@genai-fi/nanogpt 0.7.2 → 0.8.0

package/dist/{random_width-sZORGo5k.js → random_width-D8Pwy_na.js}

@@ -1,25 +1,25 @@
-import { B as T, C as I, E as O, bE as Oa, bF as Ma, bG as Ci, l as b, bH as Ii, X as L, bI as Di, bJ as $i, bK as Ti, bL as zi, h as Ei, bM as Li, bN as Fi, bO as Oi, bP as Mi, bQ as _a, bR as _i, bS as Ra, bT as Ri, bU as Ba, bV as Bi, I as Ge, k as kt, bq as Wa, bW as Wi, bX as Gi, bY as Pi, y as Pe, c as V, a as w, bZ as Ga, b_ as Ui, b$ as Vi, c0 as ji, p as ce, aU as pt, bv as Pa, c1 as Ki, c2 as Hi, c3 as qi, c4 as Ji, c5 as Zi, bx as Xi, c6 as Yi, c7 as Qi, F as Ua, al as Va, aq as ja, c8 as tr, c9 as Ka, o as z, ca as Ps, cb as Ha, cc as qa, j as pn, cd as Us, ce as Ja, cf as Za, cg as Xa, ch as Ya, ci as Qa, cj as tl, ck as el, bf as nl, cl as sl, q as he, b as et, n as U, cm as il, bm as rl, aw as ht, cn as ol, z as Q, co as al, cp as ll, cq as ul, cr as cl, cs as hl, ct as pl, cu as dl, cv as fl, J as ml, cw as gl, bj as bl, bp as yl, cx as wl, Q as kl, cy as xl, a8 as Nl, cz as vl, cA as Al, cB as Sl, as as Cl, cC as Il, ae as Dl, aV as $l, br as Tl, ak as zl, bs as El, G as Ll, aX as Fl, am as Ol, cD as Ml, cE as _l, cF as Rl, an as Bl, af as Wl, cG as Gl, cH as Pl, cI as Ul, U as Vl, bt as jl, cJ as Kl, cK as Hl, aS as ql, b1 as Jl, cL as Zl, cM as Xl, bu as Yl, a$ as Ql, P as tu, cN as eu, x as rs, ap as nu, bw as su, az as iu, V as ru, au as ou, at as au, K as lu, bc as uu, cO as cu, bd as hu, cP as pu, b3 as du, aQ as fu, ar as mu, cQ as gu, a9 as bu, W as yu, S as wu, L as ku, bz as xu, cR as Nu, bA as vu, av as Au, bC as Su, Y as Cu, cS as Iu, N as Du, b5 as $u, b4 as Tu, cT as Oe, cU as zu, i as Eu, aj as Vs, cV as Lu, t as x, aT as $e, cW as S, cX as He, cY as qe, ab as Vt, d as J, ac as Fu, cZ as js, w as Jt, M as Ou, T as Te, O as Mu, c_ as _u, m as Ks, c$ as Ru, d0 as Hs, d1 as Bu } from "./index-BoWRt-10.js";
-import { k as ke, c as Nt, o as ze, s as er, b as nr, d as Wu, m as sr, t as In, l as ir, v as os, a as Gu, S as Pu, p as Uu, w as as, x as rr, y as Je, z as Vu, A as ju } from "./selu_util-BJEXVvjX.js";
-import { n as wt, w as ne, a as Ze, g as Xe, b as ls, t as K, c as Ce, d as Xt, e as Ku, u as dn, f as ye, h as fn, i as Hu, l as qu, s as us, m as or, j as qt, k as Ju } from "./ops-BFGCx8Ri.js";
-import { r as N } from "./reshape-CdBq1WJ6.js";
-import { s as W } from "./sum-66ew2byf.js";
-import { m as ct } from "./mat_mul-8m8pfdcx.js";
-import { s as Qt } from "./split-B-dikLRw.js";
-import { s as Zu, c as ar } from "./sin-BeA3tsEd.js";
-import { e as Hn, g as lr, h as cs, c as Xu } from "./axis_util-BzbKo31C.js";
-import { a as se, e as ie, l as Yu } from "./log_sum_exp-DbjkV734.js";
-import { s as Dn } from "./stack-B17UN2nn.js";
-import { o as xe } from "./ones-Dj0SDhHf.js";
-import { s as Dt } from "./slice-BiOsknYS.js";
-import { M as Qu, f as ur, r as tc, d as ec, a as $n } from "./dropout-DYs5QFGQ.js";
-import { z as vt } from "./zeros--BdLQ3oG.js";
-import { c as pe } from "./concat-CsxrgovM.js";
-import { g as cr } from "./gather-CMMy2KEG.js";
-import { s as hr } from "./softmax-Bv_6lyMX.js";
-import { m as Ee } from "./max-Ddnnb5xe.js";
-import { t as nc } from "./tensor-JwS7ZYY6.js";
-import { r as sc } from "./range-CRuAh-gd.js";
-import { v as ic } from "./variable-BuddVFLa.js";
+import { B as T, C as I, E as O, bG as Oa, bH as Ma, bI as Ci, n as b, bJ as Ii, $ as L, bK as Di, bL as $i, bM as Ti, bN as zi, h as Ei, bO as Li, bP as Fi, bQ as Oi, bR as Mi, bS as _a, bT as _i, bU as Ra, bV as Ri, bW as Ba, bX as Bi, N as Ge, l as kt, bs as Wa, bY as Wi, bZ as Gi, b_ as Pi, I as Pe, c as V, a as w, b$ as Ga, c0 as Ui, c1 as Vi, c2 as ji, p as ce, aW as pt, bx as Pa, c3 as Ki, c4 as Hi, c5 as qi, c6 as Zi, c7 as Ji, bz as Xi, c8 as Yi, c9 as Qi, L as Ua, an as Va, as as ja, ca as tr, cb as Ka, q as z, cc as Ps, cd as Ha, ce as qa, j as pn, cf as Us, cg as Za, ch as Ja, ci as Xa, cj as Ya, ck as Qa, cl as tl, cm as el, bh as nl, cn as sl, w as he, b as et, o as U, co as il, bo as rl, ay as ht, cp as ol, K as Q, cq as al, cr as ll, cs as ul, ct as cl, cu as hl, cv as pl, cw as dl, cx as fl, Q as ml, cy as gl, bl, br as yl, cz as wl, H as kl, cA as xl, ab as Nl, cB as vl, cC as Al, cD as Sl, au as Cl, cE as Il, ag as Dl, aX as $l, bt as Tl, am as zl, bu as El, G as Ll, aZ as Fl, ao as Ol, cF as Ml, cG as _l, cH as Rl, ap as Bl, ah as Wl, cI as Gl, cJ as Pl, cK as Ul, X as Vl, bv as jl, cL as Kl, cM as Hl, aU as ql, b3 as Zl, cN as Jl, Y as Xl, bw as Yl, b1 as Ql, P as tu, cO as eu, x as rs, ar as nu, by as su, aB as iu, Z as ru, aw as ou, av as au, U as lu, be as uu, cP as cu, bf as hu, cQ as pu, b5 as du, aS as fu, at as mu, cR as gu, ac as bu, _ as yu, S as wu, V as ku, bB as xu, cS as Nu, bC as vu, ax as Au, bE as Su, a0 as Cu, cT as Iu, W as Du, b7 as $u, b6 as Tu, cU as Oe, cV as zu, i as Eu, al as Vs, cW as Lu, t as x, aV as $e, cX as S, cY as He, cZ as qe, ae as Vt, d as Z, af as Fu, c_ as js, k as Zt, F as Ou, T as Te, O as Mu, c$ as _u, m as Ks, d0 as Ru, d1 as Hs, d2 as Bu } from "./index-CUQrfsw_.js";
+import { k as ke, c as Nt, o as ze, s as er, b as nr, d as Wu, m as sr, t as In, l as ir, v as os, a as Gu, S as Pu, p as Uu, w as as, x as rr, y as Ze, z as Vu, A as ju } from "./selu_util-8vv5JxQV.js";
+import { n as wt, w as ne, a as Je, g as Xe, b as ls, t as K, c as Ce, d as Xt, e as Ku, u as dn, f as ye, h as fn, i as Hu, l as qu, s as us, m as or, j as qt, k as Zu } from "./ops-C_1K_-35.js";
+import { r as N } from "./reshape-DEfQGSin.js";
+import { s as W } from "./sum-l_0SqM4h.js";
+import { m as ct } from "./mat_mul-DMkduNJu.js";
+import { s as Qt } from "./split-By_n4TKP.js";
+import { s as Ju, c as ar } from "./sin-D2CKKmyR.js";
+import { e as Hn, g as lr, h as cs, c as Xu } from "./axis_util-DubwyOhW.js";
+import { a as se, e as ie, l as Yu } from "./log_sum_exp-C142qZqY.js";
+import { s as Dn } from "./stack-DkdFLq37.js";
+import { o as xe } from "./ones-ZdgQGBCP.js";
+import { s as Dt } from "./slice-BnyE-M_7.js";
+import { M as Qu, f as ur, r as tc, d as ec, a as $n } from "./dropout-sx0sjVAT.js";
+import { z as vt } from "./zeros-VZ72lWXM.js";
+import { c as pe } from "./concat-BmDqqFsa.js";
+import { g as cr } from "./gather-C1siEkdp.js";
+import { s as hr } from "./softmax-DLoZWYBx.js";
+import { m as Ee } from "./max-B3JOcNGb.js";
+import { t as nc } from "./tensor-BAQdLqoU.js";
+import { r as sc } from "./range-LVHrSLdi.js";
+import { v as ic } from "./variable-C9hihzDB.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.
