@genai-fi/nanogpt 0.9.1 → 0.10.1

package/dist/index-BzFyqcy-.js

@@ -1,4457 +0,0 @@
-import { g as Gt } from "./index-D5v913EJ.js";
-import { p as Q } from "./index-xuotMAFm.js";
-import { B as pt } from "./index-Tf7vU29b.js";
-/**
- * @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.
- * =============================================================================
- */
-const Fe = 1e-7, Re = 1e-4;
-class Rs {
-  constructor(t, n) {
-    this.backend = t, this.dataMover = n, this.data = /* @__PURE__ */ new WeakMap(), this.dataIdsCount = 0;
-  }
-  get(t) {
-    return this.data.has(t) || this.dataMover.moveData(this.backend, t), this.data.get(t);
-  }
-  set(t, n) {
-    this.dataIdsCount++, this.data.set(t, n);
-  }
-  has(t) {
-    return this.data.has(t);
-  }
-  delete(t) {
-    return this.dataIdsCount--, this.data.delete(t);
-  }
-  numDataIds() {
-    return this.dataIdsCount;
-  }
-}
-class xe {
-  refCount(t) {
-    return B("refCount");
-  }
-  incRef(t) {
-    return B("incRef");
-  }
-  timerAvailable() {
-    return !0;
-  }
-  time(t) {
-    return B("time");
-  }
-  read(t) {
-    return B("read");
-  }
-  readSync(t) {
-    return B("readSync");
-  }
-  readToGPU(t, n) {
-    return B("readToGPU");
-  }
-  numDataIds() {
-    return B("numDataIds");
-  }
-  disposeData(t, n) {
-    return B("disposeData");
-  }
-  write(t, n, s) {
-    return B("write");
-  }
-  move(t, n, s, r, a) {
-    return B("move");
-  }
-  createTensorFromGPUData(t, n, s) {
-    return B("createTensorFromGPUData");
-  }
-  memory() {
-    return B("memory");
-  }
-  /** Returns the highest precision for floats in bits (e.g. 16 or 32) */
-  floatPrecision() {
-    return B("floatPrecision");
-  }
-  /** Returns the smallest representable number.  */
-  epsilon() {
-    return this.floatPrecision() === 32 ? Fe : Re;
-  }
-  dispose() {
-    return B("dispose");
-  }
-}
-function B(e) {
-  throw new Error(`'${e}' not yet implemented or not found in the registry. This kernel may not be supported by the tfjs backend you have chosen`);
-}
-/**
- * @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.
- * =============================================================================
- */
-function xs(e) {
-  let t = e.length, n = 0;
-  for (; t > 0; )
-    n = Math.random() * t | 0, t--, Ne(e, t, n);
-}
-function Ns(e, t, n) {
-  return Math.max(e, Math.min(t, n));
-}
-function $s(e) {
-  return e % 2 === 0 ? e : e + 1;
-}
-function Ne(e, t, n) {
-  const s = e[t];
-  e[t] = e[n], e[n] = s;
-}
-function Cs(e) {
-  let t = 0;
-  for (let n = 0; n < e.length; n++)
-    t += e[n];
-  return t;
-}
-function p(e, t) {
-  if (!e)
-    throw new Error(typeof t == "string" ? t : t());
-}
-function te(e, t, n = "") {
-  p(xt(e, t), () => n + ` Shapes ${e} and ${t} must match`);
-}
-function Ps(e) {
-  p(e != null, () => "The input to the tensor constructor must be a non-null value.");
-}
-function z(e) {
-  if (e.length === 0)
-    return 1;
-  let t = e[0];
-  for (let n = 1; n < e.length; n++)
-    t *= e[n];
-  return t;
-}
-function Ls(e) {
-  return e.length === 0;
-}
-function xt(e, t) {
-  if (e === t)
-    return !0;
-  if (e == null || t == null || e.length !== t.length)
-    return !1;
-  for (let n = 0; n < e.length; n++)
-    if (e[n] !== t[n])
-      return !1;
-  return !0;
-}
-function $e(e) {
-  return e % 1 === 0;
-}
-function _s(e) {
-  const t = Math.ceil(Math.sqrt(e));
-  return [t, Math.ceil(e / t)];
-}
-function ht(e, t) {
-  return t <= e.length ? e : e + " ".repeat(t - e.length);
-}
-function Os(e, t = (r) => 0, n, s) {
-  return new Promise((r, a) => {
-    let o = 0;
-    const i = () => {
-      if (e()) {
-        r();
-        return;
-      }
-      o++;
-      const c = t(o);
-      if (n != null && o >= n) {
-        a();
-        return;
-      }
-      s != null ? s(i, c) : setTimeout(i, c);
-    };
-    i();
-  });
-}
-function Gs(e, t) {
-  let n = 1, s = -1;
-  for (let a = 0; a < e.length; ++a)
-    if (e[a] >= 0)
-      n *= e[a];
-    else if (e[a] === -1) {
-      if (s !== -1)
-        throw Error(`Shapes can only have 1 implicit size. Found -1 at dim ${s} and dim ${a}`);
-      s = a;
-    } else if (e[a] < 0)
-      throw Error(`Shapes can not be < 0. Found ${e[a]} at dim ${a}`);
-  if (s === -1) {
-    if (t > 0 && t !== n)
-      throw Error(`Size(${t}) must match the product of shape ${e}`);
-    return e;
-  }
-  if (n === 0)
-    throw Error(`Cannot infer the missing size in [${e}] when there are 0 elements`);
-  if (t % n !== 0)
-    throw Error(`The implicit shape can't be a fractional number. Got ${t} / ${n}`);
-  const r = e.slice();
-  return r[s] = t / n, r;
-}
-function Ce(e, t) {
-  const n = t.length;
-  return e = e == null ? t.map((s, r) => r) : [].concat(e), p(e.every((s) => s >= -n && s < n), () => `All values in axis param must be in range [-${n}, ${n}) but got axis ${e}`), p(e.every((s) => $e(s)), () => `All values in axis param must be integers but got axis ${e}`), e.map((s) => s < 0 ? n + s : s);
-}
-function Us(e, t) {
-  const n = [], s = [], r = t != null && Array.isArray(t) && t.length === 0, a = t == null || r ? null : Ce(t, e).sort();
-  let o = 0;
-  for (let i = 0; i < e.length; ++i) {
-    if (a != null) {
-      if (a[o] === i && e[i] !== 1)
-        throw new Error(`Can't squeeze axis ${i} since its dim '${e[i]}' is not 1`);
-      (a[o] == null || a[o] > i) && e[i] === 1 && (n.push(e[i]), s.push(i)), a[o] <= i && o++;
-    }
-    e[i] !== 1 && (n.push(e[i]), s.push(i));
-  }
-  return { newShape: n, keptDims: s };
-}
-function zs(e, t) {
-  return ee(e, t);
-}
-function ee(e, t) {
-  let n = null;
-  if (e == null || e === "float32")
-    n = new Float32Array(t);
-  else if (e === "int32")
-    n = new Int32Array(t);
-  else if (e === "bool")
-    n = new Uint8Array(t);
-  else if (e === "string")
-    n = new Array(t);
-  else
-    throw new Error(`Unknown data type ${e}`);
-  return n;
-}
-function Pe(e, t) {
-  for (let n = 0; n < e.length; n++) {
-    const s = e[n];
-    if (isNaN(s) || !isFinite(s))
-      throw Error(`A tensor of type ${t} being uploaded contains ${s}.`);
-  }
-}
-function Le(e) {
-  return e === "bool" || e === "complex64" || e === "float32" || e === "int32" || e === "string";
-}
-function Vs(e, t) {
-  return !(t === "complex64" || t === "float32" && e !== "complex64" || t === "int32" && e !== "float32" && e !== "complex64" || t === "bool" && e === "bool");
-}
-function It(e) {
-  if (e === "float32" || e === "int32")
-    return 4;
-  if (e === "complex64")
-    return 8;
-  if (e === "bool")
-    return 1;
-  throw new Error(`Unknown dtype ${e}`);
-}
-function _e(e) {
-  if (e == null)
-    return 0;
-  let t = 0;
-  return e.forEach((n) => t += n.length), t;
-}
-function Nt(e) {
-  return typeof e == "string" || e instanceof String;
-}
-function Oe(e) {
-  return typeof e == "boolean";
-}
-function Ge(e) {
-  return typeof e == "number";
-}
-function yt(e) {
-  return Array.isArray(e) ? yt(e[0]) : e instanceof Float32Array ? "float32" : e instanceof Int32Array || e instanceof Uint8Array || e instanceof Uint8ClampedArray ? "int32" : Ge(e) ? "float32" : Nt(e) ? "string" : Oe(e) ? "bool" : "float32";
-}
-function lt(e) {
-  return !!(e && e.constructor && e.call && e.apply);
-}
-function Ws(e, t) {
-  for (let n = t; n < e; ++n)
-    if (e % n === 0)
-      return n;
-  return e;
-}
-function $t(e) {
-  const t = e.length;
-  if (t < 2)
-    return [];
-  const n = new Array(t - 1);
-  n[t - 2] = e[t - 1];
-  for (let s = t - 3; s >= 0; --s)
-    n[s] = n[s + 1] * e[s + 1];
-  return n;
-}
-function ne(e, t, n, s = !1) {
-  const r = new Array();
-  if (t.length === 1) {
-    const a = t[0] * (s ? 2 : 1);
-    for (let o = 0; o < a; o++)
-      r[o] = n[e + o];
-  } else {
-    const a = t[0], o = t.slice(1), i = o.reduce((c, l) => c * l) * (s ? 2 : 1);
-    for (let c = 0; c < a; c++)
-      r[c] = ne(e + c * i, o, n, s);
-  }
-  return r;
-}
-function ct(e, t, n = !1) {
-  if (e.length === 0)
-    return t[0];
-  const s = e.reduce((r, a) => r * a) * (n ? 2 : 1);
-  if (s === 0)
-    return [];
-  if (s !== t.length)
-    throw new Error(`[${e}] does not match the input size ${t.length}${n ? " for a complex tensor" : ""}.`);
-  return ne(0, e, t, n);
-}
-function js(e, t) {
-  if (Array.isArray(e))
-    return e;
-  if (t === "float32")
-    return e instanceof Float32Array ? e : new Float32Array(e);
-  if (t === "int32")
-    return e instanceof Int32Array ? e : new Int32Array(e);
-  if (t === "bool" || t === "string")
-    return Uint8Array.from(new Int32Array(e));
-  throw new Error(`Unknown dtype ${t}`);
-}
-function Ue(e, t) {
-  const n = se(e, t);
-  for (let s = 0; s < n.length; s++)
-    n[s] = 1;
-  return n;
-}
-function se(e, t) {
-  if (t == null || t === "float32" || t === "complex64")
-    return new Float32Array(e);
-  if (t === "int32")
-    return new Int32Array(e);
-  if (t === "bool")
-    return new Uint8Array(e);
-  throw new Error(`Unknown data type ${t}`);
-}
-function qs(e, t) {
-  const n = e.reduce((s, r) => s * r, 1);
-  if (t == null || t === "float32")
-    return ct(e, new Float32Array(n));
-  if (t === "int32")
-    return ct(e, new Int32Array(n));
-  if (t === "bool")
-    return ct(e, new Uint8Array(n));
-  throw new Error(`Unknown data type ${t}`);
-}
-function Ct(e) {
-  e.forEach((t) => {
-    p(Number.isInteger(t) && t >= 0, () => `Tensor must have a shape comprised of positive integers but got shape [${e}].`);
-  });
-}
-function Ks(e, t, n) {
-  if (t === 0)
-    return 0;
-  if (t === 1)
-    return e[0];
-  let s = e[e.length - 1];
-  for (let r = 0; r < e.length - 1; ++r)
-    s += n[r] * e[r];
-  return s;
-}
-function Hs(e, t, n) {
-  if (t === 0)
-    return [];
-  if (t === 1)
-    return [e];
-  const s = new Array(t);
-  for (let r = 0; r < s.length - 1; ++r)
-    s[r] = Math.floor(e / n[r]), e -= s[r] * n[r];
-  return s[s.length - 1] = e, s;
-}
-function Pt(e) {
-  return e && e.then && typeof e.then == "function";
-}
-/**
- * @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.
- * =============================================================================
- */
-const Ut = "tfjsflags";
-class ze {
-  // tslint:disable-next-line: no-any
-  constructor(t) {
-    this.global = t, this.flags = {}, this.flagRegistry = {}, this.urlFlags = {}, this.getQueryParams = Ve, this.populateURLFlags();
-  }
-  setPlatform(t, n) {
-    this.platform != null && (S().getBool("IS_TEST") || S().getBool("PROD") || console.warn(`Platform ${this.platformName} has already been set. Overwriting the platform with ${t}.`)), this.platformName = t, this.platform = n;
-  }
-  registerFlag(t, n, s) {
-    if (this.flagRegistry[t] = { evaluationFn: n, setHook: s }, this.urlFlags[t] != null) {
-      const r = this.urlFlags[t];
-      S().getBool("IS_TEST") || S().getBool("PROD") || console.warn(`Setting feature override from URL ${t}: ${r}.`), this.set(t, r);
-    }
-  }
-  async getAsync(t) {
-    return t in this.flags ? this.flags[t] : (this.flags[t] = await this.evaluateFlag(t), this.flags[t]);
-  }
-  get(t) {
-    if (t in this.flags)
-      return this.flags[t];
-    const n = this.evaluateFlag(t);
-    if (Pt(n))
-      throw new Error(`Flag ${t} cannot be synchronously evaluated. Please use getAsync() instead.`);
-    return this.flags[t] = n, this.flags[t];
-  }
-  getNumber(t) {
-    return this.get(t);
-  }
-  getBool(t) {
-    return this.get(t);
-  }
-  getString(t) {
-    return this.get(t);
-  }
-  getFlags() {
-    return this.flags;
-  }
-  // For backwards compatibility.
-  get features() {
-    return this.flags;
-  }
-  set(t, n) {
-    if (this.flagRegistry[t] == null)
-      throw new Error(`Cannot set flag ${t} as it has not been registered.`);
-    this.flags[t] = n, this.flagRegistry[t].setHook != null && this.flagRegistry[t].setHook(n);
-  }
-  evaluateFlag(t) {
-    if (this.flagRegistry[t] == null)
-      throw new Error(`Cannot evaluate flag '${t}': no evaluation function found.`);
-    return this.flagRegistry[t].evaluationFn();
-  }
-  setFlags(t) {
-    this.flags = Object.assign({}, t);
-  }
-  reset() {
-    this.flags = {}, this.urlFlags = {}, this.populateURLFlags();
-  }
-  populateURLFlags() {
-    if (typeof this.global > "u" || typeof this.global.location > "u" || typeof this.global.location.search > "u")
-      return;
-    const t = this.getQueryParams(this.global.location.search);
-    Ut in t && t[Ut].split(",").forEach((s) => {
-      const [r, a] = s.split(":");
-      this.urlFlags[r] = je(r, a);
-    });
-  }
-}
-function Ve(e) {
-  const t = {};
-  return e.replace(/[?&]([^=?&]+)(?:=([^&]*))?/g, (n, ...s) => (We(t, s[0], s[1]), s.join("="))), t;
-}
-function We(e, t, n) {
-  e[decodeURIComponent(t)] = decodeURIComponent(n || "");
-}
-function je(e, t) {
-  const n = t.toLowerCase();
-  return n === "true" || n === "false" ? n === "true" : `${+n}` === n ? +n : t;
-}
-function S() {
-  return re;
-}
-let re = null;
-function qe(e) {
-  re = e;
-}
-/**
- * @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.
