask==null&&(h.mask=x),b=bt(p.call(g,h)),w=bt(p.computeMask(g,x));if(p.activityRegularizer)throw new Se("LayersModel invocation with concrete Tensor value(s) in the presence of activity regularizer(s) is not supported yet.");for(let _=0;_<f.length;++_){let I=f[_],E=b[_],D=w[_];n[I.id]=[E,D]}}}}let s=[],a=[],i=[];for(let l of this.outputs){Gn(l.id in n,`Could not compute output ${l.name} : ${l.id}`);let[u,c]=n[l.id];i.push(u.shape),s.push(u),a.push(c)}return[s,a,i]}buildNodeConversionMap(e){let t={},n;for(let o of this.layers){n=o instanceof Wn?1:0;for(let s=0;s<o.inboundNodes.length;s++){let a=Wn.nodeKey(o,s);this.containerNodes.has(a)&&(t[a]=n,n+=1)}}return t}getLayer(e,t){if(t!=null){if(this.layers.length<=t)throw new z(`Was asked to retrieve layer at index ${t}, but model only has ${this.layers.length} layer(s).`);return this.layers[t]}else if(e==null)throw new z("Provide either a layer name or layer index");for(let n of this.layers)if(n.name===e)return n;throw new z(`No such layer: ${e}`)}calculateLosses(){return V(()=>{let e=[];for(let t of this.layers)for(let n=0;n<t.inboundNodes.length;++n){let o=Wn.nodeKey(t,n);this.containerNodes.has(o)&&e.push(...t.calculateLosses())}return e})}getConfig(){let e={name:this.name},t=this.buildNodeConversionMap(this.layers),n=[];for(let a of this.layers){let i=a.getClassName(),l=a.getConfig(),u=[];for(let p=0;p<a.inboundNodes.length;p++){let m=a.inboundNodes[p],f=Wn.nodeKey(a,p),d={};if(this.containerNodes.has(f)){if(m.callArgs)try{JSON.stringify(m.callArgs),d=m.callArgs}catch(h){console.warn(`Layer ${a.name} was passed non-serializable keyword arguments: ${m.callArgs}. They will not be included in the serialized model (and thus will be missing at deserialization time).`),d={}}if(m.inboundLayers.length>0){let h=[];for(let g=0;g<m.inboundLayers.length;g++){let x=m.inboundLayers[g],b=m.nodeIndices[g],w=m.tensorIndices[g],_=Wn.nodeKey(x,b),I=t[_];I==null&&(I=0),h.push([x.name,I,w,d])}u.push(h)}}}let c={};c.name=a.name,c.className=i,c.config=l,c.inboundNodes=u,n.push(c)}e.layers=n;let o=[];for(let a=0;a<this.inputLayers.length;a++){let i=this.inputLayers[a],l=this.inputLayersNodeIndices[a],u=Wn.nodeKey(i,l);if(!this.containerNodes.has(u))continue;let c=t[u];c==null&&(c=0);let p=this.inputLayersTensorIndices[a];o.push([i.name,c,p])}e.inputLayers=o;let s=[];for(let a=0;a<this.outputLayers.length;a++){let i=this.outputLayers[a],l=this.outputLayersNodeIndices[a],u=Wn.nodeKey(i,l);if(!this.containerNodes.has(u))continue;let c=t[u];c==null&&(c=0);let p=this.outputLayersTensorIndices[a];s.push([i.name,c,p])}return e.outputLayers=s,e}static fromConfig(e,t,n={},o=!1){let s={},a={};function i(g,x){g.name in a?a[g.name].push(x):a[g.name]=[x]}function l(g,x){let b=[],w;for(let _ of x){let I=_[0],E=_[1],D=_[2];if(w=_[3]==null?{}:_[3],!(I in s)){i(g,x);return}let $=s[I];if($.inboundNodes.length<=E){i(g,x);return}let R=$.inboundNodes[E];b.push(R.outputTensors[D])}b.length>0&&g.apply(wr(b),w)}function u(g){let x=g.name,b=an(g,t.customObjects!=null?t.customObjects:{});b.setFastWeightInitDuringBuild(o),s[x]=b,g.inboundNodes.forEach(_=>{if(!(_ instanceof Array))throw new z(`Corrupted configuration, expected array for nodeData: ${_}`);i(b,_)})}let c=t.name,p=t.layers;for(let g of p)u(g);for(;!Kz(a);)for(let g of p){let x=s[g.name];if(x.name in a){let b=a[x.name];delete a[x.name];for(let w of b)l(x,w)}}let m=[],f=[],d=t.inputLayers;for(let g of d){let x=g[0],b=g[1],w=g[2];Gn(x in s);let I=s[x].inboundNodes[b].outputTensors;m.push(I[w])}let h=t.outputLayers;for(let g of h){let x=g[0],b=g[1],w=g[2];Gn(x in s);let I=s[x].inboundNodes[b].outputTensors;f.push(I[w])}return new e({inputs:m,outputs:f,name:c})}get stateful(){if(this._stateful)throw new z("Container instance unexpectedly has _stateful = true. The statefulness of a Container is determined by the Layers it contains. Its _stateful property must remain the default false.");for(let e of this.layers)if(e.stateful)return!0;return!1}resetStates(){V(()=>{this.layers.forEach(e=>{e.stateful&&e.resetStates()})})}};function ste(r,e,t){let n=e.length;if(r==null||Array.isArray(r)&&r.length===0)return e.map(o=>null);if(n===1)return Array.isArray(r)&&r.length===1?r:typeof r=="object"&&e[0]in r?[r[e[0]]]:[r];if(Array.isArray(r)){if(r.length!==n)throw new Error(`Provided ${t} is an array of ${r.length} element(s), but the model has ${n} outputs. Make sure a set of weights is provided for each model output.`);return r}else if(typeof r=="object"&&Object.keys(r).length>0&&typeof r[Object.keys(r)[0]]=="object"){let o=[];return e.forEach(s=>{s in r?o.push(r[s]):o.push(null)}),o}else throw new Error(`The model has multiple (${n}) outputs, so ${t} must be either an array with ${n} elements or an object with ${e} keys. Provided ${t} not understood: ${JSON.stringify(r)}`)}function ry(r,e){return ste(r,e,"classWeight")}async function ny(r,e,t,n){if(e!=null||n!=null)throw new Error("Support sampleWeight is not implemented yet");if(t!=null){let o=V(()=>{if(r.shape.length===1)return r.clone();if(r.shape.length===2)if(r.shape[1]>1){let i=1;return r.argMax(i)}else{if(r.shape[1]===1)return r.reshape([r.shape[0]]);throw new Error(`Encountered unexpected last-dimension size (${r.shape[1]}) during handling of class weights. The size is expected to be >= 1.`)}else throw new Error(`Unexpected rank of target (y) tensor (${r.rank}) during handling of class weights. The rank is expected to be 1 or 2.`)}),s=Array.from(await o.data());Ee(o);let a=[];return s.forEach(i=>{if(t[i]==null)throw new Error(`classWeight must contain all classes in the training data. The class ${i} exists in the data but not in classWeight`);a.push(t[i])}),Ct(a,"float32")}else return null}function R3(r,e){return O(r,e)}var ite=32;function F3(r,e){let t,n,o=e;t=o.xs,n=o.ys,y.assert(t!=null&&n!=null,()=>`A Dataset iterator for fitDataset() is expected to generate objects of the form \`{xs: xVal, ys: yVal}\`, where the two values may be \`tf.Tensor\`, an array of Tensors, or a map of string to Tensor. The provided Dataset instead generates ${e}`);let s=O3("input",r.inputNames,t),a=O3("output",r.outputNames,n),i=s[0].shape[0];y.assert(s.length===r.inputs.length,()=>`LayersModel has ${r.inputs.length} inputs, but the dataset provides ${s.length} inputs. (Expected input keys: ${JSON.stringify(r.inputNames)})`),y.assert(a.length===r.outputs.length,()=>`LayersModel has ${r.outputs.length} outputs, but the dataset provides ${a.length} outputs. (Expected output keys: ${JSON.stringify(r.outputNames)})`);for(let l=0;l<s.length;l++)y.assert(s[l].shape[0]===i,()=>`Batch size mismatch: input ${r.inputNames[l]} has ${s[l].shape[0]}; expected ${i} based on input ${r.inputNames[0]}.`);for(let l=0;l<a.length;l++)y.assert(a[l].shape[0]===i,()=>`Batch size mismatch: output ${r.outputNames[l]} has ${a[l].shape[0]}; expected ${i} based on input ${r.inputNames[0]}.`);return{xs:s,ys:a}}function O3(r,e,t){if(t instanceof Pe)return[t];if(Array.isArray(t))return y.assert(t.length===e.length,()=>`Received an array of ${t.length} Tensors, but expected ${e.length} to match the ${r} keys ${e}.`),t;{let n=[];for(let o of e){if(t[o]==null)throw new z(`The feature data generated by the dataset lacks the required ${r} key '${o}'.`);n.push(t[o])}return n}}function ate(r){if(r.length===3)throw new Se("Validation with sample weights is not implemented yet.");return{xs:r[0],ys:r[1]}}async function P3(r,e,t){let n=t.batchesPerEpoch!=null;if(y.assert(r.optimizer!=null,()=>"You must compile a model before training/testing. Use LayersModel.compile(modelCompileConfig)."),y.assert(t!=null,()=>"For fitDataset(), the 2nd argument (config) is required, but it is not provided in this call."),y.assert(t.epochs!=null&&t.epochs>0&&Number.isInteger(t.epochs),()=>`For fitDataset(), config.epochs is expected to be a positive integer, but got ${t.epochs}`),y.assert(!n||t.batchesPerEpoch>0&&Number.isInteger(t.batchesPerEpoch),()=>`For fitDataset(), config.batchesPerEpoch is expected to be a positive integer if specified, but got ${t.batchesPerEpoch}`),y.assert(t.validationSplit==null,()=>"`validationSplit` is not supported by `fitDataset()`. Use validationData instead."),r.isTraining)throw new Error("Cannot start training because another fit() call is ongoing.");r.isTraining=!0;try{let o=t.validationData!=null,s,a;if(o)if(M3(t.validationData))y.assert(t.validationBatches==null||t.validationBatches>0&&Number.isInteger(t.validationBatches),()=>`For fitDataset() with dataset-based validation, config.validationBatches is expected not to be provided, or to be a positive integer, but got ${t.validationBatches}`);else{let g=ate(t.validationData);s=g.xs,a=g.ys}let i=r.makeTrainFunction(),l=r.getDedupedMetricsNames(),u;o?u=l.slice().concat(l.map(g=>"val_"+g)):u=l.slice();let c=Kx(t.callbacks,t.yieldEvery),p=t.verbose==null?1:t.verbose,{callbackList:m,history:f}=Xx(c,p,t.epochs,null,null,lte(e,t),null,o,u);m.setModel(r),r.history=f,await m.onTrainBegin(),r.stopTraining_=!1;let d=t.initialEpoch==null?0:t.initialEpoch,h=await e.iterator();for(;d<t.epochs;){let g={};await m.onEpochBegin(d);let x=0,b=0;for(n||(h=await e.iterator());n?x<t.batchesPerEpoch:!0;){let w=await h.next();if(n&&w.done){console.warn(`You provided \`batchesPerEpoch\` as ${t.batchesPerEpoch}, but your dataset iterator ran out of data after ${x} batches; interrupting training. Make sure that your dataset can generate at least \`batchesPerEpoch * epochs\` batches (in this case, ${t.batchesPerEpoch*t.epochs} batches). You may need to use the repeat() function when building your dataset.`);break}if(w.value!=null){let{xs:_,ys:I}=F3(r,w.value),E={};E.batch=b,E.size=_[0].shape[0],await m.onBatchBegin(b,E);let D=[];if(t.classWeight!=null){let M=ry(t.classWeight,r.outputNames);for(let G=0;G<M.length;++G)D.push(await ny(I[G],null,M[G]))}let $=_.concat(I).concat(D),R=i($);Ee($);for(let M=0;M<l.length;++M){let G=l[M],j=R[M];E[G]=j,$t(j)}await m.onBatchEnd(b,E),qx(E),b++,x++}if(n?x>=t.batchesPerEpoch:w.done){if(o){let _;M3(t.validationData)?_=bt(await r.evaluateDataset(t.validationData,{batches:t.validationBatches})):_=bt(r.evaluate(s,a,{batchSize:t.validationBatchSize==null?ite:t.validationBatchSize,verbose:0}));for(let I=0;I<r.metricsNames.length;++I)g[`val_${r.metricsNames[I]}`]=_[I]}break}if(r.stopTraining_)break}if(await m.onEpochEnd(d,g),d++,r.stopTraining_)break}return await m.onTrainEnd(),await r.history.syncData(),r.history}finally{r.isTraining=!1}}function lte(r,e){let t=null;return e.batchesPerEpoch!=null?t=e.batchesPerEpoch:Number.isFinite(r.size)&&(t=r.size),t}function M3(r){return typeof r.iterator=="function"}function ute(r){return typeof r.next=="function"}async function L3(r,e,t){t=t||{};let n=t.batches!=null,o=r.testFunction,s=[];if(t.verbose>0)throw new Se("Verbose mode is not implemented yet.");y.assert(!n||t.batches>0&&Number.isInteger(t.batches),()=>`Test loop expects \`batches\` to be a positive integer, but received ${JSON.stringify(t.batches)}`);let a=ute(e)?e:await e.iterator(),i=0,l=0;for(;n?l<t.batches:!0;){let u=await a.next();if(s=V(()=>{if(u.value){let{xs:c,ys:p}=F3(r,u.value),m=c.concat(p),f=V(()=>o(m));if(Ee(m),l===0)for(let h=0;h<f.length;++h)s.push(ue(0));let d=m[0].shape[0];for(let h=0;h<f.length;++h){let g=f[h],x=s[h];s[h]=V(()=>J(s[h],O(d,g))),l>0&&Ee(x)}Ee(f),i+=d,++l}return s}),u.done){n&&console.warn(`Your dataset iterator ran out of data during evaluateDataset(). Interrupting evalution. Make sure that your dataset can generate at least \`batches\` batches (in this case, ${t.batches} batches). You may need to use the repeat() function when building your dataset.`);break}}for(let u=0;u<s.length;++u){let c=s[u];s[u]=pe(s[u],i),Ee(c)}return wr(s)}function oy(r){y.assert(r>0&&Number.isInteger(r),()=>`batchSize is required to be a positive integer, but got ${r}`)}function em(r,e,t){return r==null?[null]:Array.isArray(r)?r.map(n=>za(n,e,t-e)):za(r,e,t-e)}function sy(r,e){return V(()=>r==null?null:Array.isArray(r)?r.map(t=>sy(t,e)):Bx(r,e.dtype==="int32"?e:e.toInt()))}function iy(r,e){let t=[],n=0,o=null;for(;n<r;)o=n+e,o>=r&&(o=r),t.push([n,o]),n=o;return t}async function cte(r,e,t,n,o,s,a,i,l,u,c,p,m,f,d){o==null&&(o=32),s==null&&(s=1),c==null&&(c=!0),m==null&&(m=0);let h=!1;if(l!=null&&u!=null&&(h=!0),d!=null&&(h=!0,f==null))throw new z("Can only use `validationSteps` when doing step-wise training, i.e., `stepsPerEpoch` must be set.");let g=r.checkNumSamples(t,o,f,"steps_per_epoch"),x;g!=null&&(x=Wr(0,g)),a==null&&(a=1);let{callbackList:b,history:w}=Xx(i,a,s,m,g,f,o,h,p);b.setModel(r),r.history=w,await b.onTrainBegin(),r.stopTraining_=!1;for(let _=m;_<s;++_){await b.onEpochBegin(_);let I={};if(f!=null)throw new Se("stepsPerEpoch mode is not implemented yet.");{if(c==="batch")throw new Se("batch shuffling is not implemneted yet");c&&y.shuffle(x);let E=Ct(x),D=iy(g,o);for(let $=0;$<D.length;++$){let R={};if(await b.onBatchBegin($,R),V(()=>{let M=D[$][0],G=D[$][1],j=za(E,M,G-M);R.batch=$,R.size=G-M;let U=sy(t,j),H=e(U);for(let q=0;q<n.length;++q){let X=n[q],ne=H[q];R[X]=ne,$t(ne)}if($===D.length-1&&h){let q=r.testLoop(l,u,o);for(let X=0;X<n.length;++X){let ne=n[X],Y=q[X];$t(Y),I["val_"+ne]=Y}}}),await b.onBatchEnd($,R),qx(R),r.stopTraining_)break}E.dispose()}if(await b.onEpochEnd(_,I),r.stopTraining_)break}return await b.onTrainEnd(),await r.history.syncData(),r.history}async function z3(r,e,t,n={}){if(r.isTraining)throw new Error("Cannot start training because another fit() call is ongoing.");r.isTraining=!0;let o,s,a,i,l,u,c;try{let p=n.batchSize==null?32:n.batchSize;oy(p);let m=!1,f=await r.standardizeUserData(e,t,n.sampleWeight,n.classWeight,m,p);o=f[0],s=f[1],c=f[2];let d=!1,h;if(n.validationData!=null&&n.validationData.length>0){if(d=!0,n.validationData.length===2)a=n.validationData[0],i=n.validationData[1];else throw n.validationData.length===3?new Se("validationData including sample weights is not supported yet."):new z(`When passing validation data, it must contain 2 (valX, valY) or 3 (valX, valY, valSampleWeight) items; ${n.validationData} is invalid.`);let D=!0,$=await r.standardizeUserData(a,i,null,null,D,p);l=$[0],u=$[1],h=l.concat(u)}else if(n.validationSplit!=null&&n.validationSplit>0&&n.validationSplit<1){d=!0;let D=Math.floor(o[0].shape[0]*(1-n.validationSplit)),$=o[0].shape[0];l=em(o,D,$),o=em(o,0,D),u=em(s,D,$),s=em(s,0,D),h=l.concat(u)}else n.validationSteps!=null&&(d=!0);let g=o.concat(s).concat(c);r.checkTrainableWeightsConsistency();let x=r.makeTrainFunction(),b=r.getDedupedMetricsNames(),w,_;d?(r.makeTestFunction(),w=r.testFunction,_=b.slice().concat(b.map(D=>"val_"+D))):(w=null,h=[],_=b.slice());let I=Kx(n.callbacks,n.yieldEvery);return await cte(r,x,g,b,p,n.epochs,n.verbose,I,w,h,n.shuffle,_,n.initialEpoch,null,null)}finally{r.isTraining=!1,Dl(o,e),Dl(s,t),Dl(l,a),Dl(u,i),c!=null&&Ee(c)}}function c0(r){let e=[];r instanceof Pe&&(r=[r]);for(let t=0;t<r.length;++t){let n=r[t];if(n.rank===1)e.push(La(n,1));else{if(n.rank===0)throw new Error("Expected tensor to be at least 1D, but received a 0D tensor (scalar).");e.push(n)}}return e}function Dl(r,e){if(r==null)return;let t=[];if(e instanceof Pe)t.push(e.id);else if(Array.isArray(e))e.forEach(o=>t.push(o.id));else if(e!=null)for(let o in e){let s=e[o];t.push(s.id)}let n=[];if(r instanceof Pe)t.indexOf(r.id)===-1&&n.push(r);else if(Array.isArray(r))r.forEach(o=>{t.indexOf(o.id)===-1&&n.push(o)});else if(r!=null)for(let o in r){let s=r[o];t.indexOf(s.id)===-1&&n.push(s)}n.forEach(o=>{o.isDisposed||o.dispose()})}function pte(r){return r instanceof Pe}function p0(r){return Array.isArray(r)}function B3(r){return!pte(r)&&!p0(r)}function V3(r,e,t,n=!0,o=""){if(e==null||e.length===0){if(r!=null){let a=!1;if(p0(r)&&r.length>0)a=!0;else if(B3(r)){for(let i in r)if(r.hasOwnProperty(i)){a=!0;break}}else a=!0;if(a)throw new z(`Error when checking model ${o} expected no data, but got ${r}`)}return[]}if(r==null)return e.map(a=>null);let s;if(B3(r)){r=r,s=[];for(let a of e){if(r[a]==null)throw new z(`No data provided for "${a}". Need data for each key in: ${e}`);s.push(r[a])}}else if(p0(r)){if(r=r,r.length!==e.length)throw new z(`Error when checking model ${o}: the Array of Tensors that you are passing to your model is not the size the model expected. Expected to see ${e.length} Tensor(s), but instead got the following list of Tensor(s): ${r}`);s=r}else{if(r=r,e.length>1)throw new z(`The model ${o} expects ${e.length} Tensor(s), but only received one Tensor. Found: Tensor with shape ${r.shape}`);s=[r]}if(s=c0(s),t!=null)for(let a=0;a<e.length;++a){if(t[a]==null)continue;let i=s[a];if(i.shape.length!==t[a].length)throw new z(`Error when checking ${o}: expected ${e[a]} to have ${t[a].length} dimension(s). but got array with shape ${i.shape}`);for(let l=0;l<t[a].length;++l){if(l===0&&!n)continue;let u=i.shape[l],c=t[a][l];if(c!=null&&c>=0&&u!==c)throw new z(`Error when checking ${o}: expected ${e[a]} to have shape [${t[a]}], but got array with shape [${i.shape}].`)}}return s}function mte(r,e,t){let n=ho(r.map(s=>s.shape[0]));n.sort();let o=ho(e.map(s=>s.shape[0]));if(o.sort(),n.length>1)throw new z(`All input Tensors (x) should have the same number of samples. Got array shapes: ${JSON.stringify(r.map(s=>s.shape))}`);if(o.length>1)throw new z(`All target Tensors (y) should have the same number of samples. Got array shapes: ${JSON.stringify(e.map(s=>s.shape))}`);if(n.length>0&&o.length>0&&!y.arraysEqual(n,o))throw new z(`Input Tensors should have the same number of samples as target Tensors. Found ${n[0]} input sample(s) and ${o[0]} target sample(s).`)}function fte(r,e,t){let n=[hi,Qp,Yu];for(let o=0;o<r.length;++o){let s=r[o],a=e[o],i=t[o];if(a!=null){if(a===Yu&&s.shape[s.shape.length-1]===1)throw new z(`You are passing a target array of shape ${s.shape} while using a loss 'categorical_crossentropy'. 'categorical_crossentropy'expects targets to be binary matrices (1s and 0s) of shape [samples, classes].`);if(n.indexOf(a)!==-1){let l=s.shape.slice(1),u=i.slice(1);for(let c=0;c<l.length;++c){let p=l[c],m=u[c];if(m!=null&&p!==m)throw new z(`A target Tensor with shape ${s.shape} was passed for an output of shape ${i}, while using a loss function that expects targets to have the same shape as the output.`)}}}}}function G3(r,e,t,n=!0,o=""){let s;if(Array.isArray(r)){if(r.length!==e.length)throw new z(`Error when checking model ${o}: the Array of Tensors that you are passing to your model is not the size the the model expected. Expected to see ${e.length} Tensor(s), but instead got ${r.length} Tensors(s).`);s=r}else{if(e.length>1)throw new z(`The model expects ${e.length} ${o} Tensors, but only received one Tensor. Found: array with shape ${JSON.stringify(r.shape)}.`);s=[r]}if(t!=null)for(let a=0;a<e.length;++a){if(t[a]==null)continue;let i=s[a];if(i.shape.length!==t[a].length)throw new z(`Error when checking ${o}: expected ${e[a]} to have ${t[a].length} dimension(s), but got array with shape ${JSON.stringify(i.shape)}`);for(let l=0;l<t[a].length;++l){if(l===0&&!n)continue;let u=i.shape[l],c=t[a][l];if(c!=null&&c!==u)throw new z(`Error when checking ${o}: expected ${e[a]} to have shape ${JSON.stringify(t[a])} but got array with shape ${JSON.stringify(i.shape)}.`)}}}function dte(r,e){if(r==null||Array.isArray(r)&&r.length===0)return e.map(n=>[]);let t;if(typeof r=="string"||typeof r=="function")t=[r];else if(Array.isArray(r)||typeof r=="object")t=r;else throw new TypeError(`Type of metrics argument not understood. Expected an string,function, Array, or Object, found: ${r}`);if(Array.isArray(t))return e.map(n=>t);{let n=[];for(let o of e){let s=t.hasOwnProperty(o)?t[o]:[];Array.isArray(s)||(s=[s]),n.push(s)}return n}}var hte="layers-model",jn=class extends Wn{constructor(e){super(e);this.isTraining=!1}summary(e,t,n=console.log){if(!this.built)throw new z("This model has never been called, thus its weights have not been created yet. So no summary can be displayed. Build the model first (e.g., by calling it on some test data).");A3(this,e,t,n)}compile(e){if(e.loss==null&&(e.loss=[]),this.loss=e.loss,typeof e.optimizer=="string")this.optimizer_=S3(e.optimizer),this.isOptimizerOwned=!0;else{if(!(e.optimizer instanceof Br))throw new z("User-defined optimizer must be an instance of tf.Optimizer.");this.optimizer_=e.optimizer,this.isOptimizerOwned=!1}let t=[];if(!Array.isArray(e.loss)&&typeof e.loss!="string"&&typeof e.loss!="function"){e.loss=e.loss;for(let a in e.loss)if(this.outputNames.indexOf(a)===-1)throw new z(`Unknown entry in loss dictionary: "${a}". Only expected the following keys: ${this.outputNames}`);for(let a of this.outputNames)e.loss[a]==null&&console.warn(`Output "${a}" is missing from loss dictionary. We assume this was done on purpose, and we will not be expecting data to be passed to ${a} during training`),t.push(Yx(e.loss[a]))}else if(Array.isArray(e.loss)){if(e.loss.length!==this.outputs.length)throw new z(`When passing an Array as loss, it should have one entry per model output. The model has ${this.outputs.length} output(s), but you passed loss=${e.loss}.`);t=e.loss.map(i=>Yx(i))}else{let a=Yx(e.loss);this.outputs.forEach(i=>{t.push(a)})}this.lossFunctions=t,this.feedOutputNames=[],this.feedOutputShapes=[],this.feedLossFns=[];for(let a=0;a<this.outputs.length;++a){let i=this.internalOutputShapes[a],l=this.outputNames[a];this.feedOutputNames.push(l),this.feedOutputShapes.push(i),this.feedLossFns.push(this.lossFunctions[a])}let n=[];this.metrics=e.metrics,this.metricsNames=["loss"],this.metricsTensors=[],Ts("loss",()=>{for(let a=0;a<this.outputs.length;++a){if(n.indexOf(a)!==-1)continue;let i=this.lossFunctions[a];this.outputs.length>1&&(this.metricsTensors.push([i,a]),this.metricsNames.push(this.outputNames[a]+"_loss"))}});let o=dte(e.metrics,this.outputNames),s=(a,i,l)=>{this.outputNames.length>1&&(i=this.outputNames[a]+"_"+i),this.metricsNames.push(i),this.metricsTensors.push([l,a])};Ts("metric",()=>{for(let a=0;a<this.outputs.length;++a){if(n.indexOf(a)!==-1)continue;let i=o[a];(u=>{let c="",p,m,f;for(let d of u){if(typeof d=="string"&&["accuracy","acc","crossentropy","ce"].indexOf(d)!==-1){let g=this.internalOutputShapes[a];g[g.length-1]===1||this.lossFunctions[a]===Qp?["accuracy","acc"].indexOf(d)!==-1?m=dd:["crossentropy","ce"].indexOf(d)!==-1&&(m=Zx):this.lossFunctions[a]===Jp?["accuracy","acc"].indexOf(d)!==-1?m=Jx:["crossentropy","ce"].indexOf(d)!==-1&&(m=i0):["accuracy","acc"].indexOf(d)!==-1?m=hd:["crossentropy","ce"].indexOf(d)!==-1&&(m=gd);let x;["accuracy","acc"].indexOf(d)!==-1?x="acc":["crossentropy","ce"].indexOf(d)!==-1&&(x="ce"),f=m,p=c+x}else f=N3(d),p=c+xd(d);let h;Ts(p,()=>{h=f}),s(a,p,h)}})(i)}}),this.collectedTrainableWeights=this.trainableWeights}checkTrainableWeightsConsistency(){this.collectedTrainableWeights!=null&&this.trainableWeights.length!==this.collectedTrainableWeights.length&&console.warn("Discrepancy between trainableweights and collected trainable weights. Did you set `model.trainable` without calling `model.compile()` afterwards?")}evaluate(e,t,n={}){let o=n.batchSize==null?32:n.batchSize;oy(o);let s=!0,a=this.standardizeUserDataXY(e,t,s,o);try{let i=a[0].concat(a[1]);this.makeTestFunction();let l=this.testFunction,u=this.testLoop(l,i,o,n.verbose,n.steps);return wr(u)}finally{Dl(a[0],e),Dl(a[1],t)}}async evaluateDataset(e,t){return this.makeTestFunction(),L3(this,e,t)}checkNumSamples(e,t,n,o="steps"){let s;if(n!=null){if(s=null,t!=null)throw new z(`If ${o} is set, batchSize must be null or undefined.Got batchSize = ${t}`)}else if(e!=null)Array.isArray(e)?s=e[0].shape[0]:s=e.shape[0];else throw new z(`Either the input data should have a defined shape, or ${o} shoud be specified.`);return s}execute(e,t){if(Array.isArray(t)&&t.length===0)throw new z("`outputs` is an empty Array, which is not allowed.");let n=Array.isArray(t),o=n?t:[t],s=this.retrieveSymbolicTensors(o),a=new Es;if(e instanceof Pe&&(e=[e]),Array.isArray(e)){if(e.length!==this.inputs.length)throw new z(`The number of inputs provided (${e.length}) does not match the number of inputs of this model (${this.inputs.length}).`);for(let l=0;l<this.inputs.length;++l)a.add(this.inputs[l],e[l])}else for(let l of this.inputs){let u=e[l.name];if(u==null)throw new z(`No value is provided for the model's input ${l.name}`);a.add(l,u)}let i=Ju(s,a);return n?i:i[0]}retrieveSymbolicTensors(e){let t=po(null,e.length),n=e.length;for(let o of this.layers){let s=Array.isArray(o.output)?o.output:[o.output],a=s.map(i=>i.name);for(let i=0;i<e.length;++i){let l=a.indexOf(e[i]);if(l!==-1&&(t[i]=s[l],n--),n===0)break}if(n===0)break}if(n>0){let o=[];throw t.forEach((s,a)=>{s==null&&o.push(e[a])}),new z(`Cannot find SymbolicTensors for output name(s): ${JSON.stringify(o)}`)}return t}predictLoop(e,t=32,n=!1){return V(()=>{let o=this.checkNumSamples(e);if(n)throw new Se("Verbose predictLoop() is not implemented yet.");let s=iy(o,t),a=this.outputs.map(i=>[]);for(let i=0;i<s.length;++i)V(()=>{let u=s[i][0],c=s[i][1],p=em(e,u,c),m=[];if(Array.isArray(p))for(let d=0;d<p.length;++d)m.push({key:this.inputs[d],value:p[d]});else m.push({key:this.inputs[0],value:p});let f=new Es(m);return Ju(this.outputs,f)}).forEach((u,c)=>a[c].push(u));return wr(a.map(i=>Qe(i,0)))})}predict(e,t={}){let n=c0(e);G3(n,this.inputNames,this.feedInputShapes,!1);try{let o=t.batchSize==null?32:t.batchSize;return oy(o),this.predictLoop(n,o)}finally{Dl(n,e)}}predictOnBatch(e){G3(e,this.inputNames,this.feedInputShapes,!0);let t=(Array.isArray(e)?e[0]:e).shape[0];return this.predictLoop(e,t)}standardizeUserDataXY(e,t,n=!0,o){if(this.optimizer_==null)throw new Gr("You must compile a model before training/testing. Use LayersModel.compile(modelCompileArgs).");let s=[];for(let a=0;a<this.feedOutputShapes.length;++a){let i=this.feedOutputShapes[a];this.feedLossFns[a]===Jp?s.push(i.slice(0,i.length-1).concat([1])):s.push(i)}if(e=V3(e,this.feedInputNames,this.feedInputShapes,!1,"input"),t=V3(t,this.feedOutputNames,s,!1,"target"),mte(e,t,null),fte(t,this.feedLossFns,this.feedOutputShapes),this.stateful&&o!=null&&o>0&&e[0].shape[0]%o!=0)throw new z(`In a stateful network, you should only pass inputs with a number of samples that is divisible by the batch size ${o}. Found: ${e[0].shape[0]} sample(s).`);return[e,t]}async standardizeUserData(e,t,n,o,s=!0,a){let[i,l]=this.standardizeUserDataXY(e,t,s,a);if(n!=null)throw new Error("sample weight is not supported yet.");let u=null;if(o!=null){let c=ry(o,this.outputNames);u=[];for(let p=0;p<c.length;++p)u.push(await ny(l[p],null,c[p]))}return[i,l,u]}testLoop(e,t,n,o=0,s){return V(()=>{let a=this.checkNumSamples(t,n,s,"steps"),i=[];if(o>0)throw new Se("Verbose mode is not implemented yet.");if(s!=null)throw new Se("steps mode in testLoop() is not implemented yet");{let l=iy(a,n),u=Ct(Wr(0,a));for(let c=0;c<l.length;++c){let p=l[c][0],m=l[c][1],f=za(u,p,m-p),d=sy(t,f),h=e(d);if(c===0)for(let g=0;g<h.length;++g)i.push(ue(0));for(let g=0;g<h.length;++g){let x=h[g];i[g]=J(i[g],O(m-p,x))}}for(let c=0;c<i.length;++c)i[c]=pe(i[c],a)}return i})}getDedupedMetricsNames(){let e=this.metricsNames,t=[];for(let n=0;n<e.length;++n){let o=e[n],s=o;KC(e,o)>1&&(s+=`_${KC(e.slice(0,n),o)}`),t.push(s)}return t}makeTrainFunction(){return e=>{let t=[],n=e.slice(0,this.inputs.length),o=e.slice(this.inputs.length,this.inputs.length+this.outputs.length),s=e.slice(this.inputs.length+this.outputs.length,this.inputs.length+this.outputs.length*2),a=[],i=()=>{let p=[];for(let h=0;h<this.inputs.length;++h)p.push({key:this.inputs[h],value:n[h]});let m=new Es(p),f=Ju(this.outputs,m,{training:!0}),d;for(let h=0;h<this.lossFunctions.length;++h){let x=this.lossFunctions[h](o[h],f[h]);s[h]!=null&&(x=R3(x,s[h]));let b=dt(x);t.push(b),h===0?d=x:d=J(d,x)}for(let h=0;h<this.metricsTensors.length;++h){let g;if(this.outputs.length>1&&h<this.outputs.length)g=t[h];else{let x=this.metricsTensors[h][0],b=this.metricsTensors[h][1];g=dt(x(o[b],f[b]))}$t(g),a.push(g)}return d=dt(d),this.calculateLosses().forEach(h=>{d=J(d,h)}),d},l=this.collectedTrainableWeights.map(p=>p.read()),u=!0;return[this.optimizer_.minimize(i,u,l)].concat(a)}}makeTestFunction(){this.testFunction=e=>V(()=>{let t=[],n,o=e.slice(0,this.inputs.length),s=e.slice(this.inputs.length,this.inputs.length+this.outputs.length),a=[];for(let u=0;u<this.inputs.length;++u)a.push({key:this.inputs[u],value:o[u]});let i=new Es(a),l=Ju(this.outputs,i);for(let u=0;u<this.lossFunctions.length;++u){let c=this.lossFunctions[u],p=dt(c(s[u],l[u]));u===0?n=p:n=J(n,p),t.push(n)}for(let u=0;u<this.metricsTensors.length;++u){let c=this.metricsTensors[u][0],p=this.metricsTensors[u][1],m=dt(c(s[p],l[p]));t.push(m)}return t})}async fit(e,t,n={}){return z3(this,e,t,n)}async fitDataset(e,t){return P3(this,e,t)}async trainOnBatch(e,t){let n=await this.standardizeUserData(e,t),o=n[0],s=n[1],i=this.makeTrainFunction()(o.concat(s)),l=[];for(let u of i){let c=await u.data();l.push(c[0])}return Ee(i),wr(l)}getNamedWeights(e){let t=[],n=e!=null&&e.trainableOnly,o=n?this.trainableWeights:this.weights,s=this.getWeights(n);for(let a=0;a<o.length;++a)n&&!o[a].trainable||t.push({name:o[a].originalName,tensor:s[a]});return t}set stopTraining(e){this.stopTraining_=e}get stopTraining(){return this.stopTraining_}get optimizer(){return this.optimizer_}set optimizer(e){this.optimizer_!==e&&(this.optimizer_=e,this.isOptimizerOwned=!1)}dispose(){let e=super.dispose();if(e.refCountAfterDispose===0&&this.optimizer!=null&&this.isOptimizerOwned){let t=Vm().numTensors;this.optimizer_.dispose(),e.numDisposedVariables+=t-Vm().numTensors}return e}getLossIdentifiers(){let e;if(typeof this.loss=="string")e=mo(this.loss);else if(Array.isArray(this.loss)){for(let t of this.loss)if(typeof t!="string")throw new Error("Serialization of non-string loss is not supported.");e=this.loss.map(t=>mo(t))}else{let t=Object.keys(this.loss);e={};let n=this.loss;for(let o of t)if(typeof n[o]=="string")e[o]=mo(n[o]);else throw new Error("Serialization of non-string loss is not supported.")}return e}getMetricIdentifiers(){if(typeof this.metrics=="string"||typeof this.metrics=="function")return[mo(xd(this.metrics))];if(Array.isArray(this.metrics))return this.metrics.map(e=>mo(xd(e)));{let e={};for(let t in this.metrics)e[t]=mo(xd(this.metrics[t]));return e}}getTrainingConfig(){return{loss:this.getLossIdentifiers(),metrics:this.getMetricIdentifiers(),optimizer_config:{class_name:this.optimizer.getClassName(),config:this.optimizer.getConfig()}}}loadTrainingConfig(e){if(e.weighted_metrics!=null)throw new Error("Loading weight_metrics is not supported yet.");if(e.loss_weights!=null)throw new Error("Loading loss_weights is not supported yet.");if(e.sample_weight_mode!=null)throw new Error("Loading sample_weight_mode is not supported yet.");let t=Zu(e.optimizer_config),n=an(t),o;if(typeof e.loss=="string")o=Pa(e.loss);else if(Array.isArray(e.loss))o=e.loss.map(a=>Pa(a));else if(e.loss!=null){o={};for(let a in e.loss)o[a]=Pa(e.loss[a])}let s;if(Array.isArray(e.metrics))s=e.metrics.map(a=>Pa(a));else if(e.metrics!=null){s={};for(let a in e.metrics)s[a]=Pa(e.metrics[a])}this.compile({loss:o,metrics:s,optimizer:n})}async save(e,t){if(typeof e=="string"){let u=Er.getSaveHandlers(e);if(u.length===0)throw new z(`Cannot find any save handlers for URL '${e}'`);if(u.length>1)throw new z(`Found more than one (${u.length}) save handlers for URL '${e}'`);e=u[0]}if(e.save==null)throw new z("LayersModel.save() cannot proceed because the IOHandler provided does not have the `save` attribute defined.");let n=await Er.encodeWeights(this.getNamedWeights(t)),o=!1,s=null,i={modelTopology:this.toJSON(s,o),format:hte,generatedBy:`TensorFlow.js tfjs-layers v${yd}`,convertedBy:null};if((t==null?!1:t.includeOptimizer)&&this.optimizer!=null){i.trainingConfig=this.getTrainingConfig();let u="optimizer",{data:c,specs:p}=await Er.encodeWeights(await this.optimizer.getWeights(),u);n.specs.push(...p),n.data=Er.concatenateArrayBuffers([n.data,c])}if(this.userDefinedMetadata!=null){let u=!0;a0(this.userDefinedMetadata,this.name,u),i.userDefinedMetadata=this.userDefinedMetadata}return i.weightData=n.data,i.weightSpecs=n.specs,e.save(i)}setUserDefinedMetadata(e){a0(e,this.name),this.userDefinedMetadata=e}getUserDefinedMetadata(){return this.userDefinedMetadata}};jn.className="Model";Q.registerClass(jn);var m0=class extends jn{};m0.className="Functional";Q.registerClass(m0);async function W3(r,e){"modelTopology"in r||(r={modelTopology:r}),r=r;let t=r.modelTopology;t.model_config!=null&&(t=t.model_config);let n=Zu(t),o=an(n,e);if(r.weightsManifest!=null){let s=await Er.loadWeights(r.weightsManifest,r.pathPrefix,o.weights.map(i=>i.originalName)),a={};for(let i of o.weights)a[i.originalName]=s[i.originalName];o.loadWeights(a),Ee(s)}return o}async function j3(r,e){if(e==null&&(e={}),typeof r=="string"){let t=Er.getLoadHandlers(r,e);if(t.length===0)t.push(Er.browserHTTPRequest(r,e));else if(t.length>1)throw new z(`Found more than one (${t.length}) load handlers for URL '${r}'`);r=t[0]}return gte(r,void 0,e)}async function gte(r,e,t){if(t==null&&(t={}),r.load==null)throw new z("Cannot proceed with model loading because the IOHandler provided does not have the `load` method implemented.");let n=await r.load(),o=n.modelTopology;o.model_config!=null&&(o=o.model_config);let s=t.strict==null?!0:t.strict,a=n.weightData!=null&&n.weightSpecs!=null&&s,i=an(Zu(o),e,a),l=n.trainingConfig;if(l!=null&&i.loadTrainingConfig(l),n.userDefinedMetadata!=null&&i.setUserDefinedMetadata(n.userDefinedMetadata),n.weightData!=null){if(n.weightSpecs==null)throw new z("LayersModel artifacts contains weight data, but not weight specs. Therefore loading of weights cannot proceed.");let{modelWeights:u,optimizerWeights:c}=xte(n.weightData,n.weightSpecs);i.loadWeights(u,s),i.optimizer!=null&&c.length>0&&await i.optimizer.setWeights(c),Ee(u),Ee(c.map(p=>p.tensor))}return i}function xte(r,e){let t=Er.decodeWeights(r,e),n={},o=[];return e.forEach(s=>{s.group==="optimizer"?o.push({name:s.name,tensor:t[s.name]}):n[s.name]=t[s.name]}),{modelWeights:n,optimizerWeights:o}}var Ba=class extends jn{constructor(e){super({inputs:[],outputs:[]});if(e=e||{},this.trainable=!0,this.built=!1,this.name=e.name!=null?e.name:Sl("sequential_"),e.layers!=null)for(let t of e.layers)this.add(t)}checkShape(e){if(e.inboundNodes[0].outputTensors[0].shape.some(n=>n<0))throw new z(`Negative dimension size caused by adding layer ${e.name} with input shape [${e.inboundNodes[0].inputTensors[0].shape}]`)}add(e){let t=e instanceof Ba||e instanceof jn,n;if(t){if(n=e,n.outputs.length!==1)throw new z("All layers in a Sequential model should have a single output tensor. For multi-output layers, use the functional API.");if(n.inputs.length!==1)throw new z("All layers in a Sequential model should have a single input tensor. For multi-input layers, use the functional API.")