@@ -444,7 +444,7 @@ function Kc(n, t = 0, e = !1, s = !1) {
   const r = { x: I(n, "x", "cumprod") }, o = { axis: t, exclusive: e, reverse: s };
   return O.runKernel(Ra, r, o);
 }
-const Js = /* @__PURE__ */ T({ cumprod_: Kc });
+const Zs = /* @__PURE__ */ T({ cumprod_: Kc });
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.
@@ -482,13 +482,13 @@ const qc = /* @__PURE__ */ T({ cumsum_: Hc });
  * limitations under the License.
  * =============================================================================
  */
-function Jc(n, t, e, s = !1) {
+function Zc(n, t, e, s = !1) {
   const i = I(n, "x", "denseBincount"), r = I(t, "weights", "denseBincount");
   b(i.dtype === "int32", () => `Error in denseBincount: input dtype must be int32, but got ${i.dtype}`), b(i.rank <= 2, () => `Error in denseBincount: input must be at most rank 2, but got rank ${i.rank}.`), b(e >= 0, () => `size must be non-negative, but got ${e}.`), b(r.size === i.size || r.size === 0, () => `Error in denseBincount: weights must have the same shape as x or 0-length, but got x shape: ${i.shape}, weights shape: ${r.shape}.`);
   const o = { x: i, weights: r }, a = { size: e, binaryOutput: s };
   return O.runKernel(Ba, o, a);
 }
-const Zs = /* @__PURE__ */ T({ denseBincount_: Jc });
+const Js = /* @__PURE__ */ T({ denseBincount_: Zc });
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.
@@ -505,7 +505,7 @@ const Zs = /* @__PURE__ */ T({ denseBincount_: Jc });
  * limitations under the License.
  * =============================================================================
  */
-function Zc(n, t, e, s, i = "NHWC", r = [1, 1], o) {
+function Jc(n, t, e, s, i = "NHWC", r = [1, 1], o) {
   const a = I(n, "x", "depthwiseConv2d", "float32"), l = I(t, "filter", "depthwiseConv2d", "float32");
   let u = a, c = !1;
   a.rank === 3 && (c = !0, u = N(a, [1, a.shape[0], a.shape[1], a.shape[2]])), b(u.rank === 4, () => `Error in depthwiseConv2d: input must be rank 4, but got rank ${u.rank}.`), b(l.rank === 4, () => `Error in depthwiseConv2d: filter must be rank 4, but got rank ${l.rank}.`);
@@ -514,7 +514,7 @@ function Zc(n, t, e, s, i = "NHWC", r = [1, 1], o) {
   const p = { x: u, filter: l }, f = { strides: e, pad: s, dataFormat: i, dilations: r, dimRoundingMode: o }, y = O.runKernel(Bi, p, f);
   return c ? N(y, [y.shape[1], y.shape[2], y.shape[3]]) : y;
 }
-const dr = /* @__PURE__ */ T({ depthwiseConv2d_: Zc });
+const dr = /* @__PURE__ */ T({ depthwiseConv2d_: Jc });
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.
@@ -858,7 +858,7 @@ function yh(n, t, e) {
   const s = I(n, "x", "spaceToBatchND");
   b(s.rank >= 1 + t.length, () => `input rank ${s.rank} should be > than [blockShape] ${t.length}`), b(e.length === t.length, () => `paddings.shape[0] ${e.length} must be equal to [blockShape] ${t.length}`), b(s.shape.reduce((o, a, l) => l > 0 && l <= t.length ? o && (a + e[l - 1][0] + e[l - 1][1]) % t[l - 1] === 0 : o, !0), () => `input spatial dimensions ${s.shape.slice(1)} with paddings ${e.toString()} must be divisible by blockShapes ${t.toString()}`);
   const i = { x: s }, r = { blockShape: t, paddings: e };
-  return O.runKernel(Ji, i, r);
+  return O.runKernel(Zi, i, r);
 }
 const wh = /* @__PURE__ */ T({ spaceToBatchND_: yh });
 /**
@@ -879,7 +879,7 @@ const wh = /* @__PURE__ */ T({ spaceToBatchND_: yh });
  */
 function kh(n, t) {
   const s = { x: I(n, "x", "reverse") }, i = { dims: t };
-  return O.runKernel(Zi, s, i);
+  return O.runKernel(Ji, s, i);
 }
 const gn = /* @__PURE__ */ T({ reverse_: kh });
 /**
@@ -1361,7 +1361,7 @@ class Vh {
    * @doc {heading: 'Training', subheading: 'Optimizers', namespace: 'train'}
    */
   static sgd(t) {
-    return new Ja(t);
+    return new Za(t);
   }
   /**
    * Constructs a `tf.MomentumOptimizer` that uses momentum gradient
@@ -1379,7 +1379,7 @@ class Vh {
    * @doc {heading: 'Training', subheading: 'Optimizers', namespace: 'train'}
    */
   static momentum(t, e, s = !1) {
-    return new Za(t, e, s);
+    return new Ja(t, e, s);
   }
   /**
    * Constructs a `tf.RMSPropOptimizer` that uses RMSProp gradient
@@ -1605,7 +1605,7 @@ const qh = {
  * limitations under the License.
  * =============================================================================
  */
-const Jh = {
+const Zh = {
   kernelName: rl,
   inputsToSave: ["a", "b"],
   gradFunc: (n, t) => {
@@ -1637,7 +1637,7 @@ const Jh = {
  * limitations under the License.
  * =============================================================================
  */
-const Zh = {
+const Jh = {
   kernelName: ol,
   saveAllInputs: !0,
   gradFunc: (n, t) => {
@@ -2092,7 +2092,7 @@ const mp = {
   gradFunc: (n, t, e) => {
     const [s] = t, { clipValueMin: i, clipValueMax: r } = e;
     return {
-      x: () => ne(Ze(Xe(s, i), ls(s, r)), n, Q(n))
+      x: () => ne(Je(Xe(s, i), ls(s, r)), n, Q(n))
     };
   }
 };
@@ -2270,7 +2270,7 @@ const vp = {
   inputsToSave: ["x"],
   gradFunc: (n, t) => {
     const [e] = t;
-    return { x: () => w(wt(Zu(L(e, "float32"))), n) };
+    return { x: () => w(wt(Ju(L(e, "float32"))), n) };
   }
 };
 /**
@@ -2973,7 +2973,7 @@ const ti = {
  * limitations under the License.
  * =============================================================================
  */
-const Jp = {
+const Zp = {
   kernelName: jl,
   inputsToSave: ["a", "b"],
   gradFunc: (n, t) => {
@@ -2997,7 +2997,7 @@ const Jp = {
  * limitations under the License.
  * =============================================================================
  */
-function Zp(n, t, e, s, i, r, o) {
+function Jp(n, t, e, s, i, r, o) {
   const a = I(n, "dy", "maxPool3dGrad"), l = I(t, "input", "maxPool3dGrad"), u = I(e, "output", "maxPool3dGrad");
   let c = a, h = l, p = u, f = !1;
   l.rank === 4 && (f = !0, c = N(a, [1, a.shape[0], a.shape[1], a.shape[2], a.shape[3]]), h = N(l, [
@@ -3016,7 +3016,7 @@ function Zp(n, t, e, s, i, r, o) {
   const y = { dy: c, input: h, output: p }, g = { filterSize: s, strides: i, pad: r, dimRoundingMode: o }, m = O.runKernel(Kl, y, g);
   return f ? N(m, [m.shape[1], m.shape[2], m.shape[3], m.shape[4]]) : m;
 }
-const Xp = /* @__PURE__ */ T({ maxPool3dGrad_: Zp });
+const Xp = /* @__PURE__ */ T({ maxPool3dGrad_: Jp });
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.