- * =============================================================================
- */
-let bt;
-function ae() {
-  if (bt == null) {
-    let e;
-    if (typeof window < "u")
-      e = window;
-    else if (typeof Gt < "u")
-      e = Gt;
-    else if (typeof Q < "u")
-      e = Q;
-    else if (typeof self < "u")
-      e = self;
-    else
-      throw new Error("Could not find a global object");
-    bt = e;
-  }
-  return bt;
-}
-function Ke() {
-  const e = ae();
-  return e._tfGlobals == null && (e._tfGlobals = /* @__PURE__ */ new Map()), e._tfGlobals;
-}
-function Lt(e, t) {
-  const n = Ke();
-  if (n.has(e))
-    return n.get(e);
-  {
-    const s = t();
-    return n.set(e, s), n.get(e);
-  }
-}
-const He = "Abs", Xs = "Acos", Js = "Acosh", oe = "Add", Ys = "AddN", Zs = "All", Qs = "Any", tr = "ArgMax", er = "ArgMin", nr = "Asin", sr = "Asinh", rr = "Atan", ar = "Atanh", or = "Atan2", ir = "AvgPool", cr = "AvgPoolGrad", lr = "AvgPool3D", ur = "AvgPool3DGrad", hr = "BatchMatMul", dr = "BatchToSpaceND", fr = "Bincount", gr = "BitwiseAnd", mr = "BroadcastTo", pr = "BroadcastArgs", ie = "Cast", yr = "Ceil", br = "ClipByValue", wr = "Complex", Xe = "ComplexAbs", Sr = "Concat", kr = "Conv2D", Ir = "Conv2DBackpropFilter", Tr = "Conv2DBackpropInput", Ar = "Conv3D", Er = "Conv3DBackpropFilterV2", vr = "Conv3DBackpropInputV2", Br = "Cos", Mr = "Cosh", Dr = "Cumprod", Fr = "Cumsum", Rr = "CropAndResize", xr = "DenseBincount", Nr = "DepthToSpace", $r = "DepthwiseConv2dNative", Cr = "DepthwiseConv2dNativeBackpropFilter", Pr = "DepthwiseConv2dNativeBackpropInput", Lr = "Diag", _r = "Dilation2D", Or = "Dilation2DBackpropInput", Gr = "Dilation2DBackpropFilter", Ur = "Draw", Je = "RealDiv", zr = "Einsum", Vr = "Elu", Wr = "EluGrad", jr = "Erf", qr = "Equal", Kr = "Exp", Hr = "ExpandDims", Xr = "Expm1", Jr = "FFT", Ye = "Fill", Yr = "FlipLeftRight", Zr = "Floor", Ze = "FloorDiv", Qr = "FusedBatchNorm", ta = "GatherV2", ea = "GatherNd", na = "Greater", sa = "GreaterEqual", ce = "Identity", ra = "IFFT", aa = "Imag", oa = "IsFinite", ia = "IsInf", ca = "IsNan", la = "LeakyRelu", ua = "Less", ha = "LessEqual", da = "LinSpace", fa = "Log", ga = "Log1p", ma = "LogicalAnd", pa = "LogicalNot", ya = "LogicalOr", ba = "LogSoftmax", wa = "LRN", Sa = "LRNGrad", ka = "Max", Qe = "Maximum", Ia = "MaxPool", Ta = "MaxPoolGrad", Aa = "MaxPool3D", Ea = "MaxPool3DGrad", va = "MaxPoolWithArgmax", Ba = "Mean", Ma = "Min", Da = "Minimum", Fa = "MirrorPad", Ra = "Mod", xa = "Multinomial", tn = "Multiply", Na = "Neg", $a = "NotEqual", Ca = "NonMaxSuppressionV3", Pa = "NonMaxSuppressionV4", La = "NonMaxSuppressionV5", _a = "OnesLike", Oa = "OneHot", Ga = "Pack", Ua = "PadV2", en = "Pow", za = "Prelu", Va = "Prod", Wa = "RaggedGather", ja = "RaggedRange", qa = "RaggedTensorToTensor", Ka = "Range", Ha = "Real", Xa = "Reciprocal", Ja = "Relu", Ya = "Reshape", Za = "ResizeNearestNeighbor", Qa = "ResizeNearestNeighborGrad", to = "ResizeBilinear", eo = "ResizeBilinearGrad", no = "Relu6", so = "Reverse", ro = "Round", ao = "Rsqrt", oo = "ScatterNd", io = "TensorScatterUpdate", co = "SearchSorted", lo = "Select", uo = "Selu", ho = "Slice", fo = "Sin", go = "Sinh", mo = "Sign", po = "Sigmoid", yo = "Softplus", nn = "Sqrt", bo = "Sum", wo = "SpaceToBatchND", So = "SplitV", ko = "Softmax", Io = "SparseFillEmptyRows", To = "SparseReshape", Ao = "SparseSegmentMean", Eo = "SparseSegmentSum", vo = "SparseToDense", Bo = "SquaredDifference", Mo = "Square", Do = "StaticRegexReplace", Fo = "StridedSlice", Ro = "StringNGrams", xo = "StringSplit", No = "StringToHashBucketFast", sn = "Sub", $o = "Tan", Co = "Tanh", Po = "Tile", Lo = "TopK", _o = "Transform", Oo = "Transpose", Go = "Unique", Uo = "Unpack", zo = "UnsortedSegmentSum", rn = "ZerosLike", Vo = "Step", Wo = "FromPixels", jo = "RotateWithOffset", qo = "_FusedMatMul", Ko = "FusedConv2D", Ho = "FusedDepthwiseConv2D";
-/**
- * @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.
- * =============================================================================
- */
-function O(...e) {
-  S().getBool("IS_TEST") || S().getBool("PROD") || console.warn(...e);
-}
-function Xo(...e) {
-  S().getBool("IS_TEST") || S().getBool("PROD") || console.log(...e);
-}
-/**
- * @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.
- * =============================================================================
- */
-const ft = Lt("kernelRegistry", () => /* @__PURE__ */ new Map()), Tt = Lt("gradRegistry", () => /* @__PURE__ */ new Map());
-function zt(e, t) {
-  const n = le(e, t);
-  return ft.get(n);
-}
-function Vt(e) {
-  return Tt.get(e);
-}
-function Wt(e) {
-  const t = ft.entries(), n = [];
-  for (; ; ) {
-    const { done: s, value: r } = t.next();
-    if (s)
-      break;
-    const [a, o] = r, [i] = a.split("_");
-    i === e && n.push(o);
-  }
-  return n;
-}
-function Jo(e) {
-  const { kernelName: t, backendName: n } = e, s = le(t, n);
-  ft.has(s) && O(`The kernel '${t}' for backend '${n}' is already registered`), ft.set(s, e);
-}
-function Yo(e) {
-  const { kernelName: t } = e;
-  Tt.has(t) && S().getBool("DEBUG") && O(`Overriding the gradient for '${t}'`), Tt.set(t, e);
-}
-function le(e, t) {
-  return `${t}_${e}`;
-}
-/**
- * @license
- * Copyright 2023 Google LLC.
- * 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.
- * =============================================================================
- */
-function ue(e) {
-  return e instanceof Float32Array || e instanceof Int32Array || e instanceof Uint8Array || e instanceof Uint8ClampedArray;
-}
-/**
- * @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.
- * =============================================================================
- */
-function Zo(e, t) {
-  return t === "string" ? he(e) : _t([e], t);
-}
-function an(e, t) {
-  return e instanceof Float32Array && t === "float32" || e instanceof Int32Array && t === "int32" || e instanceof Uint8Array && t === "bool";
-}
-function _t(e, t) {
-  if (t === "string")
-    throw new Error("Cannot convert a string[] to a TypedArray");
-  if (Array.isArray(e) && (e = ut(e)), S().getBool("DEBUG") && Pe(e, t), an(e, t))
-    return e;
-  if (t == null || t === "float32" || t === "complex64")
-    return new Float32Array(e);
-  if (t === "int32")
-    return new Int32Array(e);
-  if (t === "bool") {
-    const n = new Uint8Array(e.length);
-    for (let s = 0; s < n.length; ++s)
-      Math.round(e[s]) !== 0 && (n[s] = 1);
-    return n;
-  } else
-    throw new Error(`Unknown data type ${t}`);
-}
-function gt() {
-  return S().platform.now();
-}
-function he(e, t = "utf-8") {
-  return t = t || "utf-8", S().platform.encode(e, t);
-}
-function jt(e, t = "utf-8") {
-  return t = t || "utf-8", S().platform.decode(e, t);
-}
-function R(e) {
-  return S().platform.isTypedArray != null ? S().platform.isTypedArray(e) : ue(e);
-}
-function ut(e, t = [], n = !1) {
-  if (t == null && (t = []), typeof e == "boolean" || typeof e == "number" || typeof e == "string" || Pt(e) || e == null || R(e) && n)
-    t.push(e);
-  else if (Array.isArray(e) || R(e))
-    for (let s = 0; s < e.length; ++s)
-      ut(e[s], t, n);
-  else {
-    let s = -1;
-    for (const r of Object.keys(e))
-      /^([1-9]+[0-9]*|0)$/.test(r) && (s = Math.max(s, Number(r)));
-    for (let r = 0; r <= s; r++)
-      ut(e[r], t, n);
-  }
-  return t;
-}
-/**
- * @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.
- * =============================================================================
- */
-class on {
-  constructor(t, n) {
-    this.backendTimer = t, this.logger = n, n == null && (this.logger = new ln());
-  }
-  profileKernel(t, n, s) {
-    let r;
-    const a = () => {
-      r = s();
-    };
-    let o;
-    const i = gt();
-    if (this.backendTimer.timerAvailable())
-      o = this.backendTimer.time(a);
-    else {
-      a();
-      for (const l of r)
-        l.dataSync();
-      o = Promise.resolve({ kernelMs: gt() - i });
-    }
-    if (S().getBool("CHECK_COMPUTATION_FOR_ERRORS"))
-      for (let l = 0; l < r.length; l++) {
-        const u = r[l];
-        u.data().then((h) => {
-          cn(h, u.dtype, t);
-        });
-      }
-    return {
-      kernelName: t,
-      outputs: r,
-      inputs: n,
-      timeMs: o.then((l) => l.kernelMs),
-      extraInfo: o.then((l) => l.getExtraProfileInfo != null ? l.getExtraProfileInfo() : "")
-    };
-  }
-  logKernelProfile(t) {
-    const { kernelName: n, outputs: s, timeMs: r, inputs: a, extraInfo: o } = t;
-    s.forEach((i) => {
-      Promise.all([i.data(), r, o]).then((c) => {
-        this.logger.logKernelProfile(n, i, c[0], c[1], a, c[2]);
-      });
-    });
-  }
-}
-function cn(e, t, n) {
-  if (t !== "float32")
-    return !1;
-  for (let s = 0; s < e.length; s++) {
-    const r = e[s];
-    if (isNaN(r) || !isFinite(r))
-      return console.warn(`Found ${r} in the result of '${n}'`), !0;
-  }
-  return !1;
-}
-class ln {
-  logKernelProfile(t, n, s, r, a, o) {
-    const i = typeof r == "number" ? ht(`${r}ms`, 9) : r.error, c = ht(t, 25), l = n.rank, u = n.size, h = ht(n.shape.toString(), 14);
-    let d = "";
-    for (const m in a) {
-      const y = a[m];
-      if (y != null) {
-        const g = y.shape || n.shape, w = g.length;
-        d += `${m}: ${w}D ${w > 0 ? g : ""} `;
-      }
-    }
-    console.log(`%c${c}	%c${i}	%c${l}D ${h}	%c${u}	%c${d}	%c${o}`, "font-weight:bold", "color:red", "color:blue", "color: orange", "color: green", "color: steelblue");
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-function un(e, t, n) {
-  const s = {}, r = {};
-  for (let c = 0; c < t.length; c++)
-    s[t[c].id] = !0;
-  for (let c = 0; c < e.length; c++) {
-    const l = e[c], u = l.inputs;
-    for (const h in u) {
-      const d = u[h];
-      let m = !1;
-      for (let y = 0; y < t.length; y++)
-        if (s[d.id]) {
-          l.outputs.forEach((g) => s[g.id] = !0), m = !0, r[l.id] = !0;
-          break;
-        }
-      if (m)
-        break;
-    }
-  }
-  const a = {};
-  a[n.id] = !0;
-  const o = {};
-  for (let c = e.length - 1; c >= 0; c--) {
-    const l = e[c], u = l.inputs;
-    for (let h = 0; h < l.outputs.length; h++)
-      if (a[l.outputs[h].id]) {
-        for (const d in u)
-          a[u[d].id] = !0, o[l.id] = !0;
-        break;
-      }
-  }
-  const i = [];
-  for (let c = 0; c < e.length; c++) {
-    const l = e[c];
-    if (r[l.id] && o[l.id]) {
-      const u = {};
-      for (const d in l.inputs) {
-        const m = l.inputs[d];
-        s[m.id] && (u[d] = m);
-      }
-      const h = Object.assign({}, l);
-      h.inputs = u, h.outputs = l.outputs, i.push(h);
-    }
-  }
-  return i;
-}
-function hn(e, t, n, s) {
-  for (let r = t.length - 1; r >= 0; r--) {
-    const a = t[r], o = [];
-    if (a.outputs.forEach((c) => {
-      const l = e[c.id];
-      l != null ? o.push(l) : o.push(null);
-    }), a.gradient == null)
-      throw new Error(`Cannot compute gradient: gradient function not found for ${a.kernelName}.`);
-    const i = a.gradient(o);
-    for (const c in a.inputs) {
-      if (!(c in i))
-        throw new Error(`Cannot backprop through input ${c}. Available gradients found: ${Object.keys(i)}.`);
-      const l = n(() => i[c]());
-      if (l.dtype !== "float32")
-        throw new Error(`Error in gradient for op ${a.kernelName}. The gradient of input ${c} must have 'float32' dtype, but has '${l.dtype}'`);
-      const u = a.inputs[c];
-      if (!xt(l.shape, u.shape))
-        throw new Error(`Error in gradient for op ${a.kernelName}. The gradient of input '${c}' has shape '${l.shape}', which does not match the shape of the input '${u.shape}'`);
-      if (e[u.id] == null)
-        e[u.id] = l;
-      else {
-        const h = e[u.id];
-        e[u.id] = s(h, l), h.dispose();
-      }
-    }
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-const qt = 20, at = 3, wt = 7;
-function dn(e, t, n, s) {
-  const r = $t(t), a = fn(e, t, n, r), o = t.length, i = dt(e, t, n, r, a), c = ["Tensor"];
-  return s && (c.push(`  dtype: ${n}`), c.push(`  rank: ${o}`), c.push(`  shape: [${t}]`), c.push("  values:")), c.push(i.map((l) => "    " + l).join(`
-`)), c.join(`
-`);
-}
-function fn(e, t, n, s) {
-  const r = z(t), a = s[s.length - 1], o = new Array(a).fill(0), i = t.length, c = n === "complex64" ? it(e) : e;
-  if (i > 1)
-    for (let l = 0; l < r / a; l++) {
-      const u = l * a;
-      for (let h = 0; h < a; h++)
-        o[h] = Math.max(o[h], ot(c[u + h], 0, n).length);
-    }
-  return o;
-}
-function ot(e, t, n) {
-  let s;
-  return Array.isArray(e) ? s = `${parseFloat(e[0].toFixed(wt))} + ${parseFloat(e[1].toFixed(wt))}j` : Nt(e) ? s = `'${e}'` : n === "bool" ? s = de(e) : s = parseFloat(e.toFixed(wt)).toString(), ht(s, t);
-}
-function de(e) {
-  return e === 0 ? "false" : "true";
-}
-function dt(e, t, n, s, r, a = !0) {
-  const o = n === "complex64" ? 2 : 1, i = t[0], c = t.length;
-  if (c === 0) {
-    if (n === "complex64") {
-      const g = it(e);
-      return [ot(g[0], 0, n)];
-    }
-    return n === "bool" ? [de(e[0])] : [e[0].toString()];
-  }
-  if (c === 1) {
-    if (i > qt) {
-      const w = at * o;
-      let I = Array.from(e.slice(0, w)), _ = Array.from(e.slice((i - at) * o, i * o));
-      return n === "complex64" && (I = it(I), _ = it(_)), [
-        "[" + I.map((X, J) => ot(X, r[J], n)).join(", ") + ", ..., " + _.map((X, J) => ot(X, r[i - at + J], n)).join(", ") + "]"
-      ];
-    }
-    return [
-      "[" + (n === "complex64" ? it(e) : Array.from(e)).map((w, I) => ot(w, r[I], n)).join(", ") + "]"
-    ];
-  }
-  const l = t.slice(1), u = s.slice(1), h = s[0] * o, d = [];
-  if (i > qt) {
-    for (let g = 0; g < at; g++) {
-      const w = g * h, I = w + h;
-      d.push(...dt(
-        e.slice(w, I),
-        l,
-        n,
-        u,
-        r,
-        !1
-        /* isLast */
-      ));
-    }
-    d.push("...");
-    for (let g = i - at; g < i; g++) {
-      const w = g * h, I = w + h;
-      d.push(...dt(
-        e.slice(w, I),
-        l,
-        n,
-        u,
-        r,
-        g === i - 1
-        /* isLast */
-      ));
-    }
-  } else
-    for (let g = 0; g < i; g++) {
-      const w = g * h, I = w + h;
-      d.push(...dt(
-        e.slice(w, I),
-        l,
-        n,
-        u,
-        r,
-        g === i - 1
-        /* isLast */
-      ));
-    }
-  const m = c === 2 ? "," : "";
-  d[0] = "[" + (i > 0 ? d[0] + m : "");
-  for (let g = 1; g < d.length - 1; g++)
-    d[g] = " " + d[g] + m;
-  let y = `,
-`;
-  for (let g = 2; g < c; g++)
-    y += `
-`;
-  return d[d.length - 1] = " " + d[d.length - 1] + "]" + (a ? "" : y), d;
-}
-function it(e) {
-  const t = [];
-  for (let n = 0; n < e.length; n += 2)
-    t.push([e[n], e[n + 1]]);
-  return t;
-}
-/**
- * @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.
- * =============================================================================
- */
-class gn {
-  constructor(t, n, s) {
-    if (this.dtype = n, this.shape = t.slice(), this.size = z(t), s != null) {
-      const r = s.length;
-      p(r === this.size, () => `Length of values '${r}' does not match the size inferred by the shape '${this.size}'.`);
-    }
-    if (n === "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 = s || ee(n, this.size), this.strides = $t(t);
-  }
-  /**
-   * Sets a value in the buffer at a given location.
-   *
-   * @param value The value to set.
-   * @param locs  The location indices.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Creation'}
-   */
-  set(t, ...n) {
-    n.length === 0 && (n = [0]), p(n.length === this.rank, () => `The number of provided coordinates (${n.length}) must match the rank (${this.rank})`);
-    const s = this.locToIndex(n);
-    this.values[s] = t;
-  }
-  /**
-   * Returns the value in the buffer at the provided location.
-   *
-   * @param locs The location indices.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Creation'}
-   */
-  get(...t) {
-    t.length === 0 && (t = [0]);
-    let n = 0;
-    for (const r of t) {
-      if (r < 0 || r >= this.shape[n]) {
-        const a = `Requested out of range element at ${t}.   Buffer shape=${this.shape}`;
-        throw new Error(a);
-      }
-      n++;
-    }
-    let s = t[t.length - 1];
-    for (let r = 0; r < t.length - 1; ++r)
-      s += this.strides[r] * t[r];
-    return this.values[s];
-  }
-  locToIndex(t) {
-    if (this.rank === 0)
-      return 0;
-    if (this.rank === 1)
-      return t[0];
-    let n = t[t.length - 1];
-    for (let s = 0; s < t.length - 1; ++s)
-      n += this.strides[s] * t[s];
-    return n;
-  }
-  indexToLoc(t) {
-    if (this.rank === 0)
-      return [];
-    if (this.rank === 1)
-      return [t];
-    const n = new Array(this.shape.length);
-    for (let s = 0; s < n.length - 1; ++s)
-      n[s] = Math.floor(t / this.strides[s]), t -= n[s] * this.strides[s];
-    return n[n.length - 1] = t, n;
-  }
-  get rank() {
-    return this.shape.length;
-  }
-  /**
-   * Creates an immutable `tf.Tensor` object from the buffer.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Creation'}
-   */
-  toTensor() {
-    return x().makeTensor(this.values, this.shape, this.dtype);
-  }
-}
-let x = null, Y = null;
-function mn(e) {
-  x = e;
-}
-function pn(e) {
-  Y = e;
-}
-class M {
-  constructor(t, n, s, r) {
-    this.kept = !1, this.isDisposedInternal = !1, this.shape = t.slice(), this.dtype = n || "float32", this.size = z(t), this.strides = $t(t), this.dataId = s, this.id = r, this.rankType = this.rank < 5 ? this.rank.toString() : "higher";
-  }
-  get rank() {
-    return this.shape.length;
-  }
-  /**
-   * Returns a promise of `tf.TensorBuffer` that holds the underlying data.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  async buffer() {
-    const t = await this.data();
-    return Y.buffer(this.shape, this.dtype, t);
-  }
-  /**
-   * Returns a `tf.TensorBuffer` that holds the underlying data.