}if(this.outputs.length===0){if(e.inboundNodes.length===0){if(e.batchInputShape==null)throw new z("The first layer in a Sequential model must get an `inputShape` or `batchInputShape` argument.");let o=Hx({batchShape:e.batchInputShape,dtype:e.dtype,name:e.name+"_input"});e.apply(o)}if(t)this.outputs=n.outputs,this.inputs=n.inputs;else{if(e.inboundNodes.length!==1)throw new z(`A layer added to a Sequential model must not already be connected somewhere else. LayersModel received layer ${e.name} which has ${e.inboundNodes.length} pre-existing inbound connections.`);if(e.inboundNodes[0].outputTensors.length!==1)throw new z("All layers in a Sequential model should have a single output tensor. For multi-output layers, use the functional API.");this.checkShape(e),this.outputs=[e.inboundNodes[0].outputTensors[0]],this.inputs=t0(this.outputs[0])}this.inboundNodes=[],new Tl({outboundLayer:this,inboundLayers:[],nodeIndices:[],tensorIndices:[],inputTensors:this.inputs,outputTensors:this.outputs,inputMasks:po(null,this.inputs.length),outputMasks:[null],inputShapes:this.inputs.map(o=>o.shape),outputShapes:this.outputs[0].shape})}else{let o=e.apply(this.outputs[0]);if(Array.isArray(o))throw new TypeError("All layers in a Sequential model should have a single output tensor. For multi-output layers, use the functional API.");this.checkShape(e),this.outputs=[o],this.inboundNodes[0].outputTensors=this.outputs,this.inboundNodes[0].outputShapes=[this.outputs[0].shape]}this.layers.push(e),this.built=!1}pop(){if(this.layers.length===0)throw new TypeError("There are no layers in the model.");if(this.layers.pop(),this.layers.length===0)this.outputs=[],this.inboundNodes=[],this.outboundNodes=[];else{let e=this.layers.length-1;this.layers[e].outboundNodes=[],this.outputs=[this.layers[e].output],this.inboundNodes[0].outputTensors=this.outputs,this.inboundNodes[0].outputShapes=[this.outputs[0].shape]}}call(e,t){return this.model==null&&this.build(),this.model.call(e,t)}build(e){if(Xe(e),this.inputs.length===0||this.outputs.length===0)throw new TypeError("Sequential model cannot be built: model is empty. Add some layers first.");this.model=new jn({inputs:this.inputs,outputs:this.outputs[0],name:this.name+"_model"}),this.model.trainable=this.trainable,this.supportsMasking=this.model.supportsMasking,this.inputLayers=this.model.inputLayers,this.inputLayersNodeIndices=this.model.inputLayersNodeIndices,this.inputLayersTensorIndices=this.model.inputLayersTensorIndices,this.outputLayers=this.model.outputLayers,this.outputLayersNodeIndices=this.model.outputLayersNodeIndices,this.outputLayersTensorIndices=this.model.outputLayersTensorIndices,this.nodesByDepth=this.model.nodesByDepth,this.containerNodes=this.model.containerNodes,this.outputNames=this.model.outputNames,this.inputNames=this.model.inputNames,this.built=!0}countParams(){return this.built||this.build(),super.countParams()}summary(e,t,n=console.log){this.built||this.build(),super.summary(e,t,n)}setWeights(e){this.model==null&&this.build(),this.model.setWeights(e)}evaluate(e,t,n={}){if(!this.built)throw new Gr("The model needs to be compiled before being used.");return this.model.evaluate(e,t,n)}async evaluateDataset(e,t){if(!this.built)throw new Gr("The model needs to be compiled before being used.");return this.model.evaluateDataset(e,t)}predict(e,t={}){return this.model==null&&this.build(),this.model.predict(e,t)}predictOnBatch(e){return this.model==null&&this.build(),this.model.predictOnBatch(e)}compile(e){this.build(),this.model.compile(e),this.optimizer_=this.model.optimizer,this.isOptimizerOwned=this.model.isOptimizerOwned,this.loss=this.model.loss,this.metrics=this.model.metrics,this.metricsTensors=this.model.metricsTensors,this.metricsNames=this.model.metricsNames}get optimizer(){return this.model==null?void 0:this.model.optimizer}set optimizer(e){this.model.optimizer=e}async fit(e,t,n={}){if(!this.built)throw new Gr("The model needs to be compiled before being used.");return this.model.fit(e,t,n)}async fitDataset(e,t){if(!this.built)throw new Gr("The model needs to be compiled before being used.");return this.model.fitDataset(e,t)}async trainOnBatch(e,t){return this.model.trainOnBatch(e,t)}static fromConfig(e,t,n={},o=!1){let s,a={};if(t instanceof Array){if(t[0].className==null||t[0].className==="Merge")throw new z("Legacy serialization format not supported yet.");s=t}else y.assert(t.layers!=null,()=>"When the config data for a Sequential model is not an Array, it must be an Object that contains the 'layers' field."),s=t.layers,delete t.layers,a=t;let i=new e(a);if(!(i instanceof Ba))throw new Se(`Sequential.fromConfig called on non-Sequential input: ${i}`);for(let l of s){let c=an(l,void 0,o);o&&c.setFastWeightInitDuringBuild(!0),i.add(c)}return i}set stopTraining(e){if(this.model==null)throw new z("Cannot set the stopTraining property of a sequential model before it is compiled.");this.model.stopTraining=e}get stopTraining(){if(this.model==null)throw new z("Cannot get the stopTraining property of a sequential model before it is compiled.");return this.model.stopTraining}getConfig(){let e=[];for(let t of this.layers){let n={};n.className=t.getClassName(),n.config=t.getConfig(),e.push(n)}return{name:this.name,layers:e}}};Ba.className="Sequential";Q.registerClass(Ba);function yte(r){return new jn(r)}function bte(r){return new Ba(r)}function wte(r,e){return e==null&&(e={}),j3(r,e)}function f0(r){return Hx(r)}function _te(r,e){wn.registerCallbackConstructor(r,e)}var ln=class extends Q.Serializable{getConfig(){return{}}},d0=class extends ln{apply(e,t=1){return f3(e,t)}};d0.className="elu";Q.registerClass(d0);var h0=class extends ln{apply(e){return yu(e)}};h0.className="selu";Q.registerClass(h0);var g0=class extends ln{apply(e){return $r(e)}};g0.className="relu";Q.registerClass(g0);var x0=class extends ln{apply(e){return V(()=>_s(6,$r(e)))}};x0.className="relu6";Q.registerClass(x0);var y0=class extends ln{apply(e){return e}};y0.className="linear";Q.registerClass(y0);var b0=class extends ln{apply(e){return Dr(e)}};b0.className="sigmoid";Q.registerClass(b0);var w0=class extends ln{apply(e){return h3(e)}};w0.className="hardSigmoid";Q.registerClass(w0);var _0=class extends ln{apply(e){return eo(e)}};_0.className="softplus";Q.registerClass(_0);var k0=class extends ln{apply(e){return d3(e)}};k0.className="softsign";Q.registerClass(k0);var v0=class extends ln{apply(e){return gs(e)}};v0.className="tanh";Q.registerClass(v0);var bd=class extends ln{apply(e,t=-1){return Ia(e,t)}};bd.className="softmax";Q.registerClass(bd);var C0=class extends ln{apply(e,t=-1){return pu(e,t)}};C0.className="logSoftmax";Q.registerClass(C0);var I0=class extends ln{apply(e,t=1){return V(()=>Dr(e.mul(t)).mul(e))}};I0.className="swish";Q.registerClass(I0);var N0=class extends ln{apply(e){return V(()=>O(e,gs(eo(e))))}};N0.className="mish";Q.registerClass(N0);function Ds(r){return r.getClassName()}function S0(r,e={}){return pi(r,Q.SerializationMap.getMap().classNameMap,e,"activation")}function $s(r){if(r==null){let e={};return e.className="linear",e.config={},S0(e)}if(typeof r=="string"){let e={};return e.className=r,e.config={},S0(e)}else return r instanceof ln?r:S0(r)}function T0(r){if(r!=null&&typeof r!="object")throw new Error(`Argument to L1L2 regularizer's constructor is expected to be an object, but received: ${r}`)}var A0=class extends Q.Serializable{},Qu=class extends A0{constructor(e){super();T0(e),this.l1=e==null||e.l1==null?.01:e.l1,this.l2=e==null||e.l2==null?.01:e.l2,this.hasL1=this.l1!==0,this.hasL2=this.l2!==0}apply(e){return V(()=>{let t=ht([1]);return this.hasL1&&(t=J(t,fe(O(this.l1,Tt(e))))),this.hasL2&&(t=J(t,fe(O(this.l2,Ku(e))))),t.asScalar()})}getConfig(){return{l1:this.l1,l2:this.l2}}static fromConfig(e,t){return new e({l1:t.l1,l2:t.l2})}};Qu.className="L1L2";Q.registerClass(Qu);function U3(r){return T0(r),new Qu({l1:r!=null?r.l1:null,l2:0})}function H3(r){return T0(r),new Qu({l2:r!=null?r.l2:null,l1:0})}var q3={l1l2:"L1L2"};function it(r){return Mp(r)}function K3(r,e={}){return pi(r,Q.SerializationMap.getMap().classNameMap,e,"regularizer")}function wt(r){if(r==null)return null;if(typeof r=="string"){let t={className:r in q3?q3[r]:r,config:{}};return K3(t)}else return r instanceof A0?r:K3(r)}var wd=class extends Le{constructor(e){super(e==null?{}:e);this.supportsMasking=!0,e!=null&&(this.maxValue=e.maxValue)}call(e,t){e=Fe(e);let n=$r(e);return this.maxValue!=null&&(n=ur(n,0,this.maxValue)),n}computeOutputShape(e){return e}getConfig(){let e={maxValue:this.maxValue},t=super.getConfig();return Object.assign(e,t),e}};wd.className="ReLU";Q.registerClass(wd);var _d=class extends Le{constructor(e){super(e==null?{}:e);this.DEFAULT_ALPHA=.3,e==null&&(e={}),this.alpha=e.alpha==null?this.DEFAULT_ALPHA:e.alpha}call(e,t){let n=Fe(e);return ba(n,this.alpha)}computeOutputShape(e){return e}getConfig(){let e={alpha:this.alpha},t=super.getConfig();return Object.assign(e,t),e}};_d.className="LeakyReLU";Q.registerClass(_d);var kd=class extends Le{constructor(e){super(e==null?{}:e);if(this.DEFAULT_ALPHA_INITIALIZER="zeros",e==null&&(e={}),this.supportsMasking=!0,this.alphaInitializer=pt(e.alphaInitializer||this.DEFAULT_ALPHA_INITIALIZER),this.alphaRegularizer=wt(e.alphaRegularizer),this.alphaConstraint=Lt(e.alphaConstraint),e.sharedAxes==null)this.sharedAxes=null;else if(Array.isArray(e.sharedAxes))this.sharedAxes=e.sharedAxes;else if(typeof e.sharedAxes=="number")this.sharedAxes=[e.sharedAxes];else throw new z(`Expected sharedAxes to be a number or an array of numbers, but got ${e.sharedAxes}`)}build(e){e=Xe(e);let t=e.slice(1);if(this.sharedAxes!=null)for(let o of this.sharedAxes)t[o-1]=1;this.alpha=this.addWeight("alpha",t,"float32",this.alphaInitializer,this.alphaRegularizer,!0,this.alphaConstraint);let n={};if(this.sharedAxes!=null)for(let o=1;o<e.length;++o)n[o]=e[o];this.inputSpec=[new Nt({ndim:e.length,axes:n})],this.built=!0}call(e,t){return e=Fe(e),va(e,this.alpha.read())}getConfig(){let e={alphaInitializer:It(this.alphaInitializer),alphaRegularizer:it(this.alphaRegularizer),alphaConstraint:Mt(this.alphaConstraint),sharedAxes:this.sharedAxes},t=super.getConfig();return Object.assign(e,t),e}};kd.className="PReLU";Q.registerClass(kd);var vd=class extends Le{constructor(e){super(e==null?{}:e);if(this.DEFAULT_ALPHA=1,e==null&&(e={}),e.alpha!=null&&e.alpha!==this.DEFAULT_ALPHA)throw new Se(`Non-default alpha value (${e.alpha}) is not supported by the ELU layer yet.`);this.alpha=e.alpha==null?this.DEFAULT_ALPHA:e.alpha}call(e,t){let n=Fe(e);return ys(n)}computeOutputShape(e){return e}getConfig(){let e={alpha:this.alpha},t=super.getConfig();return Object.assign(e,t),e}};vd.className="ELU";Q.registerClass(vd);var Cd=class extends Le{constructor(e){super(e==null?{}:e);this.DEFAULT_THETA=1,e==null&&(e={}),this.theta=e.theta==null?this.DEFAULT_THETA:e.theta}call(e,t){let n=Fe(e);return n.mul(Ma(n.greater(this.theta),"float32"))}computeOutputShape(e){return e}getConfig(){let e={theta:this.theta},t=super.getConfig();return Object.assign(e,t),e}};Cd.className="ThresholdedReLU";Q.registerClass(Cd);var Id=class extends Le{constructor(e){super(e==null?{}:e);this.DEFAULT_AXIS=1,e==null&&(e={}),this.softmax=new bd().apply,this.axis=e.axis==null?this.DEFAULT_AXIS:e.axis}call(e,t){let n=Fe(e);return this.softmax(n,this.axis)}computeOutputShape(e){return e}getConfig(){let e={axis:this.axis},t=super.getConfig();return Object.assign(e,t),e}};Id.className="Softmax";Q.registerClass(Id);function $l(r,e,t){if(typeof r=="number")return po(r,e);if(r.length!==e)throw new z(`The ${t} argument must be an integer or tuple of ${e} integers. Received: ${r.length} elements.`);for(let n=0;n<e;++n){let o=r[n];if(!l3(o))throw new z(`The ${t} argument must be an integer or tuple of ${e} integers. Received: ${JSON.stringify(r)} including a non-integer number ${o}`)}return r}function _n(r,e,t,n,o=1){if(r==null)return r;let s=e+(e-1)*(o-1),a;return t==="same"?a=r:a=r-s+1,Math.floor((a+n-1)/n)}function Rs(r,e,t,n){if(r==null)return null;if(n==="valid")r=r*e+As([t-e,0]);else if(n==="same")r=r*e;else throw new z(`Unsupport padding mode: ${n}.`);return r}function Nd(r,e){return V(()=>(Ot(e),e==="channelsFirst"?Ue(r,[0,2,3,1]):r))}function E0(r,e){return V(()=>(Ot(e),e==="channelsFirst"?Ue(r,[0,2,3,4,1]):r))}function kte(r,e,t,n=1,o="valid",s,a=1){return V(()=>{if(s==null&&(s=rn()),Ot(s),r.shape.length!==3)throw new z(`The input of a conv1dWithBias operation should be 3, but is ${r.shape.length} instead.`);if(e.shape.length!==3)throw new z(`The kernel for a conv1dWithBias operation should be 3, but is ${e.shape.length} instead`);if(t!=null&&t.shape.length!==1)throw new z(`The bias for a conv1dWithBias operation should be 1, but is ${e.shape.length} instead`);if(s==="channelsFirst"&&(r=Ue(r,[0,2,1])),o==="causal")throw new Se("The support for CAUSAL padding mode in conv1dWithBias is not implemented yet.");let i=ou(r,e,n,o==="same"?"same":"valid","NWC",a);return t!=null&&(i=on(i,t)),i})}function X3(r,e,t,n=[1,1],o="valid",s,a,i=null){return V(()=>{if(s==null&&(s=rn()),Ot(s),r.rank!==3&&r.rank!==4)throw new z(`conv2dWithBiasActivation expects input to be of rank 3 or 4, but received ${r.rank}.`);if(e.rank!==3&&e.rank!==4)throw new z(`conv2dWithBiasActivation expects kernel to be of rank 3 or 4, but received ${r.rank}.`);let l=Nd(r,s);if(o==="causal")throw new Se("The support for CAUSAL padding mode in conv1dWithBias is not implemented yet.");return l=no.conv2d({x:l,filter:e,strides:n,pad:o==="same"?"same":"valid",dilations:a,dataFormat:"NHWC",bias:t,activation:i}),s==="channelsFirst"&&(l=Ue(l,[0,3,1,2])),l})}function vte(r,e,t,n=[1,1,1],o="valid",s,a){return V(()=>{if(s==null&&(s=rn()),Ot(s),r.rank!==4&&r.rank!==5)throw new z(`conv3dWithBias expects input to be of rank 4 or 5, but received ${r.rank}.`);if(e.rank!==4&&e.rank!==5)throw new z(`conv3dWithBias expects kernel to be of rank 4 or 5, but received ${r.rank}.`);let i=E0(r,s);if(o==="causal")throw new Se("The support for CAUSAL padding mode in conv3dWithBias is not implemented yet.");return i=Qm(i,e,n,o==="same"?"same":"valid","NDHWC",a),t!=null&&(i=on(i,t)),s==="channelsFirst"&&(i=Ue(i,[0,4,1,2,3])),i})}var tm=class extends Le{constructor(e,t){super(t);if(this.bias=null,this.DEFAULT_KERNEL_INITIALIZER="glorotNormal",this.DEFAULT_BIAS_INITIALIZER="zeros",tm.verifyArgs(t),this.rank=e,Ht(this.rank,"rank"),this.rank!==1&&this.rank!==2&&this.rank!==3)throw new Se(`Convolution layer for rank other than 1, 2, or 3 (${this.rank}) is not implemented yet.`);if(this.kernelSize=$l(t.kernelSize,e,"kernelSize"),this.strides=$l(t.strides==null?1:t.strides,e,"strides"),this.padding=t.padding==null?"valid":t.padding,nn(this.padding),this.dataFormat=t.dataFormat==null?"channelsLast":t.dataFormat,Ot(this.dataFormat),this.activation=$s(t.activation),this.useBias=t.useBias==null?!0:t.useBias,this.biasInitializer=pt(t.biasInitializer||this.DEFAULT_BIAS_INITIALIZER),this.biasConstraint=Lt(t.biasConstraint),this.biasRegularizer=wt(t.biasRegularizer),this.activityRegularizer=wt(t.activityRegularizer),this.dilationRate=$l(t.dilationRate==null?1:t.dilationRate,e,"dilationRate"),this.rank===1&&Array.isArray(this.dilationRate)&&this.dilationRate.length!==1)throw new z(`dilationRate must be a number or an array of a single number for 1D convolution, but received ${JSON.stringify(this.dilationRate)}`);if(this.rank===2){if(typeof this.dilationRate=="number")this.dilationRate=[this.dilationRate,this.dilationRate];else if(this.dilationRate.length!==2)throw new z(`dilationRate must be a number or array of two numbers for 2D convolution, but received ${JSON.stringify(this.dilationRate)}`)}else if(this.rank===3){if(typeof this.dilationRate=="number")this.dilationRate=[this.dilationRate,this.dilationRate,this.dilationRate];else if(this.dilationRate.length!==3)throw new z(`dilationRate must be a number or array of three numbers for 3D convolution, but received ${JSON.stringify(this.dilationRate)}`)}}static verifyArgs(e){if(Gn("kernelSize"in e,"required key 'kernelSize' not in config"),typeof e.kernelSize!="number"&&!Ox(e.kernelSize,"number",1,3))throw new z(`BaseConv expects config.kernelSize to be number or number[] with length 1, 2, or 3, but received ${JSON.stringify(e.kernelSize)}.`)}getConfig(){let e={kernelSize:this.kernelSize,strides:this.strides,padding:this.padding,dataFormat:this.dataFormat,dilationRate:this.dilationRate,activation:Ds(this.activation),useBias:this.useBias,biasInitializer:It(this.biasInitializer),biasRegularizer:it(this.biasRegularizer),activityRegularizer:it(this.activityRegularizer),biasConstraint:Mt(this.biasConstraint)},t=super.getConfig();return Object.assign(e,t),e}},ec=class extends tm{constructor(e,t){super(e,t);this.kernel=null,ec.verifyArgs(t),this.filters=t.filters,Ht(this.filters,"filters"),this.kernelInitializer=pt(t.kernelInitializer||this.DEFAULT_KERNEL_INITIALIZER),this.kernelConstraint=Lt(t.kernelConstraint),this.kernelRegularizer=wt(t.kernelRegularizer)}build(e){e=Xe(e);let t=this.dataFormat==="channelsFirst"?1:e.length-1;if(e[t]==null)throw new z(`The channel dimension of the input should be defined. Found ${e[t]}`);let n=e[t],o=this.kernelSize.concat([n,this.filters]);this.kernel=this.addWeight("kernel",o,null,this.kernelInitializer,this.kernelRegularizer,!0,this.kernelConstraint),this.useBias&&(this.bias=this.addWeight("bias",[this.filters],null,this.biasInitializer,this.biasRegularizer,!0,this.biasConstraint)),this.inputSpec=[{ndim:this.rank+2,axes:{[t]:n}}],this.built=!0}call(e,t){return V(()=>{e=Fe(e);let n,o=this.bias==null?null:this.bias.read(),s=Px(this.activation.getClassName());if(s!=null&&this.rank===2)n=X3(e,this.kernel.read(),o,this.strides,this.padding,this.dataFormat,this.dilationRate,s);else{if(this.rank===1)n=kte(e,this.kernel.read(),o,this.strides[0],this.padding,this.dataFormat,this.dilationRate[0]);else if(this.rank===2)n=X3(e,this.kernel.read(),o,this.strides,this.padding,this.dataFormat,this.dilationRate);else if(this.rank===3)n=vte(e,this.kernel.read(),o,this.strides,this.padding,this.dataFormat,this.dilationRate);else throw new Se("convolutions greater than 3D are not implemented yet.");this.activation!=null&&(n=this.activation.apply(n))}return n})}computeOutputShape(e){e=Xe(e);let t=[],n=this.dataFormat==="channelsLast"?e.slice(1,e.length-1):e.slice(2);for(let s=0;s<n.length;++s){let a=_n(n[s],this.kernelSize[s],this.padding,this.strides[s],typeof this.dilationRate=="number"?this.dilationRate:this.dilationRate[s]);t.push(a)}let o=[e[0]];return this.dataFormat==="channelsLast"?(o=o.concat(t),o.push(this.filters)):(o.push(this.filters),o=o.concat(t)),o}getConfig(){let e={filters:this.filters,kernelInitializer:It(this.kernelInitializer),kernelRegularizer:it(this.kernelRegularizer),kernelConstraint:Mt(this.kernelConstraint)},t=super.getConfig();return Object.assign(e,t),e}static verifyArgs(e){if(!("filters"in e)||typeof e.filters!="number"||e.filters<1)throw new z(`Convolution layer expected config.filters to be a 'number' > 0 but got ${JSON.stringify(e.filters)}`)}},Rl=class extends ec{constructor(e){super(2,e);Rl.verifyArgs(e)}getConfig(){let e=super.getConfig();return delete e.rank,e}static verifyArgs(e){if(typeof e.kernelSize!="number"&&!Ox(e.kernelSize,"number",1,2))throw new z(`Conv2D expects config.kernelSize to be number or number[] with length 1 or 2, but received ${JSON.stringify(e.kernelSize)}.`)}};Rl.className="Conv2D";Q.registerClass(Rl);var Fl=class extends ec{constructor(e){super(3,e);Fl.verifyArgs(e)}getConfig(){let e=super.getConfig();return delete e.rank,e}static verifyArgs(e){if(typeof e.kernelSize!="number"&&!(Array.isArray(e.kernelSize)&&(e.kernelSize.length===1||e.kernelSize.length===3)))throw new z(`Conv3D expects config.kernelSize to be number or [number, number, number], but received ${JSON.stringify(e.kernelSize)}.`)}};Fl.className="Conv3D";Q.registerClass(Fl);var Sd=class extends Rl{constructor(e){super(e);if(this.inputSpec=[new Nt({ndim:4})],this.padding!=="same"&&this.padding!=="valid")throw new z(`Conv2DTranspose currently supports only padding modes 'same' and 'valid', but received padding mode ${this.padding}`)}build(e){if(e=Xe(e),e.length!==4)throw new z("Input should have rank 4; Received input shape: "+JSON.stringify(e));let t=this.dataFormat==="channelsFirst"?1:e.length-1;if(e[t]==null)throw new z("The channel dimension of the inputs should be defined. Found `None`.");let n=e[t],o=this.kernelSize.concat([this.filters,n]);this.kernel=this.addWeight("kernel",o,"float32",this.kernelInitializer,this.kernelRegularizer,!0,this.kernelConstraint),this.useBias&&(this.bias=this.addWeight("bias",[this.filters],"float32",this.biasInitializer,this.biasRegularizer,!0,this.biasConstraint)),this.inputSpec=[new Nt({ndim:4,axes:{[t]:n}})],this.built=!0}call(e,t){return V(()=>{let n=Fe(e);if(n.shape.length!==4)throw new z(`Conv2DTranspose.call() expects input tensor to be rank-4, but received a tensor of rank-${n.shape.length}`);let o=n.shape,s=o[0],a,i;this.dataFormat==="channelsFirst"?(a=2,i=3):(a=1,i=2);let l=o[a],u=o[i],c=this.kernelSize[0],p=this.kernelSize[1],m=this.strides[0],f=this.strides[1],d=Rs(l,m,c,this.padding),h=Rs(u,f,p,this.padding),g=[s,d,h,this.filters];this.dataFormat!=="channelsLast"&&(n=Ue(n,[0,2,3,1]));let x=su(n,this.kernel.read(),g,this.strides,this.padding);return this.dataFormat!=="channelsLast"&&(x=Ue(x,[0,3,1,2])),this.bias!=null&&(x=on(x,this.bias.read(),this.dataFormat)),this.activation!=null&&(x=this.activation.apply(x)),x})}computeOutputShape(e){e=Xe(e);let t=e.slice(),n,o,s;this.dataFormat==="channelsFirst"?(n=1,o=2,s=3):(n=3,o=1,s=2);let a=this.kernelSize[0],i=this.kernelSize[1],l=this.strides[0],u=this.strides[1];return t[n]=this.filters,t[o]=Rs(t[o],l,a,this.padding),t[s]=Rs(t[s],u,i,this.padding),t}getConfig(){let e=super.getConfig();return delete e.dilationRate,e}};Sd.className="Conv2DTranspose";Q.registerClass(Sd);var Td=class extends Fl{constructor(e){super(e);if(this.inputSpec=[new Nt({ndim:5})],this.padding!=="same"&&this.padding!=="valid")throw new z(`Conv3DTranspose currently supports only padding modes 'same' and 'valid', but received padding mode ${this.padding}`)}build(e){if(e=Xe(e),e.length!==5)throw new z("Input should have rank 5; Received input shape: "+JSON.stringify(e));let t=this.dataFormat==="channelsFirst"?1:e.length-1;if(e[t]==null)throw new z("The channel dimension of the inputs should be defined. Found `None`.");let n=e[t],o=this.kernelSize.concat([this.filters,n]);this.kernel=this.addWeight("kernel",o,"float32",this.kernelInitializer,this.kernelRegularizer,!0,this.kernelConstraint),this.useBias&&(this.bias=this.addWeight("bias",[this.filters],"float32",this.biasInitializer,this.biasRegularizer,!0,this.biasConstraint)),this.inputSpec=[new Nt({ndim:5,axes:{[t]:n}})],this.built=!0}call(e,t){return V(()=>{let n=Fe(e);if(n.shape.length!==5)throw new z(`Conv3DTranspose.call() expects input tensor to be rank-4, but received a tensor of rank-${n.shape.length}`);let o=n.shape,s=o[0],a,i,l;this.dataFormat==="channelsFirst"?(l=2,a=3,i=4):(l=1,a=2,i=3);let u=o[l],c=o[a],p=o[i],m=this.kernelSize[0],f=this.kernelSize[1],d=this.kernelSize[2],h=this.strides[0],g=this.strides[1],x=this.strides[2],b=Rs(u,h,m,this.padding),w=Rs(c,g,f,this.padding),_=Rs(p,x,d,this.padding),I=[s,b,w,_,this.filters];this.dataFormat!=="channelsLast"&&(n=Ue(n,[0,2,3,4,1]));let E=Qw(n,this.kernel.read(),I,this.strides,this.padding);return this.dataFormat!=="channelsLast"&&(E=Ue(E,[0,4,1,2,3])),this.bias!==null&&(E=on(E,this.bias.read(),this.dataFormat)),this.activation!==null&&(E=this.activation.apply(E)),E})}computeOutputShape(e){e=Xe(e);let t=e.slice(),n,o,s,a;this.dataFormat==="channelsFirst"?(n=1,o=2,s=3,a=4):(n=4,o=1,s=2,a=3);let i=this.kernelSize[0],l=this.kernelSize[1],u=this.kernelSize[2],c=this.strides[0],p=this.strides[1],m=this.strides[2];return t[n]=this.filters,t[o]=Rs(t[o],c,i,this.padding),t[s]=Rs(t[s],p,l,this.padding),t[a]=Rs(t[a],m,u,this.padding),t}getConfig(){let e=super.getConfig();return delete e.dilationRate,e}};Td.className="Conv3DTranspose";Q.registerClass(Td);var D0=class extends ec{constructor(e,t){super(e,t);if(this.DEFAULT_DEPTHWISE_INITIALIZER="glorotUniform",this.DEFAULT_POINTWISE_INITIALIZER="glorotUniform",this.depthwiseKernel=null,this.pointwiseKernel=null,t.filters==null)throw new z("The `filters` configuration field is required by SeparableConv, but is unspecified.");if(t.kernelInitializer!=null||t.kernelRegularizer!=null||t.kernelConstraint!=null)throw new z("Fields kernelInitializer, kernelRegularizer and kernelConstraint are invalid for SeparableConv2D. Use depthwiseInitializer, depthwiseRegularizer, depthwiseConstraint, pointwiseInitializer, pointwiseRegularizer and pointwiseConstraint instead.");if(t.padding!=null&&t.padding!=="same"&&t.padding!=="valid")throw new z(`SeparableConv${this.rank}D supports only padding modes: 'same' and 'valid', but received ${JSON.stringify(t.padding)}`);this.depthMultiplier=t.depthMultiplier==null?1:t.depthMultiplier,this.depthwiseInitializer=pt(t.depthwiseInitializer||this.DEFAULT_DEPTHWISE_INITIALIZER),this.depthwiseRegularizer=wt(t.depthwiseRegularizer),this.depthwiseConstraint=Lt(t.depthwiseConstraint),this.pointwiseInitializer=pt(t.depthwiseInitializer||this.DEFAULT_POINTWISE_INITIALIZER),this.pointwiseRegularizer=wt(t.pointwiseRegularizer),this.pointwiseConstraint=Lt(t.pointwiseConstraint)}build(e){if(e=Xe(e),e.length<this.rank+2)throw new z(`Inputs to SeparableConv${this.rank}D should have rank ${this.rank+2}, but received input shape: ${JSON.stringify(e)}`);let t=this.dataFormat==="channelsFirst"?1:e.length-1;if(e[t]==null||e[t]<0)throw new z(`The channel dimension of the inputs should be defined, but found ${JSON.stringify(e[t])}`);let n=e[t],o=this.kernelSize.concat([n,this.depthMultiplier]),s=[];for(let i=0;i<this.rank;++i)s.push(1);s.push(n*this.depthMultiplier,this.filters);let a=!0;this.depthwiseKernel=this.addWeight("depthwise_kernel",o,"float32",this.depthwiseInitializer,this.depthwiseRegularizer,a,this.depthwiseConstraint),this.pointwiseKernel=this.addWeight("pointwise_kernel",s,"float32",this.pointwiseInitializer,this.pointwiseRegularizer,a,this.pointwiseConstraint),this.useBias?this.bias=this.addWeight("bias",[this.filters],"float32",this.biasInitializer,this.biasRegularizer,a,this.biasConstraint):this.bias=null,this.inputSpec=[new Nt({ndim:this.rank+2,axes:{[t]:n}})],this.built=!0}call(e,t){return V(()=>{e=Fe(e);let n;if(this.rank===1)throw new Se("1D separable convolution is not implemented yet.");return this.rank===2&&(this.dataFormat==="channelsFirst"&&(e=Ue(e,[0,2,3,1])),n=df(e,this.depthwiseKernel.read(),this.pointwiseKernel.read(),this.strides,this.padding,this.dilationRate,"NHWC")),this.useBias&&(n=on(n,this.bias.read(),this.dataFormat)),this.activation!=null&&(n=this.activation.apply(n)),this.dataFormat==="channelsFirst"&&(n=Ue(n,[0,3,1,2])),n})}getConfig(){let e=super.getConfig();return delete e.rank,delete e.kernelInitializer,delete e.kernelRegularizer,delete e.kernelConstraint,e.depthwiseInitializer=It(this.depthwiseInitializer),e.pointwiseInitializer=It(this.pointwiseInitializer),e.depthwiseRegularizer=it(this.depthwiseRegularizer),e.pointwiseRegularizer=it(this.pointwiseRegularizer),e.depthwiseConstraint=Mt(this.depthwiseConstraint),e.pointwiseConstraint=Mt(this.pointwiseConstraint),e}};D0.className="SeparableConv";var Ad=class extends D0{constructor(e){super(2,e)}};Ad.className="SeparableConv2D";Q.registerClass(Ad);var tc=class extends ec{constructor(e){super(1,e);tc.verifyArgs(e),this.inputSpec=[{ndim:3}]}getConfig(){let e=super.getConfig();return delete e.rank,delete e.dataFormat,e}static verifyArgs(e){if(typeof e.kernelSize!="number"&&!Ox(e.kernelSize,"number",1,1))throw new z(`Conv1D expects config.kernelSize to be number or number[] with length 1, but received ${JSON.stringify(e.kernelSize)}.`)}};tc.className="Conv1D";Q.registerClass(tc);var Ed=class extends Le{constructor(e){super(e);typeof e.cropping=="number"?this.cropping=[[e.cropping,e.cropping],[e.cropping,e.cropping]]:typeof e.cropping[0]=="number"?this.cropping=[[e.cropping[0],e.cropping[0]],[e.cropping[1],e.cropping[1]]]:this.cropping=e.cropping,this.dataFormat=e.dataFormat===void 0?"channelsLast":e.dataFormat,this.inputSpec=[{ndim:4}]}computeOutputShape(e){return this.dataFormat==="channelsFirst"?[e[0],e[1],e[2]-this.cropping[0][0]-this.cropping[0][1],e[3]-this.cropping[1][0]-this.cropping[1][1]]:[e[0],e[1]-this.cropping[0][0]-this.cropping[0][1],e[2]-this.cropping[1][0]-this.cropping[1][1],e[3]]}call(e,t){return V(()=>{if(e=Fe(e),this.dataFormat==="channelsLast"){let n=rd(e,this.cropping[0][0],e.shape[1]-this.cropping[0][0]-this.cropping[0][1],2);return rd(n,this.cropping[1][0],e.shape[2]-this.cropping[1][1]-this.cropping[1][0],3)}else{let n=rd(e,this.cropping[0][0],e.shape[2]-this.cropping[0][0]-this.cropping[0][1],3);return rd(n,this.cropping[1][0],e.shape[3]-this.cropping[1][1]-this.cropping[1][0],4)}})}getConfig(){let e={cropping:this.cropping,dataFormat:this.dataFormat},t=super.getConfig();return Object.assign(e,t),e}};Ed.className="Cropping2D";Q.registerClass(Ed);var Dd=class extends Le{constructor(e){super(e);this.DEFAULT_SIZE=[2,2],this.inputSpec=[{ndim:4}],this.size=e.size==null?this.DEFAULT_SIZE:e.size,this.dataFormat=e.dataFormat==null?"channelsLast":e.dataFormat,Ot(this.dataFormat),this.interpolation=e.interpolation==null?"nearest":e.interpolation,s3(this.interpolation)}computeOutputShape(e){if(this.dataFormat==="channelsFirst"){let t=e[2]==null?null:this.size[0]*e[2],n=e[3]==null?null:this.size[1]*e[3];return[e[0],e[1],t,n]}else{let t=e[1]==null?null:this.size[0]*e[1],n=e[2]==null?null:this.size[1]*e[2];return[e[0],t,n,e[3]]}}call(e,t){return V(()=>{let n=Fe(e),o=n.shape;if(this.dataFormat==="channelsFirst"){n=Ue(n,[0,2,3,1]);let s=this.size[0]*o[2],a=this.size[1]*o[3],i=this.interpolation==="nearest"?n.resizeNearestNeighbor([s,a]):n.resizeBilinear([s,a]);return Ue(i,[0,3,1,2])}else{let s=this.size[0]*o[1],a=this.size[1]*o[2];return this.interpolation==="nearest"?n.resizeNearestNeighbor([s,a]):n.resizeBilinear([s,a])}})}getConfig(){let e={size:this.size,dataFormat:this.dataFormat},t=super.getConfig();return Object.assign(e,t),e}};Dd.className="UpSampling2D";Q.registerClass(Dd);function Cte(r,e,t=[1,1],n="valid",o,s){return V(()=>{o==null&&(o=rn()),Ot(o);let a=Nd(r,o);if(r.rank!==4)throw new z(`Input for depthwiseConv2d is required to be 4-D, but is instead ${r.rank}-D`);if(e.rank!==4)throw new z(`depthwiseKernel is required to be 4-D, but is instead ${e.rank}-D`);return a=xs(a,e,t,n==="same"?"same":"valid","NHWC",s),o==="channelsFirst"&&(a=Ue(a,[0,3,1,2])),a})}var $d=class extends tm{constructor(e){super(2,e);this.depthwiseKernel=null,this.depthMultiplier=e.depthMultiplier==null?1:e.depthMultiplier,this.depthwiseInitializer=pt(e.depthwiseInitializer||this.DEFAULT_KERNEL_INITIALIZER),this.depthwiseConstraint=Lt(e.depthwiseConstraint),this.depthwiseRegularizer=wt(e.depthwiseRegularizer)}build(e){if(e=Xe(e),e.length<4)throw new z(`Inputs to DepthwiseConv2D should have rank 4. Received input shape: ${JSON.stringify(e)}.`);let t=this.dataFormat==="channelsFirst"?1:3;if(e[t]==null||e[t]<0)throw new z(`The channel dimension of the inputs to DepthwiseConv2D should be defined, but is not (${e[t]}).`);let n=e[t],o=[this.kernelSize[0],this.kernelSize[1],n,this.depthMultiplier];this.depthwiseKernel=this.addWeight("depthwise_kernel",o,null,this.depthwiseInitializer,this.depthwiseRegularizer,!0,this.depthwiseConstraint),this.useBias?this.bias=this.addWeight("bias",[n*this.depthMultiplier],null,this.biasInitializer,this.biasRegularizer,!0,this.biasConstraint):this.bias=null,this.built=!0}call(e,t){return V(()=>{e=Fe(e);let n=Cte(e,this.depthwiseKernel.read(),this.strides,this.padding,this.dataFormat,null);return this.useBias&&(n=on(n,this.bias.read(),this.dataFormat)),this.activation!=null&&(n=this.activation.apply(n)),n})}computeOutputShape(e){e=Xe(e);let t=this.dataFormat==="channelsFirst"?e[2]:e[1],n=this.dataFormat==="channelsFirst"?e[3]:e[2],o=this.dataFormat==="channelsFirst"?e[1]*this.depthMultiplier:e[3]*this.depthMultiplier,s=_n(t,this.kernelSize[0],this.padding,this.strides[0]),a=_n(n,this.kernelSize[1],this.padding,this.strides[1]);return this.dataFormat==="channelsFirst"?[e[0],o,s,a]:[e[0],s,a,o]}getConfig(){let e=super.getConfig();return e.depthMultiplier=this.depthMultiplier,e.depthwiseInitializer=It(this.depthwiseInitializer),e.depthwiseRegularizer=it(this.depthwiseRegularizer),e.depthwiseConstraint=Mt(this.depthwiseRegularizer),e}};$d.className="DepthwiseConv2D";Q.registerClass($d);function $0(r,e,t,n){if(Array.isArray(r)){if(e!=null||t!=null)throw new z("When inputs is an array, neither initialState or constants should be provided");n!=null&&(t=r.slice(r.length-n,r.length),r=r.slice(0,r.length-n)),r.length>1&&(e=r.slice(1,r.length)),r=r[0]}function o(s){return s==null||Array.isArray(s)?s:[s]}return e=o(e),t=o(t),{inputs:r,initialState:e,constants:t}}function R0(r,e,t,n=!1,o,s,a=!1,i=!1){return V(()=>{let l=e.shape.length;if(l<3)throw new z(`Input should be at least 3D, but is ${l}D.`);let u=[1,0].concat(Wr(2,l));if(e=Ue(e,u),s!=null)throw new Se("The rnn() functoin of the deeplearn.js backend does not support constants yet.");a&&console.warn("Backend rnn(): the unroll = true option is not applicable to the imperative deeplearn.js backend."),o!=null&&(o=o.asType("bool").asType("float32"),o.rank===l-1&&(o=cr(o,-1)),o=Ue(o,u)),n&&(e=Kt(e,0),o!=null&&(o=Kt(o,0)));let c=[],p,m=t,f=e.shape[0],d=fr(e),h;o!=null&&(h=fr(o));for(let x=0;x<f;++x){let b=d[x],w=V(()=>r(b,m));if(o==null)p=w[0],m=w[1];else{let _=V(()=>{let I=h[x],E=sr(I).sub(I),D=w[0].mul(I).add(m[0].mul(E)),$=m.map((R,M)=>w[1][M].mul(I).add(R.mul(E)));return{output:D,newStates:$}});p=_.output,m=_.newStates}i&&c.push(p)}let g;return i&&(g=Gt(c,1)),[p,g,m]})}var $n=class extends Le{constructor(e){super(e);let t;if(e.cell==null)throw new z("cell property is missing for the constructor of RNN.");if(Array.isArray(e.cell)?t=new om({cells:e.cell}):t=e.cell,t.stateSize==null)throw new z("The RNN cell should have an attribute `stateSize` (tuple of integers, one integer per RNN state).");this.cell=t,this.returnSequences=e.returnSequences==null?!1:e.returnSequences,this.returnState=e.returnState==null?!1:e.returnState,this.goBackwards=e.goBackwards==null?!1:e.goBackwards,this._stateful=e.stateful==null?!1:e.stateful,this.unroll=e.unroll==null?!1:e.unroll,this.supportsMasking=!0,this.inputSpec=[new Nt({ndim:3})],this.stateSpec=null,this.states_=null,this.numConstants=null,this.keptStates=[]}getStates(){if(this.states_==null){let e=Array.isArray(this.cell.stateSize)?this.cell.stateSize.length:1;return Wr(0,e).map(t=>null)}else return this.states_}setStates(e){this.states_=e}computeOutputShape(e){jx(e)&&(e=e[0]),e=e;let t=this.cell.stateSize;Array.isArray(t)||(t=[t]);let n=t[0],o;if(this.returnSequences?o=[e[0],e[1],n]:o=[e[0],n],this.returnState){let s=[];for(let a of t)s.push([e[0],a]);return[o].concat(s)}else return o}computeMask(e,t){return V(()=>{Array.isArray(t)&&(t=t[0]);let n=this.returnSequences?t:null;if(this.returnState){let o=this.states.map(s=>null);return[n].concat(o)}else return n})}get states(){if(this.states_==null){let e=Array.isArray(this.cell.stateSize)?this.cell.stateSize.length:1,t=[];for(let n=0;n<e;++n)t.push(null);return t}else return this.states_}set states(e){this.states_=e}build(e){let t=null;if(this.numConstants!