@@ -3169,7 +3169,7 @@ const sd = {
  * =============================================================================
  */
 const id = {
-  kernelName: Jl,
+  kernelName: Zl,
   inputsToSave: ["a", "b"],
   gradFunc: (n, t) => {
     const [e, s] = t;
@@ -3193,7 +3193,7 @@ const id = {
  * =============================================================================
  */
 const rd = {
-  kernelName: Zl,
+  kernelName: Jl,
   inputsToSave: ["x"],
   gradFunc: (n, t, e) => {
     const s = t[0], { paddings: i } = e, r = i.map((o) => o[0]);
@@ -3457,7 +3457,7 @@ const dd = {
 function fd(n, t, e) {
   const s = n.shape.slice();
   s[e] = 1;
-  const i = N(t, s), r = Js(n, e, !0, !1), o = Js(n, e, !0, !0), a = w(r, o);
+  const i = N(t, s), r = Zs(n, e, !0, !1), o = Zs(n, e, !0, !0), a = w(r, o);
   return w(i, a);
 }
 function md(n, t, e) {
@@ -3683,7 +3683,7 @@ const vd = {
  * =============================================================================
  */
 const Ad = {
-  kernelName: Zi,
+  kernelName: Ji,
   gradFunc: (n, t, e) => {
     const { dims: s } = e, i = ce(s, n.shape);
     return { x: () => gn(n, i) };
@@ -3977,7 +3977,7 @@ const Od = {
  * =============================================================================
  */
 const ni = {
-  kernelName: Ji,
+  kernelName: Zi,
   gradFunc: (n, t, e) => {
     const { blockShape: s, paddings: i } = e;
     return { x: () => mc(n, s, i) };
@@ -4337,7 +4337,7 @@ function qd(n, t) {
   const r = s.rank - i.rank;
   for (let a = 0; a < r; ++a)
     i = ye(i, a + 1);
-  i = Ze(i, xe(s.shape, "bool"));
+  i = Je(i, xe(s.shape, "bool"));
   const o = Q(s);
   return ne(i, s, o);
 }
@@ -4357,7 +4357,7 @@ function qd(n, t) {
  * limitations under the License.
  * =============================================================================
  */
-const Jd = {
+const Zd = {
   kernelName: zu,
   gradFunc: (n) => ({ x: () => Q(n) })
 };
@@ -4377,12 +4377,12 @@ const Jd = {
  * limitations under the License.
  * =============================================================================
  */
-const Zd = [
+const Jd = [
   br,
   Hh,
   qh,
-  Jh,
   Zh,
+  Jh,
   Xh,
   Yh,
   Qh,
@@ -4430,7 +4430,7 @@ const Zd = [
   qp,
   ti,
   ti,
-  Jp,
+  Zp,
   Yp,
   ed,
   nd,
@@ -4482,9 +4482,9 @@ const Zd = [
   jd,
   Kd,
   Hd,
-  Jd
+  Zd
 ];
-for (const n of Zd)
+for (const n of Jd)
   Eu(n);
 /**
  * @license
@@ -4624,15 +4624,15 @@ function ks(n) {
   const t = {};
   return t.className = n.getClassName(), t.config = n.getConfig(), t;
 }
-function Jn(n) {
+function Zn(n) {
   if (!(n == null || typeof n != "object"))
     if (Array.isArray(n))
-      n.forEach((t) => Jn(t));
+      n.forEach((t) => Zn(t));
     else {
       const t = Object.keys(n);
       for (const e of t) {
         const s = n[e];
-        s != null && typeof s == "object" && (!Array.isArray(s) && s.type === "ndarray" && typeof s.value == "number" ? n[e] = s.value : Jn(s));
+        s != null && typeof s == "object" && (!Array.isArray(s) && s.type === "ndarray" && typeof s.value == "number" ? n[e] = s.value : Zn(s));
       }
     }
 }
@@ -4671,7 +4671,7 @@ function Ye(n, t = {}, e = {}, s = "object", i = !1) {
       const h = Object.assign({}, Ct);
       for (const f of Object.keys(e))
         Ct[f] = e[f];
-      Jn(r.config);
+      Zn(r.config);
       const p = l(a, r.config, e, i);
       return Ct = Object.assign({}, h), p;
     } else {
@@ -4894,7 +4894,7 @@ function hf(n, t) {
     if (n.shape.length !== 2)
       throw new d(`repeat() expects a rank-2 tensor, but received a rank-${n.shape.length} tensor.`);
     const e = Qe(n, 1);
-    return Zn(e, [1, t, 1]);
+    return Jn(e, [1, t, 1]);
   });
 }
 function pf(n) {
@@ -5017,7 +5017,7 @@ function oi(n, t) {
       throw new d(`concatAlongFirstAxis() received an unsupported tensor rank: ${n.rank}`);
   }
 }
-function Zn(n, t) {
+function Jn(n, t) {
   if (Array.isArray(t) || (t = [t]), n.rank !== t.length)
     throw new d(`The length of input n (${t.length}) does not match the number of dimensions in input x (${n.rank})`);
   return Ce(n, t);
@@ -5443,7 +5443,7 @@ function li(n, t = {}) {
 function Y(n) {
   return ks(n);
 }
-function Z(n) {
+function J(n) {
   if (typeof n == "string") {
     const t = n in ai ? ai[n] : n;
     if (t === "GlorotNormal")
@@ -6119,7 +6119,7 @@ class G extends He {
   addWeight(t, e, s, i, r, o, a, l) {
     if (this._addedWeightNames.indexOf(t) !== -1)
       throw new d(`Duplicate weight name ${t} for layer ${this.name}`);
-    this._addedWeightNames.push(t), s == null && (s = "float32"), this.fastWeightInitDuringBuild && (i = l != null ? l() : Z("zeros"));
+    this._addedWeightNames.push(t), s == null && (s = "float32"), this.fastWeightInitDuringBuild && (i = l != null ? l() : J("zeros"));
     const u = i.apply(e, s), c = new Nf(u, s, t, o, a);
     return u.dispose(), r != null && this.addLoss(() => r.apply(c.read())), o == null && (o = !0), o ? this._trainableWeights.push(c) : this._nonTrainableWeights.push(c), c;
   }
@@ -6533,7 +6533,7 @@ class Yt {
   }
   /** Dispose all mask Tensors held by this object. */
   disposeMasks() {
-    this.id2Mask != null && J(this.id2Mask);
+    this.id2Mask != null && Z(this.id2Mask);
   }
 }
 const kn = new wr(), xn = new wr();
@@ -6572,7 +6572,7 @@ function Be(n, t, e, s) {
       const M = a.indexOf(E[F].name);
       M !== -1 && (l[M] = D[F]);
     }
-    i || J(C);
+    i || Z(C);
   }
   return f.disposeMasks(), r ? l : l[0];
 }
@@ -6707,7 +6707,7 @@ Rr.className = "UnitNorm";
 S(Rr);
 class Br extends sn {
   apply(t) {
-    return Je(t);
+    return Ze(t);
   }
 }
 Br.className = "NonNeg";
@@ -6777,7 +6777,7 @@ async function fe(n) {
     const i = await Promise.all(t);
     for (let r = 0; r < i.length; ++r)
       n[e[r]] = i[r][0];
-    J(s);
+    Z(s);
   }
 }
 function Gr(n) {
@@ -6943,7 +6943,7 @@ class Bf extends Ue {
       for (const s of this.params.metrics)
         this.totals[s] != null && (typeof this.totals[s] == "number" ? e[s] = this.totals[s] / this.seen : x(() => {
           const i = w(U(1, this.seen), this.totals[s]);
-          e[s] = i, this.totals[s].dispose(), Jt(e[s]);
+          e[s] = i, this.totals[s].dispose(), Zt(e[s]);
         }));
   }
 }
@@ -7087,7 +7087,7 @@ function Ur(n, t, e, s, i, r, o, a, l) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function Zt(n, t = {}, e = !1) {
+function Jt(n, t = {}, e = !1) {
   return Ye(n, qe.getMap().classNameMap, t, "layer", e);
 }
 /**
@@ -7171,7 +7171,7 @@ function Hf(n, t) {
   if (!Vt(n.shape, t.shape))
     throw new d(`logits and labels must have the same shape, but got shapes ${JSON.stringify(n.shape)} and ${JSON.stringify(t.shape)}`);
   return x(() => {
-    const e = Je(t), s = wt($e(t));
+    const e = Ze(t), s = wt($e(t));
     return z(V(e, w(t, n)), eh(ie(s)));
   });
 }
@@ -7187,7 +7187,7 @@ function qf(n, t) {
     return W(w(n, se(U(e, s))), -1);
   });
 }
-function Jf(n, t) {
+function Zf(n, t) {
   return x(() => {
     const e = se(z(st(), t));
     return nt(V(t, w(n, e)), -1);
@@ -7212,7 +7212,7 @@ const An = {
   sparseCategoricalCrossentropy: vn,
   binaryCrossentropy: Fn,
   kullbackLeiblerDivergence: qf,
-  poisson: Jf,
+  poisson: Zf,
   cosineProximity: Vr
 };
 function Un(n) {
@@ -7242,15 +7242,15 @@ function jr(n, t) {
 function Kr(n, t) {
   return x(() => Pt(re(mn(n, -1), mn(t, -1)), "float32"));
 }
-function Zf(n, t) {
-  return x(() => L(W(Ze(re(n, 1), re(t, 1))), "float32"));
+function Jf(n, t) {
+  return x(() => L(W(Je(re(n, 1), re(t, 1))), "float32"));
 }
 function Xf(n, t) {
-  return x(() => L(W(Ze(re(n, 0), re(t, 1))), "float32"));
+  return x(() => L(W(Je(re(n, 0), re(t, 1))), "float32"));
 }
 function Yf(n, t) {
   return x(() => {
-    const e = Zf(n, t), s = Xf(n, t), i = z(e, s);
+    const e = Jf(n, t), s = Xf(n, t), i = z(e, s);
     return L(ne(Xt(i, 0), U(e, i), 0), "float32");
   });
 }
@@ -7485,7 +7485,7 @@ function fm(n, t, e, s) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function Jr(n, t, e) {
+function Zr(n, t, e) {
   return (n === "inboundNodes" || n === "outputLayers" || n === "inputLayers") && t === 0 && typeof e == "string";
 }
 function es(n, t) {
@@ -7499,7 +7499,7 @@ function es(n, t) {
     const e = [], s = n.length;
     for (let i = 0; i < s; ++i) {
       const r = n[i];
-      Jr(t, i, r) ? e.push(r) : e.push(es(r, t));
+      Zr(t, i, r) ? e.push(r) : e.push(es(r, t));
     }
     return e;
   } else {
@@ -7527,7 +7527,7 @@ function ns(n, t) {
     const e = [], s = n.length;
     for (let i = 0; i < s; ++i) {
       const r = n[i];
-      Jr(t, i, r) ? e.push(r) : e.push(ns(r, t));
+      Zr(t, i, r) ? e.push(r) : e.push(ns(r, t));
     }
     return e;
   } else {
@@ -7540,7 +7540,7 @@ function ns(n, t) {
   }
 }
 /** @license See the LICENSE file. */
-const Zr = "4.22.0";
+const Jr = "4.22.0";
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -7804,7 +7804,7 @@ class Lt extends G {
    */
   updatedConfig() {
     const t = this.getConfig(), e = {};
-    return e.className = this.getClassName(), e.config = t, e.kerasVersion = `tfjs-layers ${Zr}`, e.backend = "TensorFlow.js", e;
+    return e.className = this.getClassName(), e.config = t, e.kerasVersion = `tfjs-layers ${Jr}`, e.backend = "TensorFlow.js", e;
   }
   /**
    * Returns a JSON string containing the network configuration.