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  bufferSync() {
-    return Y.buffer(this.shape, this.dtype, this.dataSync());
-  }
-  /**
-   * Returns the tensor data as a nested array. The transfer of data is done
-   * asynchronously.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  async array() {
-    const t = await this.data();
-    return ct(this.shape, t, this.dtype === "complex64");
-  }
-  /**
-   * Returns the tensor data as a nested array. The transfer of data is done
-   * synchronously.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  arraySync() {
-    return ct(this.shape, this.dataSync(), this.dtype === "complex64");
-  }
-  /**
-   * Asynchronously downloads the values from the `tf.Tensor`. Returns a
-   * promise of `TypedArray` that resolves when the computation has finished.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  async data() {
-    this.throwIfDisposed();
-    const t = x().read(this.dataId);
-    if (this.dtype === "string") {
-      const n = await t;
-      try {
-        return n.map((s) => jt(s));
-      } catch {
-        throw new Error("Failed to decode the string bytes into utf-8. To get the original bytes, call tensor.bytes().");
-      }
-    }
-    return t;
-  }
-  /**
-   * Copy the tensor's data to a new GPU resource. Comparing to the `dataSync()`
-   * and `data()`, this method prevents data from being downloaded to CPU.
-   *
-   * For WebGL backend, the data will be stored on a densely packed texture.
-   * This means that the texture will use the RGBA channels to store value.
-   *
-   * For WebGPU backend, the data will be stored on a buffer. There is no
-   * parameter, so can not use a user-defined size to create the buffer.
-   *
-   * @param options:
-   *     For WebGL,
-   *         - customTexShape: Optional. If set, will use the user defined
-   *     texture shape to create the texture.
-   *
-   * @returns For WebGL backend, a GPUData contains the new texture and
-   *     its information.
-   *     {
-   *        tensorRef: The tensor that is associated with this texture,
-   *        texture: WebGLTexture,
-   *        texShape: [number, number] // [height, width]
-   *     }
-   *
-   *     For WebGPU backend, a GPUData contains the new buffer.
-   *     {
-   *        tensorRef: The tensor that is associated with this buffer,
-   *        buffer: GPUBuffer,
-   *     }
-   *
-   *     Remember to dispose the GPUData after it is used by
-   *     `res.tensorRef.dispose()`.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  dataToGPU(t) {
-    return this.throwIfDisposed(), x().readToGPU(this.dataId, t);
-  }
-  /**
-   * Synchronously downloads the values from the `tf.Tensor`. This blocks the
-   * UI thread until the values are ready, which can cause performance issues.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  dataSync() {
-    this.throwIfDisposed();
-    const t = x().readSync(this.dataId);
-    if (this.dtype === "string")
-      try {
-        return t.map((n) => jt(n));
-      } catch {
-        throw new Error("Failed to decode the string bytes into utf-8. To get the original bytes, call tensor.bytes().");
-      }
-    return t;
-  }
-  /** Returns the underlying bytes of the tensor's data. */
-  async bytes() {
-    this.throwIfDisposed();
-    const t = await x().read(this.dataId);
-    return this.dtype === "string" ? t : new Uint8Array(t.buffer);
-  }
-  /**
-   * Disposes `tf.Tensor` from memory.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  dispose() {
-    this.isDisposed || (this.kerasMask && this.kerasMask.dispose(), x().disposeTensor(this), this.isDisposedInternal = !0);
-  }
-  get isDisposed() {
-    return this.isDisposedInternal;
-  }
-  throwIfDisposed() {
-    if (this.isDisposed)
-      throw new Error("Tensor is disposed.");
-  }
-  /**
-   * Prints the `tf.Tensor`. See `tf.print` for details.
-   *
-   * @param verbose Whether to print verbose information about the tensor,
-   *    including dtype and size.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  print(t = !1) {
-    return Y.print(this, t);
-  }
-  /**
-   * Returns a copy of the tensor. See `tf.clone` for details.
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  clone() {
-    return this.throwIfDisposed(), Y.clone(this);
-  }
-  /**
-   * Returns a human-readable description of the tensor. Useful for logging.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  toString(t = !1) {
-    const n = this.dataSync();
-    return dn(n, this.shape, this.dtype, t);
-  }
-  cast(t) {
-    return this.throwIfDisposed(), Y.cast(this, t);
-  }
-  variable(t = !0, n, s) {
-    return this.throwIfDisposed(), x().makeVariable(this, t, n, s);
-  }
-}
-Object.defineProperty(M, Symbol.hasInstance, {
-  value: (e) => !!e && e.data != null && e.dataSync != null && e.throwIfDisposed != null
-});
-function fe() {
-  return Lt("Tensor", () => M);
-}
-fe();
-class mt extends M {
-  constructor(t, n, s, r) {
-    super(t.shape, t.dtype, t.dataId, r), this.trainable = n, this.name = s;
-  }
-  /**
-   * Assign a new `tf.Tensor` to this variable. The new `tf.Tensor` must have
-   * the same shape and dtype as the old `tf.Tensor`.
-   *
-   * @param newValue New tensor to be assigned to this variable.
-   *
-   * @doc {heading: 'Tensors', subheading: 'Classes'}
-   */
-  assign(t) {
-    if (t.dtype !== this.dtype)
-      throw new Error(`dtype of the new value (${t.dtype}) and previous value (${this.dtype}) must match`);
-    if (!xt(t.shape, this.shape))
-      throw new Error(`shape of the new value (${t.shape}) and previous value (${this.shape}) must match`);
-    x().disposeTensor(this), this.dataId = t.dataId, x().incRef(
-      this,
-      null
-      /* backend */
-    );
-  }
-  dispose() {
-    x().disposeVariable(this), this.isDisposedInternal = !0;
-  }
-}
-Object.defineProperty(mt, Symbol.hasInstance, {
-  value: (e) => e instanceof M && e.assign != null && e.assign instanceof Function
-});
-/**
- * @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.
- * =============================================================================
- */
-var Kt;
-(function(e) {
-  e.R0 = "R0", e.R1 = "R1", e.R2 = "R2", e.R3 = "R3", e.R4 = "R4", e.R5 = "R5", e.R6 = "R6";
-})(Kt || (Kt = {}));
-var At;
-(function(e) {
-  e.float32 = "float32", e.int32 = "int32", e.bool = "int32", e.complex64 = "complex64";
-})(At || (At = {}));
-var Et;
-(function(e) {
-  e.float32 = "float32", e.int32 = "int32", e.bool = "bool", e.complex64 = "complex64";
-})(Et || (Et = {}));
-var vt;
-(function(e) {
-  e.float32 = "float32", e.int32 = "float32", e.bool = "float32", e.complex64 = "complex64";
-})(vt || (vt = {}));
-var Bt;
-(function(e) {
-  e.float32 = "complex64", e.int32 = "complex64", e.bool = "complex64", e.complex64 = "complex64";
-})(Bt || (Bt = {}));
-const yn = {
-  float32: vt,
-  int32: At,
-  bool: Et,
-  complex64: Bt
-};
-function ge(e, t) {
-  if (e === "string" || t === "string") {
-    if (e === "string" && t === "string")
-      return "string";
-    throw new Error(`Can not upcast ${e} with ${t}`);
-  }
-  return yn[e][t];
-}
-function Qo(e) {
-  return ge(e, "int32");
-}
-function me(e) {
-  return e != null && typeof e == "object" && "texture" in e && e.texture instanceof WebGLTexture;
-}
-function pe(e) {
-  return typeof GPUBuffer < "u" && e != null && typeof e == "object" && "buffer" in e && e.buffer instanceof GPUBuffer;
-}
-/**
- * @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.
- * =============================================================================
- */
-function K(e, t) {
-  if (e.dtype === t.dtype)
-    return [e, t];
-  const n = ge(e.dtype, t.dtype);
-  return [e.cast(n), t.cast(n)];
-}
-function ti(e, t) {
-  return t.some((n) => n.id === e.id);
-}
-function ye(e) {
-  const t = [];
-  return be(e, t, /* @__PURE__ */ new Set()), t;
-}
-function be(e, t, n) {
-  if (e == null)
-    return;
-  if (e instanceof M) {
-    t.push(e);
-    return;
-  }
-  if (!bn(e))
-    return;
-  const s = e;
-  for (const r in s) {
-    const a = s[r];
-    n.has(a) || (n.add(a), be(a, t, n));
-  }
-}
-function bn(e) {
-  return Array.isArray(e) || typeof e == "object";
-}
-/**
- * @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.
- * =============================================================================
- */
-function St(e) {
-  return e.kernelName != null;
-}
-class Ht {
-  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 = /* @__PURE__ */ 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((t) => t.name)));
-      }
-    };
-  }
-  dispose() {
-    for (const t in this.registeredVariables)
-      this.registeredVariables[t].dispose();
-  }
-}
-class et {
-  constructor(t) {
-    this.ENV = t, this.registry = {}, this.registryFactory = {}, this.pendingBackendInitId = 0, this.state = new Ht();
-  }
-  async ready() {
-    if (this.pendingBackendInit != null)
-      return this.pendingBackendInit.then(() => {
-      });
-    if (this.backendInstance != null)
-      return;
-    const t = this.getSortedBackends();
-    for (let n = 0; n < t.length; n++) {
-      const s = t[n];
-      if (await this.initializeBackend(s).success) {
-        await this.setBackend(s);
-        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) {
-      const { name: t, asyncInit: n } = this.initializeBackendsAndReturnBest();
-      if (n)
-        throw new Error(`The highest priority backend '${t}' has not yet been initialized. Make sure to await tf.ready() or await tf.setBackend() before calling other methods`);
-      this.setBackend(t);
-    }
-    return this.backendInstance;
-  }
-  backendNames() {
-    return Object.keys(this.registryFactory);
-  }
-  findBackend(t) {
-    if (!(t in this.registry))
-      if (t in this.registryFactory) {
-        const { asyncInit: n } = this.initializeBackend(t);
-        if (n)
-          return null;
-      } else
-        return null;
-    return this.registry[t];
-  }
-  findBackendFactory(t) {
-    return t in this.registryFactory ? this.registryFactory[t].factory : null;
-  }
-  registerBackend(t, n, s = 1) {
-    return t in this.registryFactory ? (O(`${t} backend was already registered. Reusing existing backend factory.`), !1) : (this.registryFactory[t] = { factory: n, priority: s }, !0);
-  }
-  async setBackend(t) {
-    if (this.registryFactory[t] == null)
-      throw new Error(`Backend name '${t}' not found in registry`);
-    if (this.backendName = t, this.registry[t] == null) {
-      this.backendInstance = null;
-      const { success: n, asyncInit: s } = this.initializeBackend(t);
-      if (!(s ? await n : n))
-        return !1;
-    }
-    return this.backendInstance = this.registry[t], this.setupRegisteredKernels(), this.profiler = new on(this.backendInstance), !0;
-  }
-  setupRegisteredKernels() {
-    Wt(this.backendName).forEach((n) => {
-      n.setupFunc != null && n.setupFunc(this.backendInstance);
-    });
-  }
-  disposeRegisteredKernels(t) {
-    Wt(t).forEach((s) => {
-      s.disposeFunc != null && s.disposeFunc(this.registry[t]);
-    });
-  }
-  /**
-   * Initializes a backend by looking up the backend name in the factory
-   * registry and calling the factory method. Returns a boolean representing
-   * whether the initialization of the backend succeeded. Throws an error if
-   * there is no backend in the factory registry.
-   */
-  initializeBackend(t) {
-    const n = this.registryFactory[t];
-    if (n == null)
-      throw new Error(`Cannot initialize backend ${t}, no registration found.`);
-    try {
-      const s = n.factory();
-      if (s && !(s instanceof xe) && typeof s.then == "function") {
-        const r = ++this.pendingBackendInitId, a = s.then((o) => r < this.pendingBackendInitId ? !1 : (this.registry[t] = o, this.pendingBackendInit = null, !0)).catch((o) => (r < this.pendingBackendInitId || (this.pendingBackendInit = null, O(`Initialization of backend ${t} failed`), O(o.stack || o.message)), !1));
-        return this.pendingBackendInit = a, { success: a, asyncInit: !0 };
-      } else
-        return this.registry[t] = s, { success: !0, asyncInit: !1 };
-    } catch (s) {
-      return O(`Initialization of backend ${t} failed`), O(s.stack || s.message), { success: !1, asyncInit: !1 };
-    }
-  }
-  removeBackend(t) {
-    if (!(t in this.registryFactory))
-      throw new Error(`${t} backend not found in registry`);
-    this.backendName === t && this.pendingBackendInit != null && this.pendingBackendInitId++, t in this.registry && (this.disposeRegisteredKernels(t), this.registry[t].dispose(), delete this.registry[t]), delete this.registryFactory[t], this.backendName === t && (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((t, n) => this.registryFactory[n].priority - this.registryFactory[t].priority);
-  }
-  initializeBackendsAndReturnBest() {
-    const t = this.getSortedBackends();
-    for (let n = 0; n < t.length; n++) {
-      const s = t[n], { success: r, asyncInit: a } = this.initializeBackend(s);
-      if (a || r)
-        return { name: s, asyncInit: a };
-    }
-    throw new Error("Could not initialize any backends, all backend initializations failed.");
-  }
-  moveData(t, n) {
-    const s = this.state.tensorInfo.get(n), r = s.backend, a = this.readSync(n), o = r.refCount(n);
-    r.disposeData(n, !0), s.backend = t, t.move(n, a, s.shape, s.dtype, o), this.shouldCheckForMemLeaks() && this.state.numDataMovesStack[this.state.numDataMovesStack.length - 1]++;
-  }
-  tidy(t, n) {
-    let s = null;
-    if (n == null) {
-      if (typeof t != "function")
-        throw new Error("Please provide a function to tidy()");
-      n = t;
-    } else {
-      if (typeof t != "string" && !(t instanceof String))
-        throw new Error("When calling with two arguments, the first argument to tidy() must be a string");
-      if (typeof n != "function")
-        throw new Error("When calling with two arguments, the 2nd argument to tidy() must be a function");
-      s = t;
-    }
-    let r;
-    return this.scopedRun(() => this.startScope(s), () => this.endScope(r), () => (r = n(), r instanceof Promise && console.error("Cannot return a Promise inside of tidy."), r));
-  }
-  scopedRun(t, n, s) {
-    t();
-    try {
-      const r = s();
-      return n(), r;
-    } catch (r) {
-      throw n(), r;
-    }
-  }
-  nextTensorId() {
-    return et.nextTensorId++;
-  }
-  nextVariableId() {
-    return et.nextVariableId++;
-  }
-  /**
-   * This method is called instead of the public-facing tensor.clone() when
-   * saving a tensor for backwards pass. It makes sure to add the clone
-   * operation to the tape regardless of being called inside a kernel
-   * execution.
-   */
-  clone(t) {
-    const n = f.runKernel(ce, { x: t }), s = { x: t }, r = (o) => ({
-      x: () => {
-        const i = "float32", c = { x: o }, l = { dtype: i };
-        return f.runKernel(
-          ie,
-          c,
-          // tslint:disable-next-line: no-unnecessary-type-assertion
-          l
-        );
-      }
-    }), a = [];
-    return this.addTapeNode(this.state.activeScope.name, s, [n], r, a, {}), n;
-  }
-  /**
-   * Execute a kernel with the given name and return the output tensor.
-   *
-   * @param kernelName The name of the kernel to execute.
-   * @param inputs A map of input names to tensors.
-   * @param attrs A map of attribute names to their values. An attribute is a
-   *     primitive (non-tensor) input to the kernel.
-   * @param inputsToSave A list of tensors, inputs to save for the backprop
-   *     computation.
-   * @param outputsToSave A list of booleans, specifying which output to save
-   *     for the backprop computation. These are booleans since the output
-   * tensors are not visible to the user.
-   */
-  runKernel(t, n, s) {
-    if (this.backendName == null && this.backend, !(zt(t, this.backendName) != null))
-      throw new Error(`Kernel '${t}' not registered for backend '${this.backendName}'`);
-    return this.runKernelFunc({ kernelName: t, inputs: n, attrs: s });
-  }
-  shouldCheckForMemLeaks() {
-    return this.ENV.getBool("IS_TEST");
-  }
-  checkKernelForMemLeak(t, n, s) {
-    const r = this.backend.numDataIds();
-    let a = 0;
-    s.forEach((c) => {
-      a += c.dtype === "complex64" ? 3 : 1;
-    });
-    const o = this.state.numDataMovesStack[this.state.numDataMovesStack.length - 1], i = r - n - a - o;
-    if (i > 0)
-      throw new Error(`Backend '${this.backendName}' has an internal memory leak (${i} data ids) after running '${t}'`);
-  }
-  /**
-   * Internal helper method to execute a kernel Func
-   *
-   * Use `runKernel` to execute kernels from outside of engine.