=null)throw new Se("Constants support is not implemented in RNN yet.");jx(e)&&(e=e[0]),e=e;let n=this.stateful?e[0]:null,o=e.slice(2);this.inputSpec[0]=new Nt({shape:[n,null,...o]});let s=[e[0]].concat(e.slice(2));if(t!=null)throw new Se("Constants support is not implemented in RNN yet.");this.cell.build(s);let a;if(Array.isArray(this.cell.stateSize)?a=this.cell.stateSize:a=[this.cell.stateSize],this.stateSpec!=null){if(!y.arraysEqual(this.stateSpec.map(i=>i.shape[i.shape.length-1]),a))throw new z(`An initialState was passed that is not compatible with cell.stateSize. Received stateSpec=${this.stateSpec}; However cell.stateSize is ${this.cell.stateSize}`)}else this.stateSpec=a.map(i=>new Nt({shape:[null,i]}));this.stateful&&this.resetStates()}resetStates(e,t=!1){V(()=>{if(!this.stateful)throw new Dn("Cannot call resetStates() on an RNN Layer that is not stateful.");let n=this.inputSpec[0].shape[0];if(n==null)throw new z("If an RNN is stateful, it needs to know its batch size. Specify the batch size of your input tensors: \n- If using a Sequential model, specify the batch size by passing a `batchInputShape` option to your first layer.\n- If using the functional API, specify the batch size by passing a `batchShape` option to your Input layer.");if(this.states_==null)Array.isArray(this.cell.stateSize)?this.states_=this.cell.stateSize.map(o=>ht([n,o])):this.states_=[ht([n,this.cell.stateSize])];else if(e==null)Ee(this.states_),this.keptStates!=null&&(Ee(this.keptStates),this.keptStates=[]),Array.isArray(this.cell.stateSize)?this.states_=this.cell.stateSize.map(o=>ht([n,o])):this.states_[0]=ht([n,this.cell.stateSize]);else{if(Array.isArray(e)||(e=[e]),e.length!==this.states_.length)throw new z(`Layer ${this.name} expects ${this.states_.length} state(s), but it received ${e.length} state value(s). Input received: ${e}`);t===!0?this.keptStates.push(this.states_.slice()):Ee(this.states_);for(let o=0;o<this.states_.length;++o){let s=e[o],a=Array.isArray(this.cell.stateSize)?this.cell.stateSize[o]:this.cell.stateSize,i=[n,a];if(!y.arraysEqual(s.shape,i))throw new z(`State ${o} is incompatible with layer ${this.name}: expected shape=${i}, received shape=${s.shape}`);this.states_[o]=s}}this.states_=this.states_.map(o=>$t(o.clone()))})}apply(e,t){let n=t==null?null:t.initialState,o=t==null?null:t.constants;t==null&&(t={});let s=$0(e,n,o,this.numConstants);e=s.inputs,n=s.initialState,o=s.constants;let a=[],i=[];if(n!=null){t.initialState=n,a=a.concat(n),this.stateSpec=[];for(let u of n)this.stateSpec.push(new Nt({shape:u.shape}));i=i.concat(this.stateSpec)}if(o!=null&&(t.constants=o,a=a.concat(o),this.numConstants=o.length),a[0]instanceof sn){let u=[e].concat(a),c=this.inputSpec.concat(i),p=this.inputSpec;this.inputSpec=c;let m=super.apply(u,t);return this.inputSpec=p,m}else return super.apply(e,t)}call(e,t){return V(()=>{let n=t==null?null:t.mask,o=t==null?null:t.training,s=t==null?null:t.initialState;e=Fe(e),s==null&&(this.stateful?s=this.states_:s=this.getInitialState(e));let a=Array.isArray(this.cell.stateSize)?this.cell.stateSize.length:1;if(s.length!==a)throw new z(`RNN Layer has ${a} state(s) but was passed ${s.length} initial state(s).`);this.unroll&&console.warn("Ignoring unroll = true for RNN layer, due to imperative backend.");let i={training:o},u=R0((d,h)=>{let g=this.cell.call([d].concat(h),i);return[g[0],g.slice(1)]},e,s,this.goBackwards,n,null,this.unroll,this.returnSequences),c=u[0],p=u[1],m=u[2];this.stateful&&this.resetStates(m,o);let f=this.returnSequences?p:c;return this.returnState?[f].concat(m):f})}getInitialState(e){return V(()=>{let t=ht(e.shape);return t=fe(t,[1,2]),t=La(t),Array.isArray(this.cell.stateSize)?this.cell.stateSize.map(n=>n>1?zx(t,[1,n]):t):this.cell.stateSize>1?[zx(t,[1,this.cell.stateSize])]:[t]})}get trainableWeights(){return this.trainable?this.cell.trainableWeights:[]}get nonTrainableWeights(){return this.trainable?this.cell.nonTrainableWeights:this.cell.weights}setFastWeightInitDuringBuild(e){super.setFastWeightInitDuringBuild(e),this.cell!=null&&this.cell.setFastWeightInitDuringBuild(e)}getConfig(){let e=super.getConfig(),t={returnSequences:this.returnSequences,returnState:this.returnState,goBackwards:this.goBackwards,stateful:this.stateful,unroll:this.unroll};this.numConstants!=null&&(t.numConstants=this.numConstants);let n=this.cell.getConfig();return this.getClassName()===$n.className&&(t.cell={className:this.cell.getClassName(),config:n}),Object.assign({},n,e,t)}static fromConfig(e,t,n={}){let o=t.cell,s=an(o,n);return new e(Object.assign(t,{cell:s}))}};$n.className="RNN";Q.registerClass($n);var Ol=class extends Le{},rm=class extends Ol{constructor(e){super(e);this.DEFAULT_ACTIVATION="tanh",this.DEFAULT_KERNEL_INITIALIZER="glorotNormal",this.DEFAULT_RECURRENT_INITIALIZER="orthogonal",this.DEFAULT_BIAS_INITIALIZER="zeros",this.units=e.units,Ht(this.units,"units"),this.activation=$s(e.activation==null?this.DEFAULT_ACTIVATION:e.activation),this.useBias=e.useBias==null?!0:e.useBias,this.kernelInitializer=pt(e.kernelInitializer||this.DEFAULT_KERNEL_INITIALIZER),this.recurrentInitializer=pt(e.recurrentInitializer||this.DEFAULT_RECURRENT_INITIALIZER),this.biasInitializer=pt(e.biasInitializer||this.DEFAULT_BIAS_INITIALIZER),this.kernelRegularizer=wt(e.kernelRegularizer),this.recurrentRegularizer=wt(e.recurrentRegularizer),this.biasRegularizer=wt(e.biasRegularizer),this.kernelConstraint=Lt(e.kernelConstraint),this.recurrentConstraint=Lt(e.recurrentConstraint),this.biasConstraint=Lt(e.biasConstraint),this.dropout=qu([1,As([0,e.dropout==null?0:e.dropout])]),this.recurrentDropout=qu([1,As([0,e.recurrentDropout==null?0:e.recurrentDropout])]),this.stateSize=this.units,this.dropoutMask=null,this.recurrentDropoutMask=null}build(e){e=Xe(e),this.kernel=this.addWeight("kernel",[e[e.length-1],this.units],null,this.kernelInitializer,this.kernelRegularizer,!0,this.kernelConstraint),this.recurrentKernel=this.addWeight("recurrent_kernel",[this.units,this.units],null,this.recurrentInitializer,this.recurrentRegularizer,!0,this.recurrentConstraint),this.useBias?this.bias=this.addWeight("bias",[this.units],null,this.biasInitializer,this.biasRegularizer,!0,this.biasConstraint):this.bias=null,this.built=!0}call(e,t){return V(()=>{if(e=e,e.length!==2)throw new z(`SimpleRNNCell expects 2 input Tensors, got ${e.length}.`);let n=e[1];e=e[0];let o=t.training==null?!1:t.training;0<this.dropout&&this.dropout<1&&this.dropoutMask==null&&(this.dropoutMask=Va({ones:()=>sr(e),rate:this.dropout,training:o})),0<this.recurrentDropout&&this.recurrentDropout<1&&this.recurrentDropoutMask==null&&(this.recurrentDropoutMask=Va({ones:()=>sr(n),rate:this.recurrentDropout,training:o}));let s,a=this.dropoutMask,i=this.recurrentDropoutMask;a!=null?s=xo(O(e,a),this.kernel.read()):s=xo(e,this.kernel.read()),this.bias!=null&&(s=on(s,this.bias.read())),i!=null&&(n=O(n,i));let l=J(s,xo(n,this.recurrentKernel.read()));return this.activation!=null&&(l=this.activation.apply(l)),[l,l]})}getConfig(){let e=super.getConfig(),t={units:this.units,activation:Ds(this.activation),useBias:this.useBias,kernelInitializer:It(this.kernelInitializer),recurrentInitializer:It(this.recurrentInitializer),biasInitializer:It(this.biasInitializer),kernelRegularizer:it(this.kernelRegularizer),recurrentRegularizer:it(this.recurrentRegularizer),biasRegularizer:it(this.biasRegularizer),activityRegularizer:it(this.activityRegularizer),kernelConstraint:Mt(this.kernelConstraint),recurrentConstraint:Mt(this.recurrentConstraint),biasConstraint:Mt(this.biasConstraint),dropout:this.dropout,recurrentDropout:this.recurrentDropout};return Object.assign({},e,t)}};rm.className="SimpleRNNCell";Q.registerClass(rm);var Rd=class extends $n{constructor(e){e.cell=new rm(e),super(e)}call(e,t){return V(()=>{this.cell.dropoutMask!=null&&(Ee(this.cell.dropoutMask),this.cell.dropoutMask=null),this.cell.recurrentDropoutMask!=null&&(Ee(this.cell.recurrentDropoutMask),this.cell.recurrentDropoutMask=null);let n=t==null?null:t.mask,o=t==null?null:t.training,s=t==null?null:t.initialState;return super.call(e,{mask:n,training:o,initialState:s})})}static fromConfig(e,t){return new e(t)}};Rd.className="SimpleRNN";Q.registerClass(Rd);var nm=class extends Ol{constructor(e){super(e);if(this.DEFAULT_ACTIVATION="tanh",this.DEFAULT_RECURRENT_ACTIVATION="hardSigmoid",this.DEFAULT_KERNEL_INITIALIZER="glorotNormal",this.DEFAULT_RECURRENT_INITIALIZER="orthogonal",this.DEFAULT_BIAS_INITIALIZER="zeros",e.resetAfter)throw new z("GRUCell does not support reset_after parameter set to true.");this.units=e.units,Ht(this.units,"units"),this.activation=$s(e.activation===void 0?this.DEFAULT_ACTIVATION:e.activation),this.recurrentActivation=$s(e.recurrentActivation===void 0?this.DEFAULT_RECURRENT_ACTIVATION:e.recurrentActivation),this.useBias=e.useBias==null?!0:e.useBias,this.kernelInitializer=pt(e.kernelInitializer||this.DEFAULT_KERNEL_INITIALIZER),this.recurrentInitializer=pt(e.recurrentInitializer||this.DEFAULT_RECURRENT_INITIALIZER),this.biasInitializer=pt(e.biasInitializer||this.DEFAULT_BIAS_INITIALIZER),this.kernelRegularizer=wt(e.kernelRegularizer),this.recurrentRegularizer=wt(e.recurrentRegularizer),this.biasRegularizer=wt(e.biasRegularizer),this.kernelConstraint=Lt(e.kernelConstraint),this.recurrentConstraint=Lt(e.recurrentConstraint),this.biasConstraint=Lt(e.biasConstraint),this.dropout=qu([1,As([0,e.dropout==null?0:e.dropout])]),this.recurrentDropout=qu([1,As([0,e.recurrentDropout==null?0:e.recurrentDropout])]),this.implementation=e.implementation,this.stateSize=this.units,this.dropoutMask=null,this.recurrentDropoutMask=null}build(e){e=Xe(e);let t=e[e.length-1];this.kernel=this.addWeight("kernel",[t,this.units*3],null,this.kernelInitializer,this.kernelRegularizer,!0,this.kernelConstraint),this.recurrentKernel=this.addWeight("recurrent_kernel",[this.units,this.units*3],null,this.recurrentInitializer,this.recurrentRegularizer,!0,this.recurrentConstraint),this.useBias?this.bias=this.addWeight("bias",[this.units*3],null,this.biasInitializer,this.biasRegularizer,!0,this.biasConstraint):this.bias=null,this.built=!0}call(e,t){return V(()=>{if(e=e,e.length!==2)throw new z(`GRUCell expects 2 input Tensors (inputs, h, c), got ${e.length}.`);let n=t.training==null?!1:t.training,o=e[1];e=e[0],0<this.dropout&&this.dropout<1&&this.dropoutMask==null&&(this.dropoutMask=Va({ones:()=>sr(e),rate:this.dropout,training:n,count:3})),0<this.recurrentDropout&&this.recurrentDropout<1&&this.recurrentDropoutMask==null&&(this.recurrentDropoutMask=Va({ones:()=>sr(o),rate:this.recurrentDropout,training:n,count:3}));let s=this.dropoutMask,a=this.recurrentDropoutMask,i,l,u;0<this.dropout&&this.dropout<1&&(e=O(e,s[0]));let c=xo(e,this.kernel.read());this.useBias&&(c=on(c,this.bias.read())),0<this.recurrentDropout&&this.recurrentDropout<1&&(o=O(o,a[0]));let p=this.recurrentKernel.read(),[m,f]=tr(p,[2*this.units,this.units],p.rank-1),d=xo(o,m),[h,g,x]=tr(c,3,c.rank-1),[b,w]=tr(d,2,d.rank-1);i=this.recurrentActivation.apply(J(h,b)),l=this.recurrentActivation.apply(J(g,w));let _=xo(O(l,o),f);u=this.activation.apply(J(x,_));let I=J(O(i,o),O(J(1,qe(i)),u));return[I,I]})}getConfig(){let e=super.getConfig(),t={units:this.units,activation:Ds(this.activation),recurrentActivation:Ds(this.recurrentActivation),useBias:this.useBias,kernelInitializer:It(this.kernelInitializer),recurrentInitializer:It(this.recurrentInitializer),biasInitializer:It(this.biasInitializer),kernelRegularizer:it(this.kernelRegularizer),recurrentRegularizer:it(this.recurrentRegularizer),biasRegularizer:it(this.biasRegularizer),activityRegularizer:it(this.activityRegularizer),kernelConstraint:Mt(this.kernelConstraint),recurrentConstraint:Mt(this.recurrentConstraint),biasConstraint:Mt(this.biasConstraint),dropout:this.dropout,recurrentDropout:this.recurrentDropout,implementation:this.implementation,resetAfter:!1};return Object.assign({},e,t)}};nm.className="GRUCell";Q.registerClass(nm);var Fd=class extends $n{constructor(e){e.implementation===0&&console.warn("`implementation=0` has been deprecated, and now defaults to `implementation=1`. Please update your layer call."),e.cell=new nm(e),super(e)}call(e,t){return V(()=>{this.cell.dropoutMask!=null&&(Ee(this.cell.dropoutMask),this.cell.dropoutMask=null),this.cell.recurrentDropoutMask!=null&&(Ee(this.cell.recurrentDropoutMask),this.cell.recurrentDropoutMask=null);let n=t==null?null:t.mask,o=t==null?null:t.training,s=t==null?null:t.initialState;return super.call(e,{mask:n,training:o,initialState:s})})}static fromConfig(e,t){return t.implmentation===0&&(t.implementation=1),new e(t)}};Fd.className="GRU";Q.registerClass(Fd);var Pl=class extends Ol{constructor(e){super(e);this.DEFAULT_ACTIVATION="tanh",this.DEFAULT_RECURRENT_ACTIVATION="hardSigmoid",this.DEFAULT_KERNEL_INITIALIZER="glorotNormal",this.DEFAULT_RECURRENT_INITIALIZER="orthogonal",this.DEFAULT_BIAS_INITIALIZER="zeros",this.units=e.units,Ht(this.units,"units"),this.activation=$s(e.activation===void 0?this.DEFAULT_ACTIVATION:e.activation),this.recurrentActivation=$s(e.recurrentActivation===void 0?this.DEFAULT_RECURRENT_ACTIVATION:e.recurrentActivation),this.useBias=e.useBias==null?!0:e.useBias,this.kernelInitializer=pt(e.kernelInitializer||this.DEFAULT_KERNEL_INITIALIZER),this.recurrentInitializer=pt(e.recurrentInitializer||this.DEFAULT_RECURRENT_INITIALIZER),this.biasInitializer=pt(e.biasInitializer||this.DEFAULT_BIAS_INITIALIZER),this.unitForgetBias=e.unitForgetBias,this.kernelRegularizer=wt(e.kernelRegularizer),this.recurrentRegularizer=wt(e.recurrentRegularizer),this.biasRegularizer=wt(e.biasRegularizer),this.kernelConstraint=Lt(e.kernelConstraint),this.recurrentConstraint=Lt(e.recurrentConstraint),this.biasConstraint=Lt(e.biasConstraint),this.dropout=qu([1,As([0,e.dropout==null?0:e.dropout])]),this.recurrentDropout=qu([1,As([0,e.recurrentDropout==null?0:e.recurrentDropout])]),this.implementation=e.implementation,this.stateSize=[this.units,this.units],this.dropoutMask=null,this.recurrentDropoutMask=null}build(e){var t;e=Xe(e);let n=e[e.length-1];this.kernel=this.addWeight("kernel",[n,this.units*4],null,this.kernelInitializer,this.kernelRegularizer,!0,this.kernelConstraint),this.recurrentKernel=this.addWeight("recurrent_kernel",[this.units,this.units*4],null,this.recurrentInitializer,this.recurrentRegularizer,!0,this.recurrentConstraint);let o;if(this.useBias){if(this.unitForgetBias){let s=this.biasInitializer,a=this.units;o=new(t=class extends bn{apply(l,u){let c=s.apply([a]),p=new Xu().apply([a]),m=s.apply([a*2]);return QC(QC(c,p),m)}},t.className="CustomInit",t)}else o=this.biasInitializer;this.bias=this.addWeight("bias",[this.units*4],null,o,this.biasRegularizer,!0,this.biasConstraint)}else this.bias=null;this.built=!0}call(e,t){return V(()=>{let n=t.training==null?!1:t.training;if(e=e,e.length!==3)throw new z(`LSTMCell expects 3 input Tensors (inputs, h, c), got ${e.length}.`);let o=e[1],s=e[2];e=e[0],0<this.dropout&&this.dropout<1&&this.dropoutMask==null&&(this.dropoutMask=Va({ones:()=>sr(e),rate:this.dropout,training:n,count:4})),0<this.recurrentDropout&&this.recurrentDropout<1&&this.recurrentDropoutMask==null&&(this.recurrentDropoutMask=Va({ones:()=>sr(o),rate:this.recurrentDropout,training:n,count:4}));let a=this.dropoutMask,i=this.recurrentDropoutMask,l,u,c,p;0<this.dropout&&this.dropout<1&&(e=O(e,a[0]));let m=xo(e,this.kernel.read());0<this.recurrentDropout&&this.recurrentDropout<1&&(o=O(o,i[0])),m=J(m,xo(o,this.recurrentKernel.read())),this.useBias&&(m=on(m,this.bias.read()));let[f,d,h,g]=tr(m,4,m.rank-1);l=this.recurrentActivation.apply(f),u=this.recurrentActivation.apply(d),c=J(O(u,s),O(l,this.activation.apply(h))),p=this.recurrentActivation.apply(g);let x=O(p,this.activation.apply(c));return[x,x,c]})}getConfig(){let e=super.getConfig(),t={units:this.units,activation:Ds(this.activation),recurrentActivation:Ds(this.recurrentActivation),useBias:this.useBias,kernelInitializer:It(this.kernelInitializer),recurrentInitializer:It(this.recurrentInitializer),biasInitializer:It(this.biasInitializer),unitForgetBias:this.unitForgetBias,kernelRegularizer:it(this.kernelRegularizer),recurrentRegularizer:it(this.recurrentRegularizer),biasRegularizer:it(this.biasRegularizer),activityRegularizer:it(this.activityRegularizer),kernelConstraint:Mt(this.kernelConstraint),recurrentConstraint:Mt(this.recurrentConstraint),biasConstraint:Mt(this.biasConstraint),dropout:this.dropout,recurrentDropout:this.recurrentDropout,implementation:this.implementation};return Object.assign({},e,t)}};Pl.className="LSTMCell";Q.registerClass(Pl);var Od=class extends $n{constructor(e){e.implementation===0&&console.warn("`implementation=0` has been deprecated, and now defaults to `implementation=1`. Please update your layer call."),e.cell=new Pl(e),super(e)}call(e,t){return V(()=>{this.cell.dropoutMask!=null&&(Ee(this.cell.dropoutMask),this.cell.dropoutMask=null),this.cell.recurrentDropoutMask!=null&&(Ee(this.cell.recurrentDropoutMask),this.cell.recurrentDropoutMask=null);let n=t==null?null:t.mask,o=t==null?null:t.training,s=t==null?null:t.initialState;return super.call(e,{mask:n,training:o,initialState:s})})}static fromConfig(e,t){return t.implmentation===0&&(t.implementation=1),new e(t)}};Od.className="LSTM";Q.registerClass(Od);var om=class extends Ol{constructor(e){super(e);this.cells=e.cells}get stateSize(){let e=[];for(let t of this.cells.slice().reverse())Array.isArray(t.stateSize)?e.push(...t.stateSize):e.push(t.stateSize);return e}call(e,t){return V(()=>{e=e;let n=e.slice(1),o=[];for(let i of this.cells.slice().reverse())Array.isArray(i.stateSize)?o.push(n.splice(0,i.stateSize.length)):o.push(n.splice(0,1));o.reverse();let s=[],a;for(let i=0;i<this.cells.length;++i){let l=this.cells[i];n=o[i],i===0?a=[e[0]].concat(n):a=[a[0]].concat(n),a=l.call(a,t),s.push(a.slice(1))}n=[];for(let i of s.slice().reverse())n.push(...i);return[a[0]].concat(n)})}build(e){jx(e)&&(e=e[0]),e=e;let t;this.cells.forEach((n,o)=>{Ts(`RNNCell_${o}`,()=>{n.build(e),Array.isArray(n.stateSize)?t=n.stateSize[0]:t=n.stateSize,e=[e[0],t]})}),this.built=!0}getConfig(){let e=super.getConfig(),t=s=>({className:s.getClassName(),config:s.getConfig()}),o={cells:this.cells.map(t)};return Object.assign({},e,o)}static fromConfig(e,t,n={}){let o=[];for(let s of t.cells)o.push(an(s,n));return new e({cells:o})}get trainableWeights(){if(!this.trainable)return[];let e=[];for(let t of this.cells)e.push(...t.trainableWeights);return e}get nonTrainableWeights(){let e=[];for(let t of this.cells)e.push(...t.nonTrainableWeights);if(!this.trainable){let t=[];for(let n of this.cells)t.push(...n.trainableWeights);return t.concat(e)}return e}getWeights(){let e=[];for(let t of this.cells)e.push(...t.weights);return cd(e)}setWeights(e){let t=[];for(let n of this.cells){let o=n.weights.length,s=e.splice(o);for(let a=0;a<n.weights.length;++a)t.push([n.weights[a],s[a]])}Yp(t)}};om.className="StackedRNNCells";Q.registerClass(om);function Va(r){let{ones:e,rate:t,training:n=!1,count:o=1}=r,s=()=>Vx(e(),t),a=()=>Nl(s,e,n);return!o||o<=1?$t(a().clone()):Array(o).fill(void 0).map(a).map(l=>$t(l.clone()))}var Ite=function(r,e){var t={};for(var n in r)Object.prototype.hasOwnProperty.call(r,n)&&e.indexOf(n)<0&&(t[n]=r[n]);if(r!=null&&typeof Object.getOwnPropertySymbols=="function")for(var o=0,n=Object.getOwnPropertySymbols(r);o<n.length;o++)e.indexOf(n[o])<0&&Object.prototype.propertyIsEnumerable.call(r,n[o])&&(t[n[o]]=r[n[o]]);return t};var F0=class extends $n{constructor(e){if(e.unroll)throw new Se("Unrolling is not possible with convolutional RNNs.");if(Array.isArray(e.cell))throw new Se("It is not possible at the moment to stack convolutional cells.");super(e);this.inputSpec=[new Nt({ndim:5})]}call(e,t){return V(()=>{if(this.cell.dropoutMask!=null&&(Ee(this.cell.dropoutMask),this.cell.dropoutMask=null),this.cell.recurrentDropoutMask!=null&&(Ee(this.cell.recurrentDropoutMask),this.cell.recurrentDropoutMask=null),t&&t.constants)throw new z("ConvRNN2D cell does not support constants");let n=t==null?null:t.mask,o=t==null?null:t.training,s=t==null?null:t.initialState;return super.call(e,{mask:n,training:o,initialState:s})})}computeOutputShape(e){let t=this.computeSingleOutputShape(e);return this.returnSequences||(t=[t[0],...t.slice(2)]),this.returnState&&(t=[t,...Array(2).fill([e[0],...t.slice(-3)])]),t}getInitialState(e){return V(()=>{let{stateSize:t}=this.cell,n=e.shape,o=this.computeSingleOutputShape(n),s=[o[0],...o.slice(2)],a=ht(s);return Array.isArray(t)?Array(t.length).fill(a):[a]})}resetStates(e,t=!1){V(()=>{if(!this.stateful)throw new Dn("Cannot call resetStates() on an RNN Layer that is not stateful.");let n=this.inputSpec[0].shape,o=this.computeSingleOutputShape(n),s=[o[0],...o.slice(2)];if(n[0]==null)throw new z("If an RNN is stateful, it needs to know its batch size. Specify the batch size of your input tensors: \n- If using a Sequential model, specify the batch size by passing a `batchInputShape` option to your first layer.\n- If using the functional API, specify the batch size by passing a `batchShape` option to your Input layer.");if(this.getStates()==null)Array.isArray(this.cell.stateSize)?this.states_=this.cell.stateSize.map(()=>ht(s)):this.states_=[ht(s)];else if(e==null)Ee(this.states_),this.keptStates!=null&&(Ee(this.keptStates),this.keptStates=[]),Array.isArray(this.cell.stateSize)?this.states_=this.cell.stateSize.map(()=>ht(s)):this.states_[0]=ht(s);else{if(Array.isArray(e)||(e=[e]),e.length!==this.states_.length)throw new z(`Layer ${this.name} expects ${this.states_.length} state(s), but it received ${e.length} state value(s). Input received: ${e}`);t?this.keptStates.push(this.states_.slice()):Ee(this.states_);for(let i=0;i<this.states_.length;++i){let l=e[i],u=s;if(!y.arraysEqual(l.shape,u))throw new z(`State ${i} is incompatible with layer ${this.name}: expected shape=${u}, received shape=${l.shape}`);this.states_[i]=l}}this.states_=this.states_.map(i=>$t(i.clone()))})}computeSingleOutputShape(e){let{dataFormat:t,filters:n,kernelSize:o,padding:s,strides:a,dilationRate:i}=this.cell,l=t==="channelsFirst",u=e[l?3:2],c=e[l?4:3],p=_n(u,o[0],s,a[0],i[0]),m=_n(c,o[1],s,a[1],i[1]);return[...e.slice(0,2),...l?[n,p,m]:[p,m,n]]}};F0.className="ConvRNN2D";var sm=class extends Pl{constructor(e){let{filters:t,kernelSize:n,strides:o,padding:s,dataFormat:a,dilationRate:i}=e;super(Object.assign({},e,{units:t}));this.filters=t,Ht(this.filters,"filters"),this.kernelSize=$l(n,2,"kernelSize"),this.kernelSize.forEach(l=>Ht(l,"kernelSize")),this.strides=$l(o||1,2,"strides"),this.strides.forEach(l=>Ht(l,"strides")),this.padding=s||"valid",nn(this.padding),this.dataFormat=a||"channelsLast",Ot(this.dataFormat),this.dilationRate=$l(i||1,2,"dilationRate"),this.dilationRate.forEach(l=>Ht(l,"dilationRate"))}build(e){var t;e=Xe(e);let n=this.dataFormat==="channelsFirst"?1:e.length-1;if(e[n]==null)throw new z(`The channel dimension of the input should be defined. Found ${e[n]}`);let o=e[n],s=4,a=this.kernelSize.concat([o,this.filters*s]);this.kernel=this.addWeight("kernel",a,null,this.kernelInitializer,this.kernelRegularizer,!0,this.kernelConstraint);let i=this.kernelSize.concat([this.filters,this.filters*s]);if(this.recurrentKernel=this.addWeight("recurrent_kernel",i,null,this.recurrentInitializer,this.recurrentRegularizer,!0,this.recurrentConstraint),this.useBias){let l;if(this.unitForgetBias){let u=this.biasInitializer,c=this.filters;l=new(t=class extends bn{apply(m,f){let d=u.apply([c]),h=er([c]),g=u.apply([c*2]);return Bp([d,h,g])}},t.className="CustomInit",t)}else l=this.biasInitializer;this.bias=this.addWeight("bias",[this.filters*s],null,l,this.biasRegularizer,!0,this.biasConstraint)}this.built=!0}call(e,t){return V(()=>{if(e.length!==3)throw new z(`ConvLSTM2DCell expects 3 input Tensors (inputs, h, c), got ${e.length}.`);let n=t.training||!1,o=e[0],s=e[1],a=e[2],i=4;0<this.dropout&&this.dropout<1&&this.dropoutMask==null&&(this.dropoutMask=Va({ones:()=>sr(o),rate:this.dropout,training:n,count:i}));let l=this.dropoutMask,u=(ee,ie,me)=>!ie||!ie[me]?ee:O(ie[me],ee),c=u(o,l,0),p=u(o,l,1),m=u(o,l,2),f=u(o,l,3);0<this.recurrentDropout&&this.recurrentDropout<1&&this.recurrentDropoutMask==null&&(this.recurrentDropoutMask=Va({ones:()=>sr(s),rate:this.recurrentDropout,training:n,count:i}));let d=this.recurrentDropoutMask,h=u(s,d,0),g=u(s,d,1),x=u(s,d,2),b=u(s,d,3),w=3,[_,I,E,D]=tr(this.kernel.read(),i,w),[$,R,M,G]=this.useBias?tr(this.bias.read(),i):[null,null,null,null];c=this.inputConv(c,_,$,this.padding),p=this.inputConv(p,I,R,this.padding),m=this.inputConv(m,E,M,this.padding),f=this.inputConv(f,D,G,this.padding);let[j,U,H,q]=tr(this.recurrentKernel.read(),i,w);h=this.recurrentConv(h,j),g=this.recurrentConv(g,U),x=this.recurrentConv(x,H),b=this.recurrentConv(b,q);let X=this.recurrentActivation.apply(J(c,h)),ne=this.recurrentActivation.apply(J(p,g)),Y=J(O(ne,a),O(X,this.activation.apply(J(m,x)))),re=O(this.recurrentActivation.apply(J(f,b)),this.activation.apply(Y));return[re,re,Y]})}getConfig(){let e=super.getConfig(),{units:t}=e,n=Ite(e,["units"]),o={filters:this.filters,kernelSize:this.kernelSize,padding:this.padding,dataFormat:this.dataFormat,dilationRate:this.dilationRate,strides:this.strides};return Object.assign({},n,o)}inputConv(e,t,n,o){let s=Jr(e,t,this.strides,o||"valid",this.dataFormat==="channelsFirst"?"NCHW":"NHWC",this.dilationRate);return n?on(s,n,this.dataFormat):s}recurrentConv(e,t){return Jr(e,t,1,"same",this.dataFormat==="channelsFirst"?"NCHW":"NHWC")}};sm.className="ConvLSTM2DCell";Q.registerClass(sm);var Pd=class extends F0{constructor(e){let t=new sm(e);super(Object.assign({},e,{cell:t}))}static fromConfig(e,t){return new e(t)}};Pd.className="ConvLSTM2D";Q.registerClass(Pd);var im=class extends Le{constructor(e){super(e);this.rate=Math.max(Math.min(e.rate,1),0),this.noiseShape=e.noiseShape,this.seed=e.seed,this.supportsMasking=!0}getNoiseShape(e){if(this.noiseShape==null)return this.noiseShape;let t=e.shape,n=[];for(let o=0;o<this.noiseShape.length;++o)n.push(this.noiseShape[o]==null?t[o]:this.noiseShape[o]);return n}call(e,t){return V(()=>{this.invokeCallHook(e,t);let n=Fe(e);if(0<this.rate&&this.rate<1){let o=t.training==null?!1:t.training,s=this.getNoiseShape(n);return Nl(()=>Vx(n,this.rate,s,this.seed),()=>n,o)}return e})}getConfig(){let e={rate:this.rate,noiseShape:this.noiseShape,seed:this.seed},t=super.getConfig();return Object.assign(e,t),e}dispose(){return super.dispose()}};im.className="Dropout";Q.registerClass(im);var Md=class extends im{constructor(e){super(e);this.inputSpec=[{ndim:3}]}getNoiseShape(e){let t=e.shape;return[t[0],1,t[2]]}};Md.className="SpatialDropout1D";Q.registerClass(Md);var Ld=class extends Le{constructor(e){super(e);if(this.activation=null,this.useBias=!0,this.kernel=null,this.bias=null,this.DEFAULT_KERNEL_INITIALIZER="glorotNormal",this.DEFAULT_BIAS_INITIALIZER="zeros",e.batchInputShape==null&&e.inputShape==null&&e.inputDim!=null){let t=null;e.batchSize!=null&&(t=e.batchSize),this.batchInputShape=[t,e.inputDim]}this.units=e.units,Ht(this.units,"units"),this.activation=$s(e.activation),e.useBias!=null&&(this.useBias=e.useBias),this.kernelInitializer=pt(e.kernelInitializer||this.DEFAULT_KERNEL_INITIALIZER),this.biasInitializer=pt(e.biasInitializer||this.DEFAULT_BIAS_INITIALIZER),this.kernelConstraint=Lt(e.kernelConstraint),this.biasConstraint=Lt(e.biasConstraint),this.kernelRegularizer=wt(e.kernelRegularizer),this.biasRegularizer=wt(e.biasRegularizer),this.activityRegularizer=wt(e.activityRegularizer),this.supportsMasking=!0,this.inputSpec=[{minNDim:2}]}build(e){e=Xe(e);let t=e[e.length-1];this.kernel==null&&(this.kernel=this.addWeight("kernel",[t,this.units],null,this.kernelInitializer,this.kernelRegularizer,!0,this.kernelConstraint),this.useBias&&(this.bias=this.addWeight("bias",[this.units],null,this.biasInitializer,this.biasRegularizer,!0,this.biasConstraint))),this.inputSpec=[{minNDim:2,axes:{[-1]:t}}],this.built=!0}computeOutputShape(e){e=Xe(e);let t=e.slice();return t[t.length-1]=this.units,t}call(e,t){return V(()=>{this.invokeCallHook(e,t);let n=Fe(e),o=Px(this.activation.getClassName()),s;return o!=null?s=xo(n,this.kernel.read(),o,this.bias?this.bias.read():null):(s=xo(n,this.kernel.read()),this.bias!=null&&(s=on(s,this.bias.read())),this.activation!=null&&(s=this.activation.apply(s))),s})}getConfig(){let e={units:this.units,activation:Ds(this.activation),useBias:this.useBias,kernelInitializer:It(this.kernelInitializer),biasInitializer:It(this.biasInitializer),kernelRegularizer:it(this.kernelRegularizer),biasRegularizer:it(this.biasRegularizer),activityRegularizer:it(this.activityRegularizer),kernelConstraint:Mt(this.kernelConstraint),biasConstraint:Mt(this.biasConstraint)},t=super.getConfig();return Object.assign(e,t),e}};Ld.className="Dense";Q.registerClass(Ld);var zd=class extends Le{constructor(e){e=e||{},super(e),this.inputSpec=[{minNDim:3}],this.dataFormat=e.dataFormat}computeOutputShape(e){e=Xe(e);for(let t of e.slice(1))if(t==null)throw new z(`The shape of the input to "Flatten" is not fully defined (got ${e.slice(1)}). Make sure to pass a complete "input_shape" or "batch_input_shape" argument to the first layer in your model.`);return[e[0],go(e,1)]}call(e,t){return V(()=>{this.invokeCallHook(e,t);let n=Fe(e);if(this.dataFormat==="channelsFirst"&&n.rank>1){let o=[0];for(let s=2;s<n.rank;++s)o.push(s);o.push(1),n=n.transpose(o)}return m3(n)})}getConfig(){let e={};this.dataFormat!=null&&(e.dataFormat=this.dataFormat);let t=super.getConfig();return Object.assign(e,t),e}};zd.className="Flatten";Q.registerClass(zd);var Bd=class extends Le{constructor(e){super(e);this.supportsMasking=!0,this.activation=$s(e.activation)}call(e,t){return V(()=>{this.invokeCallHook(e,t);let n=Fe(e);return this.activation.apply(n)})}getConfig(){let e={activation:Ds(this.activation)},t=super.getConfig();return Object.assign(e,t),e}};Bd.className="Activation";Q.registerClass(Bd);var Vd=class extends Le{constructor(e){super(e);this.n=e.n,this.inputSpec=[{ndim:2}]}computeOutputShape(e){return[e[0],this.n,e[1]]}call(e,t){return V(()=>(e=Fe(e),c3(e,this.n)))}getConfig(){let e={n:this.n},t=super.getConfig();return Object.assign(e,t),e}};Vd.className="RepeatVector";Q.registerClass(Vd);var Gd=class extends Le{constructor(e){super(e);this.targetShape=e.targetShape;for(let t=0;t<this.targetShape.length;++t)this.isUnknown(this.targetShape[t])&&(this.targetShape[t]=null)}isUnknown(e){return e<0||e==null}fixUnknownDimension(e,t){let n="Total size of new array must be unchanged.",o=t.slice(),s=1,a=null;for(let l=0;l<o.length;++l){let u=o[l];if(this.isUnknown(u))if(a===null)a=l;else throw new z("Can only specifiy one unknown dimension.");else s*=u}let i=go(e);if(a!==null){if(s===0||i%s!=0)throw new z(n);o[a]=i/s}else if(i!==s)throw new z(n);return o}computeOutputShape(e){let t=!1;for(let n=0;n<e.length;++n)if(this.isUnknown(e[n])){t=!0;break}return t?e.slice(0,1).concat(this.targetShape):e.slice(0,1).concat(this.fixUnknownDimension(e.slice(1),this.targetShape))}call(e,t){return V(()=>{this.invokeCallHook(e,t);let n=Fe(e),o=n.shape,s=o.slice(0,1).concat(this.fixUnknownDimension(o.slice(1),this.targetShape));return n.reshape(s)})}getConfig(){let e={targetShape:this.targetShape},t=super.getConfig();return Object.assign(e,t),e}};Gd.className="Reshape";Q.registerClass(Gd);var Wd=class extends Le{constructor(e){super(e);if(e.dims==null)throw new Error("Required configuration field `dims` is missing during Permute constructor call.");if(!Array.isArray(e.dims))throw new Error(`Permute constructor requires \`dims\` to be an Array, but received ${e.dims} instead.`);let t=Wr(1,e.dims.length+1);if(!y.arraysEqual(e.dims.slice().sort(),t))throw new Error("Invalid permutation `dims`: "+JSON.stringify(e.dims)+" `dims` must contain consecutive integers starting from 1.");this.dims=e.dims,this.dimsIncludingBatch=[0].concat(this.dims),this.inputSpec=[new Nt({ndim:this.dims.length+1})]}computeOutputShape(e){e=Xe(e);let t=e.slice();return this.dims.forEach((n,o)=>{t[o+1]=e[n]}),t}call(e,t){return Ue(Fe(e),this.dimsIncludingBatch)}getConfig(){let e={dims:this.dims},t=super.getConfig();return Object.assign(e,t),e}};Wd.className="Permute";Q.registerClass(Wd);var jd=class extends Le{constructor(e){super(e==null?{}:e);this.supportsMasking=!0,e!=null?this.maskValue=e.maskValue==null?0:e.maskValue:this.maskValue=0}computeOutputShape(e){return e}getConfig(){let e=super.getConfig(),t={maskValue:this.maskValue};return Object.assign(t,e),t}computeMask(e,t){let n=Fe(e),o=-1;return dl(ro(n,this.maskValue),o)}call(e,t){return V(()=>{this.invokeCallHook(e,t);let n=Fe(e),o=-1,s=!0,a=dl(ro(n,this.maskValue),o,s);return n.mul(a.asType(n.dtype))})}};jd.className="Masking";Q.registerClass(jd);var Ud=class extends Le{constructor(e){super(e);if(this.embeddings=null,this.DEFAULT_EMBEDDINGS_INITIALIZER="randomUniform",e.batchInputShape==null&&e.inputShape==null){let t=null;e.batchSize!=null&&(t=e.batchSize),e.inputLength==null?this.batchInputShape=[t,null]:this.batchInputShape=[t].concat(bt(e.inputLength))}this.inputDim=e.inputDim,Ht(this.inputDim,"inputDim"),this.outputDim=e.outputDim,Ht(this.outputDim,"outputDim"),this.embeddingsInitializer=pt(e.embeddingsInitializer||this.DEFAULT_EMBEDDINGS_INITIALIZER),this.embeddingsRegularizer=wt(e.embeddingsRegularizer),this.activityRegularizer=wt(e.activityRegularizer),this.embeddingsConstraint=Lt(e.embeddingsConstraint),this.maskZero=e.maskZero,this.supportsMasking=e.maskZero,this.inputLength=e.inputLength}build(e){this.embeddings=this.addWeight("embeddings",[this.inputDim,this.outputDim],this.dtype,this.embeddingsInitializer,this.embeddingsRegularizer,!0,this.embeddingsConstraint),this.built=!0}warnOnIncompatibleInputShape(e){}computeMask(e,t){return V(()=>this.maskZero?(e=Fe(e),ro(e,Ie(e))):null)}computeOutputShape(e){if(e=Xe(e),this.inputLength==null)return[...e,this.outputDim];let t=bt(this.inputLength);if(t.length!==e.length-1)throw new z(`"inputLength" is ${this.inputLength}, but received input shape has shape ${e}`);{let n=0;for(let o=0;o<t.length;++o){let s=t[o],a=e[o+1];if(s!=null&&a!=null&&s!