@@ -8100,7 +8100,7 @@ class Lt extends G {
       k.length > 0 && m.apply(yt(k), C);
     }
     function u(m) {
-      const A = m.name, k = Zt(m, e.customObjects != null ? e.customObjects : {});
+      const A = m.name, k = Jt(m, e.customObjects != null ? e.customObjects : {});
       k.setFastWeightInitDuringBuild(i), r[A] = k, m.inboundNodes.forEach((v) => {
         if (!(v instanceof Array))
           throw new d(`Corrupted configuration, expected array for nodeData: ${v}`);
@@ -8208,7 +8208,7 @@ async function Yr(n, t, e, s) {
       } else
         throw new Error(`Unexpected rank of target (y) tensor (${n.rank}) during handling of class weights. The rank is expected to be 1 or 2.`);
     }), r = Array.from(await i.data());
-    J(i);
+    Z(i);
     const o = [];
     return r.forEach((a) => {
       if (e[a] == null)
@@ -8319,10 +8319,10 @@ async function km(n, t, e) {
               R.push(await Yr(D[P], null, M[P]));
           }
           const E = v.concat(D).concat(R), F = a(E);
-          J(E);
+          Z(E);
           for (let M = 0; M < l.length; ++M) {
             const P = l[M], ut = F[M];
-            $[P] = ut, Jt(ut);
+            $[P] = ut, Zt(ut);
           }
           await p.onBatchEnd(k, $), Gr($), k++, A++;
         }
@@ -8373,15 +8373,15 @@ async function vm(n, t, e) {
     if (r = x(() => {
       if (u.value) {
         const { xs: c, ys: h } = Qr(n, u.value), p = c.concat(h), f = x(() => i(p));
-        if (J(p), l === 0)
+        if (Z(p), l === 0)
           for (let g = 0; g < f.length; ++g)
             r.push(et(0));
         const y = p[0].shape[0];
         for (let g = 0; g < f.length; ++g) {
           const m = f[g], A = r[g];
-          r[g] = x(() => z(r[g], w(y, m))), l > 0 && J(A);
+          r[g] = x(() => z(r[g], w(y, m))), l > 0 && Z(A);
         }
-        J(f), a += y, ++l;
+        Z(f), a += y, ++l;
       }
       return r;
     }), u.done) {
@@ -8391,7 +8391,7 @@ async function vm(n, t, e) {
   }
   for (let u = 0; u < r.length; ++u) {
     const c = r[u];
-    r[u] = U(r[u], a), J(c);
+    r[u] = U(r[u], a), Z(c);
   }
   return yt(r);
 }
@@ -9089,7 +9089,7 @@ class Ie extends Lt {
             const A = this.metricsTensors[g][0], k = this.metricsTensors[g][1];
             m = nt(A(i[k], f[k]));
           }
-          Jt(m), o.push(m);
+          Zt(m), o.push(m);
         }
         return y = nt(y), this.calculateLosses().forEach((g) => {
           y = z(y, g);
@@ -9194,7 +9194,7 @@ class Ie extends Lt {
       const R = Pr(s.callbacks, s.yieldEvery);
       return await this.fitLoop(C, k, v, f, s.epochs, s.verbose, R, D, A, s.shuffle, $, s.initialEpoch, null, null);
     } finally {
-      this.isTraining = !1, Et(i, t), Et(r, e), Et(o, t), Et(a, e), Et(c, l), Et(h, u), p != null && J(p);
+      this.isTraining = !1, Et(i, t), Et(r, e), Et(o, t), Et(a, e), Et(c, l), Et(h, u), p != null && Z(p);
     }
   }
   /**
@@ -9252,13 +9252,13 @@ class Ie extends Lt {
             const ft = ss(e, ut), mt = t(ft);
             for (let at = 0; at < s.length; ++at) {
               const bt = s[at], gt = mt[at];
-              F[bt] = gt, Jt(gt);
+              F[bt] = gt, Zt(gt);
             }
             if (E === R.length - 1 && g) {
               const at = this.testLoop(l, u, i);
               for (let bt = 0; bt < s.length; ++bt) {
                 const gt = s[bt], St = at[bt];
-                Jt(St), D["val_" + gt] = St;
+                Zt(St), D["val_" + gt] = St;
               }
             }
           }), await k.onBatchEnd(E, F), Gr(F), this.stopTraining_)
@@ -9326,7 +9326,7 @@ class Ie extends Lt {
       const c = await u.data();
       l.push(c[0]);
     }
-    return J(a), Et(s[0], t), Et(s[1], e), yt(l);
+    return Z(a), Et(s[0], t), Et(s[1], e), yt(l);
   }
   /**
    * Extract weight values of the model.