-   */
-  runKernelFunc(t) {
-    let n, s = [];
-    const r = this.isTapeOn(), a = this.state.numBytes, o = this.state.numTensors;
-    this.shouldCheckForMemLeaks() && this.state.numDataMovesStack.push(0);
-    let i;
-    this.backendName == null && this.backend;
-    let c;
-    const l = St(t) ? t.kernelName : this.state.activeScope != null ? this.state.activeScope.name : "";
-    if (St(t)) {
-      const { kernelName: y, inputs: g, attrs: w } = t;
-      this.backendName == null && this.backend;
-      const I = zt(y, this.backendName);
-      p(I != null, () => `Cannot find registered kernel '${y}' for backend '${this.backendName}'`), i = () => {
-        const _ = this.backend.numDataIds();
-        c = I.kernelFunc({ inputs: g, attrs: w, backend: this.backend });
-        const X = Array.isArray(c) ? c : [c];
-        this.shouldCheckForMemLeaks() && this.checkKernelForMemLeak(y, _, X);
-        const J = X.map((rt) => rt.rank != null ? rt : this.makeTensorFromTensorInfo(rt));
-        if (r) {
-          const rt = this.getTensorsForGradient(y, g, J);
-          s = this.saveTensorsForBackwardMode(rt);
-        }
-        return J;
-      };
-    } else {
-      const { forwardFunc: y } = t, g = (w) => {
-        r && (s = w.map((I) => this.keep(this.clone(I))));
-      };
-      i = () => {
-        const w = this.backend.numDataIds();
-        c = this.tidy(() => y(this.backend, g));
-        const I = Array.isArray(c) ? c : [c];
-        return this.shouldCheckForMemLeaks() && this.checkKernelForMemLeak(l, w, I), I;
-      };
-    }
-    const { inputs: u, attrs: h } = t, d = St(t) ? null : t.backwardsFunc;
-    let m;
-    return this.scopedRun(
-      // Stop recording to a tape when running a kernel.
-      () => this.state.kernelDepth++,
-      () => this.state.kernelDepth--,
-      () => {
-        !this.ENV.getBool("DEBUG") && !this.state.profiling ? n = i() : (m = this.profiler.profileKernel(l, u, () => i()), this.ENV.getBool("DEBUG") && this.profiler.logKernelProfile(m), n = m.outputs);
-      }
-    ), r && this.addTapeNode(l, u, n, d, s, h), this.state.profiling && this.state.activeProfile.kernels.push({
-      name: l,
-      bytesAdded: this.state.numBytes - a,
-      totalBytesSnapshot: this.state.numBytes,
-      tensorsAdded: this.state.numTensors - o,
-      totalTensorsSnapshot: this.state.numTensors,
-      inputShapes: Object.keys(u).map((y) => u[y] != null ? u[y].shape : null),
-      outputShapes: n.map((y) => y.shape),
-      kernelTimeMs: m.timeMs,
-      extraInfo: m.extraInfo
-    }), Array.isArray(c) ? n : n[0];
-  }
-  /**
-   * Saves tensors used in forward mode for use in backward mode.
-   *
-   * @param tensors the list of tensors to save.
-   */
-  saveTensorsForBackwardMode(t) {
-    return t.map((s) => this.keep(this.clone(s)));
-  }
-  /**
-   * Returns a list of tensors to save for a given gradient calculation.
-   *
-   * @param kernelName name of kernel to look up gradient for.
-   * @param inputs a map of input tensors.
-   * @param outputs an array of output tensors from forward mode of kernel.
-   */
-  getTensorsForGradient(t, n, s) {
-    const r = Vt(t);
-    if (r != null) {
-      const a = r.inputsToSave || [], o = r.outputsToSave || [];
-      let i;
-      r.saveAllInputs ? (p(Array.isArray(n), () => "saveAllInputs is true, expected inputs to be an array."), i = Object.keys(n).map((l) => n[l])) : i = a.map((l) => n[l]);
-      const c = s.filter((l, u) => o[u]);
-      return i.concat(c);
-    }
-    return [];
-  }
-  /**
-   * Internal method used by public APIs for tensor creation. Makes a new
-   * tensor with the provided shape, dtype and values. It always
-   * creates a new data id and writes the values to the underlying backend.
-   */
-  makeTensor(t, n, s, r) {
-    if (t == null)
-      throw new Error("Values passed to engine.makeTensor() are null");
-    s = s || "float32", r = r || this.backend;
-    let a = t;
-    s === "string" && Nt(t[0]) && (a = t.map((c) => he(c)));
-    const o = r.write(a, n, s), i = new M(n, s, o, this.nextTensorId());
-    if (this.trackTensor(i, r), s === "string") {
-      const c = this.state.tensorInfo.get(o), l = _e(a);
-      this.state.numBytes += l - c.bytes, c.bytes = l;
-    }
-    return i;
-  }
-  /**
-   * Internal method used by backends. Makes a new tensor
-   * that is a wrapper around an existing data id. It doesn't create
-   * a new data id, only increments the ref count used in memory tracking.
-   * @deprecated
-   */
-  makeTensorFromDataId(t, n, s, r) {
-    s = s || "float32";
-    const a = { dataId: t, shape: n, dtype: s };
-    return this.makeTensorFromTensorInfo(a, r);
-  }
-  /**
-   * Internal method used by backends. Makes a new tensor that is a wrapper
-   * around an existing data id in TensorInfo. It doesn't create a new data id,
-   * only increments the ref count used in memory tracking.
-   */
-  makeTensorFromTensorInfo(t, n) {
-    const { dataId: s, shape: r, dtype: a } = t, o = new M(r, a, s, this.nextTensorId());
-    return this.trackTensor(o, n), o;
-  }
-  makeVariable(t, n = !0, s, r) {
-    s = s || this.nextVariableId().toString(), r != null && r !== t.dtype && (t = t.cast(r));
-    const a = new mt(t, n, s, this.nextTensorId());
-    if (this.state.registeredVariables[a.name] != null)
-      throw new Error(`Variable with name ${a.name} was already registered`);
-    return this.state.registeredVariables[a.name] = a, this.incRef(a, this.backend), a;
-  }
-  trackTensor(t, n) {
-    this.state.numTensors++, t.dtype === "string" && this.state.numStringTensors++;
-    let s = 0;
-    t.dtype !== "complex64" && t.dtype !== "string" && (s = t.size * It(t.dtype)), this.state.numBytes += s, this.state.tensorInfo.has(t.dataId) || (this.state.numDataBuffers++, this.state.tensorInfo.set(t.dataId, {
-      backend: n || this.backend,
-      dtype: t.dtype,
-      shape: t.shape,
-      bytes: s
-    })), t instanceof mt || this.track(t);
-  }
-  // Track the tensor by dataId and increase the refCount for the dataId in the
-  // backend.
-  // TODO(pyu10055): This is currently used by makeVariable method, to increase
-  // refCount on the backend for the dataId. It can potentially be replaced with
-  // Identity op indead of calling backend directly.
-  incRef(t, n) {
-    this.trackTensor(t, n), this.backend.incRef(t.dataId);
-  }
-  removeDataId(t, n) {
-    this.state.tensorInfo.has(t) && this.state.tensorInfo.get(t).backend === n && (this.state.tensorInfo.delete(t), this.state.numDataBuffers--);
-  }
-  disposeTensor(t) {
-    if (!this.state.tensorInfo.has(t.dataId))
-      return;
-    const n = this.state.tensorInfo.get(t.dataId);
-    if (this.state.numTensors--, t.dtype === "string" && (this.state.numStringTensors--, this.state.numBytes -= n.bytes), t.dtype !== "complex64" && t.dtype !== "string") {
-      const s = t.size * It(t.dtype);
-      this.state.numBytes -= s;
-    }
-    n.backend.disposeData(t.dataId) && this.removeDataId(t.dataId, n.backend);
-  }
-  disposeVariables() {
-    for (const t in this.state.registeredVariables) {
-      const n = this.state.registeredVariables[t];
-      this.disposeVariable(n);
-    }
-  }
-  disposeVariable(t) {
-    this.disposeTensor(t), this.state.registeredVariables[t.name] != null && delete this.state.registeredVariables[t.name];
-  }
-  memory() {
-    const t = this.backend.memory();
-    return t.numTensors = this.state.numTensors, t.numDataBuffers = this.state.numDataBuffers, t.numBytes = this.state.numBytes, this.state.numStringTensors > 0 && (t.unreliable = !0, t.reasons == null && (t.reasons = []), t.reasons.push("Memory usage by string tensors is approximate (2 bytes per character)")), t;
-  }
-  async profile(t) {
-    this.state.profiling = !0;
-    const n = this.state.numBytes, s = this.state.numTensors;
-    this.state.activeProfile.kernels = [], this.state.activeProfile.result = await t(), this.state.profiling = !1, this.state.activeProfile.peakBytes = Math.max(...this.state.activeProfile.kernels.map((r) => r.totalBytesSnapshot)), this.state.activeProfile.newBytes = this.state.numBytes - n, this.state.activeProfile.newTensors = this.state.numTensors - s;
-    for (const r of this.state.activeProfile.kernels)
-      r.kernelTimeMs = await r.kernelTimeMs, r.extraInfo = await r.extraInfo;
-    return this.state.activeProfile;
-  }
-  isTapeOn() {
-    return this.state.gradientDepth > 0 && this.state.kernelDepth === 0;
-  }
-  addTapeNode(t, n, s, r, a, o) {
-    const i = { id: this.state.nextTapeNodeId++, kernelName: t, inputs: n, outputs: s, saved: a }, c = Vt(t);
-    c != null && (r = c.gradFunc), r != null && (i.gradient = (l) => (l = l.map((u, h) => {
-      if (u == null) {
-        const d = s[h], m = se(d.size, d.dtype);
-        return this.makeTensor(m, d.shape, d.dtype);
-      }
-      return u;
-    }), r(l.length > 1 ? l : l[0], a, o))), this.state.activeTape.push(i);
-  }
-  keep(t) {
-    return t.kept = !0, t;
-  }
-  startTape() {
-    this.state.gradientDepth === 0 && (this.state.activeTape = []), this.state.gradientDepth++;
-  }
-  endTape() {
-    this.state.gradientDepth--;
-  }
-  /**
-   * Start a scope. Use this with endScope() to achieve the same functionality
-   * as scope() without the need for a function closure.
-   */
-  startScope(t) {
-    const n = {
-      track: [],
-      name: "unnamed scope",
-      id: this.state.nextScopeId++
-    };
-    t && (n.name = t), this.state.scopeStack.push(n), this.state.activeScope = n;
-  }
-  /**
-   * End a scope. Use this with startScope() to achieve the same functionality
-   * as scope() without the need for a function closure.
-   */
-  endScope(t) {
-    const n = ye(t), s = new Set(n.map((a) => a.id));
-    for (let a = 0; a < this.state.activeScope.track.length; a++) {
-      const o = this.state.activeScope.track[a];
-      !o.kept && !s.has(o.id) && o.dispose();
-    }
-    const r = this.state.scopeStack.pop();
-    this.state.activeScope = this.state.scopeStack.length === 0 ? null : this.state.scopeStack[this.state.scopeStack.length - 1], n.forEach((a) => {
-      !a.kept && a.scopeId === r.id && this.track(a);
-    });
-  }
-  /**
-   * Returns gradients of `f` with respect to each of the `xs`. The gradients
-   * returned are of the same length as `xs`, but some might be null if `f`
-   * was not a function of that `x`. It also takes optional dy to multiply the
-   * gradient, which defaults to `1`.
-   */
-  gradients(t, n, s, r = !1) {
-    if (p(n.length > 0, () => "gradients() received an empty list of xs."), s != null && s.dtype !== "float32")
-      throw new Error(`dy must have 'float32' dtype, but has '${s.dtype}'`);
-    const a = this.scopedRun(() => this.startTape(), () => this.endTape(), () => this.tidy("forward", t));
-    p(a instanceof M, () => "The result y returned by f() must be a tensor.");
-    const o = un(this.state.activeTape, n, a);
-    if (!r && o.length === 0 && n.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", () => {
-      const i = {};
-      i[a.id] = s ?? wn(a.shape), hn(
-        i,
-        o,
-        // Pass the tidy function to avoid circular dep with `tape.ts`.
-        (l) => this.tidy(l),
-        // Pass an add function to avoide a circular dep with `tape.ts`.
-        Sn
-      );
-      const c = n.map((l) => i[l.id]);
-      return this.state.gradientDepth === 0 && (this.state.activeTape.forEach((l) => {
-        for (const u of l.saved)
-          u.dispose();
-      }), this.state.activeTape = null), { value: a, grads: c };
-    });
-  }
-  customGrad(t) {
-    return p(lt(t), () => "The f passed in customGrad(f) must be a function."), (...n) => {
-      p(n.every((i) => i instanceof M), () => "The args passed in customGrad(f)(x1, x2,...) must all be tensors");
-      let s;
-      const r = {};
-      n.forEach((i, c) => {
-        r[c] = i;
-      });
-      const a = (i, c) => (s = t(...n, c), p(s.value instanceof M, () => "The function f passed in customGrad(f) must return an object where `obj.value` is a tensor"), p(lt(s.gradFunc), () => "The function f passed in customGrad(f) must return an object where `obj.gradFunc` is a function."), s.value), o = (i, c) => {
-        const l = s.gradFunc(i, c), u = Array.isArray(l) ? l : [l];
-        p(u.length === n.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(...)."), p(u.every((d) => d instanceof M), () => "The function f passed in customGrad(f) must return an object where `obj.gradFunc` is a function that returns a list of only tensors.");
-        const h = {};
-        return u.forEach((d, m) => {
-          h[m] = () => d;
-        }), h;
-      };
-      return this.runKernelFunc({
-        forwardFunc: a,
-        backwardsFunc: o,
-        inputs: r
-      });
-    };
-  }
-  readSync(t) {
-    return this.state.tensorInfo.get(t).backend.readSync(t);
-  }
-  read(t) {
-    return this.state.tensorInfo.get(t).backend.read(t);
-  }
-  readToGPU(t, n) {
-    return this.state.tensorInfo.get(t).backend.readToGPU(t, n);
-  }
-  async time(t) {
-    const n = gt(), s = await this.backend.time(t);
-    return s.wallMs = gt() - n, s;
-  }
-  /**
-   * Tracks a Tensor in the current scope to be automatically cleaned up
-   * when the current scope ends, and returns the value.
-   *
-   * @param result The Tensor to track in the current scope.
-   */
-  track(t) {
-    return this.state.activeScope != null && (t.scopeId = this.state.activeScope.id, this.state.activeScope.track.push(t)), t;
-  }
-  get registeredVariables() {
-    return this.state.registeredVariables;
-  }
-  /**
-   * Resets the engine state. Removes all backends but does not remove
-   * registered backend factories.
-   */
-  reset() {
-    this.pendingBackendInitId++, this.state.dispose(), this.ENV.reset(), this.state = new Ht();
-    for (const t in this.registry)
-      this.disposeRegisteredKernels(t), this.registry[t].dispose(), delete this.registry[t];
-    this.backendName = null, this.backendInstance = null, this.pendingBackendInit = null;
-  }
-}
-et.nextTensorId = 0;
-et.nextVariableId = 0;
-function wn(e) {
-  const t = Ue(z(e), "float32");
-  return f.makeTensor(t, e, "float32");
-}
-function we() {
-  const e = ae();
-  if (e._tfengine == null) {
-    const t = new ze(e);
-    e._tfengine = new et(t);
-  }
-  return qe(e._tfengine.ENV), mn(() => e._tfengine), e._tfengine;
-}
-const f = we();
-function Sn(e, t) {
-  const n = { a: e, b: t };
-  return f.runKernel(oe, n);
-}
-/**
- * @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.
- * =============================================================================
- */
-function kn() {
-  return typeof navigator < "u" && navigator != null;
-}
-function ei(e) {
-  if (e || kn()) {
-    if (e || (e = navigator), e.product === "ReactNative")
-      return !0;
-    const t = e.userAgent || e.vendor || // tslint:disable-next-line:no-any
-    (typeof window < "u" ? window.opera : "");
-    if (!t) {
-      const n = e;
-      return n.userAgentData && n.userAgentData.mobile;
-    }
-    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(t) || // tslint:disable-next-line:max-line-length
-    /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(t.substr(0, 4));
-  }
-  return !1;
-}
-function In() {
-  return typeof window < "u" && window.document != null || //@ts-ignore
-  typeof WorkerGlobalScope < "u";
-}
-const v = S();
-v.registerFlag("DEBUG", () => !1, (e) => {
-  e && 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.");
-});
-v.registerFlag("IS_BROWSER", () => In());
-v.registerFlag("IS_NODE", () => typeof Q < "u" && typeof Q.versions < "u" && typeof Q.versions.node < "u");
-v.registerFlag("IS_CHROME", () => typeof navigator < "u" && navigator != null && navigator.userAgent != null && /Chrome/.test(navigator.userAgent) && /Google Inc/.test(navigator.vendor));
-v.registerFlag("IS_SAFARI", () => typeof navigator < "u" && navigator != null && navigator.userAgent != null && /Safari/.test(navigator.userAgent) && /Apple/.test(navigator.vendor));
-v.registerFlag("PROD", () => !1);
-v.registerFlag("TENSORLIKE_CHECK_SHAPE_CONSISTENCY", () => v.getBool("DEBUG"));
-v.registerFlag("DEPRECATION_WARNINGS_ENABLED", () => !0);
-v.registerFlag("IS_TEST", () => !1);
-v.registerFlag("CHECK_COMPUTATION_FOR_ERRORS", () => v.getBool("DEBUG"));
-v.registerFlag("WRAP_TO_IMAGEBITMAP", () => !1);
-v.registerFlag("CANVAS2D_WILL_READ_FREQUENTLY_FOR_GPU", () => !1);
-v.registerFlag("USE_SETTIMEOUTCUSTOM", () => !1);
-/**
- * @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.