==a)throw new z(`"inputLength" is ${this.inputLength}, but received input shape has shape ${e}`);s==null&&(t[n]=a),n++}}return[e[0],...t,this.outputDim]}call(e,t){return V(()=>{this.invokeCallHook(e,t);let n=Fe(e);return n.dtype!=="int32"&&(n=Ma(n,"int32")),Bx(this.embeddings.read(),n.as1D()).reshape(Xe(this.computeOutputShape(n.shape)))})}getConfig(){let e={inputDim:this.inputDim,outputDim:this.outputDim,embeddingsInitializer:It(this.embeddingsInitializer),embeddingsRegularizer:it(this.embeddingsRegularizer),activityRegularizer:it(this.activityRegularizer),embeddingsConstraint:Mt(this.embeddingsConstraint),maskZero:this.maskZero,inputLength:this.inputLength},t=super.getConfig();return Object.assign(e,t),e}};Ud.className="Embedding";Q.registerClass(Ud);var Ml=class extends Le{constructor(e){super(e||{});this.supportsMasking=!0}mergeFunction(e){throw new Se}computeElementwiseOpOutputShape(e,t){if(e==null||t==null)return null;if(e.length<t.length)return this.computeElementwiseOpOutputShape(t,e);if(t.length===0)return e;let n=e.slice(0,e.length-t.length);for(let o=0;o<t.length;++o){let s=e[e.length-t.length+o],a=t[o];if(s==null||a==null||s<0||a<0)n.push(null);else if(s===1)n.push(a);else if(a===1)n.push(s);else{if(s!==a)throw new z("Operands could not be broadcast together with shapes "+JSON.stringify(e)+" "+JSON.stringify(t));n.push(s)}}return n}build(e){if(Array.isArray(e)&&!Array.isArray(e[0])&&(e=[Xe(e)]),e=e,e.length<2)throw new z(`A merge layer should be called on an Array of at least 2 inputs. Got ${e.length} input(s).`);let t=[];for(let s of e)s!=null&&s[0]!==null&&t.push(s[0]);if(t=ho(t),t.length>1)throw new z(`Can not merge tensors with different batch sizes. Got tensors with shapes: ${JSON.stringify(e)}.`);let n=e[0]==null?null:e[0].slice(1);for(let s=1;s<e.length;++s){let a=e[s]==null?null:e[s].slice(1);n=this.computeElementwiseOpOutputShape(n,a)}let o=e.map(s=>s.length);e.indexOf(null)===-1&&ho(o).length===1?this.reshapeRequired=!1:this.reshapeRequired=!0}call(e,t){return V(()=>{if(e=e,this.reshapeRequired){let n=[],o=e.map(s=>s.rank);if(o.indexOf(null)===-1){let s=As(o);for(let a of e){let i=a.rank;for(let l=0;l<s-i;++l)a=La(a,1);n.push(a)}return this.mergeFunction(n)}else{let s=!1;for(let l of e){let u=l.rank;if(u==null){let c=l.shape,p=c[0],m=c.slice(1).concat([p]),f=l.reshape([p].concat(go(c.slice(1))));f=Ue(f,[1,0]),f=f.reshape(m),n.push(f),s=!0}else if(u>1){let c=Wr(1,u).concat([0]);n.push(Ue(l,c)),s=!0}else n.push(l)}let a=this.mergeFunction(n),i=a.rank;if(s){if(i==null){let l=a.shape,u=l.length,c=l[u-1],p=[c].concat(l.slice(0,l.length-1));a=Ue(a.reshape([-1,c]),[1,0]).reshape(p)}else if(i>1){let l=[i-1].concat(Wr(0,i-1));a=Ue(a,l)}}return a}}else return this.mergeFunction(e)})}computeOutputShape(e){e=e;let t;e[0]==null?t=null:t=e[0].slice(1);for(let o=1;o<e.length;++o){let s=e[o]==null?null:e[o].slice(1);t=this.computeElementwiseOpOutputShape(t,s)}let n=[];for(let o of e)o!=null&&o[0]!==null&&n.push(o[0]);return n=ho(n),n.length===1?t=n.concat(t):t=[null].concat(t),t}computeMask(e,t){return V(()=>{if(t==null)return null;if(!Array.isArray(t))throw new z("`mask` should be an Array");if(!Array.isArray(e))throw new z("`inputs` should be an Array");if(t.length!==e.length)throw new z(`The Array 'inputs' and 'mask' are expected to have the same length, but have different lengths (${e.length} vs ${t.length})`);if(t.every(o=>o==null))return null;t=t.map(o=>o==null?o:cr(o,0));let n=t[0];for(let o=1;o<t.length-1;++o)n=yr(n,t[o]);return n})}},Hd=class extends Ml{constructor(e){super(e)}mergeFunction(e){return V(()=>{let t=e[0].clone();for(let n=1;n<e.length;++n)t=J(t,e[n]);return t})}};Hd.className="Add";Q.registerClass(Hd);var qd=class extends Ml{constructor(e){super(e)}mergeFunction(e){return V(()=>{let t=e[0].clone();for(let n=1;n<e.length;++n)t=O(t,e[n]);return t})}};qd.className="Multiply";Q.registerClass(qd);var Kd=class extends Ml{constructor(e){super(e)}mergeFunction(e){return V(()=>{let t=e[0].clone();for(let n=1;n<e.length;++n)t=J(t,e[n]);return O(1/e.length,t)})}};Kd.className="Average";Q.registerClass(Kd);var Xd=class extends Ml{constructor(e){super(e)}mergeFunction(e){return V(()=>{let t=e[0];for(let n=1;n<e.length;++n)t=en(t,e[n]);return t})}};Xd.className="Maximum";Q.registerClass(Xd);var Yd=class extends Ml{constructor(e){super(e)}mergeFunction(e){return V(()=>{let t=e[0];for(let n=1;n<e.length;++n)t=_s(t,e[n]);return t})}};Yd.className="Minimum";Q.registerClass(Yd);var Zd=class extends Ml{constructor(e){super(e);this.DEFAULT_AXIS=-1,e==null&&(e={}),this.axis=e.axis==null?this.DEFAULT_AXIS:e.axis,this.supportsMasking=!0,this.reshapeRequired=!1}build(e){if(!(Array.isArray(e)&&Array.isArray(e[0]))||e.length===1)throw new z("A `Concatenate` layer should be called on a list of at least 2 inputs");e=e;let t=!0;for(let o of e)if(o!=null){t=!1;break}if(t)return;let n=[];for(let o=0;o<e.length;++o){let s=e[o].slice();s.splice(this.axis,1);let a=!1;for(let i of n)if(y.arraysEqual(i,s)){a=!0;break}a||n.push(s)}if(n.length>1)throw new z("A `Concatenate` layer requires inputs with matching shapes except for the concat axis. Got input shapes: "+JSON.stringify(e))}mergeFunction(e){return V(()=>Bp(e,this.axis))}computeOutputShape(e){if(!(Array.isArray(e)&&Array.isArray(e[0])))throw new z("A `Concatenate` layer should be called on a list of inputs.");let t=e,n=t[0].slice(),o=this.axis<0?n.length+this.axis:this.axis;for(let s of t.slice(1)){if(n[o]==null||s[o]==null){n[o]=null;break}n[o]+=s[o]}return n}computeMask(e,t){if(t==null)return null;if(!Array.isArray(t))throw new z("`mask` should be an array for Concatenate");if(!Array.isArray(e))throw new z("`inputs` should be an array for Concatenate");if(t.length!==e.length)throw new z(`Mismatch in the length of mask (${t.length}) and the legnth of inputs (${e.length})`);return V(()=>{let n=!0;if(t.forEach(a=>{if(a!=null){n=!1;return}}),n)return null;let o=[];for(let a=0;a<e.length;++a)t[a]==null?o.push(sr(e[a]).asType("bool")):t[a].rank<e[a].rank?o.push(cr(t[a],-1)):o.push(t[a]);let s=Qe(o,this.axis);return tu(s,-1,!1)})}getConfig(){let e={axis:this.axis},t=super.getConfig();return Object.assign(e,t),e}};Zd.className="Concatenate";Q.registerClass(Zd);function Jd(r,e){for(;r<0;)r+=e;return r}function Nte(r,e,t){if(r.shape.length>3||e.shape.length>3)throw new Se("batchDot is not implemented for tensors of 4D or higher rank yet");if(y.assert(r.shape.length>=2,()=>`batchDot requires the rank of x to be >= 2, but got ${r.shape.length}`),y.assert(r.shape.length>=2,()=>`batchDot requires the rank of y to be >= 2, but got ${e.shape.length}`),typeof t=="number"&&(t=[t,t]),r.dtype==="complex64"||e.dtype==="complex64")throw new Se("batchDot is not implemented for complex64-type Tensors yet.");let n=r.shape.length,o=e.shape.length;t==null&&(t=[n-1,o-2]);let s=t;return V(()=>{let a;if(n>o){a=n-o;let l=[];for(let u=0;u<a;++u)l.push(1);e=e.reshape(e.shape.concat(l))}else if(o>n){a=o-n;let l=[];for(let u=0;u<a;++u)l.push(1);r=r.reshape(r.shape.concat(l))}else a=0;let i;if(r.shape.length===2&&e.shape.length===2)s[0]===s[1]?i=r.mul(e).sum(s[0]):i=r.transpose([1,0]).mul(e).sum(s[1]);else{let l=s[0]!==r.shape.length-1,u=s[1]===e.shape.length-1;i=r.matMul(e,l,u)}if(a>0){let l;n>o?l=n+o-3:l=n-1;let u=[];for(let c=l;c<l+a;++c)u.push(c);i=i.squeeze(u)}return i.shape.length===1&&(i=i.expandDims(1)),i})}var Qd=class extends Ml{constructor(e){super(e);this.axes=e.axes,this.normalize=e.normalize==null?!1:e.normalize,this.supportsMasking=!0,this.reshapeRequired=!1}build(e){y.assert(Array.isArray(e)&&e.length===2&&Array.isArray(e[0])&&Array.isArray(e[1]),()=>"A `Dot` layer should be called on a list of exactly 2 inputs.");let t=e[0],n=e[1];if(t.length>3||n.length>3)throw new Se("Dot layer does not support tensors of 4D or higher rank yet.");let o=this.interpretAxes(t,n);if(t[o[0]]!==n[o[1]])throw new z(`Dimension incompatibility: ${t[o[0]]} !== ${n[o[1]]}`)}mergeFunction(e){if(e.length!==2)throw new z(`A \`Dot\` layer must be called on exactly 2 inputs, but received ${e.length} input(s).`);let t=e[0],n=e[1],o;return Array.isArray(this.axes)?o=this.axes.map((s,a)=>Jd(s,e[a].shape.length)):o=[Jd(this.axes,t.shape.length),Jd(this.axes,n.shape.length)],this.normalize&&(t=pd(t,o[0]),n=pd(n,o[1])),Nte(t,n,o)}interpretAxes(e,t){let n;return Array.isArray(this.axes)?n=this.axes:n=[Jd(this.axes,e.length),Jd(this.axes,t.length)],n}computeOutputShape(e){y.assert(Array.isArray(e)&&e.length===2&&Array.isArray(e[0])&&Array.isArray(e[1]),()=>"A `Dot` layer should be called on a list of exactly 2 inputs.");let t=e[0].slice(),n=e[1].slice();if(t.length>3||n.length>3)throw new Se("Dot layer does not support tensors of 4D or higher rank yet.");let o=this.interpretAxes(t,n);t.splice(o[0],1),n.splice(o[1],1),n.splice(0,1);let s=t.concat(n);return s.length===1&&s.push(1),s}computeMask(e,t){return null}getConfig(){let e={axes:this.axes,normalize:this.normalize},t=super.getConfig();return Object.assign(e,t),e}};Qd.className="Dot";Q.registerClass(Qd);var eh=class extends Le{constructor(e){super(e);this.supportsMasking=!0,this.stddev=e.stddev}computeOutputShape(e){return e}getConfig(){let e=super.getConfig(),t={stddev:this.stddev};return Object.assign(t,e),t}call(e,t){return V(()=>{this.invokeCallHook(e,t);let n=Fe(e);return Nl(()=>Vp(n.shape,0,this.stddev).add(n),()=>n,t.training||!1)})}};eh.className="GaussianNoise";Q.registerClass(eh);var th=class extends Le{constructor(e){super(e);this.supportsMasking=!0,this.rate=e.rate}computeOutputShape(e){return e}getConfig(){let e=super.getConfig(),t={rate:this.rate};return Object.assign(t,e),t}call(e,t){return V(()=>{this.invokeCallHook(e,t);let n=Fe(e);return this.rate>0&&this.rate<1?Nl(()=>{let s=Math.sqrt(this.rate/(1-this.rate));return n.mul(Vp(n.shape,1,s))},()=>n,t.training||!1):n})}};th.className="GaussianDropout";Q.registerClass(th);var rh=class extends Le{constructor(e){super(e);this.supportsMasking=!0,this.rate=e.rate,this.noiseShape=e.noiseShape}_getNoiseShape(e){return this.noiseShape||Fe(e).shape}computeOutputShape(e){return e}getConfig(){let e=super.getConfig(),t={rate:this.rate};return Object.assign(t,e),t}call(e,t){return V(()=>{if(this.rate<1&&this.rate>0){let n=this._getNoiseShape(e);return Nl(()=>{let s=Fe(e),a=1.6732632423543772,i=1.0507009873554805,l=-a*i,u=hn(ks(n),this.rate);u=Ma(u,"float32");let c=((1-this.rate)*(1+this.rate*l**2))**-.5,p=-c*l*this.rate;return s.mul(u).add(u.add(-1).mul(l)).mul(c).add(p)},()=>Fe(e),t.training||!1)}return e})}};rh.className="AlphaDropout";Q.registerClass(rh);function nh(r,e,t,n,o,s=.001){let a;if(r.rank===2)a=Hw(r,e,t,n,o,s);else if(r.rank===3)a=qw(r,e,t,n,o,s);else if(r.rank===4)a=Kw(r,e,t,n,o,s);else throw new Se(`batchNormalization is not implemented for array of rank ${r.rank} yet`);return a}function Ste(r,e,t,n,o=.001){return V(()=>{let s=ap(r,n),a=s.mean,i=s.variance;return[nh(r,a,i,t,e,o),a,i]})}function Tte(r,e,t,n,o=.001){return V(()=>{let s=ap(r,n),a=s.mean,i=s.variance,l=[];for(let d of Wr(0,r.rank))n.indexOf(d)!==-1?l.push(1):l.push(r.shape[d]);let u=a.reshape(l),c=i.reshape(l),p=e==null?null:e.reshape(l),m=t==null?null:t.reshape(l);return[nh(r,u,c,m,p,o),a,i]})}function Ate(r,e,t,n,o=.001){return y.arraysEqual(n.slice().sort(),Wr(0,r.rank-1))?Ste(r,e,t,n,o):Tte(r,e,t,n,o)}var oh=class extends Le{constructor(e){e==null&&(e={}),super(e),this.supportsMasking=!0,this.axis=e.axis==null?-1:e.axis,this.momentum=e.momentum==null?.99:e.momentum,this.epsilon=e.epsilon==null?.001:e.epsilon,this.center=e.center==null?!0:e.center,this.scale=e.scale==null?!0:e.scale,this.betaInitializer=pt(e.betaInitializer||"zeros"),this.gammaInitializer=pt(e.gammaInitializer||"ones"),this.movingMeanInitializer=pt(e.movingMeanInitializer||"zeros"),this.movingVarianceInitializer=pt(e.movingVarianceInitializer||"ones"),this.betaConstraint=Lt(e.betaConstraint),this.gammaConstraint=Lt(e.gammaConstraint),this.betaRegularizer=wt(e.betaRegularizer),this.gammaRegularizer=wt(e.gammaRegularizer)}build(e){e=Xe(e);let t=this.axis>=0?this.axis:this.axis+e.length,n=e[t];if(n==null)throw new z(`Axis ${t} of input tensor should have a defined dimension but the layer received an input with shape ${JSON.stringify(e)}.`);this.inputSpec=[new Nt({ndim:e.length,axes:{[t]:n}})];let o=[n];this.scale&&(this.gamma=this.addWeight("gamma",o,null,this.gammaInitializer,this.gammaRegularizer,!0,this.gammaConstraint)),this.center&&(this.beta=this.addWeight("beta",o,null,this.betaInitializer,this.betaRegularizer,!0,this.betaConstraint)),this.movingMean=this.addWeight("moving_mean",o,null,this.movingMeanInitializer,null,!1),this.movingVariance=this.addWeight("moving_variance",o,null,this.movingVarianceInitializer,null,!1),this.built=!0}call(e,t){return V(()=>{let n=t.training==null?!1:t.training,o=Fe(e),s=o.shape,a=s.length,i=Wr(0,a),l=this.axis>=0?this.axis:this.axis+a;i.splice(l,1);let u=po(1,a);u[l]=s[l];let c=i.slice();c.sort();let p=!y.arraysEqual(c,Wr(0,a).slice(0,a-1)),m=()=>{if(p){let b=this.movingMean.read().reshape(u),w=this.movingVariance.read().reshape(u),_=this.center?this.beta.read().reshape(u):null,I=this.scale?this.gamma.read().reshape(u):null;return nh(o,b,w,_,I,this.epsilon)}else return nh(o,this.movingMean.read(),this.movingVariance.read(),this.beta==null?null:this.beta.read(),this.gamma==null?null:this.gamma.read(),this.epsilon)};if(!n)return m();let[f,d,h]=Ate(o,this.gamma.read(),this.beta.read(),i,this.epsilon),g=(b,w,_)=>{V(()=>{let I=1-_,E=b.read(),D=E.sub(w).mul(I);b.write(E.sub(D))})};return(()=>{g(this.movingMean,d,this.momentum),g(this.movingVariance,h,this.momentum)})(),f})}getConfig(){let e={axis:this.axis,momentum:this.momentum,epsilon:this.epsilon,center:this.center,scale:this.scale,betaInitializer:It(this.betaInitializer),gammaInitializer:It(this.gammaInitializer),movingMeanInitializer:It(this.movingMeanInitializer),movingVarianceInitializer:It(this.movingVarianceInitializer),betaRegularizer:it(this.betaRegularizer),gammaRegularizer:it(this.gammaRegularizer),betaConstraint:Mt(this.betaConstraint),gammaConstraint:Mt(this.gammaConstraint)},t=super.getConfig();return Object.assign(e,t),e}};oh.className="BatchNormalization";Q.registerClass(oh);var sh=class extends Le{constructor(e){if(e==null&&(e={}),super(e),this.axis=e.axis==null?-1:e.axis,typeof this.axis=="number"){if(!Number.isInteger(this.axis))throw new Error(`Expected axis to be an integer, but received ${this.axis}`)}else if(Array.isArray(this.axis)){for(let t of this.axis)if(!Number.isInteger(t))throw new Error(`Expected axis to be an array of integers, but received ${JSON.stringify(this.axis)}`)}else throw new Error(`Expected axis to be an integer or an array of integers, but received ${JSON.stringify(this.axis)}`);this.epsilon=e.epsilon==null?.001:e.epsilon,this.center=e.center==null?!0:e.center,this.scale=e.scale==null?!0:e.scale,this.betaInitializer=pt(e.betaInitializer||"zeros"),this.gammaInitializer=pt(e.gammaInitializer||"ones"),this.betaRegularizer=wt(e.betaRegularizer),this.gammaRegularizer=wt(e.gammaRegularizer),this.supportsMasking=!0}build(e){e=Xe(e);let t=e.length;typeof this.axis=="number"&&(this.axis=[this.axis]);for(let s=0;s<this.axis.length;++s)this.axis[s]<0&&(this.axis[s]+=t);for(let s of this.axis)if(s<0||s>=t)throw new Error(`Invalid axis: ${s}`);if(this.axis.length!==ho(this.axis).length)throw new Error(`Found duplicate axes in: ${this.axis}`);let n=this.axis.map(s=>e[s]),o=!0;this.scale?this.gamma=this.addWeight("gamma",n,"float32",this.gammaInitializer,this.gammaRegularizer,o):this.gamma=null,this.center?this.beta=this.addWeight("beta",n,"float32",this.betaInitializer,this.betaRegularizer,o):this.beta=null,this.built=!0}call(e,t){let n=Fe(e),o=n.shape,s=o.length;return V(()=>{let a=!0,{mean:i,variance:l}=ap(n,this.axis,a),u=po(1,s);for(let h of this.axis)u[h]=o[h];let c=h=>h!=null&&h.shape.length!==s&&this.axis!==[s-1]?h.reshape(u):h,p=c(this.gamma.read()),m=c(this.beta.read()),f=[],d=[];for(let h=0;h<s;++h)this.axis.indexOf(h)!==-1?(f.push(o[h]),d.push(1)):(f.push(1),d.push(o[h]));return i=i.tile(f),l=l.tile(f),p=p.tile(d),m=m.tile(d),nh(n,i,l,m,p,this.epsilon)})}getConfig(){let e={axis:this.axis,epsilon:this.epsilon,center:this.center,scale:this.scale,betaInitializer:It(this.betaInitializer),gammaInitializer:It(this.gammaInitializer),betaRegularizer:it(this.betaRegularizer),gammaRegularizer:it(this.gammaRegularizer)},t=super.getConfig();return Object.assign(e,t),e}};sh.className="LayerNormalization";Q.registerClass(sh);function Ete(r,e,t){return V(()=>{if(r.rank!==4)throw new z(`temporalPadding expects input tensor to be 4-D, but received a ${r.rank}-D tensor.`);if(e==null&&(e=[[1,1],[1,1]]),e.length!==2||e[0].length!==2||e[1].length!==2)throw new z("spatial2dPadding expects `padding` to be an Array of two Arrays, each of which is an Array of two integers.");if(t==null&&(t=rn()),t!=="channelsLast"&&t!=="channelsFirst")throw new z(`Unknown data format: ${t}. Supported data formats are 'channelsLast' and 'channelsFirst.`);let n;return t==="channelsFirst"?n=[[0,0],[0,0],e[0],e[1]]:n=[[0,0],e[0],e[1],[0,0]],Lr(r,n)})}var ih=class extends Le{constructor(e){if(e==null&&(e={}),super(e),this.dataFormat=e.dataFormat==null?rn():e.dataFormat,e.padding==null)this.padding=[[1,1],[1,1]];else if(typeof e.padding=="number")this.padding=[[e.padding,e.padding],[e.padding,e.padding]];else{if(e.padding=e.padding,e.padding.length!==2)throw new z(`ZeroPadding2D expects padding to be a length-2 array, but received a length-${e.padding.length} array.`);let t,n;if(typeof e.padding[0]=="number")t=[e.padding[0],e.padding[0]],n=[e.padding[1],e.padding[1]];else{if(e.padding=e.padding,e.padding[0].length!==2)throw new z(`ZeroPadding2D expects height padding to be a length-2 array, but received a length-${e.padding[0].length} array.`);if(t=e.padding[0],e.padding[1].length!==2)throw new z(`ZeroPadding2D expects width padding to be a length-2 array, but received a length-${e.padding[1].length} array.`);n=e.padding[1]}this.padding=[t,n]}this.inputSpec=[new Nt({ndim:4})]}computeOutputShape(e){e=Xe(e);let t,n;return this.dataFormat==="channelsFirst"?(e[2]!=null&&e[2]>=0?t=e[2]+this.padding[0][0]+this.padding[0][1]:t=null,e[3]!=null&&e[3]>=0?n=e[3]+this.padding[1][0]+this.padding[1][1]:n=null,[e[0],e[1],t,n]):(e[1]!=null&&e[1]>=0?t=e[1]+this.padding[0][0]+this.padding[0][1]:t=null,e[2]!=null&&e[2]>=0?n=e[2]+this.padding[1][0]+this.padding[1][1]:n=null,[e[0],t,n,e[3]])}call(e,t){return V(()=>Ete(Fe(e),this.padding,this.dataFormat))}getConfig(){let e={padding:this.padding,dataFormat:this.dataFormat},t=super.getConfig();return Object.assign(e,t),e}};ih.className="ZeroPadding2D";Q.registerClass(ih);function ay(r,e,t,n,o,s){return V(()=>{Ot(o),ZC(s),nn(n),t==null&&(t=[1,1]),n==null&&(n="valid"),o==null&&(o=rn()),s==null&&(s="max"),r=Nd(r,o);let a,i=n==="same"?"same":"valid";return s==="max"?a=_a(r,e,t,i):a=ha(r,e,t,i),o==="channelsFirst"&&(a=Ue(a,[0,3,1,2])),a})}function Y3(r,e,t,n,o,s){return V(()=>{Ot(o),ZC(s),nn(n),t==null&&(t=[1,1,1]),n==null&&(n="valid"),o==null&&(o=rn()),s==null&&(s="max"),r=E0(r,o);let a,i=n==="same"?"same":"valid";return s==="max"?a=cf(r,e,t,i):a=Ym(r,e,t,i),o==="channelsFirst"&&(a=Ue(a,[0,4,1,2,3])),a})}var O0=class extends Le{constructor(e){if(e.poolSize==null&&(e.poolSize=2),super(e),typeof e.poolSize=="number")this.poolSize=[e.poolSize];else if(Array.isArray(e.poolSize)&&e.poolSize.length===1&&typeof e.poolSize[0]=="number")this.poolSize=e.poolSize;else throw new z(`poolSize for 1D convolutional layer must be a number or an Array of a single number, but received ${JSON.stringify(e.poolSize)}`);if(Ht(this.poolSize,"poolSize"),e.strides==null)this.strides=this.poolSize;else if(typeof e.strides=="number")this.strides=[e.strides];else if(Array.isArray(e.strides)&&e.strides.length===1&&typeof e.strides[0]=="number")this.strides=e.strides;else throw new z(`strides for 1D convolutional layer must be a number or an Array of a single number, but received ${JSON.stringify(e.strides)}`);Ht(this.strides,"strides"),this.padding=e.padding==null?"valid":e.padding,nn(this.padding),this.inputSpec=[new Nt({ndim:3})]}computeOutputShape(e){e=Xe(e);let t=_n(e[1],this.poolSize[0],this.padding,this.strides[0]);return[e[0],t,e[2]]}call(e,t){return V(()=>{this.invokeCallHook(e,t),e=La(Fe(e),2);let n=this.poolingFunction(Fe(e),[this.poolSize[0],1],[this.strides[0],1],this.padding,"channelsLast");return Tn(n,[2])})}getConfig(){let e={poolSize:this.poolSize,padding:this.padding,strides:this.strides},t=super.getConfig();return Object.assign(e,t),e}},ah=class extends O0{constructor(e){super(e)}poolingFunction(e,t,n,o,s){return Ot(s),nn(o),ay(e,t,n,o,s,"max")}};ah.className="MaxPooling1D";Q.registerClass(ah);var lh=class extends O0{constructor(e){super(e)}poolingFunction(e,t,n,o,s){return Ot(s),nn(o),ay(e,t,n,o,s,"avg")}};lh.className="AveragePooling1D";Q.registerClass(lh);var P0=class extends Le{constructor(e){if(e.poolSize==null&&(e.poolSize=[2,2]),super(e),this.poolSize=Array.isArray(e.poolSize)?e.poolSize:[e.poolSize,e.poolSize],e.strides==null)this.strides=this.poolSize;else if(Array.isArray(e.strides)){if(e.strides.length!==2)throw new z(`If the strides property of a 2D pooling layer is an Array, it is expected to have a length of 2, but received length ${e.strides.length}.`);this.strides=e.strides}else this.strides=[e.strides,e.strides];Ht(this.poolSize,"poolSize"),Ht(this.strides,"strides"),this.padding=e.padding==null?"valid":e.padding,this.dataFormat=e.dataFormat==null?"channelsLast":e.dataFormat,Ot(this.dataFormat),nn(this.padding),this.inputSpec=[new Nt({ndim:4})]}computeOutputShape(e){e=Xe(e);let t=this.dataFormat==="channelsFirst"?e[2]:e[1],n=this.dataFormat==="channelsFirst"?e[3]:e[2];return t=_n(t,this.poolSize[0],this.padding,this.strides[0]),n=_n(n,this.poolSize[1],this.padding,this.strides[1]),this.dataFormat==="channelsFirst"?[e[0],e[1],t,n]:[e[0],t,n,e[3]]}call(e,t){return V(()=>(this.invokeCallHook(e,t),this.poolingFunction(Fe(e),this.poolSize,this.strides,this.padding,this.dataFormat)))}getConfig(){let e={poolSize:this.poolSize,padding:this.padding,strides:this.strides,dataFormat:this.dataFormat},t=super.getConfig();return Object.assign(e,t),e}},uh=class extends P0{constructor(e){super(e)}poolingFunction(e,t,n,o,s){return Ot(s),nn(o),ay(e,t,n,o,s,"max")}};uh.className="MaxPooling2D";Q.registerClass(uh);var ch=class extends P0{constructor(e){super(e)}poolingFunction(e,t,n,o,s){return Ot(s),nn(o),ay(e,t,n,o,s,"avg")}};ch.className="AveragePooling2D";Q.registerClass(ch);var M0=class extends Le{constructor(e){if(e.poolSize==null&&(e.poolSize=[2,2,2]),super(e),this.poolSize=Array.isArray(e.poolSize)?e.poolSize:[e.poolSize,e.poolSize,e.poolSize],e.strides==null)this.strides=this.poolSize;else if(Array.isArray(e.strides)){if(e.strides.length!==3)throw new z(`If the strides property of a 3D pooling layer is an Array, it is expected to have a length of 3, but received length ${e.strides.length}.`);this.strides=e.strides}else this.strides=[e.strides,e.strides,e.strides];Ht(this.poolSize,"poolSize"),Ht(this.strides,"strides"),this.padding=e.padding==null?"valid":e.padding,this.dataFormat=e.dataFormat==null?"channelsLast":e.dataFormat,Ot(this.dataFormat),nn(this.padding),this.inputSpec=[new Nt({ndim:5})]}computeOutputShape(e){e=Xe(e);let t=this.dataFormat==="channelsFirst"?e[2]:e[1],n=this.dataFormat==="channelsFirst"?e[3]:e[2],o=this.dataFormat==="channelsFirst"?e[4]:e[3];return t=_n(t,this.poolSize[0],this.padding,this.strides[0]),n=_n(n,this.poolSize[1],this.padding,this.strides[1]),o=_n(o,this.poolSize[2],this.padding,this.strides[2]),this.dataFormat==="channelsFirst"?[e[0],e[1],t,n,o]:[e[0],t,n,o,e[4]]}call(e,t){return V(()=>(this.invokeCallHook(e,t),this.poolingFunction(Fe(e),this.poolSize,this.strides,this.padding,this.dataFormat)))}getConfig(){let e={poolSize:this.poolSize,padding:this.padding,strides:this.strides,dataFormat:this.dataFormat},t=super.getConfig();return Object.assign(e,t),e}},ph=class extends M0{constructor(e){super(e)}poolingFunction(e,t,n,o,s){return Ot(s),nn(o),Y3(e,t,n,o,s,"max")}};ph.className="MaxPooling3D";Q.registerClass(ph);var mh=class extends M0{constructor(e){super(e)}poolingFunction(e,t,n,o,s){return Ot(s),nn(o),Y3(e,t,n,o,s,"avg")}};mh.className="AveragePooling3D";Q.registerClass(mh);var L0=class extends Le{constructor(e){super(e);this.inputSpec=[new Nt({ndim:3})]}computeOutputShape(e){return[e[0],e[2]]}call(e,t){throw new Se}},fh=class extends L0{constructor(e){super(e||{})}call(e,t){return V(()=>{let n=Fe(e);return dt(n,1)})}};fh.className="GlobalAveragePooling1D";Q.registerClass(fh);var dh=class extends L0{constructor(e){super(e||{})}call(e,t){return V(()=>{let n=Fe(e);return mr(n,1)})}};dh.className="GlobalMaxPooling1D";Q.registerClass(dh);var z0=class extends Le{constructor(e){super(e);this.dataFormat=e.dataFormat==null?"channelsLast":e.dataFormat,Ot(this.dataFormat),this.inputSpec=[new Nt({ndim:4})]}computeOutputShape(e){return e=e,this.dataFormat==="channelsLast"?[e[0],e[3]]:[e[0],e[1]]}call(e,t){throw new Se}getConfig(){let e={dataFormat:this.dataFormat},t=super.getConfig();return Object.assign(e,t),e}},hh=class extends z0{call(e,t){return V(()=>{let n=Fe(e);return this.dataFormat==="channelsLast"?dt(n,[1,2]):dt(n,[2,3])})}};hh.className="GlobalAveragePooling2D";Q.registerClass(hh);var gh=class extends z0{call(e,t){return V(()=>{let n=Fe(e);return this.dataFormat==="channelsLast"?mr(n,[1,2]):mr(n,[2,3])})}};gh.className="GlobalMaxPooling2D";Q.registerClass(gh);var B0=class extends Le{constructor(e){super(e);this.layer=e.layer}build(e){this.built=!0}get trainable(){return this.layer!=null?this.layer.trainable:!1}set trainable(e){this.layer!=null&&(this.layer.trainable=e)}get trainableWeights(){return this.layer.trainableWeights}get nonTrainableWeights(){return this.layer.nonTrainableWeights}get updates(){return this.layer._updates}get losses(){return this.layer.losses}getWeights(){return this.layer.getWeights()}setWeights(e){this.layer.setWeights(e)}getConfig(){let e={layer:{className:this.layer.getClassName(),config:this.layer.getConfig()}},t=super.getConfig();return Object.assign(e,t),e}setFastWeightInitDuringBuild(e){super.setFastWeightInitDuringBuild(e),this.layer!=null&&this.layer.setFastWeightInitDuringBuild(e)}static fromConfig(e,t,n={}){let o=t.layer,s=an(o,n);delete t.layer;let a={layer:s};return Object.assign(a,t),new e(a)}},xh=class extends B0{constructor(e){super(e);this.supportsMasking=!0}build(e){if(e=Xe(e),e.length<3)throw new z(`TimeDistributed layer expects an input shape >= 3D, but received input shape ${JSON.stringify(e)}`);this.inputSpec=[{shape:e}];let t=[e[0]].concat(e.slice(2));this.layer.built||(this.layer.build(t),this.layer.built=!0),super.build(e)}computeOutputShape(e){e=Xe(e);let t=[e[0]].concat(e.slice(2)),n=this.layer.computeOutputShape(t),o=e[1];return[n[0],o].concat(n.slice(1))}call(e,t){return V(()=>(e=Fe(e),R0((a,i)=>[Fe(this.layer.call(a,t)),[]],e,[],!1,null,null,!1,!0)[1]))}};xh.className="TimeDistributed";Q.registerClass(xh);function Dte(r){mi(o3,"BidirectionalMergeMode",r)}var $te="concat",yh=class extends B0{constructor(e){super(e);let t=e.layer.getConfig(),n={};n.className=e.layer.getClassName(),n.config=t,this.forwardLayer=an(n),t.goBackwards=t.goBackwards!==!0;let o={};if(o.className=e.layer.getClassName(),o.config=t,this.backwardLayer=an(o),this.forwardLayer.name="forward_"+this.forwardLayer.name,this.backwardLayer.name="backward_"+this.backwardLayer.name,this.mergeMode=e.mergeMode===void 0?$te:e.mergeMode,Dte(this.mergeMode),e.weights)throw new Se("weights support is not implemented for Bidirectional layer yet.");this._stateful=e.layer.stateful,this.returnSequences=e.layer.returnSequences,this.returnState=e.layer.returnState,this.supportsMasking=!0,this._trainable=!0,this.inputSpec=e.layer.inputSpec,this.numConstants=null}get trainable(){return this._trainable}set trainable(e){this._trainable=e,this.forwardLayer!=null&&(this.forwardLayer.trainable=e),this.backwardLayer!=null&&(this.backwardLayer.trainable=e)}getWeights(){return this.forwardLayer.getWeights().concat(this.backwardLayer.getWeights())}setWeights(e){let t=e.length,n=Math.floor(t/2);this.forwardLayer.setWeights(e.slice(0,n)),this.backwardLayer.setWeights(e.slice(n))}computeOutputShape(e){let t=this.forwardLayer.computeOutputShape(e);Array.isArray(t)&&Array.isArray(t[0])||(t=[t]),t=t;let n,o,s;return this.returnState&&(s=t.slice(1)),n=t[0],n=n,this.mergeMode==="concat"?(n[n.length-1]*=2,o=[n]):this.mergeMode==null?o=[n,n.slice()]:o=[n],this.returnState?this.mergeMode==null?o.concat(s).concat(s.slice()):[n].concat(s).concat(s.slice()):wr(o)}apply(e,t){let n=t==null?null:t.initialState,o=t==null?null:t.constants;t==null&&(t={});let s=$0(e,n,o,this.numConstants);if(e=s.inputs,n=s.initialState,o=s.constants,Array.isArray(e)&&(n=e.slice(1),e=e[0]),(n==null||n.length===0)&&o==null)return super.apply(e,t);let a=[],i=[];if(n!=null){let u=n.length;if(u%2>0)throw new z("When passing `initialState` to a Bidrectional RNN, the state should be an Array containing the states of the underlying RNNs.");t.initialState=n,a.push(...n);let c=n.map(p=>new Nt({shape:p.shape}));this.forwardLayer.stateSpec=c.slice(0,u/2),this.backwardLayer.stateSpec=c.slice(u/2),i.push(...c)}if(o!=null)throw new Se("Support for constants in Bidirectional layers is not implemented yet.");let l=a[0]instanceof sn;for(let u of a)if(u instanceof sn!==l)throw new z("The initial state of a Bidirectional layer cannot be specified as a mix of symbolic and non-symbolic tensors");if(l){let u=[e].concat(a),c=this.inputSpec.concat(i),p=this.inputSpec;this.inputSpec=c;let m=super.apply(u,t);return this.inputSpec=p,m}else return super.apply(e,t)}call(e,t){return V(()=>{let n=t.initialState,o,s;if(n==null)o=this.forwardLayer.call(e,t),s=this.backwardLayer.call(e,t);else{let l=n.slice(0,n.length/2),u=n.slice(n.length/2);o=this.forwardLayer.call(e,Object.assign(t,{initialState:l})),s=this.backwardLayer.call(e,Object.assign(t,{initialState:u}))}let a;this.returnState&&(Array.isArray(o)&&(a=o.slice(1).concat(s.slice(1))),o=o[0],s=s[0]),this.returnSequences&&(s=Kt(s,1));let i;return this.mergeMode==="concat"?i=Bp([o,s]):this.mergeMode==="sum"?i=J(o,s):this.mergeMode==="ave"?i=O(.5,J(o,s)):this.mergeMode==="mul"?i=O(o,s):this.mergeMode==null&&(i=[o,s]),this.returnState?this.mergeMode==null?i.concat(a):[i].concat(a):i})}resetStates(e){this.forwardLayer.resetStates(),this.backwardLayer.resetStates()}build(e){Ts(this.forwardLayer.name,()=>{this.forwardLayer.build(e)}),Ts(this.backwardLayer.name,()=>{this.backwardLayer.build(e)}),this.built=!0}computeMask(e,t){Array.isArray(t)&&(t=t[0]);let n;if(this.returnSequences?this.mergeMode==null?n=[t,t]:n=t:this.mergeMode==null?n=[null,null]:n=null,this.returnState){let s=this.forwardLayer.states.map(a=>null);return Array.isArray(n)?n.concat(s).concat(s):[n].concat(s).concat(s)}else return n}get trainableWeights(){return this.forwardLayer.trainableWeights.concat(this.backwardLayer.trainableWeights)}get nonTrainableWeights(){return this.forwardLayer.nonTrainableWeights.concat(this.backwardLayer.nonTrainableWeights)}setFastWeightInitDuringBuild(e){super.setFastWeightInitDuringBuild(e),this.forwardLayer!=null&&this.forwardLayer.setFastWeightInitDuringBuild(e),this.backwardLayer!=null&&this.backwardLayer.setFastWeightInitDuringBuild(e)}getConfig(){let e={mergeMode:this.mergeMode},t=super.getConfig();return Object.assign(e,t),e}static fromConfig(e,t){let n=an(t.layer);if(delete t.layer,t.numConstants!=null)throw new Se("Deserialization of a Bidirectional layer with numConstants present is not supported yet.");let o=t;return o.layer=n,new e(o)}};yh.className="Bidirectional";Q.registerClass(yh);function Rte(r){return new fi(r)}function Fte(r){return new vd(r)}function Ote(r){return new wd(r)}function Pte(r){return new _d(r)}function Mte(r){return new kd(r)}function Lte(r){return new Id(r)}function zte(r){return new Cd(r)}function Bte(r){return new tc(r)}function Vte(r){return new Rl(r)}function Gte(r){return new Sd(r)}function Wte(r){return new Fl(r)}function jte(r){return new Td(r)}function Ute(r){return new Ad(r)}function Hte(r){return new Ed(r)}function qte(r){return new Dd(r)}function Kte(r){return new $d(r)}function Xte(r){return new Bd(r)}function Yte(r){return new Ld(r)}function Zte(r){return new im(r)}function Jte(r){return new Md(r)}function Qte(r){return new zd(r)}function ere(r){return new Vd(r)}function tre(r){return new Gd(r)}function rre(r){return new Wd(r)}function nre(r){return new Ud(r)}function ore(r){return new Hd(r)}function sre(r){return new Kd(r)}function ire(r){return new Zd(r)}function are(r){return new Xd(r)}function lre(r){return new Yd(r)}function ure(r){return new qd(r)}function cre(r){return new Qd(r)}function pre(r){return new oh(r)}function mre(r){return new sh(r)}function fre(r){return new ih(r)}function V0(r){return new lh(r)}function dre(r){return V0(r)}function hre(r){return V0(r)}function G0(r){return new ch(r)}function gre(r){return G0(r)}function xre(r){return G0(r)}function W0(r){return new mh(r)}function yre(r){return W0(r)}function bre(r){return W0(r)}function wre(r){return new fh(r)}function _re(r){return new hh(r)}function Z3(r){return new dh(r)}function J3(r){return new gh(r)}function Q3(r){return new ah(r)}function eB(r){return new uh(r)}function kre(r){return new ph(r)}function vre(r){return new Fd(r)}function Cre(r){return new nm(r)}function Ire(r){return new Od(r)}function Nre(r){return new Pl(r)}function Sre(r){return new Rd(r)}function Tre(r){return new rm(r)}function Are(r){return new Pd(r)}function Ere(r){return new sm(r)}function Dre(r){return new $n(r)}function $re(r){return new om(r)}function Rre(r){return new yh(r)}function Fre(r){return new xh(r)}var Ore=Z3,Pre=J3,Mre=Q3,Lre=eB;function zre(r){return new eh(r)}function Bre(r){return new th(r)}function Vre(r){return new rh(r)}function Gre(r){return new jd(r)}var rB={};We(rB,{MAPE:()=>Qre,MSE:()=>rne,binaryAccuracy:()=>Wre,binaryCrossentropy:()=>jre,categoricalAccuracy:()=>Hre,categoricalCrossentropy:()=>qre,cosineProximity:()=>Yre,mape:()=>ene,meanAbsoluteError:()=>Zre,meanAbsolutePercentageError:()=>Jre,meanSquaredError:()=>tne,mse:()=>nne,precision:()=>Kre,recall:()=>Xre,sparseCategoricalAccuracy:()=>Ure});function Wre(r,e){return dd(r,e)}function jre(r,e){return Zx(r,e)}function Ure(r,e){return Jx(r,e)}function Hre(r,e){return hd(r,e)}function qre(r,e){return gd(r,e)}function Kre(r,e){return s0(r,e)}function Xre(r,e){return I3(r,e)}function Yre(r,e){return md(r,e)}function Zre(r,e){return Zp(r,e)}function Jre(r,e){return El(r,e)}function Qre(r,e){return El(r,e)}function ene(r,e){return El(r,e)}function tne(r,e){return hi(r,e)}function rne(r,e){return hi(r,e)}function nne(r,e){return hi(r,e)}var nB={};We(nB,{modelFromJSON:()=>W3});var oB={};We(oB,{l1:()=>sne,l1l2:()=>one,l2:()=>ine});function one(r){return new Qu(r)}function sne(r){return U3(r)}function ine(r){return H3(r)}var j0=class extends Al{constructor(){super(...arguments);this.model=null}setModel(e){if(!