@@ -9443,7 +9443,7 @@ class Ie extends Lt {
       throw new Error("Loading loss_weights is not supported yet.");
     if (t.sample_weight_mode != null)
       throw new Error("Loading sample_weight_mode is not supported yet.");
-    const e = es(t.optimizer_config), s = Zt(e);
+    const e = es(t.optimizer_config), s = Jt(e);
     let i;
     if (typeof t.loss == "string")
       i = me(t.loss);
@@ -9559,7 +9559,7 @@ class Ie extends Lt {
     const s = await Hs(this.getNamedWeights(e)), a = {
       modelTopology: this.toJSON(null, !1),
       format: Dm,
-      generatedBy: `TensorFlow.js tfjs-layers v${Zr}`,
+      generatedBy: `TensorFlow.js tfjs-layers v${Jr}`,
       convertedBy: null
     };
     if ((e == null ? !1 : e.includeOptimizer) && this.optimizer != null) {
@@ -10049,7 +10049,7 @@ class je extends Ie {
     if (!(a instanceof je))
       throw new B(`Sequential.fromConfig called on non-Sequential input: ${a}`);
     for (const l of r) {
-      const c = Zt(l, void 0, i);
+      const c = Jt(l, void 0, i);
       i && c.setFastWeightInitDuringBuild(!0), a.add(c);
     }
     return a;
@@ -10142,14 +10142,14 @@ so.className = "selu";
 S(so);
 class io extends dt {
   apply(t) {
-    return Je(t);
+    return Ze(t);
   }
 }
 io.className = "relu";
 S(io);
 class ro extends dt {
   apply(t) {
-    return x(() => or(6, Je(t)));
+    return x(() => or(6, Ze(t)));
   }
 }
 ro.className = "relu6";
@@ -10378,7 +10378,7 @@ class ko extends G {
   }
   call(t, e) {
     t = _(t);
-    let s = Je(t);
+    let s = Ze(t);
     return this.maxValue != null && (s = Tt(s, 0, this.maxValue)), s;
   }
   computeOutputShape(t) {
@@ -10411,7 +10411,7 @@ xo.className = "LeakyReLU";
 S(xo);
 class No extends G {
   constructor(t) {
-    if (super(t ?? {}), this.DEFAULT_ALPHA_INITIALIZER = "zeros", t == null && (t = {}), this.supportsMasking = !0, this.alphaInitializer = Z(t.alphaInitializer || this.DEFAULT_ALPHA_INITIALIZER), this.alphaRegularizer = X(t.alphaRegularizer), this.alphaConstraint = ot(t.alphaConstraint), t.sharedAxes == null)
+    if (super(t ?? {}), this.DEFAULT_ALPHA_INITIALIZER = "zeros", t == null && (t = {}), this.supportsMasking = !0, this.alphaInitializer = J(t.alphaInitializer || this.DEFAULT_ALPHA_INITIALIZER), this.alphaRegularizer = X(t.alphaRegularizer), this.alphaConstraint = ot(t.alphaConstraint), t.sharedAxes == null)
       this.sharedAxes = null;
     else if (Array.isArray(t.sharedAxes))
       this.sharedAxes = t.sharedAxes;
@@ -10619,7 +10619,7 @@ class On extends G {
   constructor(t, e) {
     if (super(e), this.bias = null, this.DEFAULT_KERNEL_INITIALIZER = "glorotNormal", this.DEFAULT_BIAS_INITIALIZER = "zeros", On.verifyArgs(e), this.rank = t, lt(this.rank, "rank"), this.rank !== 1 && this.rank !== 2 && this.rank !== 3)
       throw new B(`Convolution layer for rank other than 1, 2, or 3 (${this.rank}) is not implemented yet.`);
-    if (this.kernelSize = De(e.kernelSize, t, "kernelSize"), this.strides = De(e.strides == null ? 1 : e.strides, t, "strides"), this.padding = e.padding == null ? "valid" : e.padding, At(this.padding), this.dataFormat = e.dataFormat == null ? "channelsLast" : e.dataFormat, tt(this.dataFormat), this.activation = le(e.activation), this.useBias = e.useBias == null ? !0 : e.useBias, this.biasInitializer = Z(e.biasInitializer || this.DEFAULT_BIAS_INITIALIZER), this.biasConstraint = ot(e.biasConstraint), this.biasRegularizer = X(e.biasRegularizer), this.activityRegularizer = X(e.activityRegularizer), this.dilationRate = De(e.dilationRate == null ? 1 : e.dilationRate, t, "dilationRate"), this.rank === 1 && Array.isArray(this.dilationRate) && this.dilationRate.length !== 1)
+    if (this.kernelSize = De(e.kernelSize, t, "kernelSize"), this.strides = De(e.strides == null ? 1 : e.strides, t, "strides"), this.padding = e.padding == null ? "valid" : e.padding, At(this.padding), this.dataFormat = e.dataFormat == null ? "channelsLast" : e.dataFormat, tt(this.dataFormat), this.activation = le(e.activation), this.useBias = e.useBias == null ? !0 : e.useBias, this.biasInitializer = J(e.biasInitializer || this.DEFAULT_BIAS_INITIALIZER), this.biasConstraint = ot(e.biasConstraint), this.biasRegularizer = X(e.biasRegularizer), this.activityRegularizer = X(e.activityRegularizer), this.dilationRate = De(e.dilationRate == null ? 1 : e.dilationRate, t, "dilationRate"), this.rank === 1 && Array.isArray(this.dilationRate) && this.dilationRate.length !== 1)
       throw new d(`dilationRate must be a number or an array of a single number for 1D convolution, but received ${JSON.stringify(this.dilationRate)}`);
     if (this.rank === 2) {
       if (typeof this.dilationRate == "number")
@@ -10656,7 +10656,7 @@ class On extends G {
 }
 class Me extends On {
   constructor(t, e) {
-    super(t, e), this.kernel = null, Me.verifyArgs(e), this.filters = e.filters, lt(this.filters, "filters"), this.kernelInitializer = Z(e.kernelInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.kernelConstraint = ot(e.kernelConstraint), this.kernelRegularizer = X(e.kernelRegularizer);
+    super(t, e), this.kernel = null, Me.verifyArgs(e), this.filters = e.filters, lt(this.filters, "filters"), this.kernelInitializer = J(e.kernelInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.kernelConstraint = ot(e.kernelConstraint), this.kernelRegularizer = X(e.kernelRegularizer);
   }
   build(t) {
     t = j(t);
@@ -10835,7 +10835,7 @@ class $o extends Me {
       throw new d("Fields kernelInitializer, kernelRegularizer and kernelConstraint are invalid for SeparableConv2D. Use depthwiseInitializer, depthwiseRegularizer, depthwiseConstraint, pointwiseInitializer, pointwiseRegularizer and pointwiseConstraint instead.");
     if (e.padding != null && e.padding !== "same" && e.padding !== "valid")
       throw new d(`SeparableConv${this.rank}D supports only padding modes: 'same' and 'valid', but received ${JSON.stringify(e.padding)}`);
-    this.depthMultiplier = e.depthMultiplier == null ? 1 : e.depthMultiplier, this.depthwiseInitializer = Z(e.depthwiseInitializer || this.DEFAULT_DEPTHWISE_INITIALIZER), this.depthwiseRegularizer = X(e.depthwiseRegularizer), this.depthwiseConstraint = ot(e.depthwiseConstraint), this.pointwiseInitializer = Z(e.depthwiseInitializer || this.DEFAULT_POINTWISE_INITIALIZER), this.pointwiseRegularizer = X(e.pointwiseRegularizer), this.pointwiseConstraint = ot(e.pointwiseConstraint);
+    this.depthMultiplier = e.depthMultiplier == null ? 1 : e.depthMultiplier, this.depthwiseInitializer = J(e.depthwiseInitializer || this.DEFAULT_DEPTHWISE_INITIALIZER), this.depthwiseRegularizer = X(e.depthwiseRegularizer), this.depthwiseConstraint = ot(e.depthwiseConstraint), this.pointwiseInitializer = J(e.depthwiseInitializer || this.DEFAULT_POINTWISE_INITIALIZER), this.pointwiseRegularizer = X(e.pointwiseRegularizer), this.pointwiseConstraint = ot(e.pointwiseConstraint);
   }
   build(t) {
     if (t = j(t), t.length < this.rank + 2)
@@ -10985,7 +10985,7 @@ function Em(n, t, e = [1, 1], s = "valid", i, r) {
 }
 class Lo extends On {
   constructor(t) {
-    super(2, t), this.depthwiseKernel = null, this.depthMultiplier = t.depthMultiplier == null ? 1 : t.depthMultiplier, this.depthwiseInitializer = Z(t.depthwiseInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.depthwiseConstraint = ot(t.depthwiseConstraint), this.depthwiseRegularizer = X(t.depthwiseRegularizer);
+    super(2, t), this.depthwiseKernel = null, this.depthMultiplier = t.depthMultiplier == null ? 1 : t.depthMultiplier, this.depthwiseInitializer = J(t.depthwiseInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.depthwiseConstraint = ot(t.depthwiseConstraint), this.depthwiseRegularizer = X(t.depthwiseRegularizer);
   }
   build(t) {
     if (t = j(t), t.length < 4)
@@ -11179,11 +11179,11 @@ class de extends G {
       if (this.states_ == null)
         Array.isArray(this.cell.stateSize) ? this.states_ = this.cell.stateSize.map((i) => vt([s, i])) : this.states_ = [vt([s, this.cell.stateSize])];
       else if (t == null)
-        J(this.states_), this.keptStates != null && (J(this.keptStates), this.keptStates = []), Array.isArray(this.cell.stateSize) ? this.states_ = this.cell.stateSize.map((i) => vt([s, i])) : this.states_[0] = vt([s, this.cell.stateSize]);