- * =============================================================================
- */
-function Tn(e, t) {
-  let n = e;
-  if (R(e))
-    return t === "string" ? [] : [e.length];
-  if (me(e)) {
-    const r = e.channels || "RGBA";
-    return [e.height, e.width * r.length];
-  } else if (pe(e))
-    return [e.buffer.size / (t == null ? 4 : It(t))];
-  if (!Array.isArray(e))
-    return [];
-  const s = [];
-  for (; Array.isArray(n) || R(n) && t !== "string"; )
-    s.push(n.length), n = n[0];
-  return Array.isArray(e) && S().getBool("TENSORLIKE_CHECK_SHAPE_CONSISTENCY") && Se(e, s, []), s;
-}
-function Se(e, t, n) {
-  if (n = n || [], !Array.isArray(e) && !R(e)) {
-    p(t.length === 0, () => `Element arr[${n.join("][")}] is a primitive, but should be an array/TypedArray of ${t[0]} elements`);
-    return;
-  }
-  p(t.length > 0, () => `Element arr[${n.join("][")}] should be a primitive, but is an array of ${e.length} elements`), p(e.length === t[0], () => `Element arr[${n.join("][")}] should have ${t[0]} elements, but has ${e.length} elements`);
-  const s = t.slice(1);
-  for (let r = 0; r < e.length; ++r)
-    Se(e[r], s, n.concat(r));
-}
-function Xt(e, t, n, s) {
-  if (e !== "string_or_numeric") {
-    if (e == null)
-      throw new Error("Expected dtype cannot be null.");
-    if (e !== "numeric" && e !== t || e === "numeric" && t === "string")
-      throw new Error(`Argument '${n}' passed to '${s}' must be ${e} tensor, but got ${t} tensor`);
-  }
-}
-function T(e, t, n, s = "numeric") {
-  if (e instanceof fe())
-    return Xt(s, e.dtype, t, n), e;
-  let r = yt(e);
-  if (r !== "string" && ["bool", "int32", "float32"].indexOf(s) >= 0 && (r = s), Xt(s, r, t, n), e == null || !R(e) && !Array.isArray(e) && typeof e != "number" && typeof e != "boolean" && typeof e != "string") {
-    const c = e == null ? "null" : e.constructor.name;
-    throw new Error(`Argument '${t}' passed to '${n}' must be a Tensor or TensorLike, but got '${c}'`);
-  }
-  const a = Tn(e, r);
-  !R(e) && !Array.isArray(e) && (e = [e]);
-  const i = r !== "string" ? _t(e, r) : ut(e, [], !0);
-  return f.makeTensor(i, a, r);
-}
-function An(e, t, n, s = "numeric") {
-  if (!Array.isArray(e))
-    throw new Error(`Argument ${t} passed to ${n} must be a \`Tensor[]\` or \`TensorLike[]\``);
-  return e.map((a, o) => T(a, `${t}[${o}]`, n, s));
-}
-/**
- * @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.
- * =============================================================================
- */
-const En = "__op";
-function F(e) {
-  const t = Object.keys(e);
-  if (t.length !== 1)
-    throw new Error(`Please provide an object with a single key (operation name) mapping to a function. Got an object with ${t.length} keys.`);
-  let n = t[0];
-  const s = e[n];
-  n.endsWith("_") && (n = n.substring(0, n.length - 1)), n = n + En;
-  const r = (...a) => {
-    f.startScope(n);
-    try {
-      const o = s(...a);
-      return Pt(o) && console.error("Cannot return a Promise inside of tidy."), f.endScope(o), o;
-    } catch (o) {
-      throw f.endScope(null), o;
-    }
-  };
-  return Object.defineProperty(r, "name", { value: n, configurable: !0 }), r;
-}
-/**
- * @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.
- * =============================================================================
- */
-function vn(e, t, n, s) {
-  if (s == null)
-    s = yt(e);
-  else if (s === "complex64")
-    throw new Error("Cannot construct a complex64 tensor directly. Please use tf.complex(real, imag).");
-  if (pe(e) || me(e)) {
-    if (s !== "float32" && s !== "int32")
-      throw new Error(`Creating tensor from GPU data only supports 'float32'|'int32' dtype, while the dtype is ${s}.`);
-    return f.backend.createTensorFromGPUData(e, t || n, s);
-  }
-  if (!R(e) && !Array.isArray(e) && typeof e != "number" && typeof e != "boolean" && typeof e != "string")
-    throw new Error("values passed to tensor(values) must be a number/boolean/string or an array of numbers/booleans/strings, or a TypedArray");
-  if (t != null) {
-    Ct(t);
-    const r = z(t), a = z(n);
-    p(r === a, () => `Based on the provided shape, [${t}], the tensor should have ${r} values but has ${a}`);
-    for (let o = 0; o < n.length; ++o) {
-      const i = n[o], c = o === n.length - 1 ? i !== z(t.slice(o)) : !0;
-      p(n[o] === t[o] || !c, () => `Error creating a new Tensor. Inferred shape (${n}) does not match the provided shape (${t}). `);
-    }
-  }
-  return !R(e) && !Array.isArray(e) && (e = [e]), t = t || n, e = s !== "string" ? _t(e, s) : ut(e, [], !0), f.makeTensor(e, t, s);
-}
-class st {
-  /**
-   * Concatenate a number of ArrayBuffers into one.
-   *
-   * @param buffers An array of ArrayBuffers to concatenate, or a single
-   *     ArrayBuffer.
-   * @returns Result of concatenating `buffers` in order.
-   */
-  static join(t) {
-    return new st(t).slice();
-  }
-  constructor(t) {
-    if (this.shards = [], this.previousShardIndex = 0, t == null || (t instanceof Array || (t = [t]), t = t.map((s) => R(s) ? s.buffer : s), t.length === 0))
-      return;
-    this.bufferUniformSize = t[0].byteLength;
-    let n = 0;
-    for (let s = 0; s < t.length; s++) {
-      const r = t[s];
-      s !== t.length - 1 && r.byteLength !== this.bufferUniformSize && (this.bufferUniformSize = void 0);
-      const a = n + r.byteLength;
-      this.shards.push({ buffer: r, start: n, end: a }), n = a;
-    }
-    this.shards.length === 0 && (this.byteLength = 0), this.byteLength = this.shards[this.shards.length - 1].end;
-  }
-  slice(t = 0, n = this.byteLength) {
-    if (this.shards.length === 0)
-      return new ArrayBuffer(0);
-    if (t = isNaN(Number(t)) ? 0 : t, n = isNaN(Number(n)) ? 0 : n, t = Math.max(0, t), n = Math.min(this.byteLength, n), n <= t)
-      return new ArrayBuffer(0);
-    const s = this.findShardForByte(t);
-    if (s === -1)
-      throw new Error(`Could not find start shard for byte ${t}`);
-    const r = n - t, a = new ArrayBuffer(r), o = new Uint8Array(a);
-    let i = 0;
-    for (let c = s; c < this.shards.length; c++) {
-      const l = this.shards[c], h = t + i - l.start, d = i, y = Math.min(n, l.end) - l.start, g = new Uint8Array(l.buffer, h, y - h);
-      if (o.set(g, d), i += g.length, n < l.end)
-        break;
-    }
-    return a;
-  }
-  /**
-   * Get the index of the shard that contains the byte at `byteIndex`.
-   */
-  findShardForByte(t) {
-    if (this.shards.length === 0 || t < 0 || t >= this.byteLength)
-      return -1;
-    if (this.bufferUniformSize != null)
-      return this.previousShardIndex = Math.floor(t / this.bufferUniformSize), this.previousShardIndex;
-    function n(r) {
-      return t < r.start ? -1 : t >= r.end ? 1 : 0;
-    }
-    if (n(this.shards[this.previousShardIndex]) === 0)
-      return this.previousShardIndex;
-    const s = Bn(this.shards, n);
-    return s === -1 ? -1 : (this.previousShardIndex = s, this.previousShardIndex);
-  }
-}
-function Bn(e, t) {
-  let n = 0, s = e.length;
-  for (; n <= s; ) {
-    const r = Math.floor((s - n) / 2) + n, a = t(e[r]);
-    if (a === 0)
-      return r;
-    a < 0 ? s = r : n = r + 1;
-  }
-  return -1;
-}
-/**
- * @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.
- * =============================================================================
- */
-function ni() {
-  f.disposeVariables();
-}
-function si() {
-  return f;
-}
-function ri() {
-  return f.memory();
-}
-function A(e, t) {
-  return f.tidy(e, t);
-}
-function D(e) {
-  ye(e).forEach((n) => n.dispose());
-}
-function Mn(e) {
-  return f.keep(e);
-}
-function ai(e) {
-  return f.setBackend(e);
-}
-function oi() {
-  return f.ready();
-}
-function ii() {
-  return f.backendName;
-}
-function ci(e, t, n = 1) {
-  return f.registerBackend(e, t, n);
-}
-function li() {
-  return f.backend;
-}
-const Jt = 4;
-async function ui(e, t) {
-  const n = [], s = [], r = Array.isArray(e) ? e.map((o) => o.name) : Object.keys(e);
-  for (let o = 0; o < r.length; ++o) {
-    const i = r[o], c = Array.isArray(e) ? e[o].tensor : e[i];
-    if (c.dtype !== "float32" && c.dtype !== "int32" && c.dtype !== "bool" && c.dtype !== "string" && c.dtype !== "complex64")
-      throw new Error(`Unsupported dtype in weight '${i}': ${c.dtype}`);
-    const l = { name: i, shape: c.shape, dtype: c.dtype };
-    if (c.dtype === "string") {
-      const u = new Promise(async (h) => {
-        const d = await c.bytes(), m = d.reduce((w, I) => w + I.length, 0) + Jt * d.length, y = new Uint8Array(m);
-        let g = 0;
-        for (let w = 0; w < d.length; w++) {
-          const I = d[w], _ = new Uint8Array(new Uint32Array([I.length]).buffer);
-          y.set(_, g), g += Jt, y.set(I, g), g += I.length;
-        }
-        h(y);
-      });
-      s.push(u);
-    } else
-      s.push(c.data());
-    t != null && (l.group = t), n.push(l);
-  }
-  const a = await Promise.all(s);
-  return { data: Dn(a), specs: n };
-}
-function Dn(e) {
-  if (e === null)
-    throw new Error(`Invalid input value: ${JSON.stringify(e)}`);
-  let t = 0;
-  const n = [];
-  e.forEach((a) => {
-    if (t += a.byteLength, n.push(a.byteLength === a.buffer.byteLength ? a : new a.constructor(a)), !(a instanceof Float32Array || a instanceof Int32Array || a instanceof Uint8Array))
-      throw new Error(`Unsupported TypedArray subtype: ${a.constructor.name}`);
-  });
-  const s = new Uint8Array(t);
-  let r = 0;
-  return n.forEach((a) => {
-    s.set(new Uint8Array(a.buffer), r), r += a.byteLength;
-  }), s.buffer;
-}
-const Ot = typeof pt < "u" && (typeof Blob > "u" || typeof atob > "u" || typeof btoa > "u");
-function Yt(e) {
-  return Ot ? pt.byteLength(e, "utf8") : new Blob([e]).size;
-}
-function Fn(e) {
-  if (Ot)
-    return pt.from(e).toString("base64");
-  const t = new Uint8Array(e);
-  let n = "";
-  for (let s = 0, r = t.length; s < r; s++)
-    n += String.fromCharCode(t[s]);
-  return btoa(n);
-}
-function Rn(e) {
-  if (Ot) {
-    const s = pt.from(e, "base64");
-    return s.buffer.slice(s.byteOffset, s.byteOffset + s.byteLength);
-  }
-  const t = atob(e), n = new Uint8Array(t.length);
-  for (let s = 0; s < t.length; ++s)
-    n.set([t.charCodeAt(s)], s);
-  return n.buffer;
-}
-function hi(e) {
-  return st.join(e);
-}
-function di(e, t) {
-  const n = {
-    modelTopology: e.modelTopology,
-    format: e.format,
-    generatedBy: e.generatedBy,
-    convertedBy: e.convertedBy,
-    weightsManifest: t
-  };
-  return e.signature != null && (n.signature = e.signature), e.userDefinedMetadata != null && (n.userDefinedMetadata = e.userDefinedMetadata), e.modelInitializer != null && (n.modelInitializer = e.modelInitializer), e.initializerSignature != null && (n.initializerSignature = e.initializerSignature), e.trainingConfig != null && (n.trainingConfig = e.trainingConfig), n;
-}
-function xn(e, t, n) {
-  const s = {
-    modelTopology: e.modelTopology,
-    format: e.format,
-    generatedBy: e.generatedBy,
-    convertedBy: e.convertedBy
-  };
-  if (e.trainingConfig != null && (s.trainingConfig = e.trainingConfig), e.weightsManifest != null) {
-    if (!t)
-      throw new Error("modelJSON has weightsManifest but weightSpecs is null");
-    if (!n)
-      throw new Error("modelJSON has weightsManifest but weightData is null");
-    s.weightSpecs = t, s.weightData = n;
-  }
-  return e.signature != null && (s.signature = e.signature), e.userDefinedMetadata != null && (s.userDefinedMetadata = e.userDefinedMetadata), e.modelInitializer != null && (s.modelInitializer = e.modelInitializer), e.initializerSignature != null && (s.initializerSignature = e.initializerSignature), s;
-}
-async function fi(e, t) {
-  let n, s;
-  return e.weightsManifest != null && ([n, s] = await t(e.weightsManifest)), xn(e, n, s);
-}
-function ke(e) {
-  if (e.modelTopology instanceof ArrayBuffer)
-    throw new Error("Expected JSON model topology, received ArrayBuffer.");
-  return {
-    dateSaved: /* @__PURE__ */ new Date(),
-    modelTopologyType: "JSON",
-    modelTopologyBytes: e.modelTopology == null ? 0 : Yt(JSON.stringify(e.modelTopology)),
-    weightSpecsBytes: e.weightSpecs == null ? 0 : Yt(JSON.stringify(e.weightSpecs)),
-    weightDataBytes: e.weightData == null ? 0 : new st(e.weightData).byteLength
-  };
-}
-function gi(e) {
-  const t = [];
-  for (const n of e)
-    t.push(...n.weights);
-  return t;
-}
-/**
- * @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.
- * =============================================================================
- */
-class E {
-  constructor() {
-    this.saveRouters = [], this.loadRouters = [];
-  }
-  static getInstance() {
-    return E.instance == null && (E.instance = new E()), E.instance;
-  }
-  /**
-   * Register a save-handler router.
-   *
-   * @param saveRouter A function that maps a URL-like string onto an instance
-   * of `IOHandler` with the `save` method defined or `null`.
-   */
-  static registerSaveRouter(t) {
-    E.getInstance().saveRouters.push(t);
-  }
-  /**
-   * Register a load-handler router.
-   *
-   * @param loadRouter A function that maps a URL-like string onto an instance
-   * of `IOHandler` with the `load` method defined or `null`.
-   */
-  static registerLoadRouter(t) {
-    E.getInstance().loadRouters.push(t);
-  }
-  /**
-   * Look up IOHandler for saving, given a URL-like string.
-   *
-   * @param url
-   * @returns If only one match is found, an instance of IOHandler with the
-   * `save` method defined. If no match is found, `null`.
-   * @throws Error, if more than one match is found.
-   */
-  static getSaveHandlers(t) {
-    return E.getHandlers(t, "save");
-  }
-  /**
-   * Look up IOHandler for loading, given a URL-like string.
-   *
-   * @param url
-   * @param loadOptions Optional, custom load options.
-   * @returns All valid handlers for `url`, given the currently registered
-   *   handler routers.
-   */
-  static getLoadHandlers(t, n) {
-    return E.getHandlers(t, "load", n);
-  }
-  static getHandlers(t, n, s) {
-    const r = [];
-    return (n === "load" ? E.getInstance().loadRouters : E.getInstance().saveRouters).forEach((o) => {
-      const i = o(t, s);
-      i !== null && r.push(i);
-    }), r;
-  }
-}
-const mi = (e) => E.getSaveHandlers(e);
-/**
- * @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.
- * =============================================================================
- */
-const Mt = "tensorflowjs", Dt = 1, U = "models_store", L = "model_info_store";
-function Ie() {
-  if (!S().getBool("IS_BROWSER"))
-    throw new Error("Failed to obtain IndexedDB factory because the current environmentis not a web browser.");
-  const e = typeof window > "u" ? self : window, t = e.indexedDB || e.mozIndexedDB || e.webkitIndexedDB || e.msIndexedDB || e.shimIndexedDB;
-  if (t == null)
-    throw new Error("The current browser does not appear to support IndexedDB.");
-  return t;
-}
-function Ft(e) {
-  const t = e.result;
-  t.createObjectStore(U, { keyPath: "modelPath" }), t.createObjectStore(L, { keyPath: "modelPath" });
-}
-class W {
-  constructor(t) {
-    if (this.indexedDB = Ie(), t == null || !t)
-      throw new Error("For IndexedDB, modelPath must not be null, undefined or empty.");
-    this.modelPath = t;
-  }
-  async save(t) {
-    if (t.modelTopology instanceof ArrayBuffer)
-      throw new Error("BrowserLocalStorage.save() does not support saving model topology in binary formats yet.");
-    return this.databaseAction(this.modelPath, t);
-  }
-  async load() {
-    return this.databaseAction(this.modelPath);
-  }
-  /**
-   * Perform database action to put model artifacts into or read model artifacts
-   * from IndexedDB object store.
-   *
-   * Whether the action is put or get depends on whether `modelArtifacts` is
-   * specified. If it is specified, the action will be put; otherwise the action
-   * will be get.
-   *
-   * @param modelPath A unique string path for the model.
-   * @param modelArtifacts If specified, it will be the model artifacts to be
-   *   stored in IndexedDB.
-   * @returns A `Promise` of `SaveResult`, if the action is put, or a `Promise`
-   *   of `ModelArtifacts`, if the action is get.