(e instanceof jn))throw new Error("model must be a LayersModel, not some other Container");this.model=e}};function ly(r,e){return r<e}function sB(r,e){return r>e}var U0=class extends j0{constructor(e){super();if(e==null&&(e={}),e.restoreBestWeights)throw new Se("restoreBestWeights = True is not implemented in EarlyStopping yet.");this.monitor=e.monitor||"val_loss",this.minDelta=Math.abs(e.minDelta||0),this.patience=e.patience||0,this.verbose=e.verbose||0,this.mode=e.mode||"auto",this.baseline=e.baseline,["auto","min","max"].indexOf(this.mode)===-1&&(console.warn(`EarlyStopping mode '${this.mode}' is invalid. Falling back to mode 'auto'.`),this.mode="auto"),this.mode==="min"?this.monitorFunc=ly:this.mode==="max"?this.monitorFunc=sB:this.monitor.indexOf("acc")!==-1?this.monitorFunc=sB:this.monitorFunc=ly,this.monitorFunc===ly&&(this.minDelta*=-1)}async onTrainBegin(e){this.wait=0,this.stoppedEpoch=0,this.baseline!=null?this.best=this.baseline:this.best=this.monitorFunc===ly?Infinity:-Infinity}async onEpochEnd(e,t){await di(t);let n=this.getMonitorValue(t);n!=null&&(this.monitorFunc(n-this.minDelta,this.best)?(this.best=n,this.wait=0):(this.wait++,this.wait>=this.patience&&(this.stoppedEpoch=e,this.model.stopTraining=!0)))}async onTrainEnd(e){this.stoppedEpoch>0&&this.verbose&&console.log(`Epoch ${this.stoppedEpoch}: early stopping.`)}getMonitorValue(e){e==null&&(e={});let t=e[this.monitor];return t==null&&console.warn(`Metric for EarlyStopping ${this.monitor} is not available. Available metrics are: ${Object.keys(e)}`),t}};function ane(r){return new U0(r)}var lne={earlyStopping:ane};var yo;(function(r){r[r.DT_INVALID=0]="DT_INVALID",r[r.DT_FLOAT=1]="DT_FLOAT",r[r.DT_DOUBLE=2]="DT_DOUBLE",r[r.DT_INT32=3]="DT_INT32",r[r.DT_UINT8=4]="DT_UINT8",r[r.DT_INT16=5]="DT_INT16",r[r.DT_INT8=6]="DT_INT8",r[r.DT_STRING=7]="DT_STRING",r[r.DT_COMPLEX64=8]="DT_COMPLEX64",r[r.DT_INT64=9]="DT_INT64",r[r.DT_BOOL=10]="DT_BOOL",r[r.DT_QINT8=11]="DT_QINT8",r[r.DT_QUINT8=12]="DT_QUINT8",r[r.DT_QINT32=13]="DT_QINT32",r[r.DT_BFLOAT16=14]="DT_BFLOAT16",r[r.DT_FLOAT_REF=101]="DT_FLOAT_REF",r[r.DT_DOUBLE_REF=102]="DT_DOUBLE_REF",r[r.DT_INT32_REF=103]="DT_INT32_REF",r[r.DT_UINT8_REF=104]="DT_UINT8_REF",r[r.DT_INT16_REF=105]="DT_INT16_REF",r[r.DT_INT8_REF=106]="DT_INT8_REF",r[r.DT_STRING_REF=107]="DT_STRING_REF",r[r.DT_COMPLEX64_REF=108]="DT_COMPLEX64_REF",r[r.DT_INT64_REF=109]="DT_INT64_REF",r[r.DT_BOOL_REF=110]="DT_BOOL_REF",r[r.DT_QINT8_REF=111]="DT_QINT8_REF",r[r.DT_QUINT8_REF=112]="DT_QUINT8_REF",r[r.DT_QINT32_REF=113]="DT_QINT32_REF",r[r.DT_BFLOAT16_REF=114]="DT_BFLOAT16_REF"})(yo||(yo={}));var iB;(function(r){let e;(function(t){t[t.LEGACY=0]="LEGACY",t[t.V1=1]="V1",t[t.V2=2]="V2"})(e=r.CheckpointFormatVersion||(r.CheckpointFormatVersion={}))})(iB||(iB={}));var H0={};function une(r,e){let t={tfOpName:r,category:"custom",inputs:[],attrs:[],customExecutor:e};H0[r]=t}function uy(r){return H0[r]}function cne(r){delete H0[r]}function v(r,e,t,n,o){let s=e.inputParams[r];if(s&&s.inputIndexStart!==void 0){let i=s.inputIndexStart,l=s.inputIndexEnd===0?void 0:s.inputIndexEnd===void 0?i+1:s.inputIndexEnd;if(s.type==="tensor")return hr(e.inputNames[s.inputIndexStart],t,n,o);if(s.type==="tensors")return e.inputNames.slice(i,l).map(m=>hr(m,t,n,o));let u=hr(e.inputNames.slice(i)[0],t,n,o),c=u.dataSync();return s.type==="number"?c[0]:y.toNestedArray(u.shape,c)}let a=e.attrParams[r];return a&&a.value}function hr(r,e,t,n){let[o,s]=un(r);if(n!=null){let i=n.getHashTableHandleByName(o);if(i!=null)return i}let a=t.currentContextIds.find(i=>!!e[cy(o,i)]);return a!==void 0?e[cy(o,a)][s]:void 0}function aB(r,e,t){return e[cy(r,t.currentContextId)]}function Fs(r,e){let[t,n]=un(r);return[cy(t,e&&e.currentContextId),n]}function cy(r,e){return e?`${r}-${e}`:r}function un(r){let e=r.split(":");return e.length===1?[r,0]:[e[0],Number(e[e.length-1])]}function bh(r,e,t){let n=v("pad",r,e,t);if(n==="explicit"){n=v("explicitPaddings",r,e,t);let o=[[0,0],[0,0],[0,0],[0,0]];for(let s=0;s<4;s++)o[s][0]=n[s*2],o[s][1]=n[s*2+1];return o}return n}function Os(r){return r.kept?r:Ln(r)}var q0={};We(q0,{json:()=>pne});var pne=[{tfOpName:"Add",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"AddV2",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"AddN",category:"arithmetic",inputs:[{start:0,end:0,name:"tensors",type:"tensors"}]},{tfOpName:"BiasAdd",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0},{tfName:"data_format",name:"dataFormat",type:"string",notSupported:!0}]},{tfOpName:"Sub",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"RealDiv",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Div",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"DivNoNan",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"FloorDiv",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Mul",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Maximum",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Minimum",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Pow",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"SquaredDifference",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Mod",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"FloorMod",category:"arithmetic",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]}];var K0={};We(K0,{json:()=>mne});var mne=[{tfOpName:"Abs",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Acos",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Asin",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Atan",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Atan2",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"y",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Ceil",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"ClipByValue",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"clipValueMin",type:"number"},{start:2,name:"clipValueMax",type:"number"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Complex",category:"basic_math",inputs:[{start:0,name:"real",type:"tensor"},{start:1,name:"imag",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"ComplexAbs",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Cos",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Cosh",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Elu",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Exp",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Floor",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Log",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Imag",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0},{tfName:"Tout",name:"outputType",type:"dtype",notSupported:!0}]},{tfOpName:"Neg",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Real",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0},{tfName:"Tout",name:"outputType",type:"dtype",notSupported:!0}]},{tfOpName:"Prelu",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"alpha",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Relu",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Relu6",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Selu",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Sigmoid",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Sin",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Sinh",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Sqrt",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Rsqrt",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Square",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Tan",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Tanh",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Sign",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Round",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Expm1",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Log1p",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Reciprocal",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Softplus",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Asinh",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Acosh",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Atanh",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Erf",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Prod",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axes",type:"number[]"}],attrs:[{tfName:"keep_dims",name:"keepDims",type:"bool",notSupported:!0},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"LeakyRelu",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"alpha",name:"alpha",type:"number",defaultValue:.2},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"IsNan",category:"basic_math",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]}];var X0={};We(X0,{json:()=>fne});var fne=[{tfOpName:"EmptyTensorList",category:"control",inputs:[{start:0,name:"elementShape",type:"shape"},{start:1,name:"maxNumElements",type:"number"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"LoopCond",category:"control",inputs:[{start:0,name:"pred",type:"tensor"}]},{tfOpName:"Switch",category:"control",inputs:[{start:0,name:"data",type:"tensor"},{start:1,name:"pred",type:"tensor"}]},{tfOpName:"Merge",category:"control",inputs:[{start:0,end:0,name:"tensors",type:"tensors"}]},{tfOpName:"Enter",category:"control",inputs:[{start:0,name:"tensor",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0},{tfName:"frame_name",name:"frameName",type:"string"},{tfName:"is_constant",name:"isConstant",type:"bool"}]},{tfOpName:"Exit",category:"control",inputs:[{start:0,name:"tensor",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"NextIteration",category:"control",inputs:[{start:0,name:"tensor",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"TensorArrayV3",category:"control",inputs:[{start:0,name:"size",type:"number"}],attrs:[{tfName:"dtype",name:"dtype",type:"dtype"},{tfName:"element_shape",name:"elementShape",type:"shape"},{tfName:"dynamic_size",name:"dynamicSize",type:"bool"},{tfName:"clear_after_read",name:"clearAfterRead",type:"bool"},{tfName:"identical_element_shapes",name:"identicalElementShapes",type:"bool"},{tfName:"tensor_array_name",name:"name",type:"string"}]},{tfOpName:"TensorArrayWriteV3",category:"control",inputs:[{start:0,name:"tensorArrayId",type:"tensor"},{start:1,name:"index",type:"number"},{start:2,name:"tensor",type:"tensor"},{start:3,name:"flowIn",type:"number"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"TensorArrayReadV3",category:"control",inputs:[{start:0,name:"tensorArrayId",type:"tensor"},{start:1,name:"index",type:"number"},{start:2,name:"flowIn",type:"number"}],attrs:[{tfName:"dtype",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"TensorArrayGatherV3",category:"control",inputs:[{start:0,name:"tensorArrayId",type:"tensor"},{start:1,name:"indices",type:"number[]"},{start:2,name:"flowIn",type:"number"}],attrs:[{tfName:"dtype",name:"dtype",type:"dtype"},{tfName:"element_shape",name:"elementShape",type:"shape"}]},{tfOpName:"TensorArrayScatterV3",category:"control",inputs:[{start:0,name:"tensorArrayId",type:"tensor"},{start:1,name:"indices",type:"number[]"},{start:2,name:"tensor",type:"tensor"},{start:3,name:"flowIn",type:"number"}],attrs:[{tfName:"T",name:"dtype",type:"dtype"}]},{tfOpName:"TensorArrayConcatV3",category:"control",inputs:[{start:0,name:"tensorArrayId",type:"tensor"},{start:1,name:"flowIn",type:"number"}],attrs:[{tfName:"dtype",name:"dtype",type:"dtype"},{tfName:"element_shape_except0",name:"elementShapeExcept0",type:"shape",notSupported:!0}]},{tfOpName:"TensorArraySplitV3",category:"control",inputs:[{start:0,name:"tensorArrayId",type:"tensor"},{start:1,name:"tensor",type:"tensor"},{start:2,name:"lengths",type:"number[]"},{start:3,name:"flowIn",type:"number"}],attrs:[{tfName:"T",name:"dtype",type:"dtype"}]},{tfOpName:"TensorArraySizeV3",category:"control",inputs:[{start:0,name:"tensorArrayId",type:"tensor"},{start:1,name:"flowIn",type:"number"}]},{tfOpName:"TensorArrayCloseV3",category:"control",inputs:[{start:0,name:"tensorArrayId",type:"tensor"}]},{tfOpName:"StatelessIf",category:"control",inputs:[{start:0,name:"cond",type:"tensor"},{start:1,end:0,name:"args",type:"tensors"}],attrs:[{tfName:"then_branch",name:"thenBranch",type:"func"},{tfName:"else_branch",name:"elseBranch",type:"func"}]},{tfOpName:"If",category:"control",inputs:[{start:0,name:"cond",type:"tensor"},{start:1,end:0,name:"args",type:"tensors"}],attrs:[{tfName:"then_branch",name:"thenBranch",type:"func"},{tfName:"else_branch",name:"elseBranch",type:"func"}]},{tfOpName:"StatelessWhile",category:"control",inputs:[{start:0,end:0,name:"args",type:"tensors"}],attrs:[{tfName:"cond",name:"cond",type:"func"},{tfName:"body",name:"body",type:"func"}]},{tfOpName:"While",category:"control",inputs:[{start:0,end:0,name:"args",type:"tensors"}],attrs:[{tfName:"cond",name:"cond",type:"func"},{tfName:"body",name:"body",type:"func"}]},{tfOpName:"TensorListScatter",category:"control",inputs:[{start:0,name:"tensor",type:"tensor"},{start:1,name:"indices",type:"number[]"},{start:2,name:"elementShape",type:"shape"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"TensorListScatterV2",category:"control",inputs:[{start:0,name:"tensor",type:"tensor"},{start:1,name:"indices",type:"number[]"},{start:2,name:"elementShape",type:"shape"},{start:3,name:"numElements",type:"number"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"TensorListGather",category:"control",inputs:[{start:0,name:"tensorListId",type:"tensor"},{start:1,name:"indices",type:"number[]"},{start:2,name:"elementShape",type:"shape"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"TensorListGetItem",category:"control",inputs:[{start:0,name:"tensorListId",type:"tensor"},{start:1,name:"index",type:"number"},{start:2,name:"elementShape",type:"shape"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"TensorListSetItem",category:"control",inputs:[{start:0,name:"tensorListId",type:"tensor"},{start:1,name:"index",type:"number"},{start:2,name:"tensor",type:"tensor"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"TensorListReserve",category:"control",inputs:[{start:0,name:"elementShape",type:"shape"},{start:1,name:"numElements",type:"number"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"TensorListFromTensor",category:"control",inputs:[{start:0,name:"tensor",type:"tensor"},{start:1,name:"elementShape",type:"shape"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"TensorListStack",category:"control",inputs:[{start:0,name:"tensorListId",type:"tensor"},{start:1,name:"elementShape",type:"shape"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"},{tfName:"num_elements",name:"numElements",type:"dtype"}]},{tfOpName:"TensorListSplit",category:"control",inputs:[{start:0,name:"tensor",type:"tensor"},{start:1,name:"elementShape",type:"shape"},{start:2,name:"lengths",type:"number[]"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"TensorListConcat",category:"control",inputs:[{start:0,name:"tensorListId",type:"tensor"}],attrs:[{tfName:"element_shape",name:"elementShape",type:"shape"},{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"TensorListPopBack",category:"control",inputs:[{start:0,name:"tensorListId",type:"tensor"},{start:1,name:"elementShape",type:"shape"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]},{tfOpName:"TensorListPushBack",category:"control",inputs:[{start:0,name:"tensorListId",type:"tensor"},{start:1,name:"tensor",type:"tensor"}],attrs:[{tfName:"element_dtype",name:"elementDType",type:"dtype"}]}];var Y0={};We(Y0,{json:()=>dne});var dne=[{tfOpName:"AvgPool",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"data_format",name:"dataFormat",type:"string",notSupported:!0},{tfName:"ksize",name:"kernelSize",type:"number[]"},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"MaxPool",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"data_format",name:"dataFormat",type:"string",notSupported:!0},{tfName:"ksize",name:"kernelSize",type:"number[]"},{tfName:"explicit_paddings",name:"explicitPaddings",type:"number[]",defaultValue:[],notSupported:!0},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"MaxPoolWithArgmax",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"ksize",name:"kernelSize",type:"number[]"},{tfName:"include_batch_in_index",name:"includeBatchInIndex",type:"bool"},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"AvgPool3D",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"data_format",name:"dataFormat",type:"string",notSupported:!0},{tfName:"ksize",name:"kernelSize",type:"number[]"},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"MaxPool3D",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"data_format",name:"dataFormat",type:"string",notSupported:!0},{tfName:"ksize",name:"kernelSize",type:"number[]"},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Conv1D",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"filter",type:"tensor"}],attrs:[{tfName:"stride",name:"stride",type:"number"},{tfName:"padding",name:"pad",type:"string"},{tfName:"data_format",name:"dataFormat",type:"string",defaultValue:"NWC"},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0},{tfName:"dilation",name:"dilation",type:"number",defaultValue:1}]},{tfOpName:"Conv2D",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"filter",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0},{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"useCudnnOnGpu",name:"useCudnnOnGpu",type:"bool"},{tfName:"data_format",name:"dataFormat",type:"string",defaultValue:"NHWC"},{tfName:"explicit_paddings",name:"explicitPaddings",type:"number[]",defaultValue:[]},{tfName:"dilations",name:"dilations",type:"number[]"}]},{tfOpName:"_FusedConv2D",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"filter",type:"tensor"},{start:2,end:0,name:"args",type:"tensors"}],attrs:[{tfName:"num_args",name:"numArgs",type:"number"},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0},{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"explicit_paddings",name:"explicitPaddings",type:"number[]",defaultValue:[]},{tfName:"use_cudnn_on_gpu",name:"useCudnnOnGpu",type:"bool",defaultValue:!0},{tfName:"data_format",name:"dataFormat",type:"string",defaultValue:"NHWC"},{tfName:"dilations",name:"dilations",type:"number[]",defaultValue:[1,1,1,1]},{tfName:"fused_ops",name:"fusedOps",type:"string[]",defaultValue:[]},{tfName:"epsilon",name:"epsilon",type:"number",defaultValue:1e-4},{tfName:"leakyrelu_alpha",name:"leakyreluAlpha",type:"number"}]},{tfOpName:"Conv2DBackpropInput",category:"convolution",inputs:[{start:2,name:"x",type:"tensor"},{start:1,name:"filter",type:"tensor"},{start:0,name:"outputShape",type:"number[]"}],attrs:[{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"data_format",name:"dataFormat",type:"string",notSupported:!0},{tfName:"explicit_paddings",name:"explicitPaddings",type:"number[]",defaultValue:[]},{tfName:"dilations",name:"dilations",type:"number[]",notSupported:!0}]},{tfOpName:"DepthwiseConv2d",category:"convolution",inputs:[{start:0,name:"input",type:"tensor"},{start:1,name:"filter",type:"tensor"}],attrs:[{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"data_format",name:"dataFormat",type:"string",defaultValue:"NHWC"},{tfName:"explicit_paddings",name:"explicitPaddings",type:"number[]",defaultValue:[]},{tfName:"dilations",name:"dilations",type:"number[]"}]},{tfOpName:"DepthwiseConv2dNative",category:"convolution",inputs:[{start:0,name:"input",type:"tensor"},{start:1,name:"filter",type:"tensor"}],attrs:[{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"data_format",name:"dataFormat",type:"string",defaultValue:"NHWC"},{tfName:"explicit_paddings",name:"explicitPaddings",type:"number[]",defaultValue:[]},{tfName:"dilations",name:"dilations",type:"number[]"}]},{tfOpName:"FusedDepthwiseConv2dNative",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"filter",type:"tensor"},{start:2,end:0,name:"args",type:"tensors"}],attrs:[{tfName:"num_args",name:"numArgs",type:"number"},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0},{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"data_format",name:"dataFormat",type:"string",defaultValue:"NHWC"},{tfName:"dilations",name:"dilations",type:"number[]",defaultValue:[1,1,1,1]},{tfName:"fused_ops",name:"fusedOps",type:"string[]",defaultValue:[]},{tfName:"explicit_paddings",name:"explicitPaddings",type:"number[]",defaultValue:[]}]},{tfOpName:"Conv3D",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"filter",type:"tensor"}],attrs:[{tfName:"strides",name:"strides",type:"number[]"},{tfName:"padding",name:"pad",type:"string"},{tfName:"data_format",name:"dataFormat",type:"string",defaultValue:"NHWC"},{tfName:"dilations",name:"dilations",type:"number[]"}]},{tfOpName:"Dilation2D",category:"convolution",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"filter",type:"tensor"}],attrs:[{tfName:"strides",name:"strides",type:"number[]"},{tfName:"rates",name:"dilations",type:"number[]"},{tfName:"padding",name:"pad",type:"string"}]}];var Z0={};We(Z0,{json:()=>hne});var hne=[{tfOpName:"Fill",category:"creation",inputs:[{start:0,name:"shape",type:"number[]"},{start:1,name:"value",type:"number"}],attrs:[{tfName:"T",name:"dtype",type:"dtype"}]},{tfOpName:"LinSpace",category:"creation",inputs:[{start:0,name:"start",type:"number"},{start:1,name:"stop",type:"number"},{start:2,name:"num",type:"number"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"OneHot",category:"creation",inputs:[{start:0,name:"indices",type:"tensor"},{start:1,name:"depth",type:"number"},{start:2,name:"onValue",type:"number",defaultValue:1},{start:3,name:"offValue",type:"number",defaultValue:0}],attrs:[{tfName:"axis",name:"axis",type:"number",notSupported:!0},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Ones",category:"creation",inputs:[{start:0,name:"shape",type:"number[]"}],attrs:[{tfName:"T",name:"dtype",type:"dtype"}]},{tfOpName:"OnesLike",category:"creation",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"dtype",name:"dtype",type:"dtype"}]},{tfOpName:"RandomUniform",category:"creation",inputs:[{start:0,name:"shape",type:"number[]"}],attrs:[{tfName:"minval",name:"minval",type:"number",defaultValue:0},{tfName:"maxval",name:"maxval",type:"number",defaultValue:1},{tfName:"dtype",name:"dtype",type:"dtype"},{tfName:"seed",name:"seed",type:"number",defaultValue:0},{tfName:"seed2",name:"seed2",type:"number",defaultValue:0,notSupported:!0},{tfName:"T",name:"T",type:"number",notSupported:!0}]},{tfOpName:"Range",category:"creation",inputs:[{start:0,name:"start",type:"number"},{start:1,name:"stop",type:"number"},{start:2,name:"step",type:"number",defaultValue:0}],attrs:[{tfName:"Tidx",name:"dtype",type:"dtype"}]},{tfOpName:"TruncatedNormal",category:"creation",inputs:[{start:0,name:"shape",type:"number[]"}],attrs:[{tfName:"means",name:"mean",type:"number",defaultValue:0},{tfName:"stddev",name:"stdDev",type:"number",defaultValue:1},{tfName:"seed",name:"seed",type:"number"},{tfName:"seed2",name:"seed2",type:"number",defaultValue:0,notSupported:!0},{tfName:"dtype",name:"dtype",type:"dtype"},{tfName:"T",name:"T",type:"number",notSupported:!0}]},{tfOpName:"Zeros",category:"creation",inputs:[{start:0,name:"shape",type:"number[]"}],attrs:[{tfName:"T",name:"dtype",type:"dtype"}]},{tfOpName:"ZerosLike",category:"creation",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype"}]},{tfOpName:"Multinomial",category:"creation",inputs:[{start:0,name:"logits",type:"tensor"},{start:1,name:"numSamples",type:"number"}],attrs:[{tfName:"seed",name:"seed",type:"number"},{tfName:"seed2",name:"seed2",type:"number"},{tfName:"T",name:"dtype",type:"dtype"},{tfName:"output_dtype",name:"output_dtype",type:"dtype"}]}];var J0={};We(J0,{json:()=>gne});var gne=[{tfOpName:"NonMaxSuppressionV2",category:"dynamic",inputs:[{start:0,name:"boxes",type:"tensor"},{start:1,name:"scores",type:"tensor"},{start:2,name:"maxOutputSize",type:"number"},{start:3,name:"iouThreshold",type:"number"}]},{tfOpName:"NonMaxSuppressionV3",category:"dynamic",inputs:[{start:0,name:"boxes",type:"tensor"},{start:1,name:"scores",type:"tensor"},{start:2,name:"maxOutputSize",type:"number"},{start:3,name:"iouThreshold",type:"number"},{start:4,name:"scoreThreshold",type:"number"}]},{tfOpName:"NonMaxSuppressionV4",category:"dynamic",inputs:[{start:0,name:"boxes",type:"tensor"},{start:1,name:"scores",type:"tensor"},{start:2,name:"maxOutputSize",type:"number"},{start:3,name:"iouThreshold",type:"number"},{start:4,name:"scoreThreshold",type:"number"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0},{tfName:"T_threshold",name:"threshold",type:"dtype",notSupported:!0},{tfName:"pad_to_max_output_size",name:"padToMaxOutputSize",type:"bool"}]},{tfOpName:"NonMaxSuppressionV5",category:"dynamic",inputs:[{start:0,name:"boxes",type:"tensor"},{start:1,name:"scores",type:"tensor"},{start:2,name:"maxOutputSize",type:"number"},{start:3,name:"iouThreshold",type:"number"},{start:4,name:"scoreThreshold",type:"number"},{start:5,name:"softNmsSigma",type:"number"}]},{tfOpName:"Where",category:"dynamic",inputs:[{start:0,name:"condition",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"ListDiff",category:"dynamic",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"y",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]}];var Q0={};We(Q0,{json:()=>xne});var xne=[{tfOpName:"TopKV2",category:"evaluation",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"k",type:"number"}],attrs:[{tfName:"sorted",name:"sorted",type:"bool"}]},{tfOpName:"Unique",category:"evaluation",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"UniqueV2",category:"evaluation",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number"}]}];var eI={};We(eI,{json:()=>yne});var yne=[{tfOpName:"PlaceholderWithDefault",category:"graph",inputs:[{start:0,name:"default",type:"tensor"}],attrs:[{tfName:"shape",name:"shape",type:"shape"},{tfName:"dtype",name:"dtype",type:"dtype"}]},{tfOpName:"Placeholder",category:"graph",attrs:[{tfName:"shape",name:"shape",type:"shape"},{tfName:"dtype",name:"dtype",type:"dtype"}]},{tfOpName:"Const",category:"graph"},{tfOpName:"Identity",category:"graph",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"IdentityN",category:"graph",inputs:[{start:0,end:0,name:"x",type:"tensors"}]},{tfOpName:"Snapshot",category:"graph",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"Rank",category:"graph",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"Size",category:"graph",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"Shape",category:"graph",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"ShapeN",category:"graph",inputs:[{start:0,end:0,name:"x",type:"tensors"}]},{tfOpName:"Print",category:"graph",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"data",type:"tensors"}],attrs:[{tfName:"message",name:"message",type:"string"},{tfName:"first_n",name:"firstN",type:"number",notSupported:!0},{tfName:"summarize",name:"summarize",type:"number",defaultValue:3}]},{tfOpName:"NoOp",category:"graph",inputs:[]},{tfOpName:"StopGradient",category:"graph",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"FakeQuantWithMinMaxVars",category:"graph",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"min",name:"min",type:"number"},{tfName:"max",name:"max",type:"number"}]}];var tI={};We(tI,{json:()=>bne});var bne=[{tfOpName:"HashTable",category:"hash_table",inputs:[],attrs:[{tfName:"shared_name",name:"sharedName",type:"string"},{tfName:"use_node_name_sharing",name:"useNodeNameSharing",type:"bool"},{tfName:"key_dtype",name:"keyDType",type:"dtype"},{tfName:"value_dtype",name:"valueDType",type:"dtype"}]},{tfOpName:"HashTableV2",category:"hash_table",inputs:[],attrs:[{tfName:"shared_name",name:"sharedName",type:"string"},{tfName:"use_node_name_sharing",name:"useNodeNameSharing",type:"bool"},{tfName:"key_dtype",name:"keyDType",type:"dtype"},{tfName:"value_dtype",name:"valueDType",type:"dtype"}]},{tfOpName:"LookupTableImport",category:"hash_table",inputs:[{start:0,name:"tableHandle",type:"tensor"},{start:1,name:"keys",type:"tensor"},{start:2,name:"values",type:"tensor"}],attrs:[{tfName:"Tin",name:"tIn",type:"dtype",notSupported:!0},{tfName:"Tout",name:"tOut",type:"dtype",notSupported:!0}]},{tfOpName:"LookupTableImportV2",category:"hash_table",inputs:[{start:0,name:"tableHandle",type:"tensor"},{start:1,name:"keys",type:"tensor"},{start:2,name:"values",type:"tensor"}],attrs:[{tfName:"Tin",name:"tIn",type:"dtype",notSupported:!0},{tfName:"Tout",name:"tOut",type:"dtype",notSupported:!0}]},{tfOpName:"LookupTableFind",category:"hash_table",inputs:[{start:0,name:"tableHandle",type:"tensor"},{start:1,name:"keys",type:"tensor"},{start:2,name:"defaultValue",type:"tensor"}],attrs:[{tfName:"Tin",name:"tIn",type:"dtype",notSupported:!0},{tfName:"Tout",name:"tOut",type:"dtype",notSupported:!0}]},{tfOpName:"LookupTableFindV2",category:"hash_table",inputs:[{start:0,name:"tableHandle",type:"tensor"},{start:1,name:"keys",type:"tensor"},{start:2,name:"defaultValue",type:"tensor"}],attrs:[{tfName:"Tin",name:"tIn",type:"dtype",notSupported:!0},{tfName:"Tout",name:"tOut",type:"dtype",notSupported:!0}]},{tfOpName:"LookupTableSize",category:"hash_table",inputs:[{start:0,name:"tableHandle",type:"tensor"}]},{tfOpName:"LookupTableSizeV2",category:"hash_table",inputs:[{start:0,name:"tableHandle",type:"tensor"}]}];var rI={};We(rI,{json:()=>wne});var wne=[{tfOpName:"ResizeBilinear",category:"image",inputs:[{start:0,name:"images",type:"tensor"},{start:1,name:"size",type:"number[]"}],attrs:[{tfName:"align_corners",name:"alignCorners",type:"bool"},{tfName:"half_pixel_centers",name:"halfPixelCenters",type:"bool"},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"ResizeNearestNeighbor",category:"image",inputs:[{start:0,name:"images",type:"tensor"},{start:1,name:"size",type:"number[]"}],attrs:[{tfName:"align_corners",name:"alignCorners",type:"bool"},{tfName:"half_pixel_centers",name:"halfPixelCenters",type:"bool"},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"CropAndResize",category:"image",inputs:[{start:0,name:"image",type:"tensor"},{start:1,name:"boxes",type:"tensor"},{start:2,name:"boxInd",type:"tensor"},{start:3,name:"cropSize",type:"number[]"}],attrs:[{tfName:"method",name:"method",type:"string"},{tfName:"extrapolation_value",name:"extrapolationValue",type:"number"}]}];var nI={};We(nI,{json:()=>_ne});var _ne=[{tfOpName:"Equal",category:"logical",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"NotEqual",category:"logical",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Greater",category:"logical",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"GreaterEqual",category:"logical",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Less",category:"logical",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"LessEqual",category:"logical",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"LogicalAnd",category:"logical",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"LogicalNot",category:"logical",inputs:[{start:0,name:"a",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"LogicalOr",category:"logical",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Select",category:"logical",inputs:[{start:0,name:"condition",type:"tensor"},{start:1,name:"a",type:"tensor"},{start:2,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"SelectV2",category:"logical",inputs:[{start:0,name:"condition",type:"tensor"},{start:1,name:"a",type:"tensor"},{start:2,name:"b",type:"tensor"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]}];var oI={};We(oI,{json:()=>kne});var kne=[{tfOpName:"_FusedMatMul",category:"matrices",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"},{start:2,end:0,name:"args",type:"tensors"}],attrs:[{tfName:"num_args",name:"numArgs",type:"number"},{tfName:"fused_ops",name:"fusedOps",type:"string[]",defaultValue:[]},{tfName:"epsilon",name:"epsilon",type:"number",defaultValue:1e-4},{tfName:"transpose_a",name:"transposeA",type:"bool",defaultValue:!1},{tfName:"transpose_b",name:"transposeB",type:"bool",defaultValue:!1},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"MatMul",category:"matrices",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"transpose_a",name:"transposeA",type:"bool",defaultValue:!1},{tfName:"transpose_b",name:"transposeB",type:"bool",defaultValue:!1},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"BatchMatMul",category:"matrices",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"adj_x",name:"transposeA",type:"bool",defaultValue:!1},{tfName:"adj_y",name:"transposeB",type:"bool",defaultValue:!1},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"BatchMatMulV2",category:"matrices",inputs:[{start:0,name:"a",type:"tensor"},{start:1,name:"b",type:"tensor"}],attrs:[{tfName:"adj_x",name:"transposeA",type:"bool",defaultValue:!1},{tfName:"adj_y",name:"transposeB",type:"bool",defaultValue:!1},{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Transpose",category:"matrices",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"perm",type:"number[]"}],attrs:[{tfName:"T",name:"dtype",type:"dtype",notSupported:!0}]},{tfOpName:"Einsum",category:"matrices",inputs:[{start:0,end:0,name:"tensors",type:"tensors"}],attrs:[{tfName:"equation",name:"equation",type:"string"},{tfName:"N",name:"n",type:"number",defaultValue:2},{tfName:"T",name:"dtype",type:"dtype"}]}];var sI={};We(sI,{json:()=>vne});var vne=[{tfOpName:"FusedBatchNorm",category:"normalization",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"scale",type:"tensor"},{start:2,name:"offset",type:"tensor"},{start:3,name:"mean",type:"tensor"},{start:4,name:"variance",type:"tensor"}],attrs:[{tfName:"epsilon",name:"epsilon",type:"number",defaultValue:.001},{tfName:"data_format",name:"dataFormat",type:"string",notSupported:!0}]},{tfOpName:"FusedBatchNormV2",category:"normalization",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"scale",type:"tensor"},{start:2,name:"offset",type:"tensor"},{start:3,name:"mean",type:"tensor"},{start:4,name:"variance",type:"tensor"}],attrs:[{tfName:"epsilon",name:"epsilon",type:"number",defaultValue:.001},{tfName:"data_format",name:"dataFormat",type:"string",notSupported:!0}]},{tfOpName:"FusedBatchNormV3",category:"normalization",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"scale",type:"tensor"},{start:2,name:"offset",type:"tensor"},{start:3,name:"mean",type:"tensor"},{start:4,name:"variance",type:"tensor"}],attrs:[{tfName:"epsilon",name:"epsilon",type:"number",defaultValue:.001},{tfName:"data_format",name:"dataFormat",type:"string",notSupported:!0}]},{tfOpName:"LRN",category:"normalization",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"depth_radius",name:"radius",type:"number",defaultValue:5},{tfName:"bias",name:"bias",type:"number",defaultValue:1},{tfName:"alpha",name:"alpha",type:"number",defaultValue:1},{tfName:"beta",name:"beta",type:"number",defaultValue:.5}]},{tfOpName:"Softmax",category:"normalization",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"LogSoftmax",category:"normalization",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"SparseToDense",category:"normalization",inputs:[{start:0,name:"sparseIndices",type:"tensor"},{start:1,name:"outputShape",type:"number[]"},{start:2,name:"sparseValues",type:"tensor"},{start:3,name:"defaultValue",type:"tensor"}],attrs:[{tfName:"validate_indices",name:"validateIndices",type:"bool",defaultValue:!0,notSupported:!0}]}];var