+        Z(this.states_), this.keptStates != null && (Z(this.keptStates), this.keptStates = []), Array.isArray(this.cell.stateSize) ? this.states_ = this.cell.stateSize.map((i) => vt([s, i])) : this.states_[0] = vt([s, this.cell.stateSize]);
       else {
         if (Array.isArray(t) || (t = [t]), t.length !== this.states_.length)
           throw new d(`Layer ${this.name} expects ${this.states_.length} state(s), but it received ${t.length} state value(s). Input received: ${t}`);
-        e === !0 ? this.keptStates.push(this.states_.slice()) : J(this.states_);
+        e === !0 ? this.keptStates.push(this.states_.slice()) : Z(this.states_);
         for (let i = 0; i < this.states_.length; ++i) {
           const r = t[i], o = Array.isArray(this.cell.stateSize) ? this.cell.stateSize[i] : this.cell.stateSize, a = [s, o];
           if (!Vt(r.shape, a))
@@ -11191,7 +11191,7 @@ class de extends G {
           this.states_[i] = r;
         }
       }
-      this.states_ = this.states_.map((i) => Jt(i.clone()));
+      this.states_ = this.states_.map((i) => Zt(i.clone()));
     });
   }
   apply(t, e) {
@@ -11236,7 +11236,7 @@ class de extends G {
   getInitialState(t) {
     return x(() => {
       let e = vt(t.shape);
-      return e = W(e, [1, 2]), e = Qe(e), Array.isArray(this.cell.stateSize) ? this.cell.stateSize.map((s) => s > 1 ? Zn(e, [1, s]) : e) : this.cell.stateSize > 1 ? [Zn(e, [1, this.cell.stateSize])] : [e];
+      return e = W(e, [1, 2]), e = Qe(e), Array.isArray(this.cell.stateSize) ? this.cell.stateSize.map((s) => s > 1 ? Jn(e, [1, s]) : e) : this.cell.stateSize > 1 ? [Jn(e, [1, this.cell.stateSize])] : [e];
     });
   }
   get trainableWeights() {
@@ -11265,7 +11265,7 @@ class de extends G {
   }
   /** @nocollapse */
   static fromConfig(t, e, s = {}) {
-    const i = e.cell, r = Zt(i, s);
+    const i = e.cell, r = Jt(i, s);
     return new t(Object.assign(e, { cell: r }));
   }
 }
@@ -11275,7 +11275,7 @@ class _n extends G {
 }
 class Ms extends _n {
   constructor(t) {
-    super(t), this.DEFAULT_ACTIVATION = "tanh", this.DEFAULT_KERNEL_INITIALIZER = "glorotNormal", this.DEFAULT_RECURRENT_INITIALIZER = "orthogonal", this.DEFAULT_BIAS_INITIALIZER = "zeros", this.units = t.units, lt(this.units, "units"), this.activation = le(t.activation == null ? this.DEFAULT_ACTIVATION : t.activation), this.useBias = t.useBias == null ? !0 : t.useBias, this.kernelInitializer = Z(t.kernelInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.recurrentInitializer = Z(t.recurrentInitializer || this.DEFAULT_RECURRENT_INITIALIZER), this.biasInitializer = Z(t.biasInitializer || this.DEFAULT_BIAS_INITIALIZER), this.kernelRegularizer = X(t.kernelRegularizer), this.recurrentRegularizer = X(t.recurrentRegularizer), this.biasRegularizer = X(t.biasRegularizer), this.kernelConstraint = ot(t.kernelConstraint), this.recurrentConstraint = ot(t.recurrentConstraint), this.biasConstraint = ot(t.biasConstraint), this.dropout = Fe([1, oe([0, t.dropout == null ? 0 : t.dropout])]), this.recurrentDropout = Fe([
+    super(t), this.DEFAULT_ACTIVATION = "tanh", this.DEFAULT_KERNEL_INITIALIZER = "glorotNormal", this.DEFAULT_RECURRENT_INITIALIZER = "orthogonal", this.DEFAULT_BIAS_INITIALIZER = "zeros", this.units = t.units, lt(this.units, "units"), this.activation = le(t.activation == null ? this.DEFAULT_ACTIVATION : t.activation), this.useBias = t.useBias == null ? !0 : t.useBias, this.kernelInitializer = J(t.kernelInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.recurrentInitializer = J(t.recurrentInitializer || this.DEFAULT_RECURRENT_INITIALIZER), this.biasInitializer = J(t.biasInitializer || this.DEFAULT_BIAS_INITIALIZER), this.kernelRegularizer = X(t.kernelRegularizer), this.recurrentRegularizer = X(t.recurrentRegularizer), this.biasRegularizer = X(t.biasRegularizer), this.kernelConstraint = ot(t.kernelConstraint), this.recurrentConstraint = ot(t.recurrentConstraint), this.biasConstraint = ot(t.biasConstraint), this.dropout = Fe([1, oe([0, t.dropout == null ? 0 : t.dropout])]), this.recurrentDropout = Fe([
       1,
       oe([0, t.recurrentDropout == null ? 0 : t.recurrentDropout])
     ]), this.dropoutFunc = t.dropoutFunc, this.stateSize = this.units, this.dropoutMask = null, this.recurrentDropoutMask = null;
@@ -11343,7 +11343,7 @@ class Mo extends de {
   }
   call(t, e) {
     return x(() => {
-      this.cell.dropoutMask != null && (J(this.cell.dropoutMask), this.cell.dropoutMask = null), this.cell.recurrentDropoutMask != null && (J(this.cell.recurrentDropoutMask), this.cell.recurrentDropoutMask = null);
+      this.cell.dropoutMask != null && (Z(this.cell.dropoutMask), this.cell.dropoutMask = null), this.cell.recurrentDropoutMask != null && (Z(this.cell.recurrentDropoutMask), this.cell.recurrentDropoutMask = null);
       const s = e == null ? null : e.mask, i = e == null ? null : e.training, r = e == null ? null : e.initialState;
       return super.call(t, { mask: s, training: i, initialState: r });
     });
@@ -11359,7 +11359,7 @@ class _s extends _n {
   constructor(t) {
     if (super(t), this.DEFAULT_ACTIVATION = "tanh", this.DEFAULT_RECURRENT_ACTIVATION = "hardSigmoid", this.DEFAULT_KERNEL_INITIALIZER = "glorotNormal", this.DEFAULT_RECURRENT_INITIALIZER = "orthogonal", this.DEFAULT_BIAS_INITIALIZER = "zeros", t.resetAfter)
       throw new d("GRUCell does not support reset_after parameter set to true.");
-    this.units = t.units, lt(this.units, "units"), this.activation = le(t.activation === void 0 ? this.DEFAULT_ACTIVATION : t.activation), this.recurrentActivation = le(t.recurrentActivation === void 0 ? this.DEFAULT_RECURRENT_ACTIVATION : t.recurrentActivation), this.useBias = t.useBias == null ? !0 : t.useBias, this.kernelInitializer = Z(t.kernelInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.recurrentInitializer = Z(t.recurrentInitializer || this.DEFAULT_RECURRENT_INITIALIZER), this.biasInitializer = Z(t.biasInitializer || this.DEFAULT_BIAS_INITIALIZER), this.kernelRegularizer = X(t.kernelRegularizer), this.recurrentRegularizer = X(t.recurrentRegularizer), this.biasRegularizer = X(t.biasRegularizer), this.kernelConstraint = ot(t.kernelConstraint), this.recurrentConstraint = ot(t.recurrentConstraint), this.biasConstraint = ot(t.biasConstraint), this.dropout = Fe([1, oe([0, t.dropout == null ? 0 : t.dropout])]), this.recurrentDropout = Fe([
+    this.units = t.units, lt(this.units, "units"), this.activation = le(t.activation === void 0 ? this.DEFAULT_ACTIVATION : t.activation), this.recurrentActivation = le(t.recurrentActivation === void 0 ? this.DEFAULT_RECURRENT_ACTIVATION : t.recurrentActivation), this.useBias = t.useBias == null ? !0 : t.useBias, this.kernelInitializer = J(t.kernelInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.recurrentInitializer = J(t.recurrentInitializer || this.DEFAULT_RECURRENT_INITIALIZER), this.biasInitializer = J(t.biasInitializer || this.DEFAULT_BIAS_INITIALIZER), this.kernelRegularizer = X(t.kernelRegularizer), this.recurrentRegularizer = X(t.recurrentRegularizer), this.biasRegularizer = X(t.biasRegularizer), this.kernelConstraint = ot(t.kernelConstraint), this.recurrentConstraint = ot(t.recurrentConstraint), this.biasConstraint = ot(t.biasConstraint), this.dropout = Fe([1, oe([0, t.dropout == null ? 0 : t.dropout])]), this.recurrentDropout = Fe([
       1,
       oe([0, t.recurrentDropout == null ? 0 : t.recurrentDropout])
     ]), this.dropoutFunc = t.dropoutFunc, this.implementation = t.implementation, this.stateSize = this.units, this.dropoutMask = null, this.recurrentDropoutMask = null;
@@ -11433,7 +11433,7 @@ class _o extends de {
   }
   call(t, e) {
     return x(() => {
-      this.cell.dropoutMask != null && (J(this.cell.dropoutMask), this.cell.dropoutMask = null), this.cell.recurrentDropoutMask != null && (J(this.cell.recurrentDropoutMask), this.cell.recurrentDropoutMask = null);
+      this.cell.dropoutMask != null && (Z(this.cell.dropoutMask), this.cell.dropoutMask = null), this.cell.recurrentDropoutMask != null && (Z(this.cell.recurrentDropoutMask), this.cell.recurrentDropoutMask = null);
       const s = e == null ? null : e.mask, i = e == null ? null : e.training, r = e == null ? null : e.initialState;