-   */
-  databaseAction(t, n) {
-    return new Promise((s, r) => {
-      const a = this.indexedDB.open(Mt, Dt);
-      a.onupgradeneeded = () => Ft(a), a.onsuccess = () => {
-        const o = a.result;
-        if (n == null) {
-          const i = o.transaction(U, "readonly"), l = i.objectStore(U).get(this.modelPath);
-          l.onsuccess = () => {
-            if (l.result == null)
-              return o.close(), r(new Error(`Cannot find model with path '${this.modelPath}' in IndexedDB.`));
-            s(l.result.modelArtifacts);
-          }, l.onerror = (u) => (o.close(), r(l.error)), i.oncomplete = () => o.close();
-        } else {
-          n.weightData = st.join(n.weightData);
-          const i = ke(n), c = o.transaction(L, "readwrite");
-          let l = c.objectStore(L), u;
-          try {
-            u = l.put({ modelPath: this.modelPath, modelArtifactsInfo: i });
-          } catch (d) {
-            return r(d);
-          }
-          let h;
-          u.onsuccess = () => {
-            h = o.transaction(U, "readwrite");
-            const d = h.objectStore(U);
-            let m;
-            try {
-              m = d.put({
-                modelPath: this.modelPath,
-                modelArtifacts: n,
-                modelArtifactsInfo: i
-              });
-            } catch (y) {
-              return r(y);
-            }
-            m.onsuccess = () => s({ modelArtifactsInfo: i }), m.onerror = (y) => {
-              l = c.objectStore(L);
-              const g = l.delete(this.modelPath);
-              g.onsuccess = () => (o.close(), r(m.error)), g.onerror = (w) => (o.close(), r(m.error));
-            };
-          }, u.onerror = (d) => (o.close(), r(u.error)), c.oncomplete = () => {
-            h == null ? o.close() : h.oncomplete = () => o.close();
-          };
-        }
-      }, a.onerror = (o) => r(a.error);
-    });
-  }
-}
-W.URL_SCHEME = "indexeddb://";
-const Te = (e) => S().getBool("IS_BROWSER") && !Array.isArray(e) && e.startsWith(W.URL_SCHEME) ? Nn(e.slice(W.URL_SCHEME.length)) : null;
-E.registerSaveRouter(Te);
-E.registerLoadRouter(Te);
-function Nn(e) {
-  return new W(e);
-}
-function $n(e) {
-  return e.startsWith(W.URL_SCHEME) ? e.slice(W.URL_SCHEME.length) : e;
-}
-class Cn {
-  constructor() {
-    this.indexedDB = Ie();
-  }
-  async listModels() {
-    return new Promise((t, n) => {
-      const s = this.indexedDB.open(Mt, Dt);
-      s.onupgradeneeded = () => Ft(s), s.onsuccess = () => {
-        const r = s.result, a = r.transaction(L, "readonly"), i = a.objectStore(L).getAll();
-        i.onsuccess = () => {
-          const c = {};
-          for (const l of i.result)
-            c[l.modelPath] = l.modelArtifactsInfo;
-          t(c);
-        }, i.onerror = (c) => (r.close(), n(i.error)), a.oncomplete = () => r.close();
-      }, s.onerror = (r) => n(s.error);
-    });
-  }
-  async removeModel(t) {
-    return t = $n(t), new Promise((n, s) => {
-      const r = this.indexedDB.open(Mt, Dt);
-      r.onupgradeneeded = () => Ft(r), r.onsuccess = () => {
-        const a = r.result, o = a.transaction(L, "readwrite"), i = o.objectStore(L), c = i.get(t);
-        let l;
-        c.onsuccess = () => {
-          if (c.result == null)
-            return a.close(), s(new Error(`Cannot find model with path '${t}' in IndexedDB.`));
-          {
-            const u = i.delete(t), h = () => {
-              l = a.transaction(U, "readwrite");
-              const m = l.objectStore(U).delete(t);
-              m.onsuccess = () => n(c.result.modelArtifactsInfo), m.onerror = (y) => s(c.error);
-            };
-            u.onsuccess = h, u.onerror = (d) => (h(), a.close(), s(c.error));
-          }
-        }, c.onerror = (u) => (a.close(), s(c.error)), o.oncomplete = () => {
-          l == null ? a.close() : l.oncomplete = () => a.close();
-        };
-      }, r.onerror = (a) => s(r.error);
-    });
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-const P = "/", Z = "tensorflowjs_models", Ae = "info", Pn = "model_topology", Ln = "weight_specs", _n = "weight_data", On = "model_metadata";
-function Ee(e) {
-  return {
-    info: [Z, e, Ae].join(P),
-    topology: [Z, e, Pn].join(P),
-    weightSpecs: [Z, e, Ln].join(P),
-    weightData: [Z, e, _n].join(P),
-    modelMetadata: [Z, e, On].join(P)
-  };
-}
-function ve(e) {
-  for (const t of Object.values(e))
-    window.localStorage.removeItem(t);
-}
-function Gn(e) {
-  const t = e.split(P);
-  if (t.length < 3)
-    throw new Error(`Invalid key format: ${e}`);
-  return t.slice(1, t.length - 1).join(P);
-}
-function Un(e) {
-  return e.startsWith(j.URL_SCHEME) ? e.slice(j.URL_SCHEME.length) : e;
-}
-class j {
-  constructor(t) {
-    if (!S().getBool("IS_BROWSER") || typeof window > "u" || typeof window.localStorage > "u")
-      throw new Error("The current environment does not support local storage.");
-    if (this.LS = window.localStorage, t == null || !t)
-      throw new Error("For local storage, modelPath must not be null, undefined or empty.");
-    this.modelPath = t, this.keys = Ee(this.modelPath);
-  }
-  /**
-   * Save model artifacts to browser local storage.
-   *
-   * See the documentation to `browserLocalStorage` for details on the saved
-   * artifacts.
-   *
-   * @param modelArtifacts The model artifacts to be stored.
-   * @returns An instance of SaveResult.
-   */
-  async save(t) {
-    if (t.modelTopology instanceof ArrayBuffer)
-      throw new Error("BrowserLocalStorage.save() does not support saving model topology in binary formats yet.");
-    {
-      const n = JSON.stringify(t.modelTopology), s = JSON.stringify(t.weightSpecs), r = ke(t), a = st.join(t.weightData);
-      try {
-        this.LS.setItem(this.keys.info, JSON.stringify(r)), this.LS.setItem(this.keys.topology, n), this.LS.setItem(this.keys.weightSpecs, s), this.LS.setItem(this.keys.weightData, Fn(a));
-        const o = {
-          format: t.format,
-          generatedBy: t.generatedBy,
-          convertedBy: t.convertedBy,
-          signature: t.signature != null ? t.signature : void 0,
-          userDefinedMetadata: t.userDefinedMetadata != null ? t.userDefinedMetadata : void 0,
-          modelInitializer: t.modelInitializer != null ? t.modelInitializer : void 0,
-          initializerSignature: t.initializerSignature != null ? t.initializerSignature : void 0,
-          trainingConfig: t.trainingConfig != null ? t.trainingConfig : void 0
-        };
-        return this.LS.setItem(this.keys.modelMetadata, JSON.stringify(o)), { modelArtifactsInfo: r };
-      } catch {
-        throw ve(this.keys), new Error(`Failed to save model '${this.modelPath}' to local storage: size quota being exceeded is a possible cause of this failure: modelTopologyBytes=${r.modelTopologyBytes}, weightSpecsBytes=${r.weightSpecsBytes}, weightDataBytes=${r.weightDataBytes}.`);
-      }
-    }
-  }
-  /**
-   * Load a model from local storage.
-   *
-   * See the documentation to `browserLocalStorage` for details on the saved
-   * artifacts.
-   *
-   * @returns The loaded model (if loading succeeds).
-   */
-  async load() {
-    const t = JSON.parse(this.LS.getItem(this.keys.info));
-    if (t == null)
-      throw new Error(`In local storage, there is no model with name '${this.modelPath}'`);
-    if (t.modelTopologyType !== "JSON")
-      throw new Error("BrowserLocalStorage does not support loading non-JSON model topology yet.");
-    const n = {}, s = JSON.parse(this.LS.getItem(this.keys.topology));
-    if (s == null)
-      throw new Error(`In local storage, the topology of model '${this.modelPath}' is missing.`);
-    n.modelTopology = s;
-    const r = JSON.parse(this.LS.getItem(this.keys.weightSpecs));
-    if (r == null)
-      throw new Error(`In local storage, the weight specs of model '${this.modelPath}' are missing.`);
-    n.weightSpecs = r;
-    const a = this.LS.getItem(this.keys.modelMetadata);
-    if (a != null) {
-      const i = JSON.parse(a);
-      n.format = i.format, n.generatedBy = i.generatedBy, n.convertedBy = i.convertedBy, i.signature != null && (n.signature = i.signature), i.userDefinedMetadata != null && (n.userDefinedMetadata = i.userDefinedMetadata), i.modelInitializer != null && (n.modelInitializer = i.modelInitializer), i.initializerSignature != null && (n.initializerSignature = i.initializerSignature), i.trainingConfig != null && (n.trainingConfig = i.trainingConfig);
-    }
-    const o = this.LS.getItem(this.keys.weightData);
-    if (o == null)
-      throw new Error(`In local storage, the binary weight values of model '${this.modelPath}' are missing.`);
-    return n.weightData = Rn(o), n;
-  }
-}
-j.URL_SCHEME = "localstorage://";
-const Be = (e) => S().getBool("IS_BROWSER") && !Array.isArray(e) && e.startsWith(j.URL_SCHEME) ? zn(e.slice(j.URL_SCHEME.length)) : null;
-E.registerSaveRouter(Be);
-E.registerLoadRouter(Be);
-function zn(e) {
-  return new j(e);
-}
-class Vn {
-  constructor() {
-    p(S().getBool("IS_BROWSER"), () => "Current environment is not a web browser"), p(typeof window > "u" || typeof window.localStorage < "u", () => "Current browser does not appear to support localStorage"), this.LS = window.localStorage;
-  }
-  async listModels() {
-    const t = {}, n = Z + P, s = P + Ae;
-    for (let r = 0; r < this.LS.length; ++r) {
-      const a = this.LS.key(r);
-      if (a.startsWith(n) && a.endsWith(s)) {
-        const o = Gn(a);
-        t[o] = JSON.parse(this.LS.getItem(a));
-      }
-    }
-    return t;
-  }
-  async removeModel(t) {
-    t = Un(t);
-    const n = Ee(t);
-    if (this.LS.getItem(n.info) == null)
-      throw new Error(`Cannot find model at path '${t}'`);
-    const s = JSON.parse(this.LS.getItem(n.info));
-    return ve(n), s;
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-const Zt = "://";
-class N {
-  constructor() {
-    this.managers = {};
-  }
-  static getInstance() {
-    return N.instance == null && (N.instance = new N()), N.instance;
-  }
-  /**
-   * Register a save-handler router.
-   *
-   * @param saveRouter A function that maps a URL-like string onto an instance
-   * of `IOHandler` with the `save` method defined or `null`.
-   */
-  static registerManager(t, n) {
-    p(t != null, () => "scheme must not be undefined or null."), t.endsWith(Zt) && (t = t.slice(0, t.indexOf(Zt))), p(t.length > 0, () => "scheme must not be an empty string.");
-    const s = N.getInstance();
-    p(s.managers[t] == null, () => `A model store manager is already registered for scheme '${t}'.`), s.managers[t] = n;
-  }
-  static getManager(t) {
-    const n = N.getInstance().managers[t];
-    if (n == null)
-      throw new Error(`Cannot find model manager for scheme '${t}'`);
-    return n;
-  }
-  static getSchemes() {
-    return Object.keys(N.getInstance().managers);
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-class Wn {
-  constructor() {
-    this.messageName = "setTimeoutCustom", this.functionRefs = [], this.handledMessageCount = 0, this.hasEventListener = !1;
-  }
-  fetch(t, n) {
-    return fetch(t, n);
-  }
-  now() {
-    return performance.now();
-  }
-  encode(t, n) {
-    if (n !== "utf-8" && n !== "utf8")
-      throw new Error(`Browser's encoder only supports utf-8, but got ${n}`);
-    return this.textEncoder == null && (this.textEncoder = new TextEncoder()), this.textEncoder.encode(t);
-  }
-  decode(t, n) {
-    return new TextDecoder(n).decode(t);
-  }
-  // If the setTimeout nesting level is greater than 5 and timeout is less
-  // than 4ms, timeout will be clamped to 4ms, which hurts the perf.
-  // Interleaving window.postMessage and setTimeout will trick the browser and
-  // avoid the clamp.
-  setTimeoutCustom(t, n) {
-    if (typeof window > "u" || !S().getBool("USE_SETTIMEOUTCUSTOM")) {
-      setTimeout(t, n);
-      return;
-    }
-    this.functionRefs.push(t), setTimeout(() => {
-      window.postMessage({ name: this.messageName, index: this.functionRefs.length - 1 }, "*");
-    }, n), this.hasEventListener || (this.hasEventListener = !0, window.addEventListener("message", (s) => {
-      if (s.source === window && s.data.name === this.messageName) {
-        s.stopPropagation();
-        const r = this.functionRefs[s.data.index];
-        r(), this.handledMessageCount++, this.handledMessageCount === this.functionRefs.length && (this.functionRefs = [], this.handledMessageCount = 0);
-      }
-    }, !0));
-  }
-  isTypedArray(t) {
-    return ue(t);
-  }
-}
-if (S().get("IS_BROWSER")) {
-  S().setPlatform("browser", new Wn());
-  try {
-    N.registerManager(j.URL_SCHEME, new Vn());
-  } catch {
-  }
-  try {
-    N.registerManager(W.URL_SCHEME, new Cn());
-  } catch {
-  }
-}
-const jn = {
-  // tslint:disable-next-line:no-require-imports
-  importFetch: () => require("node-fetch")
-};
-let kt;
-class qn {
-  constructor() {
-    this.util = require("util"), this.textEncoder = new this.util.TextEncoder();
-  }
-  fetch(t, n) {
-    return S().global.fetch != null ? S().global.fetch(t, n) : (kt == null && (kt = jn.importFetch()), kt(t, n));
-  }
-  now() {
-    const t = Q.hrtime();
-    return t[0] * 1e3 + t[1] / 1e6;
-  }
-  encode(t, n) {
-    if (n !== "utf-8" && n !== "utf8")
-      throw new Error(`Node built-in encoder only supports utf-8, but got ${n}`);
-    return this.textEncoder.encode(t);
-  }
-  decode(t, n) {
-    return t.length === 0 ? "" : new this.util.TextDecoder(n).decode(t);
-  }
-  isTypedArray(t) {
-    return this.util.types.isFloat32Array(t) || this.util.types.isInt32Array(t) || this.util.types.isUint8Array(t) || this.util.types.isUint8ClampedArray(t);
-  }
-}
-S().get("IS_NODE") && !S().get("IS_BROWSER") && S().setPlatform("node", new qn());
-/**
- * @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.
- * =============================================================================
- */
-function Kn(e, t = "float32", n) {
-  return t = t || "float32", Ct(e), new gn(e, t, n);
-}
-/**
- * @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.
- * =============================================================================
- */
-function Hn(e, t) {
-  const n = T(e, "x", "cast");
-  if (!Le(t))
-    throw new Error(`Failed to cast to unknown dtype ${t}`);
-  if (t === "string" && n.dtype !== "string" || t !== "string" && n.dtype === "string")
-    throw new Error("Only strings can be casted to strings");
-  const s = { x: n }, r = { dtype: t };
-  return f.runKernel(ie, s, r);
-}
-const Rt = /* @__PURE__ */ F({ cast_: Hn });
-/**
- * @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.
- * =============================================================================
- */
-function Xn(e) {
-  const n = { x: T(e, "x", "clone", "string_or_numeric") };
-  return f.runKernel(ce, n);
-}
-const Jn = /* @__PURE__ */ F({ clone_: Xn });
-/**
- * @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.
- * =============================================================================
- */
-function Yn(e, t = !1) {
-  console.log(e.toString(t));
-}
-/**
- * @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.
- * =============================================================================
- */
-we();
-const Zn = {
-  buffer: Kn,
-  cast: Rt,
-  clone: Jn,
-  print: Yn
-};
-pn(Zn);
-/**
- * @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.
- * =============================================================================
- */
-function Qn(e, t) {
-  let n = T(e, "a", "add"), s = T(t, "b", "add");
-  [n, s] = K(n, s);
-  const r = { a: n, b: s };
-  return f.runKernel(oe, r);
-}
-const k = /* @__PURE__ */ F({ add_: Qn });
-/**
- * @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.
- * =============================================================================
- */
-function ts(e, t) {
-  let n = T(e, "a", "floorDiv"), s = T(t, "b", "floorDiv");
-  [n, s] = K(n, s);
-  const r = { a: n, b: s };
-  return f.runKernel(Ze, r);
-}
-const es = /* @__PURE__ */ F({ floorDiv_: ts });
-/**
- * @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.
- * =============================================================================
- */
-function ns(e, t) {
-  let n = T(e, "a", "div"), s = T(t, "b", "div");
-  if ([n, s] = K(n, s), n.dtype === "int32" && s.dtype === "int32")
-    return es(n, s);
-  const r = { a: n, b: s }, a = {};
-  return f.runKernel(Je, r, a);
-}
-const $ = /* @__PURE__ */ F({ div_: ns });
-/**
- * @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.
- * =============================================================================
- */
-function ss(e, t) {
-  let n = T(e, "a", "mul"), s = T(t, "b", "mul");
-  [n, s] = K(n, s);
-  const r = { a: n, b: s };
-  return f.runKernel(tn, r);
-}
-const b = /* @__PURE__ */ F({ mul_: ss });
-/**
- * @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.
- * =============================================================================
- */
-function rs(e) {
-  const t = T(e, "x", "abs");
-  if (t.dtype === "complex64") {
-    const n = { x: t };
-    return f.runKernel(Xe, n);
-  } else {
-    const n = { x: t };
-    return f.runKernel(He, n);
-  }
-}
-const as = /* @__PURE__ */ F({ abs_: rs });
-/**
- * @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.
- * =============================================================================
- */
-function os(e, t, n) {
-  Ct(e), n = n || yt(t);
-  const s = { shape: e, value: t, dtype: n };
-  return f.runKernel(Ye, {}, s);
-}
-/**
- * @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.