iI={};We(iI,{json:()=>Cne});var Cne=[{tfOpName:"Bincount",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"size",type:"number"},{start:2,name:"weights",type:"tensor"}]},{tfOpName:"DenseBincount",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"size",type:"number"},{start:2,name:"weights",type:"tensor"}],attrs:[{tfName:"binary_output",name:"binaryOutput",type:"bool"}]},{tfOpName:"Max",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number[]"}],attrs:[{tfName:"keep_dims",name:"keepDims",type:"bool"}]},{tfOpName:"Mean",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number[]"}],attrs:[{tfName:"keep_dims",name:"keepDims",type:"bool"}]},{tfOpName:"Min",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number[]"}],attrs:[{tfName:"keep_dims",name:"keepDims",type:"bool"}]},{tfOpName:"Sum",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number[]"}],attrs:[{tfName:"keep_dims",name:"keepDims",type:"bool"}]},{tfOpName:"All",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number[]"}],attrs:[{tfName:"keep_dims",name:"keepDims",type:"bool"}]},{tfOpName:"Any",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number[]"}],attrs:[{tfName:"keep_dims",name:"keepDims",type:"bool"}]},{tfOpName:"ArgMax",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number"}]},{tfOpName:"ArgMin",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number"}]},{tfOpName:"Prod",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number[]"}],attrs:[{tfName:"keep_dims",name:"keepDims",type:"bool"}]},{tfOpName:"Cumsum",category:"reduction",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number"}],attrs:[{tfName:"exclusive",name:"exclusive",type:"bool"},{tfName:"reverse",name:"reverse",type:"bool"}]}];var aI={};We(aI,{json:()=>Ine});var Ine=[{tfOpName:"ConcatV2",category:"slice_join",inputs:[{start:0,end:-1,name:"tensors",type:"tensors"},{start:-1,name:"axis",type:"number"}],attrs:[{tfName:"N",name:"n",type:"number",defaultValue:2}]},{tfOpName:"Concat",category:"slice_join",inputs:[{start:1,end:0,name:"tensors",type:"tensors"},{start:0,name:"axis",type:"number"}],attrs:[{tfName:"N",name:"n",type:"number",defaultValue:2}]},{tfOpName:"GatherV2",category:"slice_join",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"indices",type:"tensor"},{start:2,name:"axis",type:"number",defaultValue:0}],attrs:[{tfName:"batch_dims",name:"batchDims",type:"number",defaultValue:0}]},{tfOpName:"Gather",category:"slice_join",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"indices",type:"tensor"}],attrs:[{tfName:"validate_indices",name:"validateIndices",type:"bool",notSupported:!0}]},{tfOpName:"Reverse",category:"slice_join",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"dims",type:"bool[]"}]},{tfOpName:"ReverseV2",category:"slice_join",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number[]"}]},{tfOpName:"Slice",category:"slice_join",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"begin",type:"number[]"},{start:2,name:"size",type:"number[]"}]},{tfOpName:"StridedSlice",category:"slice_join",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"begin",type:"number[]"},{start:2,name:"end",type:"number[]"},{start:3,name:"strides",type:"number[]"}],attrs:[{tfName:"begin_mask",name:"beginMask",type:"number",defaultValue:0},{tfName:"end_mask",name:"endMask",type:"number",defaultValue:0},{tfName:"new_axis_mask",name:"newAxisMask",type:"number",defaultValue:0},{tfName:"ellipsis_mask",name:"ellipsisMask",type:"number",defaultValue:0},{tfName:"shrink_axis_mask",name:"shrinkAxisMask",type:"number",defaultValue:0}]},{tfOpName:"Pack",category:"slice_join",inputs:[{start:0,end:0,name:"tensors",type:"tensors"}],attrs:[{tfName:"axis",name:"axis",type:"number",defaultValue:0}]},{tfOpName:"Unpack",category:"slice_join",inputs:[{start:0,name:"tensor",type:"tensor"}],attrs:[{tfName:"axis",name:"axis",type:"number",defaultValue:0},{tfName:"num",name:"num",type:"number",defaultValue:0,notSupported:!0}]},{tfOpName:"Tile",category:"slice_join",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"reps",type:"number[]"}]},{tfOpName:"Split",category:"slice_join",inputs:[{start:0,name:"axis",type:"number",defaultValue:0},{start:1,name:"x",type:"tensor"}],attrs:[{tfName:"num_split",name:"numOrSizeSplits",type:"number",defaultValue:1}]},{tfOpName:"SplitV",category:"slice_join",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"numOrSizeSplits",type:"number[]"},{start:2,name:"axis",type:"number",defaultValue:0}]},{tfOpName:"ScatterNd",category:"slice_join",inputs:[{start:0,name:"indices",type:"tensor"},{start:1,name:"values",type:"tensor"},{start:2,name:"shape",type:"number[]"}]},{tfOpName:"GatherNd",category:"slice_join",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"indices",type:"tensor"}]},{tfOpName:"SparseToDense",category:"slice_join",inputs:[{start:0,name:"sparseIndices",type:"tensor"},{start:1,name:"outputShape",type:"number[]"},{start:2,name:"sparseValues",type:"tensor"},{start:3,name:"defaultValue",type:"tensor"}],attrs:[{tfName:"validate_indices",name:"validateIndices",type:"bool",defaultValue:!1,notSupported:!0}]}];var lI={};We(lI,{json:()=>Nne});var Nne=[{tfOpName:"FFT",category:"spectral",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"IFFT",category:"spectral",inputs:[{start:0,name:"x",type:"tensor"}]},{tfOpName:"RFFT",category:"spectral",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"fft_length",type:"number",notSupported:!0}]},{tfOpName:"IRFFT",category:"spectral",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"fft_length",type:"number",notSupported:!0}]}];var uI={};We(uI,{json:()=>Sne});var Sne=[{tfOpName:"Cast",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"SrcT",name:"sdtype",type:"dtype",notSupported:!0},{tfName:"DstT",name:"dtype",type:"dtype"}]},{tfOpName:"ExpandDims",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"axis",type:"number"}]},{tfOpName:"MirrorPad",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"padding",type:"number[]"}],attrs:[{tfName:"mode",name:"mode",type:"string"}]},{tfOpName:"Pad",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"padding",type:"number[]"}],attrs:[{tfName:"constant_value",name:"constantValue",type:"number",defaultValue:0}]},{tfOpName:"PadV2",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"padding",type:"number[]"},{start:2,name:"constantValue",type:"number",defaultValue:0}]},{tfOpName:"Reshape",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"shape",type:"number[]"}]},{tfOpName:"Squeeze",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"axis",tfDeprecatedName:"squeeze_dims",name:"axis",type:"number[]"}]},{tfOpName:"SpaceToBatchND",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"blockShape",type:"number[]"},{start:2,name:"paddings",type:"number[]"}]},{tfOpName:"BatchToSpaceND",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"blockShape",type:"number[]"},{start:2,name:"crops",type:"number[]"}]},{tfOpName:"DepthToSpace",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"}],attrs:[{tfName:"block_size",name:"blockSize",type:"number"},{tfName:"data_format",name:"dataFormat",type:"string"}]},{tfOpName:"BroadcastTo",category:"transformation",inputs:[{start:0,name:"x",type:"tensor"},{start:1,name:"shape",type:"number[]"}],attrs:[]}];var py=class{static get Instance(){return this._instance||(this._instance=new this)}constructor(){let e=[q0,K0,X0,Y0,Z0,J0,Q0,nI,rI,eI,oI,sI,iI,aI,lI,uI,tI],t=[].concat(...e.map(n=>n.json));this.opMappers=t.reduce((n,o)=>(n[o.tfOpName]=o,n),{})}transformGraph(e,t={}){let n=e.node,o=[],s=[],a=[],i=n.reduce((h,g)=>(h[g.name]=this.mapNode(g),g.op.startsWith("Placeholder")?o.push(h[g.name]):g.op==="Const"?s.push(h[g.name]):(g.input==null||g.input.length===0)&&a.push(h[g.name]),h),{}),l=[],u=[],c={},p={};t!=null&&(c=this.mapSignatureEntries(t.inputs),p=this.mapSignatureEntries(t.outputs));let m=Object.keys(i);m.forEach(h=>{let g=i[h];g.inputNames.forEach(x=>{let[b]=Fs(x);g.inputs.push(i[b]),i[b].children.push(g)})}),Object.keys(p).length===0?m.forEach(h=>{let g=i[h];g.children.length===0&&u.push(g)}):Object.keys(p).forEach(h=>{let[g]=Fs(h),x=i[g];x!=null&&(x.signatureKey=p[h],u.push(x))}),Object.keys(c).length>0?Object.keys(c).forEach(h=>{let[g]=Fs(h),x=i[g];x&&(x.signatureKey=c[h],l.push(x))}):l=o;let f={};e.library!=null&&e.library.function!=null&&(f=e.library.function.reduce((h,g)=>(h[g.signature.name]=this.mapFunction(g),h),{}));let d={nodes:i,inputs:l,outputs:u,weights:s,placeholders:o,signature:t,functions:f};return a.length>0&&(d.initNodes=a),d}mapSignatureEntries(e){return Object.keys(e||{}).reduce((t,n)=>(t[e[n].name]=n,t),{})}mapNode(e){let t=uy(e.op)||this.opMappers[e.op]||{};e.attr==null&&(e.attr={});let n={name:e.name,op:e.op,category:t.category,inputNames:(e.input||[]).map(o=>o.startsWith("^")?o.substr(1):o),inputs:[],children:[],inputParams:{},attrParams:{},rawAttrs:e.attr};return t.inputs!=null&&(n.inputParams=t.inputs.reduce((o,s)=>(o[s.name]={type:s.type,inputIndexStart:s.start,inputIndexEnd:s.end},o),{})),t.attrs!=null&&(n.attrParams=t.attrs.reduce((o,s)=>{let a=s.type,i;switch(s.type){case"string":i=my(e.attr,s.tfName,s.defaultValue),i===void 0&&!!s.tfDeprecatedName&&(i=my(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"string[]":i=by(e.attr,s.tfName,s.defaultValue),i===void 0&&!!s.tfDeprecatedName&&(i=by(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"number":i=dy(e.attr,s.tfName,s.defaultValue||0),i===void 0&&!!s.tfDeprecatedName&&(i=dy(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"number[]":i=yy(e.attr,s.tfName,s.defaultValue),i===void 0&&!!s.tfDeprecatedName&&(i=yy(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"bool":i=fy(e.attr,s.tfName,s.defaultValue),i===void 0&&!!s.tfDeprecatedName&&(i=fy(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"bool[]":i=_y(e.attr,s.tfName,s.defaultValue),i===void 0&&!!s.tfDeprecatedName&&(i=_y(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"shape":i=xy(e.attr,s.tfName,s.defaultValue),i===void 0&&!!s.tfDeprecatedName&&(i=xy(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"shape[]":i=wy(e.attr,s.tfName,s.defaultValue),i===void 0&&!!s.tfDeprecatedName&&(i=wy(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"dtype":i=hy(e.attr,s.tfName,s.defaultValue),i===void 0&&!!s.tfDeprecatedName&&(i=hy(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"dtype[]":i=gy(e.attr,s.tfName,s.defaultValue),i===void 0&&!!s.tfDeprecatedName&&(i=gy(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"func":i=uB(e.attr,s.tfName,s.defaultValue),i===void 0&&!!s.tfDeprecatedName&&(i=uB(e.attr,s.tfDeprecatedName,s.defaultValue));break;case"tensor":case"tensors":break;default:throw new Error(`Unsupported param type: ${s.type} for op: ${e.op}`)}return o[s.name]={value:i,type:a},o},{})),n}mapFunction(e){let t=e.nodeDef,n=[],o=[],s={};t!=null&&(s=t.reduce((p,m)=>(p[m.name]=this.mapNode(m),m.op==="Const"&&o.push(p[m.name]),p),{}));let a=[],i=[];e.signature.inputArg.forEach(p=>{let[m]=Fs(p.name),f={name:m,op:"Placeholder",inputs:[],inputNames:[],category:"graph",inputParams:{},attrParams:{dtype:{value:cI(p.type),type:"dtype"}},children:[]};f.signatureKey=p.name,a.push(f),s[m]=f}),Object.keys(s).forEach(p=>{let m=s[p];m.inputNames.forEach(f=>{let[d]=Fs(f);m.inputs.push(s[d]),s[d].children.push(m)})});let u=e.ret;e.signature.outputArg.forEach(p=>{let[m,f]=Fs(u[p.name]),d=s[m];d!=null&&(d.defaultOutput=f,i.push(d))});let c=this.mapArgsToSignature(e);return{nodes:s,inputs:a,outputs:i,weights:o,placeholders:n,signature:c}}mapArgsToSignature(e){return{methodName:e.signature.name,inputs:e.signature.inputArg.reduce((t,n)=>(t[n.name]=this.mapArgToTensorInfo(n),t),{}),outputs:e.signature.outputArg.reduce((t,n)=>(t[n.name]=this.mapArgToTensorInfo(n,e.ret),t),{})}}mapArgToTensorInfo(e,t){let n=e.name;return t!=null&&(n=t[n]),{name:n,dtype:e.type}}};function Tne(r){let e=W().global;if(typeof e.atob!="undefined")return e.atob(r);if(typeof Buffer!="undefined")return new Buffer(r,"base64").toString();throw new Error("Unable to decode base64 in this environment. Missing built-in atob() or Buffer()")}function lB(r,e){let t=Array.isArray(r)?String.fromCharCode.apply(null,r):Tne(r);return e?t:t.toLowerCase()}function my(r,e,t,n=!1){let o=r[e];return o!=null?lB(o.s,n):t}function fy(r,e,t){let n=r[e];return n?n.b:t}function dy(r,e,t){let n=r[e]||{},o=n.i!=null?n.i:n.f!=null?n.f:t;return typeof o=="number"?o:parseInt(o,10)}function cI(r){switch(typeof r=="string"&&(r=yo[r]),r){case yo.DT_FLOAT:return"float32";case yo.DT_INT32:case yo.DT_INT64:case yo.DT_INT8:case yo.DT_UINT8:return"int32";case yo.DT_BOOL:return"bool";case yo.DT_DOUBLE:return"float32";case yo.DT_STRING:return"string";default:return null}}function uB(r,e,t){let n=r[e];return n&&n.func?n.func.name:t}function hy(r,e,t){let n=r[e];return n&&n.type?cI(n.type):t}function gy(r,e,t){let n=r[e];return n&&n.list&&n.list.type?n.list.type.map(o=>cI(o)):t}function cB(r){if(!r.unknownRank)return r.dim!=null?r.dim.map(e=>typeof e.size=="number"?e.size:parseInt(e.size,10)):[]}function xy(r,e,t){let n=r[e];return n&&n.shape?cB(n.shape):t}function yy(r,e,t){let n=r[e];return n?((n.list.f&&n.list.f.length?n.list.f:n.list.i)||[]).map(o=>typeof o=="number"?o:parseInt(o,10)):t}function by(r,e,t,n=!1){let o=r[e];return o&&o.list&&o.list.s?o.list.s.map(s=>lB(s,n)):t}function wy(r,e,t){let n=r[e];return n&&n.list&&n.list.shape?n.list.shape.map(o=>cB(o)):t}function _y(r,e,t){let n=r[e];return n&&n.list&&n.list.b?n.list.b:t}var pI=class{constructor(e,t,n){this.node=e,this.tensorMap=t,this.context=n,this.inputs=[],this.attrs={},this.inputs=e.inputNames.map(o=>this.getInput(o)),e.rawAttrs!=null&&(this.attrs=Object.keys(e.rawAttrs).reduce((o,s)=>(o[s]=this.getAttr(s),o),{}))}getInput(e){return hr(e,this.tensorMap,this.context)}getAttr(e,t){let n=this.node.rawAttrs[e];if(n.tensor!=null)return hr(e,this.tensorMap,this.context);if(n.i!=null||n.f!=null)return dy(this.node.rawAttrs,e,t);if(n.s!=null)return my(this.node.rawAttrs,e,t);if(n.b!=null)return fy(this.node.rawAttrs,e,t);if(n.shape!=null)return xy(this.node.rawAttrs,e,t);if(n.type!=null)return hy(this.node.rawAttrs,e,t);if(n.list!=null){if(n.list.i!=null||n.list.f!=null)return yy(this.node.rawAttrs,e,t);if(n.list.s!=null)return by(this.node.rawAttrs,e,t);if(n.list.shape!=null)return wy(this.node.rawAttrs,e,t);if(n.list.b!=null)return _y(this.node.rawAttrs,e,t);if(n.list.type!=null)return gy(this.node.rawAttrs,e,t)}return t}};var pB=(r,e,t)=>{switch(r.op){case"BiasAdd":case"AddV2":case"Add":return[J(v("a",r,e,t),v("b",r,e,t))];case"AddN":return[Gw(v("tensors",r,e,t))];case"FloorMod":case"Mod":return[mf(v("a",r,e,t),v("b",r,e,t))];case"Mul":return[O(v("a",r,e,t),v("b",r,e,t))];case"RealDiv":case"Div":return[pe(v("a",r,e,t),v("b",r,e,t))];case"DivNoNan":return[rf(v("a",r,e,t),v("b",r,e,t))];case"FloorDiv":return[eu(v("a",r,e,t),v("b",r,e,t))];case"Sub":return[le(v("a",r,e,t),v("b",r,e,t))];case"Minimum":return[_s(v("a",r,e,t),v("b",r,e,t))];case"Maximum":return[en(v("a",r,e,t),v("b",r,e,t))];case"Pow":return[zr(v("a",r,e,t),v("b",r,e,t))];case"SquaredDifference":return[ku(v("a",r,e,t),v("b",r,e,t))];default:throw TypeError(`Node type ${r.op} is not implemented`)}};var mB=(r,e,t)=>{switch(r.op){case"Abs":case"ComplexAbs":return[Tt(v("x",r,e,t))];case"Acos":return[Gm(v("x",r,e,t))];case"Acosh":return[Wm(v("x",r,e,t))];case"Asin":return[Um(v("x",r,e,t))];case"Asinh":return[Hm(v("x",r,e,t))];case"Atan":return[qm(v("x",r,e,t))];case"Atan2":return[Km(v("x",r,e,t),v("y",r,e,t))];case"Atanh":return[Xm(v("x",r,e,t))];case"Ceil":return[Jm(v("x",r,e,t))];case"Complex":return[Nn(v("real",r,e,t),v("imag",r,e,t))];case"Cos":return[ya(v("x",r,e,t))];case"Cosh":return[iu(v("x",r,e,t))];case"Elu":return[ys(v("x",r,e,t))];case"Erf":return[nf(v("x",r,e,t))];case"Exp":return[Qt(v("x",r,e,t))];case"Expm1":return[of(v("x",r,e,t))];case"Floor":return[ws(v("x",r,e,t))];case"Log":return[pr(v("x",r,e,t))];case"Log1p":return[cu(v("x",r,e,t))];case"Imag":return[lu(v("x",r,e,t))];case"Neg":return[qe(v("x",r,e,t))];case"Reciprocal":return[ff(v("x",r,e,t))];case"Real":return[gl(v("x",r,e,t))];case"Relu":return[$r(v("x",r,e,t))];case"Round":return[gu(v("x",r,e,t))];case"Selu":return[yu(v("x",r,e,t))];case"Sigmoid":return[Dr(v("x",r,e,t))];case"Sin":return[bu(v("x",r,e,t))];case"Sign":return[hf(v("x",r,e,t))];case"Sinh":return[wu(v("x",r,e,t))];case"Softplus":return[eo(v("x",r,e,t))];case"Sqrt":return[gt(v("x",r,e,t))];case"Square":return[Me(v("x",r,e,t))];case"Tanh":return[gs(v("x",r,e,t))];case"Tan":return[bf(v("x",r,e,t))];case"ClipByValue":return[ur(v("x",r,e,t),v("clipValueMin",r,e,t),v("clipValueMax",r,e,t))];case"Relu6":return[hu(v("x",r,e,t))];case"Rsqrt":return[xu(hr(r.inputNames[0],e,t))];case"Prod":return[fu(v("x",r,e,t),v("axes",r,e,t))];case"LeakyRelu":return[ba(v("x",r,e,t),v("alpha",r,e,t))];case"Prelu":return[va(v("x",r,e,t),v("alpha",r,e,t))];case"IsNan":return[sf(hr(r.inputNames[0],e,t))];default:throw TypeError(`Node type ${r.op} is not implemented`)}};function Rn(r,e,t=""){if(!(typeof r=="number"||typeof e=="number")){y.assert(r.length===e.length,()=>t+` Shapes ${r} and ${e} must match`);for(let n=0;n<r.length;n++){let o=r[n],s=e[n];y.assert(o<0||s<0||o===s,()=>t+` Shapes ${r} and ${e} must match`)}}}function fB(r){return!(typeof r=="number"||r.some(e=>e<0))}function am(r,e,t){let n=ky(r,t),o=!fB(n);if(o&&e.length===0)throw new Error(`Tried to calculate elements of an empty list with non-fully-defined elementShape: ${n}`);if(o&&e.forEach(s=>{n=ky(s.shape,n)}),!fB(n))throw new Error(`Non-fully-defined elementShape: ${n}`);return n}function ky(r,e){if(typeof r=="number")return e;if(typeof e=="number")return r;if(r.length!==e.length)throw new Error(`Incompatible ranks during merge: ${r} vs. ${e}`);let t=[];for(let n=0;n<r.length;++n){let o=r[n],s=e[n];if(o>=0&&s>=0&&o!==s)throw new Error(`Incompatible shape during merge: ${r} vs. ${e}`);t[n]=o>=0?o:s}return t}var mI=class{constructor(e,t,n,o,s,a,i){this.name=e,this.dtype=t,this.maxSize=n,this.elementShape=o,this.identicalElementShapes=s,this.dynamicSize=a,this.clearAfterRead=i,this.tensors=[],this.closed_=!1,this.idTensor=ue(0),$t(this.idTensor)}get id(){return this.idTensor.id}get closed(){return this.closed_}clearAndClose(e){this.tensors.forEach(t=>{(e==null||!e.has(t.tensor.id))&&t.tensor.dispose()}),this.tensors=[],this.closed_=!0,this.idTensor.dispose()}size(){return this.tensors.length}read(e){if(this.closed_)throw new Error(`TensorArray ${this.name} has already been closed.`);if(e<0||e>=this.size())throw new Error(`Tried to read from index ${e}, but array size is: ${this.size()}`);let t=this.tensors[e];if(t.cleared)throw new Error(`TensorArray ${this.name}: Could not read index ${e} twice because it was cleared after a previous read (perhaps try setting clear_after_read = false?).`);return this.clearAfterRead&&(t.cleared=!0),t.read=!0,t.tensor}readMany(e){return e.map(t=>this.read(t))}write(e,t){if(this.closed_)throw new Error(`TensorArray ${this.name} has already been closed.`);if(e<0||!this.dynamicSize&&e>=this.maxSize)throw new Error(`Tried to write to index ${e}, but array is not resizeable and size is: ${this.maxSize}`);let n=this.tensors[e]||{};if(t.dtype!==this.dtype)throw new Error(`TensorArray ${this.name}: Could not write to TensorArray index ${e},
|
|
because the value dtype is ${t.dtype}, but TensorArray dtype is ${this.dtype}.`);if(this.size()===0&&(this.elementShape==null||this.elementShape.length===0)&&(this.elementShape=t.shape),Rn(this.elementShape,t.shape,`TensorArray ${this.name}: Could not write to TensorArray index ${e}.`),n.read)throw new Error(`TensorArray ${this.name}: Could not write to TensorArray index ${e}, because it has already been read.`);if(n.written)throw new Error(`TensorArray ${this.name}: Could not write to TensorArray index ${e}, because it has already been written.`);n.tensor=t,$t(t),n.written=!0,this.tensors[e]=n}writeMany(e,t){if(e.length!==t.length)throw new Error(`TensorArray ${this.name}: could not write multiple tensors,because the index size: ${e.length} is not the same as tensors size: ${t.length}.`);e.forEach((n,o)=>this.write(n,t[o]))}gather(e,t){if(!!t&&t!==this.dtype)throw new Error(`TensorArray dtype is ${this.dtype} but gather requested dtype ${t}`);if(e)e=e.slice(0,this.size());else{e=[];for(let o=0;o<this.size();o++)e.push(o)}if(e.length===0)return vr([],[0].concat(this.elementShape));let n=this.readMany(e);return Rn(this.elementShape,n[0].shape,"TensorArray shape mismatch: "),Gt(n,0)}concat(e){if(!!e&&e!==this.dtype)throw new Error(`TensorArray dtype is ${this.dtype} but concat requested dtype ${e}`);if(this.size()===0)return vr([],[0].concat(this.elementShape));let t=[];for(let o=0;o<this.size();o++)t.push(o);let n=this.readMany(t);return Rn(this.elementShape,n[0].shape,`TensorArray shape mismatch: tensor array shape (${this.elementShape}) vs first tensor shape (${n[0].shape})`),Qe(n,0)}scatter(e,t){if(t.dtype!==this.dtype)throw new Error(`TensorArray dtype is ${this.dtype} but tensor has dtype ${t.dtype}`);if(e.length!==t.shape[0])throw new Error(`Expected len(indices) == tensor.shape[0], but saw: ${e.length} vs. ${t.shape[0]}`);let n=Math.max(...e);if(!this.dynamicSize&&n>=this.maxSize)throw new Error(`Max index must be < array size (${n} vs. ${this.maxSize})`);this.writeMany(e,fr(t,0))}split(e,t){if(t.dtype!==this.dtype)throw new Error(`TensorArray dtype is ${this.dtype} but tensor has dtype ${t.dtype}`);let n=0,o=e.map(l=>(n+=l,n));if(n!==t.shape[0])throw new Error(`Expected sum of lengths to be equal to
|
|
tensor.shape[0], but sum of lengths is
|
|
${n}, and tensor's shape is: ${t.shape}`);if(!this.dynamicSize&&e.length!==this.maxSize)throw new Error(`TensorArray's size is not equal to the size of lengths (${this.maxSize} vs. ${e.length}), and the TensorArray is not marked as dynamically resizeable`);let s=n===0?0:t.size/n,a=[];V(()=>{t=L(t,[1,n,s]);for(let l=0;l<e.length;++l){let u=l===0?0:o[l-1],c=[0,u,0],p=[1,e[l],s];a[l]=L(Re(t,c,p),this.elementShape)}return a});let i=[];for(let l=0;l<e.length;l++)i[l]=l;this.writeMany(i,a)}};var rc=class{constructor(e,t,n,o=-1){this.tensors=e,this.elementShape=t,this.elementDtype=n,e!=null&&e.forEach(s=>{if(n!==s.dtype)throw new Error(`Invalid data types; op elements ${n}, but list elements ${s.dtype}`);Rn(t,s.shape,"TensorList shape mismatch: "),$t(s)}),this.idTensor=ue(0),this.maxNumElements=o,$t(this.idTensor)}get id(){return this.idTensor.id}copy(){return new rc([...this.tensors],this.elementShape,this.elementDtype)}clearAndClose(e){this.tensors.forEach(t=>{(e==null||!e.has(t.id))&&t.dispose()}),this.tensors.length=0,this.idTensor.dispose()}size(){return this.tensors.length}stack(e,t,n=-1){if(t!==this.elementDtype)throw new Error(`Invalid data types; op elements ${t}, but list elements ${this.elementDtype}`);if(n!==-1&&this.tensors.length!==n)throw new Error(`Operation expected a list with ${n} elements but got a list with ${this.tensors.length} elements.`);Rn(e,this.elementShape,"TensorList shape mismatch: ");let o=am(this.elementShape,this.tensors,e);return V(()=>{let s=this.tensors.map(a=>L(a,o));return Gt(s,0)})}popBack(e,t){if(t!==this.elementDtype)throw new Error(`Invalid data types; op elements ${t}, but list elements ${this.elementDtype}`);if(this.size()===0)throw new Error("Trying to pop from an empty list.");let n=am(this.elementShape,this.tensors,e),o=this.tensors.pop();return Rn(o.shape,e,"TensorList shape mismatch: "),L(o,n)}pushBack(e){if(e.dtype!==this.elementDtype)throw new Error(`Invalid data types; op elements ${e.dtype}, but list elements ${this.elementDtype}`);if(Rn(e.shape,this.elementShape,"TensorList shape mismatch: "),this.maxNumElements===this.size())throw new Error("Trying to push element into a full list.");$t(e),this.tensors.push(e)}resize(e){if(e<0)throw new Error(`TensorListResize expects size to be non-negative. Got: ${e}`);if(this.maxNumElements!==-1&&e>this.maxNumElements)throw new Error(`TensorListResize input size ${e} is greater maxNumElement ${this.maxNumElements}.`);this.tensors.length=e}getItem(e,t,n){if(n!==this.elementDtype)throw new Error(`Invalid data types; op elements ${n}, but list elements ${this.elementDtype}`);if(e<0||e>this.tensors.length)throw new Error(`Trying to access element ${e} in a list with ${this.tensors.length} elements.`);if(this.tensors[e]==null)throw new Error(`element at index ${e} is null.`);Rn(this.tensors[e].shape,t,"TensorList shape mismatch: ");let o=am(this.elementShape,this.tensors,t);return L(this.tensors[e],o)}setItem(e,t){if(t.dtype!==this.elementDtype)throw new Error(`Invalid data types; op elements ${t.dtype}, but list elements ${this.elementDtype}`);if(e<0||this.maxNumElements!==-1&&e>=this.maxNumElements)throw new Error(`Trying to set element ${e} in a list with max ${this.maxNumElements} elements.`);Rn(this.elementShape,t.shape,"TensorList shape mismatch: "),$t(t),this.tensors[e]=t}gather(e,t,n){if(t!==this.elementDtype)throw new Error(`Invalid data types; op elements ${t}, but list elements ${this.elementDtype}`);Rn(this.elementShape,n,"TensorList shape mismatch: "),e=e.slice(0,this.size());let o=am(this.elementShape,this.tensors,n);return e.length===0?vr([],[0].concat(o)):V(()=>{let s=e.map(a=>L(this.tensors[a],o));return Gt(s,0)})}concat(e,t){if(!!e&&e!==this.elementDtype)throw new Error(`TensorList dtype is ${this.elementDtype} but concat requested dtype ${e}`);Rn(this.elementShape,t,"TensorList shape mismatch: ");let n=am(this.elementShape,this.tensors,t);return this.size()===0?vr([],[0].concat(n)):V(()=>{let o=this.tensors.map(s=>L(s,n));return Qe(o,0)})}};function dB(r,e,t){let n=r.dtype;if(r.shape.length<1)throw new Error(`Tensor must be at least a vector, but saw shape: ${r.shape}`);if(r.dtype!==t)throw new Error(`Invalid data types; op elements ${r.dtype}, but list elements ${t}`);let o=r.shape.slice(1);Rn(o,e,"TensorList shape mismatch: ");let s=fr(r);return new rc(s,e,n)}function hB(r,e,t){return new rc([],r,e,t)}function gB(r,e,t,n){if(e.length!==r.shape[0])throw new Error(`Expected len(indices) == tensor.shape[0], but saw: ${e.length} vs. ${r.shape[0]}`);let o=Math.max(...e);if(n!=null&&n!==-1&&o>=n)throw new Error(`Max index must be < array size (${o} vs. ${n})`);let s=new rc([],t,r.dtype,n),a=fr(r,0);return e.forEach((i,l)=>{s.setItem(i,a[l])}),s}function xB(r,e,t){let n=0,o=e.map(c=>(n+=c,n));if(n!==r.shape[0])throw new Error(`Expected sum of lengths to be equal to
|
|
tensor.shape[0], but sum of lengths is
|
|
${n}, and tensor's shape is: ${r.shape}`);let s=r.shape.slice(1),a=ky(s,t),i=n===0?0:r.size/n,l=V(()=>{let c=[];r=L(r,[1,n,i]);for(let p=0;p<e.length;++p){let m=p===0?0:o[p-1],f=[0,m,0],d=[1,e[p],i];c[p]=L(Re(r,f,d),a)}return r.dispose(),c}),u=new rc([],t,r.dtype,e.length);for(let c=0;c<l.length;c++)u.setItem(c,l[c]);return u}var yB=async(r,e,t)=>{switch(r.op){case"If":case"StatelessIf":{let n=v("thenBranch",r,e,t),o=v("elseBranch",r,e,t),s=v("cond",r,e,t),a=v("args",r,e,t);return(await s.data())[0]?t.functionMap[n].executeFunctionAsync(a,t.tensorArrayMap,t.tensorListMap):t.functionMap[o].executeFunctionAsync(a,t.tensorArrayMap,t.tensorListMap)}case"While":case"StatelessWhile":{let n=v("body",r,e,t),o=v("cond",r,e,t),s=v("args",r,e,t),a=await t.functionMap[o].executeFunctionAsync(s,t.tensorArrayMap,t.tensorListMap),i=s.map(c=>c.id),l=await a[0].data();a.forEach(c=>{!c.kept&&i.indexOf(c.id)===-1&&c.dispose()});let u=s;for(;l[0];){let c=u;u=await t.functionMap[n].executeFunctionAsync(u,t.tensorArrayMap,t.tensorListMap);let p=u.map(f=>f.id);c.forEach(f=>{!f.kept&&i.indexOf(f.id)===-1&&p.indexOf(f.id)===-1&&f.dispose()});let m=await t.functionMap[o].executeFunctionAsync(u,t.tensorArrayMap,t.tensorListMap);l=await m[0].data(),m.forEach(f=>{!f.kept&&i.indexOf(f.id)===-1&&p.indexOf(f.id)===-1&&f.dispose()})}return u}case"LoopCond":{let n=v("pred",r,e,t);return[Os(n)]}case"Switch":{let n=v("pred",r,e,t),o=v("data",r,e,t);return o.kept||(o=Os(o)),(await n.data())[0]?[void 0,o]:[o,void 0]}case"Merge":{let n=r.inputNames.find(o=>hr(o,e,t)!==void 0);if(n){let o=hr(n,e,t);return[Os(o)]}return}case"Enter":{let n=v("frameName",r,e,t),o=v("tensor",r,e,t);return t.enterFrame(n),[Os(o)]}case"Exit":{let n=v("tensor",r,e,t);return t.exitFrame(),[Os(n)]}case"NextIteration":{let n=v("tensor",r,e,t);return t.nextIteration(),[Os(n)]}case"TensorArrayV3":{let n=v("size",r,e,t),o=v("dtype",r,e,t),s=v("elementShape",r,e,t),a=v("dynamicSize",r,e,t),i=v("clearAfterRead",r,e,t),l=v("identicalElementShapes",r,e,t),u=v("name",r,e,t),c=new mI(u,o,n,s,l,a,i);return t.addTensorArray(c),[c.idTensor,ue(1)]}case"TensorArrayWriteV3":{let n=v("tensorArrayId",r,e,t),o=v("index",r,e,t),s=v("tensor",r,e,t),a=t.getTensorArray(n.id);return a.write(o,s),[a.idTensor]}case"TensorArrayReadV3":{let n=v("tensorArrayId",r,e,t),o=v("index",r,e,t);return[t.getTensorArray(n.id).read(o)]}case"TensorArrayGatherV3":{let n=v("tensorArrayId",r,e,t),o=v("indices",r,e,t),s=v("dtype",r,e,t);return[t.getTensorArray(n.id).gather(o,s)]}case"TensorArrayScatterV3":{let n=v("tensorArrayId",r,e,t),o=v("indices",r,e,t),s=v("tensor",r,e,t),a=t.getTensorArray(n.id);return a.scatter(o,s),[a.idTensor]}case"TensorArrayConcatV3":{let n=v("tensorArrayId",r,e,t),o=t.getTensorArray(n.id),s=v("dtype",r,e,t);return[o.concat(s)]}case"TensorArraySplitV3":{let n=v("tensorArrayId",r,e,t),o=v("tensor",r,e,t),s=v("lengths",r,e,t),a=t.getTensorArray(n.id);return a.split(s,o),[a.idTensor]}case"TensorArraySizeV3":{let n=v("tensorArrayId",r,e,t),o=t.getTensorArray(n.id);return[ue(o.size(),"int32")]}case"TensorArrayCloseV3":{let n=v("tensorArrayId",r,e,t),o=t.getTensorArray(n.id);return o.clearAndClose(),[o.idTensor]}case"TensorListSetItem":{let n=v("tensorListId",r,e,t),o=v("index",r,e,t),s=v("tensor",r,e,t),a=t.getTensorList(n.id);return a.setItem(o,s),[a.idTensor]}case"TensorListGetItem":{let n=v("tensorListId",r,e,t),o=v("index",r,e,t),s=v("elementShape",r,e,t),a=v("elementDType",r,e,t);return[t.getTensorList(n.id).getItem(o,s,a)]}case"TensorListScatterV2":case"TensorListScatter":{let n=v("indices",r,e,t),o=v("tensor",r,e,t),s=v("elementShape",r,e,t),a=v("numElements",r,e,t),i=gB(o,n,s,a);return t.addTensorList(i),[i.idTensor]}case"TensorListReserve":case"EmptyTensorList":{let n=v("elementShape",r,e,t),o=v("elementDType",r,e,t),s;r.op==="TensorListReserve"?s="numElements":s="maxNumElements";let a=v(s,r,e,t),i=hB(n,o,a);return t.addTensorList(i),[i.idTensor]}case"TensorListGather":{let n=v("tensorListId",r,e,t),o=v("indices",r,e,t),s=v("elementShape",r,e,t),a=v("elementDType",r,e,t);return[t.getTensorList(n.id).gather(o,a,s)]}case"TensorListStack":{let n=v("tensorListId",r,e,t),o=v("elementShape",r,e,t),s=v("elementDType",r,e,t),a=v("numElements",r,e,t);return[t.getTensorList(n.id).stack(o,s,a)]}case"TensorListFromTensor":{let n=v("tensor",r,e,t),o=v("elementShape",r,e,t),s=v("elementDType",r,e,t),a=dB(n,o,s);return t.addTensorList(a),[a.idTensor]}case"TensorListConcat":{let n=v("tensorListId",r,e,t),o=t.getTensorList(n.id),s=v("dtype",r,e,t),a=v("elementShape",r,e,t);return[o.concat(s,a)]}case"TensorListPushBack":{let n=v("tensorListId",r,e,t),o=v("tensor",r,e,t),s=t.getTensorList(n.id);return s.pushBack(o),[s.idTensor]}case"TensorListPopBack":{let n=v("tensorListId",r,e,t),o=v("elementShape",r,e,t),s=v("elementDType",r,e,t);return[t.getTensorList(n.id).popBack(o,s)]}case"TensorListSplit":{let n=v("tensor",r,e,t),o=v("elementShape",r,e,t),s=v("lengths",r,e,t),a=xB(n,s,o);return t.addTensorList(a),[a.idTensor]}default:throw TypeError(`Node type ${r.op} is not implemented`)}};function bB(r,e,t){let[n,o]=v("fusedOps",r,e,t),s=n==="biasadd",a=o==="prelu",i=n==="fusedbatchnorm",l=v("numArgs",r,e,t);if(s){if(a&&l!==2)throw new Error("FusedConv2d and DepthwiseConv2d with BiasAdd and Prelu must have two extra arguments: bias and alpha.");if(!a&&l!==1)throw new Error("FusedConv2d and DepthwiseConv2d with BiasAdd must have one extra argument: bias.")