       return super.call(t, { mask: s, training: i, initialState: r });
     });
@@ -11447,7 +11447,7 @@ _o.className = "GRU";
 S(_o);
 class Rn extends _n {
   constructor(t) {
-    super(t), this.DEFAULT_ACTIVATION = "tanh", this.DEFAULT_RECURRENT_ACTIVATION = "hardSigmoid", this.DEFAULT_KERNEL_INITIALIZER = "glorotNormal", this.DEFAULT_RECURRENT_INITIALIZER = "orthogonal", this.DEFAULT_BIAS_INITIALIZER = "zeros", this.units = t.units, lt(this.units, "units"), this.activation = le(t.activation === void 0 ? this.DEFAULT_ACTIVATION : t.activation), this.recurrentActivation = le(t.recurrentActivation === void 0 ? this.DEFAULT_RECURRENT_ACTIVATION : t.recurrentActivation), this.useBias = t.useBias == null ? !0 : t.useBias, this.kernelInitializer = Z(t.kernelInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.recurrentInitializer = Z(t.recurrentInitializer || this.DEFAULT_RECURRENT_INITIALIZER), this.biasInitializer = Z(t.biasInitializer || this.DEFAULT_BIAS_INITIALIZER), this.unitForgetBias = t.unitForgetBias, this.kernelRegularizer = X(t.kernelRegularizer), this.recurrentRegularizer = X(t.recurrentRegularizer), this.biasRegularizer = X(t.biasRegularizer), this.kernelConstraint = ot(t.kernelConstraint), this.recurrentConstraint = ot(t.recurrentConstraint), this.biasConstraint = ot(t.biasConstraint), this.dropout = Fe([1, oe([0, t.dropout == null ? 0 : t.dropout])]), this.recurrentDropout = Fe([
+    super(t), this.DEFAULT_ACTIVATION = "tanh", this.DEFAULT_RECURRENT_ACTIVATION = "hardSigmoid", this.DEFAULT_KERNEL_INITIALIZER = "glorotNormal", this.DEFAULT_RECURRENT_INITIALIZER = "orthogonal", this.DEFAULT_BIAS_INITIALIZER = "zeros", this.units = t.units, lt(this.units, "units"), this.activation = le(t.activation === void 0 ? this.DEFAULT_ACTIVATION : t.activation), this.recurrentActivation = le(t.recurrentActivation === void 0 ? this.DEFAULT_RECURRENT_ACTIVATION : t.recurrentActivation), this.useBias = t.useBias == null ? !0 : t.useBias, this.kernelInitializer = J(t.kernelInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.recurrentInitializer = J(t.recurrentInitializer || this.DEFAULT_RECURRENT_INITIALIZER), this.biasInitializer = J(t.biasInitializer || this.DEFAULT_BIAS_INITIALIZER), this.unitForgetBias = t.unitForgetBias, this.kernelRegularizer = X(t.kernelRegularizer), this.recurrentRegularizer = X(t.recurrentRegularizer), this.biasRegularizer = X(t.biasRegularizer), this.kernelConstraint = ot(t.kernelConstraint), this.recurrentConstraint = ot(t.recurrentConstraint), this.biasConstraint = ot(t.biasConstraint), this.dropout = Fe([1, oe([0, t.dropout == null ? 0 : t.dropout])]), this.recurrentDropout = Fe([
       1,
       oe([0, t.recurrentDropout == null ? 0 : t.recurrentDropout])
     ]), this.dropoutFunc = t.dropoutFunc, this.implementation = t.implementation, this.stateSize = [this.units, this.units], this.dropoutMask = null, this.recurrentDropoutMask = null;
@@ -11538,7 +11538,7 @@ class Ro extends de {
   }
   call(t, e) {
     return x(() => {
-      this.cell.dropoutMask != null && (J(this.cell.dropoutMask), this.cell.dropoutMask = null), this.cell.recurrentDropoutMask != null && (J(this.cell.recurrentDropoutMask), this.cell.recurrentDropoutMask = null);
+      this.cell.dropoutMask != null && (Z(this.cell.dropoutMask), this.cell.dropoutMask = null), this.cell.recurrentDropoutMask != null && (Z(this.cell.recurrentDropoutMask), this.cell.recurrentDropoutMask = null);
       const s = e == null ? null : e.mask, i = e == null ? null : e.training, r = e == null ? null : e.initialState;
       return super.call(t, { mask: s, training: i, initialState: r });
     });
@@ -11600,7 +11600,7 @@ class Rs extends _n {
   static fromConfig(t, e, s = {}) {
     const i = [];
     for (const r of e.cells)
-      i.push(Zt(r, s));
+      i.push(Jt(r, s));
     return new t({ cells: i });
   }
   get trainableWeights() {
@@ -11654,7 +11654,7 @@ Rs.className = "StackedRNNCells";
 S(Rs);
 function ue(n) {
   const { ones: t, rate: e, training: s = !1, count: i = 1, dropoutFunc: r } = n, o = () => r != null ? r(t(), e) : Dr(t(), e), a = () => en(o, t, s);
-  return !i || i <= 1 ? Jt(a().clone()) : Array(i).fill(void 0).map(a).map((u) => Jt(u.clone()));
+  return !i || i <= 1 ? Zt(a().clone()) : Array(i).fill(void 0).map(a).map((u) => Zt(u.clone()));
 }
 /**
  * @license
@@ -11683,7 +11683,7 @@ class Bo extends de {
   }
   call(t, e) {
     return x(() => {
-      if (this.cell.dropoutMask != null && (J(this.cell.dropoutMask), this.cell.dropoutMask = null), this.cell.recurrentDropoutMask != null && (J(this.cell.recurrentDropoutMask), this.cell.recurrentDropoutMask = null), e && e.constants)
+      if (this.cell.dropoutMask != null && (Z(this.cell.dropoutMask), this.cell.dropoutMask = null), this.cell.recurrentDropoutMask != null && (Z(this.cell.recurrentDropoutMask), this.cell.recurrentDropoutMask = null), e && e.constants)
         throw new d("ConvRNN2D cell does not support constants");
       const s = e == null ? null : e.mask, i = e == null ? null : e.training, r = e == null ? null : e.initialState;
       return super.call(t, { mask: s, training: i, initialState: r });
@@ -11709,11 +11709,11 @@ class Bo extends de {
       if (this.getStates() == null)
         Array.isArray(this.cell.stateSize) ? this.states_ = this.cell.stateSize.map(() => vt(r)) : this.states_ = [vt(r)];
       else if (t == null)
-        J(this.states_), this.keptStates != null && (J(this.keptStates), this.keptStates = []), Array.isArray(this.cell.stateSize) ? this.states_ = this.cell.stateSize.map(() => vt(r)) : this.states_[0] = vt(r);
+        Z(this.states_), this.keptStates != null && (Z(this.keptStates), this.keptStates = []), Array.isArray(this.cell.stateSize) ? this.states_ = this.cell.stateSize.map(() => vt(r)) : this.states_[0] = vt(r);
       else {
         if (Array.isArray(t) || (t = [t]), t.length !== this.states_.length)
           throw new d(`Layer ${this.name} expects ${this.states_.length} state(s), but it received ${t.length} state value(s). Input received: ${t}`);
-        e ? this.keptStates.push(this.states_.slice()) : J(this.states_);
+        e ? this.keptStates.push(this.states_.slice()) : Z(this.states_);
         for (let a = 0; a < this.states_.length; ++a) {
           const l = t[a], u = r;
           if (!Vt(l.shape, u))
@@ -11721,7 +11721,7 @@ class Bo extends de {
           this.states_[a] = l;
         }
       }
-      this.states_ = this.states_.map((a) => Jt(a.clone()));
+      this.states_ = this.states_.map((a) => Zt(a.clone()));
     });
   }
   computeSingleOutputShape(t) {
@@ -11891,7 +11891,7 @@ class Po extends G {
       let e = null;
       t.batchSize != null && (e = t.batchSize), this.batchInputShape = [e, t.inputDim];
     }
-    this.units = t.units, lt(this.units, "units"), this.activation = le(t.activation), t.useBias != null && (this.useBias = t.useBias), this.kernelInitializer = Z(t.kernelInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.biasInitializer = Z(t.biasInitializer || this.DEFAULT_BIAS_INITIALIZER), this.kernelConstraint = ot(t.kernelConstraint), this.biasConstraint = ot(t.biasConstraint), this.kernelRegularizer = X(t.kernelRegularizer), this.biasRegularizer = X(t.biasRegularizer), this.activityRegularizer = X(t.activityRegularizer), this.supportsMasking = !0, this.inputSpec = [{ minNDim: 2 }];
+    this.units = t.units, lt(this.units, "units"), this.activation = le(t.activation), t.useBias != null && (this.useBias = t.useBias), this.kernelInitializer = J(t.kernelInitializer || this.DEFAULT_KERNEL_INITIALIZER), this.biasInitializer = J(t.biasInitializer || this.DEFAULT_BIAS_INITIALIZER), this.kernelConstraint = ot(t.kernelConstraint), this.biasConstraint = ot(t.biasConstraint), this.kernelRegularizer = X(t.kernelRegularizer), this.biasRegularizer = X(t.biasRegularizer), this.activityRegularizer = X(t.activityRegularizer), this.supportsMasking = !0, this.inputSpec = [{ minNDim: 2 }];
   }
   build(t) {
     t = j(t);
@@ -12133,13 +12133,13 @@ S(qo);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-class Jo extends G {
+class Zo extends G {
   constructor(t) {