- * =============================================================================
- */
-function pi(e, t) {
-  const n = e.length, s = [];
-  for (let r = 0; r < n; r++) {
-    const a = n - 1 - r, o = e[a] || 1;
-    (t[t.length - 1 - r] || 1) > 1 && o === 1 && s.unshift(a);
-  }
-  return s;
-}
-function yi(e, t) {
-  const n = [];
-  for (let s = 0; s < t.length; s++) {
-    const r = e[e.length - s - 1], a = t.length - s - 1, o = t[a];
-    (r == null || r === 1 && o > 1) && n.unshift(a);
-  }
-  return n;
-}
-function is(e, t) {
-  const n = Math.max(e.length, t.length), s = new Array(n);
-  for (let r = 0; r < n; r++) {
-    let a = e[e.length - r - 1];
-    a == null && (a = 1);
-    let o = t[t.length - r - 1];
-    if (o == null && (o = 1), a === 1)
-      s[n - r - 1] = o;
-    else if (o === 1)
-      s[n - r - 1] = a;
-    else if (a !== o) {
-      const i = `Operands could not be broadcast together with shapes ${e} and ${t}.`;
-      throw Error(i);
-    } else
-      s[n - r - 1] = a;
-  }
-  return s;
-}
-/**
- * @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.
- * =============================================================================
- */
-function cs(e) {
-  const n = { x: T(e, "x", "zerosLike") };
-  return f.runKernel(rn, n);
-}
-const C = /* @__PURE__ */ F({ zerosLike_: cs });
-/**
- * @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.
- * =============================================================================
- */
-function ls(e, t) {
-  let n = T(e, "base", "pow"), s = T(t, "exp", "pow");
-  [n, s] = K(n, s);
-  const r = { a: n, b: s };
-  return f.runKernel(en, r);
-}
-const Qt = /* @__PURE__ */ F({ pow_: ls });
-/**
- * @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.
- * =============================================================================
- */
-function q(e, t) {
-  if ((R(e) && t !== "string" || Array.isArray(e)) && t !== "complex64")
-    throw new Error("Error creating a new Scalar: value must be a primitive (number|boolean|string)");
-  if (t === "string" && R(e) && !(e instanceof Uint8Array))
-    throw new Error("When making a scalar from encoded string, the value must be `Uint8Array`.");
-  return vn(e, [], [], t);
-}
-/**
- * @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.
- * =============================================================================
- */
-function us(e) {
-  const n = { x: T(e, "x", "sqrt", "float32") };
-  return f.runKernel(nn, n);
-}
-const nt = /* @__PURE__ */ F({ sqrt_: us });
-/**
- * @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.
- * =============================================================================
- */
-function hs(e) {
-  const t = T(e, "x", "square"), n = {};
-  return f.runKernel("Square", { x: t }, n);
-}
-const V = /* @__PURE__ */ F({ square_: hs });
-/**
- * @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.
- * =============================================================================
- */
-function bi(e) {
-  return p(lt(e), () => "The f passed in grads(f) must be a function"), (t, n) => {
-    p(Array.isArray(t), () => "The args passed in grads(f)(args) must be an array of `Tensor`s or `TensorLike`s");
-    const s = An(t, "args", "tf.grads", "string_or_numeric"), r = n != null ? T(n, "dy", "tf.grads") : null;
-    return f.tidy(() => {
-      const { value: a, grads: o } = f.gradients(() => e(...s), s, r);
-      return r != null && te(a.shape, r.shape, "The shape of dy passed in grads(f)([x1,...], dy) must match the shape returned by f([x1,...])"), Me(o), o;
-    });
-  };
-}
-function wi(e) {
-  return p(lt(e), () => "The f passed in valueAndGrads(f) must be a function"), (t, n) => {
-    p(Array.isArray(t) && t.every((r) => r instanceof M), () => "The args passed in valueAndGrads(f)(args) must be array of tensors"), p(n == null || n instanceof M, () => "The dy passed in valueAndGrads(f)(args, dy) must be a tensor");
-    const s = f.gradients(() => e(...t), t, n);
-    return n != null && te(s.value.shape, n.shape, "The shape of dy passed in valueAndGrads(f)([x1,...], dy) must match the shape returned by f([x1,...])"), Me(s.grads), s;
-  };
-}
-function ds(e, t) {
-  p(lt(e), () => "The f passed in variableGrads(f) must be a function"), p(t == null || Array.isArray(t) && t.every((l) => l instanceof mt), () => "The varList passed in variableGrads(f, varList) must be an array of variables");
-  const n = t != null;
-  if (!n) {
-    t = [];
-    for (const l in f.registeredVariables)
-      t.push(f.registeredVariables[l]);
-  }
-  const s = n ? t.filter((l) => !l.trainable) : null, r = t.length;
-  t = t.filter((l) => l.trainable), p(t.length > 0, () => `variableGrads() expects at least one of the input variables to be trainable, but none of the ${r} variables is trainable.`);
-  const a = !0, { value: o, grads: i } = f.gradients(e, t, null, a);
-  p(i.some((l) => l != null), () => "Cannot find a connection between any variable and the result of the loss function y=f(x). Please make sure the operations that use variables are inside the function f passed to minimize()."), p(o.rank === 0, () => `The f passed in variableGrads(f) must return a scalar, but it returned a rank-${o.rank} tensor`);
-  const c = {};
-  return t.forEach((l, u) => {
-    i[u] != null && (c[l.name] = i[u]);
-  }), s?.forEach((l) => c[l.name] = null), { value: o, grads: c };
-}
-function Si(e) {
-  return f.customGrad(e);
-}
-function Me(e) {
-  if (e.filter((n) => n == 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.`);
-}
-/**
- * @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.
- * =============================================================================
- */
-function fs(e, t) {
-  let n = T(e, "a", "sub"), s = T(t, "b", "sub");
-  [n, s] = K(n, s);
-  const r = { a: n, b: s };
-  return f.runKernel(sn, r);
-}
-const tt = /* @__PURE__ */ F({ sub_: fs });
-/**
- * @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.
- * =============================================================================
- */
-function gs(e, t) {
-  let n = T(e, "a", "maximum"), s = T(t, "b", "maximum");
-  [n, s] = K(n, s), n.dtype === "bool" && (n = Rt(n, "int32"), s = Rt(s, "int32")), is(n.shape, s.shape);
-  const r = { a: n, b: s };
-  return f.runKernel(Qe, r);
-}
-const ms = /* @__PURE__ */ F({ maximum_: gs });
-/**
- * @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.
- * =============================================================================
- */
-const ps = /* @__PURE__ */ new Map(), ys = /* @__PURE__ */ new Map();
-class bs {
-  /**
-   * Return the class name for this class to use in serialization contexts.
-   *
-   * Generally speaking this will be the same thing that constructor.name
-   * would have returned.  However, the class name needs to be robust
-   * against minification for serialization/deserialization to work properly.
-   *
-   * There's also places such as initializers.VarianceScaling, where
-   * implementation details between different languages led to different
-   * class hierarchies and a non-leaf node is used for serialization purposes.
-   */
-  getClassName() {
-    return this.constructor.className;
-  }
-  /**
-   * Creates an instance of T from a ConfigDict.
-   *
-   * This works for most descendants of serializable.  A few need to
-   * provide special handling.
-   * @param cls A Constructor for the class to instantiate.
-   * @param config The Configuration for the object.
-   */
-  /** @nocollapse */
-  static fromConfig(t, n) {
-    return new t(n);
-  }
-}
-class G {
-  constructor() {
-    this.classNameMap = {};
-  }
-  /**
-   * Returns the singleton instance of the map.
-   */
-  static getMap() {
-    return G.instance == null && (G.instance = new G()), G.instance;
-  }
-  /**
-   * Registers the class as serializable.
-   */
-  static register(t) {
-    G.getMap().classNameMap[t.className] = [t, t.fromConfig];
-  }
-}
-function ws(e, t, n) {
-  p(e.className != null, () => "Class being registered does not have the static className property defined."), p(typeof e.className == "string", () => "className is required to be a string, but got type " + typeof e.className), p(e.className.length > 0, () => "Class being registered has an empty-string as its className, which is disallowed."), typeof t > "u" && (t = "Custom"), typeof n > "u" && (n = e.className);
-  const s = n, r = t + ">" + s;
-  return G.register(e), ps.set(r, e), ys.set(e, r), e;
-}
-/**
- * @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.
- * =============================================================================
- */
-class H extends bs {
-  /**
-   * Executes `f()` and minimizes the scalar output of `f()` by computing
-   * gradients of y with respect to the list of trainable variables provided by
-   * `varList`. If no list is provided, it defaults to all trainable variables.
-   *
-   * @param f The function to execute and whose output to minimize.
-   * @param returnCost Whether to return the scalar cost value produced by
-   * executing `f()`.
-   * @param varList An optional list of variables to update. If specified, only
-   * the trainable variables in varList will be updated by minimize. Defaults to
-   * all trainable variables.
-   *
-   * @doc {heading: 'Training', subheading: 'Optimizers'}
-   */
-  minimize(t, n = !1, s) {
-    const { value: r, grads: a } = this.computeGradients(t, s);
-    if (s != null) {
-      const o = s.map((i) => ({ name: i.name, tensor: a[i.name] }));
-      this.applyGradients(o);
-    } else
-      this.applyGradients(a);
-    return D(a), n ? r : (r.dispose(), null);
-  }
-  /**
-   * The number of iterations that this optimizer instance has been invoked for.
-   */
-  get iterations() {
-    return this.iterations_ == null && (this.iterations_ = 0), this.iterations_;
-  }
-  incrementIterations() {
-    this.iterations_ = this.iterations + 1;
-  }
-  /**
-   * Executes f() and computes the gradient of the scalar output of f() with
-   * respect to the list of trainable variables provided by `varList`. If no
-   * list is provided, it defaults to all trainable variables.
-   *
-   * @param f The function to execute and whose output to use for computing
-   * gradients with respect to variables.
-   * @param varList An optional list of variables to compute gradients with
-   * respect to. If specified, only the trainable variables in varList will have
-   * gradients computed with respect to. Defaults to all trainable variables.
-   *
-   * @doc {heading: 'Training', subheading: 'Optimizers'}
-   */
-  computeGradients(t, n) {
-    return ds(t, n);
-  }
-  /**
-   * Dispose the variables (if any) owned by this optimizer instance.
-   */
-  dispose() {
-    this.iterations_ != null && D(this.iterations_);
-  }
-  async saveIterations() {
-    return this.iterations_ == null && (this.iterations_ = 0), {
-      name: "iter",
-      // TODO(cais): Use 'int64' type when available.
-      tensor: q(this.iterations_, "int32")
-    };
-  }
-  async getWeights() {
-    throw new Error("getWeights() is not implemented for this optimizer yet.");
-  }
-  async setWeights(t) {
-    throw new Error(`setWeights() is not implemented for this optimizer class ${this.getClassName()}`);
-  }
-  /**
-   * Extract the first element of the weight values and set it
-   * as the iterations counter variable of this instance of optimizer.
-   *
-   * @param weightValues
-   * @returns Weight values with the first element consumed and excluded.
-   */
-  async extractIterations(t) {
-    return this.iterations_ = (await t[0].tensor.data())[0], t.slice(1);
-  }
-}
-Object.defineProperty(H, Symbol.hasInstance, {
-  value: (e) => e.minimize != null && e.computeGradients != null && e.applyGradients != null
-});
-/**
- * @license
- * Copyright 2018 Google LLC. All Rights Reserved.
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- * http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- * =============================================================================
- */
-class Ss extends H {
-  /** @nocollapse */
-  static get className() {
-    return "Adadelta";
-  }
-  constructor(t, n, s = null) {
-    super(), this.learningRate = t, this.rho = n, this.epsilon = s, this.accumulatedGrads = [], this.accumulatedUpdates = [], s == null && (this.epsilon = f.backend.epsilon());
-  }
-  applyGradients(t) {
-    (Array.isArray(t) ? t.map((s) => s.name) : Object.keys(t)).forEach((s, r) => {
-      const a = f.registeredVariables[s], o = !1;
-      this.accumulatedGrads[r] == null && (this.accumulatedGrads[r] = {
-        originalName: `${s}/accum_grad`,
-        variable: A(() => C(a).variable(o))
-      }), this.accumulatedUpdates[r] == null && (this.accumulatedUpdates[r] = {
-        originalName: `${s}/accum_var`,
-        variable: A(() => C(a).variable(o))
-      });
-      const i = Array.isArray(t) ? t[r].tensor : t[s];
-      if (i == null)
-        return;
-      const c = this.accumulatedGrads[r].variable, l = this.accumulatedUpdates[r].variable;
-      A(() => {
-        const u = k(b(c, this.rho), b(V(i), 1 - this.rho)), h = b($(nt(k(l, this.epsilon)), nt(k(c, this.epsilon))), i), d = k(b(l, this.rho), b(V(h), 1 - this.rho));
-        c.assign(u), l.assign(d);
-        const m = k(b(h, -this.learningRate), a);
-        a.assign(m);
-      });
-    }), this.incrementIterations();
-  }
-  dispose() {
-    this.accumulatedUpdates != null && (D(this.accumulatedGrads.map((t) => t.variable)), D(this.accumulatedUpdates.map((t) => t.variable)));
-  }
-  async getWeights() {
-    const t = [...this.accumulatedGrads, ...this.accumulatedUpdates];
-    return [await this.saveIterations()].concat(t.map((n) => ({ name: n.originalName, tensor: n.variable })));
-  }
-  async setWeights(t) {
-    t = await this.extractIterations(t);
-    const n = t.length / 2, s = !1;
-    this.accumulatedGrads = t.slice(0, n).map((r) => ({
-      originalName: r.name,
-      variable: r.tensor.variable(s)
-    })), this.accumulatedUpdates = t.slice(n, n * 2).map((r) => ({
-      originalName: r.name,
-      variable: r.tensor.variable(s)
-    }));
-  }
-  getConfig() {
-    return {
-      learningRate: this.learningRate,
-      rho: this.rho,
-      epsilon: this.epsilon
-    };
-  }
-  /** @nocollapse */
-  static fromConfig(t, n) {
-    return new t(n.learningRate, n.rho, n.epsilon);
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-class ks extends H {
-  /** @nocollapse */
-  static get className() {
-    return "Adagrad";
-  }
-  constructor(t, n = 0.1) {
-    super(), this.learningRate = t, this.initialAccumulatorValue = n, this.accumulatedGrads = [];
-  }
-  applyGradients(t) {
-    (Array.isArray(t) ? t.map((s) => s.name) : Object.keys(t)).forEach((s, r) => {
-      const a = f.registeredVariables[s];
-      this.accumulatedGrads[r] == null && (this.accumulatedGrads[r] = {
-        originalName: `${s}/accumulator`,
-        variable: A(() => os(a.shape, this.initialAccumulatorValue).variable(!1))
-      });
-      const o = Array.isArray(t) ? t[r].tensor : t[s];
-      if (o == null)
-        return;
-      const i = this.accumulatedGrads[r].variable;
-      A(() => {
-        const c = k(i, V(o));
-        i.assign(c);
-        const l = k(b($(o, nt(k(c, f.backend.epsilon()))), -this.learningRate), a);
-        a.assign(l);
-      });
-    }), this.incrementIterations();
-  }
-  dispose() {
-    this.accumulatedGrads != null && D(this.accumulatedGrads.map((t) => t.variable));
-  }
-  async getWeights() {
-    return [await this.saveIterations()].concat(this.accumulatedGrads.map((t) => ({ name: t.originalName, tensor: t.variable })));
-  }
-  async setWeights(t) {
-    t = await this.extractIterations(t);
-    const n = !1;
-    this.accumulatedGrads = t.map((s) => ({ originalName: s.name, variable: s.tensor.variable(n) }));
-  }
-  getConfig() {
-    return {
-      learningRate: this.learningRate,
-      initialAccumulatorValue: this.initialAccumulatorValue
-    };
-  }
-  /** @nocollapse */
-  static fromConfig(t, n) {
-    return new t(n.learningRate, n.initialAccumulatorValue);
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-class Is extends H {
-  /** @nocollapse */
-  static get className() {
-    return "Adam";
-  }
-  constructor(t, n, s, r = null) {
-    super(), this.learningRate = t, this.beta1 = n, this.beta2 = s, this.epsilon = r, this.accumulatedFirstMoment = [], this.accumulatedSecondMoment = [], A(() => {
-      this.accBeta1 = q(n).variable(), this.accBeta2 = q(s).variable();
-    }), r == null && (this.epsilon = f.backend.epsilon());
-  }
-  applyGradients(t) {
-    const n = Array.isArray(t) ? t.map((s) => s.name) : Object.keys(t);
-    A(() => {
-      const s = tt(1, this.accBeta1), r = tt(1, this.accBeta2);
-      n.forEach((a, o) => {
-        const i = f.registeredVariables[a], c = !1;
-        this.accumulatedFirstMoment[o] == null && (this.accumulatedFirstMoment[o] = {
-          originalName: `${a}/m`,
-          variable: A(() => C(i).variable(c))
-        }), this.accumulatedSecondMoment[o] == null && (this.accumulatedSecondMoment[o] = {
-          originalName: `${a}/v`,
-          variable: A(() => C(i).variable(c))
-        });
-        const l = Array.isArray(t) ? t[o].tensor : t[a];
-        if (l == null)
-          return;
-        const u = this.accumulatedFirstMoment[o].variable, h = this.accumulatedSecondMoment[o].variable, d = k(b(u, this.beta1), b(l, 1 - this.beta1)), m = k(b(h, this.beta2), b(V(l), 1 - this.beta2)), y = $(d, s), g = $(m, r);
-        u.assign(d), h.assign(m);
-        const w = k(b($(y, k(nt(g), this.epsilon)), -this.learningRate), i);
-        i.assign(w);
-      }), this.accBeta1.assign(b(this.accBeta1, this.beta1)), this.accBeta2.assign(b(this.accBeta2, this.beta2));
-    }), this.incrementIterations();
-  }
-  dispose() {
-    this.accBeta1.dispose(), this.accBeta2.dispose(), this.accumulatedFirstMoment != null && D(this.accumulatedFirstMoment.map((t) => t.variable)), this.accumulatedSecondMoment != null && D(this.accumulatedSecondMoment.map((t) => t.variable));
-  }
-  async getWeights() {
-    const t = [...this.accumulatedFirstMoment, ...this.accumulatedSecondMoment];
-    return [await this.saveIterations()].concat(t.map((n) => ({ name: n.originalName, tensor: n.variable })));
-  }
-  async setWeights(t) {
-    t = await this.extractIterations(t), A(() => {
-      this.accBeta1.assign(Qt(this.beta1, this.iterations_ + 1)), this.accBeta2.assign(Qt(this.beta2, this.iterations_ + 1));
-    });
-    const n = t.length / 2, s = !1;
-    this.accumulatedFirstMoment = t.slice(0, n).map((r) => ({
-      originalName: r.name,
-      variable: r.tensor.variable(s)
-    })), this.accumulatedSecondMoment = t.slice(n, n * 2).map((r) => ({
-      originalName: r.name,
-      variable: r.tensor.variable(s)
-    }));
-  }
-  getConfig() {
-    return {
-      learningRate: this.learningRate,
-      beta1: this.beta1,
-      beta2: this.beta2,
-      epsilon: this.epsilon
-    };
-  }
-  /** @nocollapse */
-  static fromConfig(t, n) {
-    return new t(n.learningRate, n.beta1, n.beta2, n.epsilon);
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-class Ts extends H {
-  /** @nocollapse */
-  static get className() {
-    return "Adamax";
-  }
-  constructor(t, n, s, r = null, a = 0) {
-    super(), this.learningRate = t, this.beta1 = n, this.beta2 = s, this.epsilon = r, this.decay = a, this.accumulatedFirstMoment = [], this.accumulatedWeightedInfNorm = [], A(() => {
-      this.iteration = q(0).variable(), this.accBeta1 = q(n).variable();
-    }), r == null && (this.epsilon = f.backend.epsilon());
-  }
-  applyGradients(t) {
-    const n = Array.isArray(t) ? t.map((s) => s.name) : Object.keys(t);
-    A(() => {
-      const s = tt(1, this.accBeta1), r = $(-this.learningRate, k(b(this.iteration, this.decay), 1));
-      n.forEach((a, o) => {
-        const i = f.registeredVariables[a], c = !1;
-        this.accumulatedFirstMoment[o] == null && (this.accumulatedFirstMoment[o] = {
-          originalName: `${a}/m`,
-          variable: C(i).variable(c)
-        }), this.accumulatedWeightedInfNorm[o] == null && (this.accumulatedWeightedInfNorm[o] = {
-          originalName: `${a}/v`,
-          variable: C(i).variable(c)
-        });
-        const l = Array.isArray(t) ? t[o].tensor : t[a];
-        if (l == null)
-          return;
-        const u = this.accumulatedFirstMoment[o].variable, h = this.accumulatedWeightedInfNorm[o].variable, d = k(b(u, this.beta1), b(l, 1 - this.beta1)), m = b(h, this.beta2), y = as(l), g = ms(m, y);
-        u.assign(d), h.assign(g);
-        const w = k(b($(r, s), $(d, k(g, this.epsilon))), i);
-        i.assign(w);
-      }), this.iteration.assign(k(this.iteration, 1)), this.accBeta1.assign(b(this.accBeta1, this.beta1));
-    }), this.incrementIterations();
-  }
-  dispose() {
-    this.accBeta1.dispose(), this.iteration.dispose(), this.accumulatedFirstMoment != null && D(this.accumulatedFirstMoment.map((t) => t.variable)), this.accumulatedWeightedInfNorm != null && D(this.accumulatedWeightedInfNorm.map((t) => t.variable));
-  }
-  async getWeights() {
-    throw new Error("getWeights() is not implemented for Adamax yet.");
-  }
-  async setWeights(t) {
-    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
-    };
-  }
-  /** @nocollapse */
-  static fromConfig(t, n) {
-    return new t(n.learningRate, n.beta1, n.beta2, n.epsilon, n.decay);
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-class De extends H {
-  /** @nocollapse */
-  static get className() {
-    return "SGD";
-  }
-  constructor(t) {
-    super(), this.learningRate = t, this.setLearningRate(t);
-  }
-  applyGradients(t) {
-    (Array.isArray(t) ? t.map((s) => s.name) : Object.keys(t)).forEach((s, r) => {
-      const a = Array.isArray(t) ? t[r].tensor : t[s];
-      if (a == null)
-        return;
-      const o = f.registeredVariables[s];
-      A(() => {
-        const i = k(b(this.c, a), o);
-        o.assign(i);
-      });
-    }), this.incrementIterations();
-  }
-  /**
-   * Sets the learning rate of the optimizer.