}if(i)throw new Error("FusedConv2d and DepthwiseConv2d with FusedBatchNorm is not supported");let u=v("strides",r,e,t),c=bh(r,e,t),p=v("dataFormat",r,e,t).toUpperCase(),m=v("dilations",r,e,t),[f,d]=v("args",r,e,t),h=v("leakyreluAlpha",r,e,t);return{stride:u,pad:c,dataFormat:p,dilations:m,biasArg:f,preluArg:d,activationFunc:o,leakyreluAlpha:h}}var wB=(r,e,t)=>{switch(r.op){case"Conv1D":{let n=v("stride",r,e,t),o=v("pad",r,e,t),s=v("dataFormat",r,e,t).toUpperCase(),a=v("dilation",r,e,t);return[ou(v("x",r,e,t),v("filter",r,e,t),n,o,s,a)]}case"Conv2D":{let n=v("strides",r,e,t),o=bh(r,e,t),s=v("dataFormat",r,e,t).toUpperCase(),a=v("dilations",r,e,t);return[Jr(v("x",r,e,t),v("filter",r,e,t),[n[1],n[2]],o,s,[a[1],a[2]])]}case"_FusedConv2D":{let{stride:n,pad:o,dataFormat:s,dilations:a,biasArg:i,preluArg:l,activationFunc:u,leakyreluAlpha:c}=bB(r,e,t);return[no.conv2d({x:v("x",r,e,t),filter:v("filter",r,e,t),strides:[n[1],n[2]],pad:o,dataFormat:s,dilations:[a[1],a[2]],bias:i,activation:u,preluActivationWeights:l,leakyreluAlpha:c})]}case"FusedDepthwiseConv2dNative":{let{stride:n,pad:o,dataFormat:s,dilations:a,biasArg:i,preluArg:l,activationFunc:u,leakyreluAlpha:c}=bB(r,e,t);return[no.depthwiseConv2d({x:v("x",r,e,t),filter:v("filter",r,e,t),strides:[n[1],n[2]],pad:o,dataFormat:s,dilations:[a[1],a[2]],bias:i,activation:u,preluActivationWeights:l,leakyreluAlpha:c})]}case"Conv2DBackpropInput":case"Conv2dTranspose":{let n=v("outputShape",r,e,t),o=v("strides",r,e,t),s=bh(r,e,t);return[su(v("x",r,e,t),v("filter",r,e,t),n,[o[1],o[2]],s)]}case"DepthwiseConv2dNative":case"DepthwiseConv2d":{let n=v("strides",r,e,t),o=bh(r,e,t),s=v("dilations",r,e,t),a=v("dataFormat",r,e,t).toUpperCase();return[xs(v("input",r,e,t),v("filter",r,e,t),[n[1],n[2]],o,a,[s[1],s[2]])]}case"Conv3D":{let n=v("strides",r,e,t),o=v("pad",r,e,t),s=v("dataFormat",r,e,t).toUpperCase(),a=v("dilations",r,e,t);return[Qm(v("x",r,e,t),v("filter",r,e,t),[n[1],n[2],n[3]],o,s,[a[1],a[2],a[3]])]}case"AvgPool":{let n=v("strides",r,e,t),o=v("pad",r,e,t),s=v("kernelSize",r,e,t);return[ha(v("x",r,e,t),[s[1],s[2]],[n[1],n[2]],o)]}case"MaxPool":{let n=v("strides",r,e,t),o=v("pad",r,e,t),s=v("kernelSize",r,e,t);return[_a(v("x",r,e,t),[s[1],s[2]],[n[1],n[2]],o)]}case"MaxPoolWithArgmax":{let n=v("strides",r,e,t),o=v("pad",r,e,t),s=v("kernelSize",r,e,t),a=v("includeBatchInIndex",r,e,t),{result:i,indexes:l}=p_(v("x",r,e,t),[s[1],s[2]],[n[1],n[2]],o,a);return[i,l]}case"AvgPool3D":{let n=v("strides",r,e,t),o=v("pad",r,e,t),s=v("kernelSize",r,e,t);return[Ym(v("x",r,e,t),[s[1],s[2],s[3]],[n[1],n[2],n[3]],o)]}case"MaxPool3D":{let n=v("strides",r,e,t),o=v("pad",r,e,t),s=v("kernelSize",r,e,t);return[cf(v("x",r,e,t),[s[1],s[2],s[3]],[n[1],n[2],n[3]],o)]}case"Dilation2D":{let n=v("strides",r,e,t),o=v("pad",r,e,t),s=v("dilations",r,e,t),a=n[1],i=n[2],l=s[1],u=s[2];return[tf(v("x",r,e,t),v("filter",r,e,t),[a,i],o,[l,u],"NHWC")]}default:throw TypeError(`Node type ${r.op} is not implemented`)}};var _B=(r,e,t)=>{switch(r.op){case"Fill":{let n=v("shape",r,e,t),o=v("dtype",r,e,t),s=v("value",r,e,t);return[bs(n,s,o)]}case"LinSpace":{let n=v("start",r,e,t),o=v("stop",r,e,t),s=v("num",r,e,t);return[s_(n,o,s)]}case"Multinomial":{let n=v("logits",r,e,t),o=v("numSamples",r,e,t),s=v("seed",r,e,t);return[m_(n,o,s)]}case"OneHot":{let n=v("indices",r,e,t),o=v("depth",r,e,t),s=v("onValue",r,e,t),a=v("offValue",r,e,t);return[ds(n,o,s,a)]}case"Ones":return[er(v("shape",r,e,t),v("dtype",r,e,t))];case"OnesLike":return[sr(v("x",r,e,t))];case"RandomUniform":return[ks(v("shape",r,e,t),v("minval",r,e,t),v("maxval",r,e,t),v("dtype",r,e,t))];case"Range":{let n=v("start",r,e,t),o=v("stop",r,e,t),s=v("step",r,e,t);return[Ca(n,o,s,v("dtype",r,e,t))]}case"TruncatedNormal":{let n=v("shape",r,e,t),o=v("mean",r,e,t),s=v("stdDev",r,e,t),a=v("seed",r,e,t);return[vu(n,o,s,v("dtype",r,e,t),a)]}case"Zeros":return[ht(v("shape",r,e,t),v("dtype",r,e,t))];case"ZerosLike":return[Ie(v("x",r,e,t))];default:throw TypeError(`Node type ${r.op} is not implemented`)}};function fI(r,e,t){let n=v("boxes",r,e,t),o=v("scores",r,e,t),s=v("maxOutputSize",r,e,t),a=v("iouThreshold",r,e,t),i=v("scoreThreshold",r,e,t),l=v("softNmsSigma",r,e,t);return{boxes:n,scores:o,maxOutputSize:s,iouThreshold:a,scoreThreshold:i,softNmsSigma:l}}var kB=async(r,e,t)=>{switch(r.op){case"NonMaxSuppressionV5":{let{boxes:n,scores:o,maxOutputSize:s,iouThreshold:a,scoreThreshold:i,softNmsSigma:l}=fI(r,e,t),u=await ai.nonMaxSuppressionWithScoreAsync(n,o,s,a,i,l);return[u.selectedIndices,u.selectedScores]}case"NonMaxSuppressionV4":{let{boxes:n,scores:o,maxOutputSize:s,iouThreshold:a,scoreThreshold:i}=fI(r,e,t),l=v("padToMaxOutputSize",r,e,t),u=await ai.nonMaxSuppressionPaddedAsync(n,o,s,a,i,l);return[u.selectedIndices,u.validOutputs]}case"NonMaxSuppressionV3":case"NonMaxSuppressionV2":{let{boxes:n,scores:o,maxOutputSize:s,iouThreshold:a,scoreThreshold:i}=fI(r,e,t);return[await ai.nonMaxSuppressionAsync(n,o,s,a,i)]}case"Where":{let n=oe(v("condition",r,e,t),"bool"),o=[await kf(n)];return n.dispose(),o}case"ListDiff":return C_(v("x",r,e,t),v("y",r,e,t));default:throw TypeError(`Node type ${r.op} is not implemented`)}};var vB=(r,e,t)=>{switch(r.op){case"TopKV2":{let n=v("x",r,e,t),o=v("k",r,e,t),s=v("sorted",r,e,t),a=wf(n,o,s);return[a.values,a.indices]}case"Unique":{let n=v("x",r,e,t),o=cp(n);return[o.values,o.indices]}case"UniqueV2":{let n=v("x",r,e,t),o=v("axis",r,e,t),s=cp(n,o);return[s.values,s.indices]}default:throw TypeError(`Node type ${r.op} is not implemented`)}};var CB=(r,e,t)=>{switch(r.op){case"Const":return e[r.name];case"PlaceholderWithDefault":let n=v("default",r,e,t);return[hr(r.name,e,t)||n];case"Placeholder":return[hr(r.name,e,t)];case"Identity":case"StopGradient":case"FakeQuantWithMinMaxVars":{let u=v("x",r,e,t);return[Os(u)]}case"IdentityN":return v("x",r,e,t).map(u=>Os(u));case"Snapshot":let o=v("x",r,e,t);return[Os(o)];case"Shape":return[Ct(v("x",r,e,t).shape,"int32")];case"ShapeN":return v("x",r,e,t).map(u=>Ct(u.shape));case"Size":return[ue(v("x",r,e,t).size,"int32")];case"Rank":return[ue(v("x",r,e,t).rank,"int32")];case"NoOp":return[ue(1)];case"Print":let s=v("x",r,e,t),a=v("data",r,e,t),i=v("message",r,e,t),l=v("summarize",r,e,t);console.warn("The graph has a tf.print() operation,usually used for debugging, which slows down performance."),console.log(i);for(let u=0;u<a.length;u++)console.log(Array.prototype.slice.call(a[u].dataSync()).slice(0,l));return[s];default:throw TypeError(`Node type ${r.op} is not implemented`)}};var dI=class{constructor(e,t){this.keyDType=e,this.valueDType=t,this.handle=ue(0),this.tensorMap=new Map,$t(this.handle)}get id(){return this.handle.id}clearAndClose(){this.tensorMap.forEach(e=>e.dispose()),this.tensorMap.clear(),this.handle.dispose()}size(){return this.tensorMap.size}tensorSize(){return ue(this.size(),"int32")}async import(e,t){this.checkKeyAndValueTensor(e,t);let n=await e.data();return this.tensorMap.forEach(o=>o.dispose()),this.tensorMap.clear(),V(()=>{let o=fr(t),s=n.length,a=o.length;y.assert(s===a,()=>`The number of elements doesn't match, keys has ${s} elements, the values has ${a} elements.`);for(let i=0;i<s;i++){let l=n[i],u=o[i];$t(u),this.tensorMap.set(l,u)}return this.handle})}async find(e,t){this.checkKeyAndValueTensor(e,t);let n=await e.data();return V(()=>{let o=[];for(let s=0;s<n.length;s++){let a=n[s],i=this.findWithDefault(a,t);o.push(i)}return Gt(o)})}findWithDefault(e,t){let n=this.tensorMap.get(e);return n!=null?n:t}checkKeyAndValueTensor(e,t){if(e.dtype!==this.keyDType)throw new Error(`Expect key dtype ${this.keyDType}, but got ${e.dtype}`);if(t.dtype!==this.valueDType)throw new Error(`Expect value dtype ${this.valueDType}, but got ${t.dtype}`)}};var IB=async(r,e,t,n)=>{switch(r.op){case"HashTable":case"HashTableV2":{let o=v("keyDType",r,e,t),s=v("valueDType",r,e,t),a=new dI(o,s);return n.addHashTable(r.name,a),[a.handle]}case"LookupTableImport":case"LookupTableImportV2":{let o=v("tableHandle",r,e,t,n),s=v("keys",r,e,t),a=v("values",r,e,t);return[await n.getHashTableById(o.id).import(s,a)]}case"LookupTableFind":case"LookupTableFindV2":{let o=v("tableHandle",r,e,t,n),s=v("keys",r,e,t),a=v("defaultValue",r,e,t);return[await n.getHashTableById(o.id).find(s,a)]}case"LookupTableSize":case"LookupTableSizeV2":{let o=v("tableHandle",r,e,t,n);return[n.getHashTableById(o.id).tensorSize()]}default:throw TypeError(`Node type ${r.op} is not implemented`)}};var NB=(r,e,t)=>{switch(r.op){case"ResizeBilinear":{let n=v("images",r,e,t),o=v("size",r,e,t),s=v("alignCorners",r,e,t),a=v("halfPixelCenters",r,e,t);return[ai.resizeBilinear(n,[o[0],o[1]],s,a)]}case"ResizeNearestNeighbor":{let n=v("images",r,e,t),o=v("size",r,e,t),s=v("alignCorners",r,e,t),a=v("halfPixelCenters",r,e,t);return[ai.resizeNearestNeighbor(n,[o[0],o[1]],s,a)]}case"CropAndResize":{let n=v("image",r,e,t),o=v("boxes",r,e,t),s=v("boxInd",r,e,t),a=v("cropSize",r,e,t),i=v("method",r,e,t),l=v("extrapolationValue",r,e,t);return[ai.cropAndResize(n,o,s,a,i,l)]}default:throw TypeError(`Node type ${r.op} is not implemented`)}};var SB=(r,e,t)=>{switch(r.op){case"Equal":return[Sn(v("a",r,e,t),v("b",r,e,t))];case"NotEqual":return[ro(v("a",r,e,t),v("b",r,e,t))];case"Greater":return[Vt(v("a",r,e,t),v("b",r,e,t))];case"GreaterEqual":return[hn(v("a",r,e,t),v("b",r,e,t))];case"Less":return[uu(v("a",r,e,t),v("b",r,e,t))];case"LessEqual":return[gn(v("a",r,e,t),v("b",r,e,t))];case"LogicalAnd":return[yr(v("a",r,e,t),v("b",r,e,t))];case"LogicalNot":return[wa(v("a",r,e,t))];case"LogicalOr":return[mu(v("a",r,e,t),v("b",r,e,t))];case"Select":case"SelectV2":return[vt(v("condition",r,e,t),v("a",r,e,t),v("b",r,e,t))];default:throw TypeError(`Node type ${r.op} is not implemented`)}};var TB=(r,e,t)=>{switch(r.op){case"BatchMatMul":case"BatchMatMulV2":case"MatMul":return[ze(v("a",r,e,t),v("b",r,e,t),v("transposeA",r,e,t),v("transposeB",r,e,t))];case"Einsum":return[r_(v("equation",r,e,t),...v("tensors",r,e,t))];case"Transpose":return[Ue(v("x",r,e,t),v("perm",r,e,t))];case"_FusedMatMul":let[n,o]=v("fusedOps",r,e,t),s=n==="biasadd",a=o==="prelu",i=v("numArgs",r,e,t),l=v("leakyreluAlpha",r,e,t);if(s){if(a&&i!==2)throw new Error("Fused MatMul with BiasAdd and Prelu must have two extra arguments: bias and alpha.");if(!a&&i!==1)throw new Error("Fused MatMul with BiasAdd must have one extra argument: bias.")}let[u,c]=v("args",r,e,t);return[no.matMul({a:v("a",r,e,t),b:v("b",r,e,t),transposeA:v("transposeA",r,e,t),transposeB:v("transposeB",r,e,t),bias:u,activation:o,preluActivationWeights:c,leakyreluAlpha:l})];default:throw TypeError(`Node type ${r.op} is not implemented`)}};var AB=(r,e,t)=>{switch(r.op){case"FusedBatchNorm":case"FusedBatchNormV2":return[Jn(v("x",r,e,t),v("mean",r,e,t),v("variance",r,e,t),v("offset",r,e,t),v("scale",r,e,t),v("epsilon",r,e,t))];case"FusedBatchNormV3":return[Jn(v("x",r,e,t),v("mean",r,e,t),v("variance",r,e,t),v("offset",r,e,t),v("scale",r,e,t),v("epsilon",r,e,t))];case"LRN":return[af(v("x",r,e,t),v("radius",r,e,t),v("bias",r,e,t),v("alpha",r,e,t),v("beta",r,e,t))];case"Softmax":return[Ia(v("x",r,e,t))];case"LogSoftmax":return[pu(v("x",r,e,t))];case"SparseToDense":return[Ig(v("sparseIndices",r,e,t),v("outputShape",r,e,t),v("sparseValues",r,e,t),v("defaultValue",r,e,t))];default:throw TypeError(`Node type ${r.op} is not implemented`)}};var EB=(r,e,t)=>{switch(r.op){case"Max":{let a=v("axis",r,e,t),i=v("keepDims",r,e,t);return[mr(v("x",r,e,t),a,i)]}case"Mean":{let a=v("axis",r,e,t),i=v("keepDims",r,e,t);return[dt(v("x",r,e,t),a,i)]}case"Min":{let a=v("axis",r,e,t),i=v("keepDims",r,e,t);return[oi(v("x",r,e,t),a,i)]}case"Sum":{let a=v("axis",r,e,t),i=v("keepDims",r,e,t);return[fe(v("x",r,e,t),a,i)]}case"All":{let a=v("axis",r,e,t),i=v("keepDims",r,e,t);return[tu(v("x",r,e,t),a,i)]}case"Any":{let a=v("axis",r,e,t),i=v("keepDims",r,e,t);return[dl(v("x",r,e,t),a,i)]}case"ArgMax":{let a=v("axis",r,e,t);return[hl(v("x",r,e,t),a)]}case"ArgMin":{let a=v("axis",r,e,t);return[jm(v("x",r,e,t),a)]}case"Prod":{let a=v("axis",r,e,t),i=v("keepDims",r,e,t);return[fu(v("x",r,e,t),a,i)]}case"Cumsum":{let a=v("axis",r,e,t),i=v("exclusive",r,e,t),l=v("reverse",r,e,t);return[au(v("x",r,e,t),a,i,l)]}case"Bincount":let n=v("x",r,e,t),o=v("weights",r,e,t),s=v("size",r,e,t);return[Zm(n,o,s)];case"DenseBincount":{let a=v("x",r,e,t),i=v("weights",r,e,t),l=v("size",r,e,t),u=v("binaryOutput",r,e,t);return[e_(a,i,l,u)]}default:throw TypeError(`Node type ${r.op} is not implemented`)}};var DB=(r,e,t)=>{switch(r.op){case"ConcatV2":case"Concat":{let n=v("n",r,e,t),o=v("axis",r,e,t),s=v("tensors",r,e,t);return s=s.slice(0,n),[Qe(s,o)]}case"Gather":{let n=v("x",r,e,t),o=v("indices",r,e,t);return[Qn(n,oe(o,"int32"),0)]}case"GatherV2":{let n=v("axis",r,e,t),o=v("batchDims",r,e,t),s=v("x",r,e,t),a=v("indices",r,e,t);return[Qn(s,oe(a,"int32"),n,o)]}case"Reverse":{let n=v("dims",r,e,t),o=[];for(let a=0;a<n.length;a++)n[a]&&o.push(a);let s=v("x",r,e,t);return[Kt(s,o)]}case"ReverseV2":{let n=v("axis",r,e,t),o=v("x",r,e,t);return[Kt(o,n)]}case"Slice":{let n=v("begin",r,e,t),o=v("size",r,e,t);return[Re(v("x",r,e,t),n,o)]}case"StridedSlice":{let n=v("begin",r,e,t),o=v("end",r,e,t),s=v("strides",r,e,t),a=v("beginMask",r,e,t),i=v("endMask",r,e,t),l=v("ellipsisMask",r,e,t),u=v("newAxisMask",r,e,t),c=v("shrinkAxisMask",r,e,t),p=v("x",r,e,t);return[yf(p,n,o,s,a,i,l,u,c)]}case"Pack":return V(()=>{let n=v("axis",r,e,t),o=v("tensors",r,e,t),s=o[0].shape,a=Tn(o[0]).shape,i=o.map(l=>{let u=y.arraysEqual(l.shape,s);if(!u&&!y.arraysEqual(Tn(l).shape,a))throw new Error("the input tensors shape does not match");return u?l:L(l,s)});return[Gt(i,n)]});case"Unpack":{let n=v("axis",r,e,t),o=v("tensor",r,e,t);return fr(o,n)}case"Tile":{let n=v("reps",r,e,t);return[Bn(v("x",r,e,t),n)]}case"Split":case"SplitV":{let n=v("axis",r,e,t),o=v("numOrSizeSplits",r,e,t),s=v("x",r,e,t);return tr(s,o,n)}case"ScatterNd":{let n=v("indices",r,e,t),o=v("values",r,e,t),s=v("shape",r,e,t);return[o1(n,o,s)]}case"GatherNd":{let n=v("x",r,e,t),o=v("indices",r,e,t);return[i1(n,o)]}case"SparseToDense":{let n=v("sparseIndices",r,e,t),o=v("outputShape",r,e,t),s=v("sparseValues",r,e,t),a=v("defaultValue",r,e,t);return[Ig(n,s,o,s.dtype===a.dtype?a:oe(a,s.dtype))]}default:throw TypeError(`Node type ${r.op} is not implemented`)}};var $B=(r,e,t)=>{switch(r.op){case"SparseReshape":{let{outputIndices:n,outputShape:o}=W1.sparseReshape(v("inputIndices",r,e,t),v("inputShape",r,e,t),v("newShape",r,e,t));return[n,o]}default:throw TypeError(`Node type ${r.op} is not implemented`)}};var RB=(r,e,t)=>{switch(r.op){case"FFT":return[Na(v("x",r,e,t))];case"IFFT":return[si(v("x",r,e,t))];case"RFFT":return[Sa(v("x",r,e,t))];case"IRFFT":return[_u(v("x",r,e,t))];default:throw TypeError(`Node type ${r.op} is not implemented`)}};var FB=(r,e,t)=>{switch(r.op){case"Cast":return[oe(v("x",r,e,t),v("dtype",r,e,t))];case"ExpandDims":{let n=v("axis",r,e,t);return[cr(v("x",r,e,t),n)]}case"Squeeze":{let n=v("axis",r,e,t);return[Tn(v("x",r,e,t),n)]}case"Reshape":return[L(v("x",r,e,t),v("shape",r,e,t))];case"MirrorPad":return[pf(v("x",r,e,t),v("padding",r,e,t),v("mode",r,e,t))];case"PadV2":case"Pad":return[Lr(v("x",r,e,t),v("padding",r,e,t),v("constantValue",r,e,t))];case"SpaceToBatchND":{let n=v("blockShape",r,e,t),o=v("paddings",r,e,t);return[ka(v("x",r,e,t),n,o)]}case"BatchToSpaceND":{let n=v("blockShape",r,e,t),o=v("crops",r,e,t);return[ga(v("x",r,e,t),n,o)]}case"DepthToSpace":{let n=v("blockSize",r,e,t),o=v("dataFormat",r,e,t).toUpperCase();return[ef(v("x",r,e,t),n,o)]}case"BroadcastTo":return[xa(v("x",r,e,t),v("shape",r,e,t))];default:throw TypeError(`Node type ${r.op} is not implemented`)}};function hI(r,e,t,n){let o=((s,a,i)=>{switch(s.category){case"arithmetic":return V(()=>pB(s,a,i));case"basic_math":return V(()=>mB(s,a,i));case"control":return yB(s,a,i);case"convolution":return V(()=>wB(s,a,i));case"creation":return V(()=>_B(s,a,i));case"dynamic":return kB(s,a,i);case"evaluation":return V(()=>vB(s,a,i));case"image":return V(()=>NB(s,a,i));case"graph":return V(()=>CB(s,a,i));case"logical":return V(()=>SB(s,a,i));case"matrices":return V(()=>TB(s,a,i));case"normalization":return V(()=>AB(s,a,i));case"reduction":return V(()=>EB(s,a,i));case"slice_join":return V(()=>DB(s,a,i));case"sparse":return V(()=>$B(s,a,i));case"spectral":return V(()=>RB(s,a,i));case"transformation":return V(()=>FB(s,a,i));case"hash_table":return IB(s,a,i,n);case"custom":let l=uy(s.op);if(l&&l.customExecutor)return l.customExecutor(new pI(s,a,i));throw TypeError(`Custom op ${s.op} is not registered.`);default:throw TypeError(`Unknown op '${s.op}'. File an issue at https://github.com/tensorflow/tfjs/issues so we can add it, or register a custom execution with tf.registerOp()`)}})(r,e,t);return y.isPromise(o)?o.then(s=>[].concat(s)):[].concat(o)}var vy=class{constructor(e={},t={},n={},o={}){this.weightMap=e,this.tensorArrayMap=t,this.tensorListMap=n,this.functionMap=o,this.rootContext={id:0,frameName:"",iterationId:0},this.contexts=[this.rootContext],this.lastId=0,this.generateCurrentContextIds()}newFrame(e,t){return{id:e,frameName:t,iterationId:0}}set currentContext(e){this.contexts!==e&&(this.contexts=e,this.generateCurrentContextIds())}get currentContext(){return this.contexts}get currentContextId(){return this._currentContextIds[0]}get currentContextIds(){return this._currentContextIds}generateCurrentContextIds(){let e=[];for(let t=0;t<this.contexts.length-1;t++){let n=this.contexts.slice(0,this.contexts.length-t);e.push(this.contextIdforContexts(n))}e.push(""),this._currentContextIds=e}contextIdforContexts(e){return e?e.map(t=>t.id===0&&t.iterationId===0?"":`${t.frameName}-${t.iterationId}`).join("/"):""}enterFrame(e){this.contexts&&(this.lastId++,this.contexts=this.contexts.slice(),this.contexts.push(this.newFrame(this.lastId,e)),this._currentContextIds.unshift(this.contextIdforContexts(this.contexts)))}exitFrame(){if(this.contexts&&this.contexts.length>1)this.contexts=this.contexts.slice(),this.contexts.splice(-1),this.currentContextIds.shift();else throw new Error("Cannot exit frame, the context is empty")}nextIteration(){if(this.contexts&&this.contexts.length>0){this.contexts=this.contexts.slice(),this.lastId++;let e=Object.assign({},this.contexts[this.contexts.length-1]);e.iterationId+=1,e.id=this.lastId,this.contexts.splice(-1,1,e),this._currentContextIds.splice(0,1,this.contextIdforContexts(this.contexts))}else throw new Error("Cannot increase frame iteration, the context is empty")}getWeight(e){return this.weightMap[e]}addTensorArray(e){this.tensorArrayMap[e.id]=e}getTensorArray(e){return this.tensorArrayMap[e]}addTensorList(e){this.tensorListMap[e.id]=e}getTensorList(e){return this.tensorListMap[e]}dispose(e){for(let t in this.tensorArrayMap)this.tensorArrayMap[t].clearAndClose(e);for(let t in this.tensorListMap)this.tensorListMap[t].clearAndClose(e)}};function gI(r,e,t,n){let o=new Set,s=[],a=null,i=null,l=new Set,u=Object.keys(r).map(m=>un(m)[0]),c=[];n!=null&&(c=n.map(m=>un(m.name)[0]));let p=[...e];for(;p.length>0;){let m=p.pop();if((xI(m)||$ne(m)||Rne(m))&&a==null&&(a=m,i=a.children.map(f=>f.name).filter(f=>o.has(f))),o.add(m.name),t[m.name]==null&&u.indexOf(m.name)===-1&&c.indexOf(m.name)===-1){if(m.inputs.length===0){s.push(m.name);continue}m.inputs.forEach(f=>{l.has(f.name)||(l.add(f.name),p.push(f))})}}return{inputs:r,outputs:e,usedNodes:o,missingInputs:s,dynamicNode:a,syncInputs:i}}function OB(r,e,t){let{usedNodes:n,inputs:o}=t,s=[],a=Object.keys(o).map(c=>un(c)[0]).map(c=>r.nodes[c]),i=r.initNodes;a.forEach(c=>{n.has(c.name)&&s.push(c)}),r.weights.forEach(c=>{n.has(c.name)&&s.push(c)}),i!=null&&i.forEach(c=>{n.has(c.name)&&s.push(c)});let l=new Set,u=[];for(;s.length>0;){let c=s.pop();l.add(c.name),e[c.name]||u.push(c),c.children.forEach(p=>{!l.has(p.name)&&n.has(p.name)&&p.inputs.every(m=>l.has(m.name))&&s.push(p)})}return u}var Ane=["Switch","Merge","Enter","Exit","NextIteration","StatelessIf","StatelessWhile","if","While"],Ene=["NonMaxSuppressionV2","NonMaxSuppressionV3","NonMaxSuppressionV5","Where"],Dne=["HashTable","HashTableV2","LookupTableImport","LookupTableImportV2","LookupTableFind","LookupTableFindV2","LookupTableSize","LookupTableSizeV2"];function xI(r){return Ane.indexOf(r.op)>=0}function $ne(r){return Ene.indexOf(r.op)>=0}function Rne(r){return Dne.indexOf(r.op)>=0}var lm=class{constructor(e,t){this.graph=e,this.parent=t,this.compiledMap=new Map,this._weightMap={},this.SEPERATOR=",",this._functions={},this._functionExecutorMap={},this._outputs=e.outputs,this._inputs=e.inputs,this._initNodes=e.initNodes,this._signature=e.signature,this._functions=e.functions,e.functions!=null&&Object.keys(e.functions).forEach(n=>{this._functionExecutorMap[n]=new lm(e.functions[n],this)})}get weightIds(){return this.parent?this.parent.weightIds:this._weightIds}get functionExecutorMap(){return this.parent?this.parent.functionExecutorMap:this._functionExecutorMap}get weightMap(){return this.parent?this.parent.weightMap:this._weightMap}set weightMap(e){let t=Object.keys(e).map(n=>e[n].map(o=>o.id));this._weightIds=[].concat(...t),this._weightMap=e}set resourceManager(e){this._resourceManager=e}get inputs(){return this._inputs.map(e=>({name:e.name,shape:e.attrParams.shape?e.attrParams.shape.value:void 0,dtype:e.attrParams.dtype?e.attrParams.dtype.value:void 0}))}get outputs(){return this._outputs.map(e=>({name:e.name,shape:e.attrParams.shape?e.attrParams.shape.value:void 0,dtype:e.attrParams.dtype?e.attrParams.dtype.value:void 0}))}get inputNodes(){return this._inputs.map(e=>e.signatureKey||e.name)}get outputNodes(){return this._outputs.map(e=>{let t=e.signatureKey||e.name;return e.defaultOutput?`${t}:${e.defaultOutput}`:t})}get functions(){return Object.keys(this._functions).reduce((e,t)=>(e[t]=this._functions[t].signature,e),{})}getCompilationKey(e,t){let n=e.map(s=>s.name).sort(),o=t.map(s=>s.name).sort();return n.join(this.SEPERATOR)+"--"+o.join(this.SEPERATOR)}compile(e,t){let n=gI(e,t,this.weightMap,this._initNodes),{missingInputs:o,dynamicNode:s,syncInputs:a}=n;if(s!=null)throw new Error(`This execution contains the node '${s.name}', which has the dynamic op '${s.op}'. Please use model.executeAsync() instead. Alternatively, to avoid the dynamic ops, specify the inputs [${a}]`);if(o.length>0){let i=t.map(u=>u.name),l=Object.keys(e);throw new Error(`Cannot compute the outputs [${i}] from the provided inputs [${l}]. Missing the following inputs: [${o}]`)}return OB(this.graph,this.weightMap,n)}execute(e,t){e=this.mapInputs(e);let n=Object.keys(e).sort();this.checkInputs(e),this.checkInputShapeAndType(e),t=this.mapOutputs(t),this.checkOutputs(t);let o=n.map(p=>this.graph.nodes[un(p)[0]]),s=t.map(p=>un(p)[0]),a=s.map(p=>this.graph.nodes[p]);a.length===0&&(a=this._outputs);let i=this.getCompilationKey(o,a),l=this.compiledMap.get(i);l==null&&(l=this.compile(e,a),this.compiledMap.set(i,l));let u={},c={};return V(()=>{let p=new vy(this.weightMap,u,c,this.functionExecutorMap),m=Object.assign({},this.weightMap);Object.keys(e).forEach(h=>{let[g,x]=un(h),b=[];b[x]=e[h],m[g]=b});let f=this.getFrozenTensorIds(m),d={};for(let h=0;h<l.length;h++){let g=l[h];if(!m[g.name]){let x=hI(g,m,p,this._resourceManager);if(y.isPromise(x))throw new Error(`The execution of the op '${g.op}' returned a promise. Please use model.executeAsync() instead.`);m[g.name]=x,this.checkTensorForDisposal(g.name,g,m,p,f,s,d)}}return this.parent==null&&p.dispose(f),t.map(h=>hr(h,m,p))})}getFrozenTensorIds(e){let t=[].concat.apply([],Object.keys(e).map(n=>e[n]).map(n=>n.map(o=>o.id)));return new Set(t)}checkTensorForDisposal(e,t,n,o,s,a,i){t.category==="control"||a.indexOf(e)!==-1||(n[e].forEach(l=>{l!=null&&(i[l.id]=(i[l.id]||0)+t.children.length)}),t.inputs.forEach(l=>{if(l.category!=="control"){let u=aB(l.name,n,o);u!=null&&u.forEach(c=>{if(c&&!c.kept&&!s.has(c.id)){let p=i[c.id];p===1?(c.dispose(),delete i[c.id]):p!=null&&i[c.id]--}})}}))}async executeAsync(e,t){return this._executeAsync(e,t)}async _executeAsync(e,t,n=!1,o={},s={}){n||(e=this.mapInputs(e),this.checkInputs(e),this.checkInputShapeAndType(e),t=this.mapOutputs(t),this.checkOutputs(t));let a=new vy(this.weightMap,o,s,this.functionExecutorMap),i=await this.executeWithControlFlow(e,a,t,n),l=t.map(m=>hr(m,i,a)),u=l.map(m=>m.id),c=Object.keys(e).map(m=>e[m].id),p=new Set([...u,...c,...this.weightIds]);return Object.keys(i).forEach(m=>{i[m].forEach(d=>{d&&!d.kept&&!d.isDisposed&&!p.has(d.id)&&d.dispose()})}),this.parent==null&&a.dispose(p),l}async executeFunctionAsync(e,t,n){let o=e.reduce((s,a,i)=>(s[this.inputs[i].name]=a,s),{});return this._executeAsync(o,this.outputNodes,!0,t,n)}async executeWithControlFlow(e,t,n,o){let s=Object.keys(e),a=s.map(w=>this.graph.nodes[un(w)[0]]),i=n.map(w=>un(w)[0]),l=i.map(w=>this.graph.nodes[w]);l.length===0&&(l=this._outputs);let{usedNodes:u,missingInputs:c,dynamicNode:p,syncInputs:m}=gI(e,l,this.weightMap,this._initNodes),f=[...a,...this.graph.weights,...this._initNodes||[]].map(w=>({node:w,contexts:t.currentContext})),d=Object.assign({},this.weightMap);Object.keys(e).forEach(w=>{let[_,I]=un(w),E=[];E[I]=e[w],d[_]=E});let h={},g=this.getFrozenTensorIds(d),x={};for(;f.length>0;){let w=this.processStack(a,f,t,d,x,g,i,h,u);await Promise.all(w)}p==null&&!o&&console.warn("This model execution did not contain any nodes with control flow or dynamic output shapes. You can use model.execute() instead.");let b=l.filter(w=>!xI(w)&&!hr(w.name,d,t)).map(w=>w.name);if(b.length>0){let w="";throw p!=null&&(w=`Alternatively, to avoid the dynamic ops, use model.execute() and specify the inputs [${m}]`),new Error(`Cannot compute the outputs [${b}] from the provided inputs [${s}]. Consider providing the following inputs: [${c}]. ${w}`)}return d}processStack(e,t,n,o,s,a,i,l,u){let c=[];for(;t.length>0;){let p=t.pop();n.currentContext=p.contexts;let m="";if(p.node.op==="Enter"&&v("isConstant",p.node,o,n)&&([m]=Fs(p.node.name,n)),o[p.node.name]==null){let f=hI(p.node,o,n,this._resourceManager);m||([m]=Fs(p.node.name,n));let d=n.currentContext;y.isPromise(f)?c.push(f.then(h=>(o[m]=h,n.currentContext=d,this.checkTensorForDisposal(m,p.node,o,n,a,i,l),this.processChildNodes(p.node,t,n,o,s,u),h))):(o[m]=f,this.checkTensorForDisposal(m,p.node,o,n,a,i,l),this.processChildNodes(p.node,t,n,o,s,u))}else this.processChildNodes(p.node,t,n,o,s,u)}return c}processChildNodes(e,t,n,o,s,a){e.children.forEach(i=>{let[l]=Fs(i.name,n);s[l]||!a.has(i.name)||(i.op==="Merge"?i.inputNames.some(u=>!!hr(u,o,n))&&(s[l]=!0,t.push({contexts:n.currentContext,node:i})):i.inputNames.every(u=>!!hr(u,o,n))&&(s[l]=!0,t.push({contexts:n.currentContext,node:i})))})}dispose(){Object.keys(this.weightMap).forEach(e=>this.weightMap[e].forEach(t=>t.dispose()))}checkInputShapeAndType(e){Object.keys(e).forEach(t=>{let n=e[t],[o]=un(t),s=this.graph.nodes[o];if(s.attrParams.shape&&s.attrParams.shape.value){let a=s.attrParams.shape.value,i=a.length===n.shape.length&&n.shape.every((l,u)=>a[u]===-1||a[u]===l);y.assert(i,()=>`The shape of dict['${s.name}'] provided in model.execute(dict) must be [${a}], but was [${n.shape}]`)}s.attrParams.dtype&&s.attrParams.dtype.value&&y.assert(n.dtype===s.attrParams.dtype.value,()=>`The dtype of dict['${s.name}'] provided in model.execute(dict) must be ${s.attrParams.dtype.value}, but was ${n.dtype}`)})}mapInputs(e){let t={};for(let n in e)if(this._signature!=null&&this._signature.inputs!=null&&this._signature.inputs[n]!=null){let o=this._signature.inputs[n];t[o.name]=e[n]}else t[n]=e[n];return t}checkInputs(e){let t=Object.keys(e).filter(n=>{let[o]=un(n);return this.graph.nodes[o]==null});if(t.length>0)throw new Error(`The dict provided in model.execute(dict) has keys: [${t}] that are not part of graph`)}mapOutputs(e){return e.map(t=>this._signature!=null&&this._signature.outputs!=null&&this._signature.outputs[t]!=null?this._signature.outputs[t].name:t,{})}checkOutputs(e){e.forEach(t=>{let[n]=un(t);if(!this.graph.nodes[n])throw new Error(`The output '${t}' is not found in the graph`)})}};var yI=class{constructor(e={},t={}){this.hashTableNameToHandle=e,this.hashTableMap=t}addHashTable(e,t){this.hashTableNameToHandle[e]=t.handle,this.hashTableMap[t.id]=t}getHashTableHandleByName(e){return this.hashTableNameToHandle[e]}getHashTableById(e){return this.hashTableMap[e]}dispose(){for(let e in this.hashTableMap)this.hashTableMap[e].clearAndClose(),delete this.hashTableMap[e];for(let e in this.hashTableNameToHandle)this.hashTableNameToHandle[e].dispose(),delete this.hashTableNameToHandle[e]}};var Fne="?tfjs-format=file",One="model.json",bI=class{constructor(e,t={}){this.modelUrl=e,this.loadOptions=t,this.version="n/a",t==null&&(this.loadOptions={}),this.resourceManager=new yI}get modelVersion(){return this.version}get inputNodes(){return this.executor.inputNodes}get outputNodes(){return this.executor.outputNodes}get inputs(){return this.executor.inputs}get outputs(){return this.executor.outputs}get weights(){return this.executor.weightMap}get metadata(){return this.artifacts.userDefinedMetadata}get modelSignature(){return this.signature}findIOHandler(){let e=this.modelUrl;if(e.load!=null)this.handler=e;else if(this.loadOptions.requestInit!=null)this.handler=Er.browserHTTPRequest(e,this.loadOptions);else{let t=Er.getLoadHandlers(e,this.loadOptions);if(t.length===0)t.push(Er.browserHTTPRequest(e,this.loadOptions));else if(t.length>1)throw new Error(`Found more than one (${t.length}) load handlers for URL '${[e]}'`);this.handler=t[0]}}async load(){if(this.findIOHandler(),this.handler.load==null)throw new Error("Cannot proceed with model loading because the IOHandler provided does not have the `load` method implemented.");let e=await this.handler.load();return this.loadSync(e)}loadSync(e){this.artifacts=e;let t=this.artifacts.modelTopology,n;this.artifacts.userDefinedMetadata!=null&&this.artifacts.userDefinedMetadata.signature!=null?n=this.artifacts.userDefinedMetadata.signature:n=this.artifacts.signature,this.signature=n,this.version=`${t.versions.producer}.${t.versions.minConsumer}`;let o=Er.decodeWeights(this.artifacts.weightData,this.artifacts.weightSpecs);if(this.executor=new lm(py.Instance.transformGraph(t,this.signature)),this.executor.weightMap=this.convertTensorMapToTensorsMap(o),this.executor.resourceManager=this.resourceManager,e.modelInitializer!=null&&e.modelInitializer.node!=null){let s=py.Instance.transformGraph(e.modelInitializer);this.initializer=new lm(s),this.initializer.weightMap=this.executor.weightMap,this.initializer.resourceManager=this.resourceManager,this.initializer.executeAsync({},[])}return!0}async save(e,t){if(typeof e=="string"){let n=Er.getSaveHandlers(e);if(n.length===0)throw new Error(`Cannot find any save handlers for URL '${e}'`);if(n.length>1)throw new Error(`Found more than one (${n.length}) save handlers for URL '${e}'`);e=n[0]}if(e.save==null)throw new Error("GraphModel.save() cannot proceed because the IOHandler provided does not have the `save` attribute defined.");return e.save(this.artifacts)}predict(e,t){return this.execute(e,this.outputNodes)}normalizeInputs(e){if(!(e instanceof Pe)&&!Array.isArray(e))return e;if(e=Array.isArray(e)?e:[e],e.length!==this.inputNodes.length)throw new Error(`Input tensor count mismatch,the graph model has ${this.inputNodes.length} placeholders, while there are ${e.length} input tensors.`);return this.inputNodes.reduce((t,n,o)=>(t[n]=e[o],t),{})}normalizeOutputs(e){return e=e||this.outputNodes,Array.isArray(e)?e:[e]}execute(e,t){e=this.normalizeInputs(e),t=this.normalizeOutputs(t);let n=this.executor.execute(e,t);return n.length>1?n:n[0]}async executeAsync(e,t){e=this.normalizeInputs(e),t=this.normalizeOutputs(t);let n=await this.executor.executeAsync(e,t);return n.length>1?n:n[0]}convertTensorMapToTensorsMap(e){return Object.keys(e).reduce((t,n)=>(t[n]=[e[n]],t),{})}dispose(){this.executor.dispose(),this.initializer&&this.initializer.dispose(),this.resourceManager.dispose()}};async function Pne(r,e={}){if(r==null)throw new Error("modelUrl in loadGraphModel() cannot be null. Please provide a url or an IOHandler that loads the model");e==null&&(e={}),e.fromTFHub&&r.load==null&&(r.endsWith("/")||(r=r+"/"),r=`${r}${One}${Fne}`);let t=new bI(r,e);return await t.load(),t}var Mne="3.6.0";var FI={};We(FI,{CSVDataset:()=>Ch,Dataset:()=>gi,FileDataSource:()=>Ah,TextLineDataset:()=>kh,URLDataSource:()=>Eh,array:()=>yV,csv:()=>TV,func:()=>AV,generator:()=>EV,microphone:()=>$V,version_data:()=>RV,webcam:()=>DV,zip:()=>bV});var xV=hc(NI());var rV=hc(NI());function ZB(r,e){return Iy(r,e)}function Iy(r,e,t=new Map,n=new Set){if(r==null)return null;if(n.has(r))throw new Error("Circular references are not supported.");if(t.has(r))return t.get(r);let o=e(r);if(o.recurse&&o.value!==null)throw new Error("A deep map function may not return both a value and recurse=true.");if(o.recurse)if(Ll(r)){let s=Array.isArray(r)?[]:{};n.add(r);for(let a in r){let i=r[a],l=Iy(i,e,t,n);s[a]=l}return n.delete(r),s}else throw new Error(`Can't recurse into non-iterable type: ${r}`);else return t.set(r,o.value),o.value}function JB(r,e=SI){return QB(r,e)}function QB(r,e,t=new Set){let n=r[0];if(t.has(n))throw new Error("Circular references are not supported.");let o=e(r);if(o.recurse&&o.value!==null)throw new Error("A deep zip function may not return both a value and recurse=true.");if(o.recurse)if(Ll(n)){let s=Array.isArray(n)?[]:{};t.add(n);for(let a in n){let i=r.map(u=>u[a]),l=QB(i,e,t);s[a]=l}return t.delete(n),s}else throw new Error(`Can't recurse into non-iterable type: ${n}`);else return o.value}function SI(r){return r===null?null:Ll(r[0])?{value:null,recurse:!0}:{value:r,recurse:!1}}async function Ny(r,e){let t=new Map;Iy(r,e,t);for(let o of Array.from(t.keys())){let s=t.get(o);if(y.isPromise(s)){let a=await s;t.set(o,a)}}return Iy(r,e,t)}function Ll(r){return r!=null&&!ArrayBuffer.isView(r)&&(Array.isArray(r)||typeof r=="object"&&!(r instanceof Pe))}function eV(r){return r==null||jne(r)||Array.isArray(r)||typeof r=="object"&&r instanceof Pe||y.isTypedArray(r)}function jne(r){return r===null||typeof r!="object"&&typeof r!="function"}function tV(r){return ZB(r,Une)}function Une(r){return r instanceof Pe?{value:r.clone(),recurse:!1}:Ll(r)?{value:null,recurse:!0}:{value:r,recurse:!1}}var wh=class{constructor(e){if(this.capacity=e,this.begin=0,this.end=0,e==null)throw new RangeError("Can't create a ring buffer of unknown capacity.");if(e<1)throw new RangeError("Can't create ring buffer of capacity < 1.");this.data=new Array(e),this.doubledCapacity=2*e}wrap(e){for(;e<0;)e+=this.doubledCapacity;return e%this.doubledCapacity}get(e){if(e<0)throw new RangeError("Can't get item at a negative index.");return this.data[e%this.capacity]}set(e,t){if(e<0)throw new RangeError("Can't set item at a negative index.");this.data[e%this.capacity]=t}length(){let e=this.end-this.begin;return e<0&&(e=this.doubledCapacity+e),e}isFull(){return this.length()===this.capacity}isEmpty(){return this.length()===0}push(e){if(this.isFull())throw new RangeError("Ring buffer is full.");this.set(this.end,e),this.end=this.wrap(this.end+1)}pushAll(e){for(let t of e)this.push(t)}pop(){if(this.isEmpty())throw new RangeError("Ring buffer is empty.");this.end=this.wrap(this.end-1);let e=this.get(this.end);return this.set(this.end,void 0),e}unshift(e){if(this.isFull())throw new RangeError("Ring buffer is full.");this.begin=this.wrap(this.begin-1),this.set(this.begin,e)}shift(){if(this.isEmpty())throw new RangeError("Ring buffer is empty.");let e=this.get(this.begin);return this.set(this.begin,void 0),this.begin=this.wrap(this.begin+1),e}shuffleExcise(e){if(this.isEmpty())throw new RangeError("Ring buffer is empty.");let t=this.wrap(this.begin+e),n=this.get(t);return this.set(t,this.pop()),n}};var um=class extends wh{constructor(){super(um.INITIAL_CAPACITY)}isFull(){return!1}push(e){super.isFull()&&this.expand(),super.push(e)}unshift(e){super.isFull()&&this.expand(),super.unshift(e)}expand(){let e=this.capacity*2,t=new Array(e),n=this.length();for(let o=0;o<n;o++)t[o]=this.get(this.wrap(this.begin+o));this.data=t,this.capacity=e,this.doubledCapacity=2*this.capacity,this.begin=0,this.end=n}};um.INITIAL_CAPACITY=32;function TI(r){return new sV(r)}function _h(r){return new iV(r)}function nV(r,e){return new EI(r,e)}function oV(r,e=Ga.FAIL){return new hV(r,e)}var Yt=class{async toArray(){let e=[],t=await this.next();for(;!t.done;)e.push(t.value),t=await this.next();return e}async toArrayForTest(){let e=this.prefetch(100),t=[],n=await e.next();for(;!n.done;)t.push(n.value),n=await e.next();return t}async resolveFully(){let e=await this.next();for(;!e.done;)e=await this.next()}async resolveWhile(e){let t=await this.next(),n=e(t.value);for(;!t.done&&n;)t=await this.next(),n=e(t.value)}handleErrors(e){return new fV(this,e)}filter(e){return new pV(this,e)}map(e){return new mV(this,e)}mapAsync(e){return new AI(this,e)}serialMapAsync(e){return new AI(this,e).serial()}flatmap(e){return new dV(this,e)}async forEachAsync(e){return this.map(e).resolveFully()}async serialForEach(e){return this.serialMapAsync(e).resolveWhile(t=>t===!0)}rowMajorBatch(e,t=!0){return new cV(this,e,t)}columnMajorBatch(e,t=!0,n=SI){return this.rowMajorBatch(e,t).map(s=>JB(s,n))}concatenate(e,t){return new EI(TI([this,e]),t)}take(e){return e<0||e==null?this:new uV(this,e)}skip(e){return e<0||e==null?this:new lV(this,e)}prefetch(e){return new DI(this,e)}shuffle(e,t){return new gV(this,e,t)}serial(){return new aV(this)}},sV=class extends Yt{constructor(e){super();this.items=e,this.trav=0}summary(){return`Array of ${this.items.length} items`}async next(){if(this.trav>=this.items.length)return{value:null,done:!0};let e=this.items[this.trav];return this.trav++,{value:tV(e),done:!1}}},iV=class extends Yt{constructor(e){super();this.nextFn=e}summary(){return"Function call"}async next(){try{return this.nextFn()}catch(e){throw e.message=`Error thrown while iterating through a dataset: ${e.message}`,e}}},aV=class extends Yt{constructor(e){super();this.upstream=e,this.lastRead=Promise.resolve({value:null,done:!1})}summary(){return`${this.upstream.summary()} -> Serial`}async next(){return this.lastRead=this.lastRead.then(()=>this.serialNext()),this.lastRead}async serialNext(){return this.upstream.next()}},lV=class extends Yt{constructor(e,t){super();this.upstream=e,this.maxCount=t,this.count=0,this.lastRead=Promise.resolve({value:null,done:!1})}summary(){return`${this.upstream.summary()} -> Skip`}async next(){return this.lastRead=this.lastRead.then(()=>this.serialNext()),this.lastRead}async serialNext(){for(;this.count++<this.maxCount;){let e=await this.upstream.next();if(e.done)return e;Ee(e.value)}return this.upstream.next()}},uV=class extends Yt{constructor(e,t){super();this.upstream=e,this.maxCount=t,this.count=0}summary(){return`${this.upstream.summary()} -> Take`}async next(){return this.count++>=this.maxCount?