     if (super(t), this.embeddings = null, this.DEFAULT_EMBEDDINGS_INITIALIZER = "randomUniform", t.batchInputShape == null && t.inputShape == null) {
       let e = null;
       t.batchSize != null && (e = t.batchSize), t.inputLength == null ? this.batchInputShape = [e, null] : this.batchInputShape = [e].concat(H(t.inputLength));
     }
-    this.inputDim = t.inputDim, lt(this.inputDim, "inputDim"), this.outputDim = t.outputDim, lt(this.outputDim, "outputDim"), this.embeddingsInitializer = Z(t.embeddingsInitializer || this.DEFAULT_EMBEDDINGS_INITIALIZER), this.embeddingsRegularizer = X(t.embeddingsRegularizer), this.activityRegularizer = X(t.activityRegularizer), this.embeddingsConstraint = ot(t.embeddingsConstraint), this.maskZero = t.maskZero, this.supportsMasking = t.maskZero, this.inputLength = t.inputLength;
+    this.inputDim = t.inputDim, lt(this.inputDim, "inputDim"), this.outputDim = t.outputDim, lt(this.outputDim, "outputDim"), this.embeddingsInitializer = J(t.embeddingsInitializer || this.DEFAULT_EMBEDDINGS_INITIALIZER), this.embeddingsRegularizer = X(t.embeddingsRegularizer), this.activityRegularizer = X(t.activityRegularizer), this.embeddingsConstraint = ot(t.embeddingsConstraint), this.maskZero = t.maskZero, this.supportsMasking = t.maskZero, this.inputLength = t.inputLength;
   }
   build(t) {
     this.embeddings = this.addWeight("embeddings", [this.inputDim, this.outputDim], this.dtype, this.embeddingsInitializer, this.embeddingsRegularizer, !0, this.embeddingsConstraint), this.built = !0;
@@ -12191,8 +12191,8 @@ class Jo extends G {
     return Object.assign(t, e), t;
   }
 }
-Jo.className = "Embedding";
-S(Jo);
+Zo.className = "Embedding";
+S(Zo);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -12335,12 +12335,12 @@ class ve extends G {
       e = e.map((i) => i == null ? i : ye(i, 0));
       let s = e[0];
       for (let i = 1; i < e.length - 1; ++i)
-        s = Ze(s, e[i]);
+        s = Je(s, e[i]);
       return s;
     });
   }
 }
-class Zo extends ve {
+class Jo extends ve {
   constructor(t) {
     super(t);
   }
@@ -12353,8 +12353,8 @@ class Zo extends ve {
     });
   }
 }
-Zo.className = "Add";
-S(Zo);
+Jo.className = "Add";
+S(Jo);
 class Xo extends ve {
   constructor(t) {
     super(t);
@@ -12722,7 +12722,7 @@ function _m(n, t, e, s, i = 1e-3) {
 }
 class oa extends G {
   constructor(t) {
-    t == null && (t = {}), super(t), this.supportsMasking = !0, this.axis = t.axis == null ? -1 : t.axis, this.momentum = t.momentum == null ? 0.99 : t.momentum, this.epsilon = t.epsilon == null ? 1e-3 : t.epsilon, this.center = t.center == null ? !0 : t.center, this.scale = t.scale == null ? !0 : t.scale, this.betaInitializer = Z(t.betaInitializer || "zeros"), this.gammaInitializer = Z(t.gammaInitializer || "ones"), this.movingMeanInitializer = Z(t.movingMeanInitializer || "zeros"), this.movingVarianceInitializer = Z(t.movingVarianceInitializer || "ones"), this.betaConstraint = ot(t.betaConstraint), this.gammaConstraint = ot(t.gammaConstraint), this.betaRegularizer = X(t.betaRegularizer), this.gammaRegularizer = X(t.gammaRegularizer);
+    t == null && (t = {}), super(t), this.supportsMasking = !0, this.axis = t.axis == null ? -1 : t.axis, this.momentum = t.momentum == null ? 0.99 : t.momentum, this.epsilon = t.epsilon == null ? 1e-3 : t.epsilon, this.center = t.center == null ? !0 : t.center, this.scale = t.scale == null ? !0 : t.scale, this.betaInitializer = J(t.betaInitializer || "zeros"), this.gammaInitializer = J(t.gammaInitializer || "ones"), this.movingMeanInitializer = J(t.movingMeanInitializer || "zeros"), this.movingVarianceInitializer = J(t.movingVarianceInitializer || "ones"), this.betaConstraint = ot(t.betaConstraint), this.gammaConstraint = ot(t.gammaConstraint), this.betaRegularizer = X(t.betaRegularizer), this.gammaRegularizer = X(t.gammaRegularizer);
   }
   build(t) {
     t = j(t);
@@ -12793,7 +12793,7 @@ class aa extends G {
           throw new Error(`Expected axis to be an array of integers, but received ${JSON.stringify(this.axis)}`);
     } else
       throw new Error(`Expected axis to be an integer or an array of integers, but received ${JSON.stringify(this.axis)}`);
-    this.epsilon = t.epsilon == null ? 1e-3 : t.epsilon, this.center = t.center == null ? !0 : t.center, this.scale = t.scale == null ? !0 : t.scale, this.betaInitializer = Z(t.betaInitializer || "zeros"), this.gammaInitializer = Z(t.gammaInitializer || "ones"), this.betaRegularizer = X(t.betaRegularizer), this.gammaRegularizer = X(t.gammaRegularizer), this.supportsMasking = !0;
+    this.epsilon = t.epsilon == null ? 1e-3 : t.epsilon, this.center = t.center == null ? !0 : t.center, this.scale = t.scale == null ? !0 : t.scale, this.betaInitializer = J(t.betaInitializer || "zeros"), this.gammaInitializer = J(t.gammaInitializer || "ones"), this.betaRegularizer = X(t.betaRegularizer), this.gammaRegularizer = X(t.gammaRegularizer), this.supportsMasking = !0;
   }
   build(t) {
     t = j(t);
@@ -13229,7 +13229,7 @@ class Sa extends G {
   }
   /** @nocollapse */
   static fromConfig(t, e, s = {}) {
-    const i = e.layer, r = Zt(i, s);
+    const i = e.layer, r = Jt(i, s);
     delete e.layer;
     const o = { layer: r };
     return Object.assign(o, e), new t(o);
@@ -13275,9 +13275,9 @@ class Ia extends Sa {
   constructor(t) {
     super(t);
     const e = t.layer.getConfig(), s = {};
-    s.className = t.layer.getClassName(), s.config = e, this.forwardLayer = Zt(s), e.goBackwards = e.goBackwards !== !0;
+    s.className = t.layer.getClassName(), s.config = e, this.forwardLayer = Jt(s), e.goBackwards = e.goBackwards !== !0;
     const i = {};
-    if (i.className = t.layer.getClassName(), i.config = e, this.backwardLayer = Zt(i), this.forwardLayer.name = "forward_" + this.forwardLayer.name, this.backwardLayer.name = "backward_" + this.backwardLayer.name, this.mergeMode = t.mergeMode === void 0 ? Wm : t.mergeMode, Bm(this.mergeMode), t.weights)
+    if (i.className = t.layer.getClassName(), i.config = e, this.backwardLayer = Jt(i), this.forwardLayer.name = "forward_" + this.forwardLayer.name, this.backwardLayer.name = "backward_" + this.backwardLayer.name, this.mergeMode = t.mergeMode === void 0 ? Wm : t.mergeMode, Bm(this.mergeMode), t.weights)
       throw new B("weights support is not implemented for Bidirectional layer yet.");
     this._stateful = t.layer.stateful, this.returnSequences = t.layer.returnSequences, this.returnState = t.layer.returnState, this.supportsMasking = !0, this._trainable = !0, this.inputSpec = t.layer.inputSpec, this.numConstants = null;
   }
@@ -13382,7 +13382,7 @@ class Ia extends Sa {
   }
   /** @nocollapse */
   static fromConfig(t, e) {
-    const s = Zt(e.layer);
+    const s = Jt(e.layer);
     if (delete e.layer, e.numConstants != null)
       throw new B("Deserialization of a Bidirectional layer with numConstants present is not supported yet.");
     const i = e;
@@ -13490,7 +13490,7 @@ function Um(n, t, e, s) {
     throw new d(`When outputMode is not int, maximum output rank is 2 Received outputMode ${t} and input shape ${r} which would result in output rank ${i.rank}.`);
   const o = ["multiHot", "oneHot"].includes(t), a = i;
   let l;
-  if (typeof s < "u" && t === "count" ? l = Zs(a, s, e, o) : l = Zs(a, [], e, o), t !== "tfIdf")
+  if (typeof s < "u" && t === "count" ? l = Js(a, s, e, o) : l = Js(a, [], e, o), t !== "tfIdf")
     return l;
   if (s)
     return w(l, s);
@@ -13529,7 +13529,7 @@ class Ta extends G {
               Received countWeights=${e.countWeights}`);
         s = _(e.countWeights);
       }
-      const i = Ee(t), r = Ju(t), o = Xt(this.numTokens, i).bufferSync().get(0), a = Xe(r, 0).bufferSync().get(0);
+      const i = Ee(t), r = Zu(t), o = Xt(this.numTokens, i).bufferSync().get(0), a = Xe(r, 0).bufferSync().get(0);
       if (!(o && a))
         throw new d(`Input values must be between 0 < values <= numTokens with numTokens=${this.numTokens}`);
       return Um(t, this.outputMode, this.numTokens, s);
@@ -13699,7 +13699,7 @@ export {
   mh as B,
   Ot as C,
   Ws as D,
-  Jo as E,
+  Zo as E,
   Nh as F,
   Ah as G,
   Ch as H,
@@ -13727,7 +13727,7 @@ export {
   jc as q,
   gn as r,
   fs as s,
-  Js as t,
+  Zs as t,
   qc as u,
   dr as v,
   Qc as w,