-   */
-  setLearningRate(t) {
-    this.learningRate = t, this.c != null && this.c.dispose(), this.c = Mn(q(-t));
-  }
-  dispose() {
-    this.c.dispose();
-  }
-  async getWeights() {
-    return [await this.saveIterations()];
-  }
-  async setWeights(t) {
-    if (t = await this.extractIterations(t), t.length !== 0)
-      throw new Error("SGD optimizer does not have settable weights.");
-  }
-  getConfig() {
-    return { learningRate: this.learningRate };
-  }
-  /** @nocollapse */
-  static fromConfig(t, n) {
-    return new t(n.learningRate);
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-class As extends De {
-  /** @nocollapse */
-  // Name matters for Python compatibility.
-  static get className() {
-    return "Momentum";
-  }
-  constructor(t, n, s = !1) {
-    super(t), this.learningRate = t, this.momentum = n, this.useNesterov = s, this.accumulations = [], this.m = q(this.momentum);
-  }
-  applyGradients(t) {
-    (Array.isArray(t) ? t.map((s) => s.name) : Object.keys(t)).forEach((s, r) => {
-      const a = f.registeredVariables[s];
-      this.accumulations[r] == null && (this.accumulations[r] = {
-        originalName: `${s}/momentum`,
-        variable: A(() => C(a).variable(!1))
-      });
-      const o = this.accumulations[r].variable, i = Array.isArray(t) ? t[r].tensor : t[s];
-      i != null && A(() => {
-        let c;
-        const l = k(b(this.m, o), i);
-        this.useNesterov ? c = k(b(this.c, k(i, b(l, this.m))), a) : c = k(b(this.c, l), a), o.assign(l), a.assign(c);
-      });
-    }), this.incrementIterations();
-  }
-  dispose() {
-    this.m.dispose(), this.accumulations != null && D(this.accumulations.map((t) => t.variable));
-  }
-  /**
-   * Sets the momentum of the optimizer.
-   *
-   * @param momentum
-   */
-  setMomentum(t) {
-    this.momentum = t;
-  }
-  async getWeights() {
-    return [await this.saveIterations()].concat(this.accumulations.map((t) => ({ name: t.originalName, tensor: t.variable })));
-  }
-  async setWeights(t) {
-    t = await this.extractIterations(t);
-    const n = !1;
-    this.accumulations = t.map((s) => ({ originalName: s.name, variable: s.tensor.variable(n) }));
-  }
-  getConfig() {
-    return {
-      learningRate: this.learningRate,
-      momentum: this.momentum,
-      useNesterov: this.useNesterov
-    };
-  }
-  /** @nocollapse */
-  static fromConfig(t, n) {
-    return new t(n.learningRate, n.momentum, n.useNesterov);
-  }
-}
-/**
- * @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.
- * =============================================================================
- */
-class Es extends H {
-  /** @nocollapse */
-  static get className() {
-    return "RMSProp";
-  }
-  constructor(t, n = 0.9, s = 0, r = null, a = !1) {
-    if (super(), this.learningRate = t, this.decay = n, this.momentum = s, this.epsilon = r, this.accumulatedMeanSquares = [], this.accumulatedMoments = [], this.accumulatedMeanGrads = [], this.centered = a, r == null && (this.epsilon = f.backend.epsilon()), t == null)
-      throw new Error("learningRate for RMSPropOptimizer must be defined.");
-  }
-  applyGradients(t) {
-    (Array.isArray(t) ? t.map((s) => s.name) : Object.keys(t)).forEach((s, r) => {
-      const a = f.registeredVariables[s], o = !1;
-      this.accumulatedMeanSquares[r] == null && (this.accumulatedMeanSquares[r] = {
-        originalName: `${s}/rms`,
-        variable: A(() => C(a).variable(o))
-      }), this.accumulatedMoments[r] == null && (this.accumulatedMoments[r] = {
-        originalName: `${s}/momentum`,
-        variable: A(() => C(a).variable(o))
-      }), this.accumulatedMeanGrads[r] == null && this.centered && (this.accumulatedMeanGrads[r] = {
-        originalName: `${s}/mg`,
-        variable: A(() => C(a).variable(o))
-      });
-      const i = Array.isArray(t) ? t[r].tensor : t[s];
-      if (i == null)
-        return;
-      const c = this.accumulatedMeanSquares[r].variable, l = this.accumulatedMoments[r].variable;
-      A(() => {
-        const u = k(b(c, this.decay), b(V(i), 1 - this.decay));
-        if (this.centered) {
-          const h = this.accumulatedMeanGrads[r].variable, d = k(b(h, this.decay), b(i, 1 - this.decay)), m = $(b(i, this.learningRate), nt(tt(u, k(V(d), this.epsilon)))), y = k(b(l, this.momentum), m);
-          c.assign(u), h.assign(d), l.assign(y);
-          const g = tt(a, y);
-          a.assign(g);
-        } else {
-          const h = k(b(c, this.decay), b(V(i), 1 - this.decay)), d = k(b(l, this.momentum), $(b(i, this.learningRate), nt(k(h, this.epsilon))));
-          c.assign(h), l.assign(d);
-          const m = tt(a, d);
-          a.assign(m);
-        }
-      });
-    }), this.incrementIterations();
-  }
-  dispose() {
-    this.accumulatedMeanSquares != null && D(this.accumulatedMeanSquares.map((t) => t.variable)), this.accumulatedMeanGrads != null && this.centered && D(this.accumulatedMeanGrads.map((t) => t.variable)), this.accumulatedMoments != null && D(this.accumulatedMoments.map((t) => t.variable));
-  }
-  async getWeights() {
-    const t = [...this.accumulatedMeanSquares, ...this.accumulatedMoments];
-    return this.centered && t.push(...this.accumulatedMeanGrads), [await this.saveIterations()].concat(t.map((n) => ({ name: n.originalName, tensor: n.variable })));
-  }
-  async setWeights(t) {
-    t = await this.extractIterations(t);
-    const n = this.centered ? t.length / 3 : t.length / 2, s = !1;
-    this.accumulatedMeanSquares = t.slice(0, n).map((r) => ({
-      originalName: r.name,
-      variable: r.tensor.variable(s)
-    })), this.accumulatedMoments = t.slice(n, n * 2).map((r) => ({
-      originalName: r.name,
-      variable: r.tensor.variable(s)
-    })), this.centered && (this.accumulatedMeanGrads = t.slice(n * 2, n * 3).map((r) => ({
-      originalName: r.name,
-      variable: r.tensor.variable(s)
-    })));
-  }
-  getConfig() {
-    return {
-      learningRate: this.learningRate,
-      decay: this.decay,
-      momentum: this.momentum,
-      epsilon: this.epsilon,
-      centered: this.centered
-    };
-  }
-  /** @nocollapse */
-  static fromConfig(t, n) {
-    return new t(n.learningRate, n.decay, n.momentum, n.epsilon, n.centered);
-  }
-}
-/**
- * @license
- * Copyright 2022 Google LLC.
- * 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.
- * =============================================================================
- */
-const vs = [
-  Ss,
-  ks,
-  Is,
-  Ts,
-  As,
-  Es,
-  De
-];
-function Bs() {
-  for (const e of vs)
-    ws(e);
-}
-/**
- * @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.
- * =============================================================================
- */
-Bs();
-export {
-  ho as $,
-  C as A,
-  Ps as B,
-  Tn as C,
-  vn as D,
-  F as E,
-  T as F,
-  f as G,
-  ta as H,
-  An as I,
-  Jn as J,
-  Sr as K,
-  bi as L,
-  Ct as M,
-  Ue as N,
-  H as O,
-  Ga as P,
-  K as Q,
-  Ka as R,
-  ko as S,
-  M as T,
-  hr as U,
-  Ya as V,
-  So as W,
-  Po as X,
-  Ra as Y,
-  Xa as Z,
-  Ja as _,
-  b as a,
-  aa as a$,
-  Rt as a0,
-  bo as a1,
-  ni as a2,
-  $t as a3,
-  te as a4,
-  Gs as a5,
-  xa as a6,
-  oo as a7,
-  ea as a8,
-  Sn as a9,
-  Zo as aA,
-  Qo as aB,
-  Je as aC,
-  E as aD,
-  st as aE,
-  di as aF,
-  ke as aG,
-  fi as aH,
-  gi as aI,
-  Us as aJ,
-  _s as aK,
-  $s as aL,
-  ht as aM,
-  Ws as aN,
-  jt as aO,
-  he as aP,
-  Xo as aQ,
-  O as aR,
-  fr as aS,
-  lo as aT,
-  zr as aU,
-  Ma as aV,
-  as as aW,
-  V as aX,
-  Hr as aY,
-  na as aZ,
-  sa as a_,
-  wr as aa,
-  se as ab,
-  Br as ac,
-  fo as ad,
-  pi as ae,
-  xt as af,
-  S as ag,
-  ka as ah,
-  Kr as ai,
-  fa as aj,
-  R as ak,
-  ye as al,
-  ti as am,
-  gt as an,
-  Zr as ao,
-  Kn as ap,
-  $e as aq,
-  po as ar,
-  Vr as as,
-  la as at,
-  za as au,
-  no as av,
-  Vo as aw,
-  yi as ax,
-  Ns as ay,
-  ce as az,
-  q as b,
-  jr as b$,
-  ua as b0,
-  ha as b1,
-  ga as b2,
-  Na as b3,
-  ma as b4,
-  Ba as b5,
-  Da as b6,
-  $a as b7,
-  Ha as b8,
-  ro as b9,
-  Qe as bA,
-  tn as bB,
-  Va as bC,
-  ao as bD,
-  gn as bE,
-  nn as bF,
-  Do as bG,
-  sn as bH,
-  Ne as bI,
-  Zs as bJ,
-  Qs as bK,
-  tr as bL,
-  ir as bM,
-  lr as bN,
-  Co as bO,
-  dr as bP,
-  Qr as bQ,
-  br as bR,
-  kr as bS,
-  Tr as bT,
-  Ar as bU,
-  vr as bV,
-  Mr as bW,
-  Dr as bX,
-  Fr as bY,
-  xr as bZ,
-  $r as b_,
-  Bo as ba,
-  Uo as bb,
-  Oo as bc,
-  Rr as bd,
-  Yr as be,
-  jo as bf,
-  Ca as bg,
-  La as bh,
-  Pa as bi,
-  to as bj,
-  Za as bk,
-  _o as bl,
-  He as bm,
-  zs as bn,
-  Hs as bo,
-  Ks as bp,
-  ie as bq,
-  Vs as br,
-  _t as bs,
-  oe as bt,
-  gr as bu,
-  ee as bv,
-  yr as bw,
-  qr as bx,
-  Xr as by,
-  Ze as bz,
-  tt as c,
-  Ge as c$,
-  yo as c0,
-  pa as c1,
-  Ia as c2,
-  Aa as c3,
-  Oa as c4,
-  _a as c5,
-  Ua as c6,
-  wo as c7,
-  so as c8,
-  uo as c9,
-  Xe as cA,
-  Er as cB,
-  _r as cC,
-  Gr as cD,
-  Or as cE,
-  Wr as cF,
-  oa as cG,
-  ia as cH,
-  ca as cI,
-  ba as cJ,
-  Sa as cK,
-  wa as cL,
-  Ea as cM,
-  Ta as cN,
-  Fa as cO,
-  en as cP,
-  eo as cQ,
-  Qa as cR,
-  mo as cS,
-  Mo as cT,
-  $o as cU,
-  ms as cV,
-  rn as cW,
-  li as cX,
-  ws as cY,
-  bs as cZ,
-  G as c_,
-  go as ca,
-  zo as cb,
-  Ir as cc,
-  Ko as cd,
-  Cr as ce,
-  Pr as cf,
-  qo as cg,
-  De as ch,
-  As as ci,
-  Es as cj,
-  Is as ck,
-  Ss as cl,
-  Ts as cm,
-  ks as cn,
-  Xs as co,
-  Js as cp,
-  Ys as cq,
-  er as cr,
-  nr as cs,
-  sr as ct,
-  or as cu,
-  rr as cv,
-  ar as cw,
-  ur as cx,
-  cr as cy,
-  mr as cz,
-  D as d,
-  xs as d0,
-  mi as d1,
-  ui as d2,
-  hi as d3,
-  xe as d4,
-  Rs as d5,
-  js as d6,
-  ut as d7,
-  Cs as d8,
-  Nt as d9,
-  es as dA,
-  ct as dB,
-  qs as dC,
-  Wa as dD,
-  ja as dE,
-  qa as dF,
-  Io as dG,
-  To as dH,
-  xo as dI,
-  No as dJ,
-  ei as dK,
-  Os as dL,
-  It as dM,
-  In as dN,
-  ci as da,
-  Ye as db,
-  yt as dc,
-  Ls as dd,
-  pr as de,
-  Nr as df,
-  Lr as dg,
-  Ur as dh,
-  Jr as di,
-  Wo as dj,
-  Ho as dk,
-  ra as dl,
-  da as dm,
-  ya as dn,
-  va as dp,
-  co as dq,
-  Ao as dr,
-  Eo as ds,
-  vo as dt,
-  Fo as du,
-  Ro as dv,
-  io as dw,
-  Lo as dx,
-  Go as dy,
-  fe as dz,
-  si as e,
-  Jo as f,
-  ii as g,
-  os as h,
-  Yo as i,
-  z as j,
-  Mn as k,
-  is as l,
-  ri as m,
-  p as n,
-  $ as o,
-  Ce as p,
-  k as q,
-  oi as r,
-  ai as s,
-  A as t,
-  ge as u,
-  ds as v,
-  nt as w,
-  Qt as x,
-  wi as y,
-  Si as z
-};