{value:null,done:!0}:this.upstream.next()}},cV=class extends Yt{constructor(e,t,n=!0){super();this.upstream=e,this.batchSize=t,this.enableSmallLastBatch=n,this.lastRead=Promise.resolve({value:null,done:!1})}summary(){return`${this.upstream.summary()} -> RowMajorBatch`}async next(){return this.lastRead=this.lastRead.then(()=>this.serialNext()),this.lastRead}async serialNext(){let e=[];for(;e.length<this.batchSize;){let t=await this.upstream.next();if(t.done)return this.enableSmallLastBatch&&e.length>0?{value:e,done:!1}:{value:null,done:!0};e.push(t.value)}return{value:e,done:!1}}},pV=class extends Yt{constructor(e,t){super();this.upstream=e,this.predicate=t,this.lastRead=Promise.resolve({value:null,done:!1})}summary(){return`${this.upstream.summary()} -> Filter`}async next(){return this.lastRead=this.lastRead.then(()=>this.serialNext()),this.lastRead}async serialNext(){for(;;){let e=await this.upstream.next();if(e.done||this.predicate(e.value))return e;Ee(e.value)}}},mV=class extends Yt{constructor(e,t){super();this.upstream=e,this.transform=t}summary(){return`${this.upstream.summary()} -> Map`}async next(){let e=await this.upstream.next();if(e.done)return{value:null,done:!0};let t=Zn.getTensorsInContainer(e.value),n=this.transform(e.value),o=Zn.getTensorsInContainer(n);for(let s of t)Zn.isTensorInList(s,o)||s.dispose();return{value:n,done:!1}}},fV=class extends Yt{constructor(e,t){super();this.upstream=e,this.handler=t,this.count=0,this.lastRead=Promise.resolve({value:null,done:!1})}summary(){return`${this.upstream.summary()} -> handleErrors`}async next(){return this.lastRead=this.lastRead.then(()=>this.serialNext()),this.lastRead}async serialNext(){for(;;)try{return await this.upstream.next()}catch(e){if(!this.handler(e))return{value:null,done:!0}}}},AI=class extends Yt{constructor(e,t){super();this.upstream=e,this.transform=t}summary(){return`${this.upstream.summary()} -> AsyncMap`}async next(){let e=await this.upstream.next();if(e.done)return{value:null,done:!0};let t=Zn.getTensorsInContainer(e.value),n=await this.transform(e.value),o=Zn.getTensorsInContainer(n);for(let s of t)Zn.isTensorInList(s,o)||s.dispose();return{value:n,done:!1}}},cm=class extends Yt{constructor(){super();this.outputQueue=new um,this.lastRead=Promise.resolve({value:null,done:!1})}async next(){return this.lastRead=this.lastRead.then(()=>this.serialNext()),this.lastRead}async serialNext(){for(;this.outputQueue.length()===0;)if(!await this.pump())return{value:null,done:!0};return{value:this.outputQueue.shift(),done:!1}}},dV=class extends cm{constructor(e,t){super();this.upstream=e,this.transform=t}summary(){return`${this.upstream.summary()} -> Flatmap`}async pump(){let e=await this.upstream.next();if(e.done)return!1;let t=Zn.getTensorsInContainer(e.value),n=this.transform(e.value),o=Zn.getTensorsInContainer(n);this.outputQueue.pushAll(n);for(let s of t)Zn.isTensorInList(s,o)||s.dispose();return!0}},EI=class extends Yt{constructor(e,t){super();this.baseErrorHandler=t,this.lastRead=null,this.iterator=null,this.moreIterators=e}summary(){return"TODO: fill in upstream of chained summaries -> Chained"}async next(){return this.lastRead=this.readFromChain(this.lastRead),this.lastRead}async readFromChain(e){if(await e,this.iterator==null){let n=await this.moreIterators.next();if(n.done)return{value:null,done:!0};this.iterator=n.value,this.baseErrorHandler!=null&&(this.iterator=this.iterator.handleErrors(this.baseErrorHandler))}let t=await this.iterator.next();return t.done?(this.iterator=null,this.readFromChain(e)):t}},Ga;(function(r){r[r.FAIL=0]="FAIL",r[r.SHORTEST=1]="SHORTEST",r[r.LONGEST=2]="LONGEST"})(Ga||(Ga={}));var hV=class extends Yt{constructor(e,t=Ga.FAIL){super();this.iterators=e,this.mismatchMode=t,this.count=0,this.currentPromise=null}summary(){return"{TODO: fill in upstream of zip summaries} -> Zip"}async nextState(e){await e;let t=0,n=0;function o(a){return a instanceof Yt?{value:a.next().then(l=>(t++,l.done&&n++,l.value)),recurse:!1}:{value:null,recurse:!0}}let s=await Ny(this.iterators,o);if(t===n)return{value:null,done:!0};if(n>0)switch(this.mismatchMode){case Ga.FAIL:throw new Error(`Zipped streams should have the same length. Mismatched at element ${this.count}.`);case Ga.SHORTEST:return{value:null,done:!0};case Ga.LONGEST:default:}return this.count++,{value:s,done:!1}}async next(){return this.currentPromise=this.nextState(this.currentPromise),this.currentPromise}},DI=class extends Yt{constructor(e,t){super();this.upstream=e,this.bufferSize=t,this.buffer=new wh(t)}summary(){return`${this.upstream.summary()} -> Prefetch`}refill(){for(;!this.buffer.isFull();){let e=this.upstream.next();this.buffer.push(e)}}next(){return this.refill(),this.buffer.shift()}},gV=class extends DI{constructor(e,t,n){super(e,t);this.upstream=e,this.windowSize=t,this.upstreamExhausted=!1,this.random=rV.alea(n||y.now().toString()),this.lastRead=Promise.resolve({value:null,done:!1})}async next(){return this.lastRead=this.lastRead.then(()=>this.serialNext()),this.lastRead}randomInt(e){return Math.floor(this.random()*e)}chooseIndex(){return this.randomInt(this.buffer.length())}async serialNext(){for(this.upstreamExhausted||this.refill();!this.buffer.isEmpty();){let e=this.chooseIndex(),t=await this.buffer.shuffleExcise(e);if(t.done)this.upstreamExhausted=!0;else return this.refill(),t}return{value:null,done:!0}}};var gi=class{constructor(){this.size=null}batch(e,t=!0){let n=this;y.assert(e>0,()=>`batchSize needs to be positive, but it is
|
|
${e}`);let o;return this.size===Infinity||this.size==null?o=this.size:t?o=Math.ceil(this.size/e):o=Math.floor(this.size/e),kn(async()=>(await n.iterator()).columnMajorBatch(e,t,Hne),o)}concatenate(e){let t=this,n;return this.size===Infinity||e.size===Infinity?n=Infinity:this.size!=null&&e.size!=null?n=this.size+e.size:n=null,kn(async()=>(await t.iterator()).concatenate(await e.iterator()),n)}filter(e){let t=this,n;return this.size===Infinity?n=Infinity:n=null,kn(async()=>(await t.iterator()).filter(o=>V(()=>e(o))),n)}async forEachAsync(e){return(await this.iterator()).forEachAsync(e)}map(e){let t=this;return kn(async()=>(await t.iterator()).map(n=>V(()=>e(n))),this.size)}mapAsync(e){let t=this;return kn(async()=>(await t.iterator()).mapAsync(e),this.size)}prefetch(e){if(e==null)throw new RangeError("`Dataset.prefetch()` requires bufferSize to be specified.");let t=this;return kn(async()=>(await t.iterator()).prefetch(e),this.size)}repeat(e){let t=this,n;return this.size!=null&&e>0?n=this.size*e:e===0?n=0:this.size!=null&&(e===void 0||e<0)?n=Infinity:n=null,kn(async()=>{let o=_h(async()=>({value:await t.iterator(),done:!1}));return nV(o.take(e))},n)}skip(e){let t=this,n;return this.size!=null&&e>=0&&this.size>=e?n=this.size-e:this.size!=null&&(this.size<e||e===void 0||e<0)?n=0:n=null,kn(async()=>(await t.iterator()).skip(e),n)}shuffle(e,t,n=!0){if(e==null||e<0)throw this.size==null?new RangeError("`Dataset.shuffle()` requires bufferSize to be specified."):new RangeError(`\`Dataset.shuffle()\` requires bufferSize to be specified. If your data fits in main memory (for regular JS objects), and/or GPU memory (for \`tf.Tensor\`s), consider setting bufferSize to the dataset size (${this.size} elements)`);let o=this,s=xV.alea(t||y.now().toString());return kn(async()=>{let a=s.int32();return n&&(a+=s.int32()),(await o.iterator()).shuffle(e,a.toString())},this.size)}take(e){let t=this,n;return this.size!=null&&this.size>e?n=e:this.size!=null&&this.size<=e?n=this.size:n=null,kn(async()=>(await t.iterator()).take(e),n)}async toArray(){if(this.size===Infinity)throw new Error("Can not convert infinite data stream to array.");return(await this.iterator()).toArray()}async toArrayForTest(){if(this.size===Infinity)throw new Error("Can not convert infinite data stream to array.");return(await this.iterator()).toArrayForTest()}};gi.MAX_BUFFER_SIZE=1e4;function kn(r,e=null){return new class extends gi{constructor(){super(...arguments);this.size=e}async iterator(){return r()}}}function yV(r){return kn(async()=>TI(r),r.length)}function bV(r){if(!Ll(r))throw new Error("The argument to zip() must be an object or array.");let e;if(Array.isArray(r))for(let t=0;t<r.length;t++)e=e==null?r[t].size:Math.min(e,r[t].size);else if(r instanceof Object)for(let t in r)e=e==null?r[t].size:Math.min(e,r[t].size);return kn(async()=>{let t=await Ny(r,n=>{if(n instanceof gi)return{value:n.iterator(),recurse:!1};if(Ll(n))return{value:null,recurse:!0};throw new Error("Leaves of the structure passed to zip() must be Datasets, not primitives.")});return oV(t,Ga.SHORTEST)},e)}function Hne(r){if(r===null)return null;let e=r[0];return eV(e)?{value:qne(r),recurse:!1}:{value:null,recurse:!0}}function qne(r){if(r.length===0)throw new Error("Can't make a batch of zero elements.");return r[0]instanceof Pe?Gt(r):vr(r)}var kh=class extends gi{constructor(e){super();this.input=e}async iterator(){return(await this.input.iterator()).decodeUTF8().split(`
|
|
`).map(o=>(o.endsWith("\r")&&(o=o.slice(0,-1)),o))}};var Sy='"',vh=Symbol("out"),wV=Symbol("field"),Ty=Symbol("quote"),$I=Symbol("quoteafterquote"),_V=Symbol("quoteinquote"),Ch=class extends gi{constructor(e,t){super();this.input=e,this.hasHeader=!0,this.fullColumnNames=null,this.columnNamesValidated=!1,this.columnConfigs=null,this.configuredColumnsOnly=!1,this.delimiter=",",this.delimWhitespace=!1,this.base=new kh(e),t||(t={}),this.hasHeader=t.hasHeader!==!1,this.fullColumnNames=t.columnNames,this.columnConfigs=t.columnConfigs,this.configuredColumnsOnly=t.configuredColumnsOnly,t.delimWhitespace?(y.assert(t.delimiter==null,()=>"Delimiter should not be provided when delimWhitespace is true."),this.delimWhitespace=!0,this.delimiter=" "):this.delimiter=t.delimiter?t.delimiter:","}async columnNames(){return this.columnNamesValidated||await this.setColumnNames(),this.configuredColumnsOnly?Object.keys(this.columnConfigs):this.fullColumnNames}async setColumnNames(){let e=await this.maybeReadHeaderLine();if(!this.fullColumnNames&&!e)throw new Error("Column names must be provided if there is no header line.");this.fullColumnNames&&e&&y.assert(e.length===this.fullColumnNames.length,()=>"The length of provided columnNames ("+this.fullColumnNames.length.toString()+") does not match the length of the header line read from file ("+e.length.toString()+")."),this.fullColumnNames||(this.fullColumnNames=e);let t=this.fullColumnNames.reduce((o,s)=>(o[s]=o[s]+1||1,o),{}),n=Object.keys(t).filter(o=>t[o]>1);if(y.assert(n.length===0,()=>"Duplicate column names found: "+n.toString()),this.columnConfigs){for(let o of Object.keys(this.columnConfigs))if(this.fullColumnNames.indexOf(o)===-1)throw new Error('The key "'+o+'" provided in columnConfigs does not match any of the column names ('+this.fullColumnNames.toString()+").")}this.columnNamesValidated=!0}async maybeReadHeaderLine(){if(this.hasHeader){let t=await(await this.base.iterator()).next();if(t.done)throw new Error("No data was found for CSV parsing.");let n=t.value;return this.parseRow(n,!1)}else return null}async iterator(){this.columnNamesValidated||await this.setColumnNames();let e=await this.base.iterator();return this.hasHeader&&(e=e.skip(1)),e.map(t=>this.makeDataElement(t))}makeDataElement(e){let t=this.parseRow(e),n={},o={};for(let s=0;s<this.fullColumnNames.length;s++){let a=this.fullColumnNames[s],i=this.columnConfigs?this.columnConfigs[a]:null;if(!(this.configuredColumnsOnly&&!i)){let l=t[s],u=null;if(l==="")if(i&&i.default!==void 0)u=i.default;else{if(i&&(i.required||i.isLabel))throw new Error(`Required column ${a} is empty in this line: ${e}`);u=void 0}else{let c=Number(l);if(isNaN(c))i&&i.dtype==="bool"?u=this.getBoolean(l):u=l;else if(!i||!i.dtype)u=c;else switch(i.dtype){case"float32":u=c;break;case"int32":u=Math.floor(c);break;case"bool":u=this.getBoolean(l);break;default:u=c}}i&&i.isLabel?o[a]=u:n[a]=u}}return Object.keys(o).length===0?n:{xs:n,ys:o}}getBoolean(e){return e==="1"||e.toLowerCase()==="true"?1:0}parseRow(e,t=!0){let n=[],o=0,s=e.length,a=vh;for(let i=0;i<s;i++)switch(a){case vh:switch(e.charAt(i)){case Sy:o=i+1,a=Ty;break;case this.delimiter:if(o=i+1,this.delimiter===" "&&this.delimWhitespace)break;n.push(""),a=vh;break;default:a=wV,o=i;break}break;case wV:switch(e.charAt(i)){case this.delimiter:n.push(e.substring(o,i)),a=vh,o=i+1;break;default:}break;case Ty:switch(e.charAt(i)){case Sy:a=$I;break;default:}break;case $I:switch(e.charAt(i)){case this.delimiter:n.push(e.substring(o,i-1)),a=vh,o=i+1;break;case Sy:a=Ty;break;default:a=_V;break}break;case _V:switch(e.charAt(i)){case Sy:a=Ty;break;default:}break;default:}if(a===$I?n.push(e.substring(o,s-1)):n.push(e.substring(o)),t&&n.length!==this.fullColumnNames.length)throw new Error(`Invalid row in csv file. Should have ${this.fullColumnNames.length} elements in a row, but got ${n}`);return n}};var Ih=class extends Yt{constructor(e){super();this.microphoneConfig=e,this.isClosed=!1,this.fftSize=e.fftSize||1024;let t=Math.log2(this.fftSize);if(this.fftSize<0||t<4||t>14||!Number.isInteger(t))throw new Error(`Invalid fftSize: it must be a power of 2 between 2 to 4 and 2 to 14, but got ${this.fftSize}`);if(this.numFrames=e.numFramesPerSpectrogram||43,this.sampleRateHz=e.sampleRateHz,this.columnTruncateLength=e.columnTruncateLength||this.fftSize,this.audioTrackConstraints=e.audioTrackConstraints,this.smoothingTimeConstant=e.smoothingTimeConstant||0,this.includeSpectrogram=e.includeSpectrogram!==!1,this.includeWaveform=e.includeWaveform===!0,!this.includeSpectrogram&&!this.includeWaveform)throw new Error("Both includeSpectrogram and includeWaveform are false. At least one type of data should be returned.")}summary(){return"microphone"}static async create(e={}){if(W().get("IS_NODE"))throw new Error("microphone API is only supported in browser environment.");let t=new Ih(e);return await t.start(),t}async start(){try{this.stream=await navigator.mediaDevices.getUserMedia({audio:this.audioTrackConstraints==null?!0:this.audioTrackConstraints,video:!1})}catch(n){throw new Error(`Error thrown while initializing video stream: ${n.message}`)}if(!this.stream)throw new Error("Could not obtain audio from microphone.");let e=window.AudioContext||window.webkitAudioContext;if(this.audioContext=new e,!this.sampleRateHz)this.sampleRateHz=this.audioContext.sampleRate;else if(this.audioContext.sampleRate!==this.sampleRateHz)throw new Error(`Mismatch in sampling rate: Expected: ${this.sampleRateHz}; Actual: ${this.audioContext.sampleRate}`);let t=this.audioContext.createMediaStreamSource(this.stream);this.analyser=this.audioContext.createAnalyser(),this.analyser.fftSize=this.fftSize*2,this.analyser.smoothingTimeConstant=this.smoothingTimeConstant,t.connect(this.analyser),this.freqData=new Float32Array(this.fftSize),this.timeData=new Float32Array(this.fftSize)}async next(){if(this.isClosed)return{value:null,done:!0};let e,t,n=await this.getAudioData();if(this.includeSpectrogram){let o=this.flattenQueue(n.freqDataQueue);e=this.getTensorFromAudioDataArray(o,[this.numFrames,this.columnTruncateLength,1])}if(this.includeWaveform){let o=this.flattenQueue(n.timeDataQueue);t=this.getTensorFromAudioDataArray(o,[this.numFrames*this.fftSize,1])}return{value:{spectrogram:e,waveform:t},done:!1}}async capture(){return(await this.next()).value}async getAudioData(){let e=[],t=[],n=0;return new Promise(o=>{let s=setInterval(()=>{this.includeSpectrogram&&(this.analyser.getFloatFrequencyData(this.freqData),this.freqData[0]===-Infinity&&o({freqDataQueue:e,timeDataQueue:t}),e.push(this.freqData.slice(0,this.columnTruncateLength))),this.includeWaveform&&(this.analyser.getFloatTimeDomainData(this.timeData),t.push(this.timeData.slice())),++n===this.numFrames&&(clearInterval(s),o({freqDataQueue:e,timeDataQueue:t}))},this.fftSize/this.sampleRateHz*1e3)})}stop(){this.isClosed||(this.isClosed=!0,this.analyser.disconnect(),this.audioContext.close(),this.stream!=null&&this.stream.getTracks().length>0&&this.stream.getTracks()[0].stop())}toArray(){throw new Error("Can not convert infinite audio stream to array.")}getSampleRate(){return this.sampleRateHz}flattenQueue(e){let t=e[0].length,n=new Float32Array(e.length*t);return e.forEach((o,s)=>n.set(o,s*t)),n}getTensorFromAudioDataArray(e,t){let n=new Float32Array(y.sizeFromShape(t));return n.set(e,n.length-e.length),vr(n,t)}};var Nh=class extends Yt{constructor(e,t){super();if(this.webcamVideoElement=e,this.webcamConfig=t,this.isClosed=!0,this.resize=!1,this.needToResize())if(this.resize=!0,this.cropSize=[this.webcamConfig.resizeHeight,this.webcamConfig.resizeWidth],this.cropBoxInd=Ct([0],"int32"),this.webcamConfig.centerCrop){let n=this.webcamConfig.resizeWidth*1/this.webcamVideoElement.width,o=this.webcamConfig.resizeHeight*1/this.webcamVideoElement.height,s=(1-n)/2,a=(1-o)/2,i=s+n,l=o+a;this.cropBox=ii([a,s,l,i],[1,4])}else this.cropBox=ii([0,0,1,1],[1,4])}summary(){return"webcam"}static async create(e,t={}){if(W().get("IS_NODE"))throw new Error("tf.data.webcam is only supported in browser environment.");if(!e){if(e=document.createElement("video"),!t.resizeWidth||!t.resizeHeight)throw new Error("Please provide webcam video element, or resizeWidth and resizeHeight to create a hidden video element.");e.width=t.resizeWidth,e.height=t.resizeHeight}let n=new Nh(e,t);return await n.start(),n}async start(){this.webcamConfig.facingMode&&y.assert(this.webcamConfig.facingMode==="user"||this.webcamConfig.facingMode==="environment",()=>`Invalid webcam facing mode: ${this.webcamConfig.facingMode}. Please provide 'user' or 'environment'`);try{this.stream=await navigator.mediaDevices.getUserMedia({video:{deviceId:this.webcamConfig.deviceId,facingMode:this.webcamConfig.facingMode?this.webcamConfig.facingMode:"user",width:this.webcamVideoElement.width,height:this.webcamVideoElement.height}})}catch(e){throw e.message=`Error thrown while initializing video stream: ${e.message}`,e}if(!this.stream)throw new Error("Could not obtain video from webcam.");try{this.webcamVideoElement.srcObject=this.stream}catch(e){console.log(e),this.webcamVideoElement.src=window.URL.createObjectURL(this.stream)}return this.webcamVideoElement.play(),this.isClosed=!1,new Promise(e=>{this.webcamVideoElement.onloadedmetadata=()=>{e()}})}async next(){if(this.isClosed)return{value:null,done:!0};let e;try{e=cg.fromPixels(this.webcamVideoElement)}catch(t){throw new Error(`Error thrown converting video to pixels: ${JSON.stringify(t)}`)}if(this.resize)try{return{value:this.cropAndResizeFrame(e),done:!1}}catch(t){throw new Error(`Error thrown cropping the video: ${t.message}`)}finally{e.dispose()}else return{value:e,done:!1}}needToResize(){return!!(this.webcamConfig.resizeWidth&&this.webcamConfig.resizeHeight&&(this.webcamVideoElement.width!==this.webcamConfig.resizeWidth||this.webcamVideoElement.height!==this.webcamConfig.resizeHeight))}cropAndResizeFrame(e){return V(()=>{let t=cr(oe(e,"float32"),0),n;n=ai.cropAndResize(t,this.cropBox,this.cropBoxInd,this.cropSize,"bilinear");let o=n.shape;return L(n,o.slice(1))})}async capture(){return(await this.next()).value}stop(){this.stream.getTracks().forEach(t=>t.stop());try{this.webcamVideoElement.srcObject=null}catch(t){console.log(t),this.webcamVideoElement.src=null}this.isClosed=!0}toArray(){throw new Error("Can not convert infinite video stream to array.")}};var Sh=class{};var Ay=class extends Yt{split(e){return new kV(this,e)}},kV=class extends Ay{constructor(e,t){super();this.upstream=e,this.impl=new vV(e,t)}summary(){return this.impl.summary()}async next(){return this.impl.next()}},vV=class extends cm{constructor(e,t){super();this.upstream=e,this.separator=t,this.carryover=""}summary(){return`${this.upstream.summary()} -> Split('${this.separator}')`}async pump(){let e=await this.upstream.next();if(e.done)return this.carryover===""?!1:(this.outputQueue.push(this.carryover),this.carryover="",!0);let t=e.value.split(this.separator);t[0]=this.carryover+t[0];for(let n of t.slice(0,-1))this.outputQueue.push(n);return this.carryover=t[t.length-1],!0}};var RI=class extends Yt{decodeUTF8(){return new IV(this)}},IV=class extends Ay{constructor(e){super();this.upstream=e,this.impl=new NV(e)}summary(){return this.impl.summary()}async next(){return this.impl.next()}},NV=class extends cm{constructor(e){super();if(this.upstream=e,W().get("IS_BROWSER"))this.decoder=new TextDecoder("utf-8");else{let{StringDecoder:t}=CV();this.decoder=new t("utf8")}}summary(){return`${this.upstream.summary()} -> Utf8`}async pump(){let e=await this.upstream.next(),t;if(e.done)return!1;t=e.value;let n;return W().get("IS_BROWSER")?n=this.decoder.decode(t,{stream:!0}):n=this.decoder.write(Buffer.from(t.buffer)),this.outputQueue.push(n),!0}};var Th=class extends RI{constructor(e,t={}){super();this.file=e,this.options=t,y.assert(e instanceof Uint8Array||(W().get("IS_BROWSER")?e instanceof File||e instanceof Blob:!1),()=>"FileChunkIterator only supports File, Blob and Uint8Array right now."),this.offset=t.offset||0,this.chunkSize=t.chunkSize||1024*1024}summary(){return`FileChunks ${this.file}`}async next(){return this.offset>=(this.file instanceof Uint8Array?this.file.byteLength:this.file.size)?{value:null,done:!0}:{value:await new Promise((t,n)=>{let o=this.offset+this.chunkSize;if(this.file instanceof Uint8Array)t(new Uint8Array(this.file.slice(this.offset,o)));else{let s=new FileReader;s.onload=i=>{let l=s.result;if(l instanceof ArrayBuffer&&(l=new Uint8Array(l)),!(l instanceof Uint8Array))return n(new TypeError("FileReader returned unknown type."));t(l)},s.onabort=i=>n(new Error("Aborted")),s.onerror=i=>n(new Error(i.type));let a=this.file.slice(this.offset,o);s.readAsArrayBuffer(a)}this.offset=o}),done:!1}}};async function SV(r,e={}){let t,n;typeof r=="string"?t=r:(t=r.url,n=Kne(r));let o=await y.fetch(t,n);if(o.ok){let s=new Uint8Array(await o.arrayBuffer());return new Th(s,e)}else throw new Error(o.statusText)}var Kne=r=>({method:r.method,headers:r.headers,body:r.body,mode:r.mode,credentials:r.credentials,cache:r.cache,redirect:r.redirect,referrer:r.referrer,integrity:r.integrity});function Ey(r){return typeof r=="string"&&r.substr(0,7)==="file://"}var Ah=class extends Sh{constructor(e,t={}){super();this.input=e,this.options=t}async iterator(){if(Ey(this.input)&&W().get("IS_NODE")){let e=dc("fs");this.input=e.readFileSync(this.input.substr(7))}return new Th(this.input,this.options)}};var Eh=class extends Sh{constructor(e,t={}){super();this.url=e,this.fileOptions=t}async iterator(){return Ey(this.url)?new Ah(this.url,this.fileOptions).iterator():SV(this.url,this.fileOptions)}};function TV(r,e={}){return new Ch(new Eh(r),e)}function AV(r){let e=_h(r);return kn(async()=>e)}function EV(r){return kn(async()=>{let e=await r();return _h(()=>e.next())})}async function DV(r,e){return Nh.create(r,e)}async function $V(r){return Ih.create(r)}var RV="3.6.0";var B4t={tfjs:(km==null?void 0:km.version)||void 0,"tfjs-core":(vm==null?void 0:vm.version)||void 0,"tfjs-data":(Cm==null?void 0:Cm.version)||void 0,"tfjs-layers":(Im==null?void 0:Im.version)||void 0,"tfjs-converter":(Nm==null?void 0:Nm.version)||void 0,"tfjs-backend-cpu":X_||void 0,"tfjs-backend-webgl":iv||void 0,"tfjs-backend-wasm":GC||void 0};export{Bs as Abs,ki as Acos,vi as Acosh,fp as AdadeltaOptimizer,dp as AdagradOptimizer,hp as AdamOptimizer,gp as AdamaxOptimizer,Pn as Add,_o as AddN,Ci as All,Ii as Any,ko as ArgMax,Ya as ArgMin,Ni as Asin,Si as Asinh,Ti as Atan,Ei as Atan2,Ai as Atanh,vo as AvgPool,Za as AvgPool3D,_c as AvgPool3DGrad,wc as AvgPoolGrad,Dx as BackendWasm,Co as BatchMatMul,Ja as BatchToSpaceND,kc as Bincount,TS as BroadcastTo,j0 as Callback,r0 as CallbackList,qn as Cast,Io as Ceil,Kn as ClipByValue,vc as Complex,Qa as ComplexAbs,Vs as Concat,No as Conv2D,Cc as Conv2DBackpropFilter,So as Conv2DBackpropInput,el as Conv3D,Ic as Conv3DBackpropFilterV2,Nc as Conv3DBackpropInputV2,To as Cos,Di as Cosh,$i as CropAndResize,Ao as Cumsum,o0 as CustomCallback,Xa as DataStorage,Sc as DenseBincount,Ri as DepthToSpace,Eo as DepthwiseConv2dNative,Tc as DepthwiseConv2dNativeBackpropFilter,Ac as DepthwiseConv2dNativeBackpropInput,Ec as Diag,tl as Dilation2D,$m as Dilation2DBackpropFilter,Dm as Dilation2DBackpropInput,iw as ENV,U0 as EarlyStopping,Dc as Einsum,Fi as Elu,$c as EluGrad,eg as Environment,Pi as Equal,Oi as Erf,$o as Exp,Gs as ExpandDims,Mi as Expm1,Rc as FFT,rl as Fill,Li as FlipLeftRight,Ro as Floor,Fo as FloorDiv,Rm as FromPixels,Oo as FusedBatchNorm,ei as FusedConv2D,ti as FusedDepthwiseConv2D,px as GPGPUContext,zi as GatherNd,Ws as GatherV2,bI as GraphModel,Bi as Greater,Po as GreaterEqual,n0 as History,Fc as IFFT,Xn as Identity,Oc as Imag,Nt as InputSpec,Vi as IsFinite,Gi as IsInf,Wi as IsNan,Ms as KernelBackend,nl as LRN,Mc as LRNGrad,Ux as LayerVariable,jn as LayersModel,Mo as LeakyRelu,ji as Less,Ui as LessEqual,Pc as LinSpace,Lo as Log,Hi as Log1p,AS as LogSoftmax,qi as LogicalAnd,Hl as LogicalNot,ql as LogicalOr,Au as MathBackendCPU,Lu as MathBackendWebGL,zo as Max,Vo as MaxPool,ol as MaxPool3D,zc as MaxPool3DGrad,Lc as MaxPoolGrad,Bc as MaxPoolWithArgmax,Bo as Maximum,Go as Mean,Wo as Min,jo as Minimum,Uo as MirrorPad,Ki as Mod,xp as MomentumOptimizer,Vc as Multinomial,Ho as Multiply,js as Neg,Yi as NonMaxSuppressionV3,Zi as NonMaxSuppressionV4,Ji as NonMaxSuppressionV5,Xi as NotEqual,WS as OP_SCOPE_SUFFIX,qo as OneHot,Us as OnesLike,Br as Optimizer,Hs as Pack,Ko as PadV2,toe as Pool,Xo as Pow,Yo as Prelu,Qi as Prod,yp as RMSPropOptimizer,$n as RNN,sl as Range,fw as Rank,Gc as Real,Do as RealDiv,ea as Reciprocal,Wt as Reduction,Zo as Relu,Qo as Relu6,qs as Reshape,Jo as ResizeBilinear,jc as ResizeBilinearGrad,il as ResizeNearestNeighbor,Wc as ResizeNearestNeighborGrad,es as Reverse,ua as RotateWithOffset,ts as Round,rs as Rsqrt,xl as SGDOptimizer,ta as ScatterNd,Ks as Select,ra as Selu,Ba as Sequential,os as Sigmoid,oa as Sign,ns as Sin,na as Sinh,Xs as Slice,as as Softmax,sa as Softplus,al as SpaceToBatchND,Uc as SparseFillEmptyRows,Hc as SparseReshape,qc as SparseToDense,Ys as SplitV,ss as Sqrt,ll as Square,ls as SquaredDifference,Yn as Step,ia as StridedSlice,us as Sub,is as Sum,sn as SymbolicTensor,cs as Tan,ps as Tanh,Pe as Tensor,ut as TensorBuffer,Mn as Tile,aa as TopK,la as Transform,ms as Transpose,Kc as Unique,Zs as Unpack,ul as UnsortedSegmentSum,pl as Variable,Js as ZerosLike,Qs as _FusedMatMul,Tt as abs,Gm as acos,Wm as acosh,J as add,Gw as addN,tu as all,dl as any,hl as argMax,jm as argMin,Um as asin,Hm as asinh,qm as atan,Km as atan2,Xm as atanh,ha as avgPool,Ym as avgPool3d,LT as backend,C as backend_util,Gj as basicLSTMCell,Jn as batchNorm,Hw as batchNorm2d,qw as batchNorm3d,Kw as batchNorm4d,ga as batchToSpaceND,Zm as bincount,IIe as booleanMaskAsync,xa as broadcastTo,cg as browser,Ce as buffer,lne as callbacks,oe as cast,Jm as ceil,ur as clipByValue,Ln as clone,Nn as complex,Qe as concat,Xw as concat1d,Yw as concat2d,Zw as concat3d,Jw as concat4d,Qz as constraints,ou as conv1d,Jr as conv2d,su as conv2dTranspose,Qm as conv3d,Qw as conv3dTranspose,aoe as copyRegisteredKernels,ya as cos,iu as cosh,Ng as cosineWindow,au as cumsum,Qr as customGrad,FI as data,e_ as denseBincount,Vw as deprecationWarn,ef as depthToSpace,xs as depthwiseConv2d,cne as deregisterOp,Jl as device_util,g4 as diag,tf as dilation2d,Lae as disableDeprecationWarnings,Ee as dispose,zae as disposeVariables,pe as div,rf as divNoNan,t_ as dot,l1 as dropout,r_ as einsum,ys as elu,Mae as enableDebugMode,Pae as enableProdMode,u1 as enclosingPowerOfTwo,hs as engine,W as env,Sn as equal,nf as erf,Qt as exp,cr as expandDims,of as expm1,ip as eye,Na as fft,bs as fill,Uae as findBackend,Hae as findBackendFactory,ws as floor,eu as floorDiv,XD as forceHalfFloat,no as fused,Qn as gather,i1 as gatherND,pg as gather_util,Wae as getBackend,uw as getGradient,Om as getKernel,tg as getKernelsForBackend,J2 as gpgpu_util,H4 as grad,q4 as grads,Vt as greater,hn as greaterEqual,si as ifft,lu as imag,ai as image,ONe as inTopKAsync,w3 as initializers,f0 as input,Er as io,_u as irfft,n_ as isFinite,o_ as isInf,sf as isNaN,$t as keep,Fr as kernel_impls,tB as layers,ba as leakyRelu,uu as less,gn as lessEqual,G1 as linalg,s_ as linspace,Pne as loadGraphModel,wte as loadLayersModel,af as localResponseNormalization,pr as log,cu as log1p,i_ as logSigmoid,pu as logSoftmax,uf as logSumExp,yr as logicalAnd,wa as logicalNot,mu as logicalOr,c_ as logicalXor,e$e as losses,ze as matMul,_T as math,mr as max,_a as maxPool,cf as maxPool3d,p_ as maxPoolWithArgmax,en as maximum,dt as mean,Vm as memory,hU as meshgrid,rB as metrics,oi as min,_s as minimum,pf as mirrorPad,mf as mod,yte as model,nB as models,ap as moments,ZIe as movingAverage,O as mul,vU as multiRNNCell,m_ as multinomial,qe as neg,vf as nextFrame,pp as norm,ro as notEqual,ds as oneHot,er as ones,sr as onesLike,S as op,TU as outerProduct,Lr as pad,DU as pad1d,RU as pad2d,OU as pad3d,MU as pad4d,f_ as pool,zr as pow,va as prelu,Dw as print,fu as prod,Bae as profile,HU as rand,eH as randomGamma,kg as randomNormal,ks as randomUniform,Ca as range,Gae as ready,gl as real,ff as reciprocal,np as registerBackend,_te as registerCallbackConstructor,ES as registerGradient,Kl as registerKernel,une as registerOp,oB as regularizers,$r as relu,hu as relu6,jae as removeBackend,L as reshape,Kt as reverse,uH as reverse1d,pH as reverse2d,fH as reverse3d,hH as reverse4d,Sa as rfft,gu as round,xu as rsqrt,ue as scalar,o1 as scatterND,fg as scatter_util,yu as selu,df as separableConv2d,bte as sequential,Q as serialization,cj as setBackend,qae as setPlatform,UQ as setWasmPath,HQ as setWasmPaths,pk as setWebGLContext,C_ as setdiff1dAsync,qg as shared,Dr as sigmoid,hf as sign,TDe as signal,bu as sin,wu as sinh,Re as slice,gf as slice1d,vg as slice2d,xf as slice3d,up as slice4d,or as slice_util,Ia as softmax,eo as softplus,ka as spaceToBatchND,W1 as sparse,Ig as sparseToDense,vDe as spectral,tr as split,gt as sqrt,Me as square,ku as squaredDifference,Tn as squeeze,Gt as stack,vs as step,yf as stridedSlice,le as sub,fe as sum,Yl as sumOutType,bf as tan,gs as tanh,vr as tensor,Ct as tensor1d,ii as tensor2d,Ow as tensor3d,VH as tensor4d,GH as tensor5d,WH as tensor6d,Zn as tensor_util,MT as test_util,V as tidy,Bn as tile,Vae as time,wf as topk,Tu as train,Ue as transpose,vu as truncatedNormal,cp as unique,ioe as unregisterGradient,soe as unregisterKernel,_f as unsortedSegmentSum,fr as unstack,lr as upcastType,y as util,K4 as valueAndGrad,X4 as valueAndGrads,I_ as variable,yg as variableGrads,B4t as version,Mne as version_converter,uj as version_core,X_ as version_cpu,yd as version_layers,GC as version_wasm,iv as version_webgl,Q7e as webgl,Z2 as webgl_util,vt as where,kf as whereAsync,ht as zeros,Ie as zerosLike};
|
|
/**
|
|
* @license
|
|
* Copyright 2017 Google LLC. All Rights Reserved.
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2018 Google LLC
|
|
*
|
|
* Use of this source code is governed by an MIT-style
|
|
* license that can be found in the LICENSE file or at
|
|
* https://opensource.org/licenses/MIT.
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2018 Google LLC. All Rights Reserved.
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
*
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2018 Google LLC. All Rights Reserved.
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2019 Google LLC
|
|
*
|
|
* Use of this source code is governed by an MIT-style
|
|
* license that can be found in the LICENSE file or at
|
|
* https://opensource.org/licenses/MIT.
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2019 Google LLC. All Rights Reserved.
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
*
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2019 Google LLC. All Rights Reserved.
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2020 Google Inc. All Rights Reserved.
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2020 Google LLC
|
|
*
|
|
* Use of this source code is governed by an MIT-style
|
|
* license that can be found in the LICENSE file or at
|
|
* https://opensource.org/licenses/MIT.
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2020 Google LLC. All Rights Reserved.
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2020 Google LLC. All Rights Reserved.
|
|
* Licensed under the Apache License, Version 2.0 (the License);
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an AS IS BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2021 Google LLC. All Rights Reserved.
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
* =============================================================================
|
|
*/
|
|
/**
|
|
* @license
|
|
* Copyright 2021 Google LLC. All Rights Reserved.
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* https://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
* =============================================================================
|
|
*/
|
|
/** @license See the LICENSE file. */
|
|
//# sourceMappingURL=tfjs.esm.js.map
|