@genai-fi/nanogpt 0.4.0 → 0.4.2

package/dist/{random_width-CMHmdbSu.js → TiedEmbedding-CnJ1bx4q.js}

@@ -1,25 +1,26 @@
-import { o as F, h as D, E as M, bb as ho, bc as po, bd as mi, j as k, b8 as Mn, be as gi, x as L, bf as bi, bg as yi, bh as wi, bi as ki, bj as xi, bk as Ni, bl as vi, bm as fo, bn as Si, bo as mo, bp as Ai, bq as go, br as Ci, p as Hn, aj as wt, bs as bo, bt as Ii, bu as Di, bv as zi, c as On, s as V, b as w, bw as yo, bx as Ti, by as $i, bz as wo, bA as Ei, bB as Li, bC as Fi, bD as Mi, bE as Oi, bF as Ri, bG as _i, bH as Bi, k as ko, ao as xo, bI as Wi, bJ as No, a5 as $, bK as Ls, bL as vo, bM as So, bN as Ao, bO as Co, bP as Io, A as Do, bQ as zo, bR as To, bS as $o, bT as S, t as x, f as tt, n as Gi, bU as Be, bV as We, aI as Ft, a as Z, ab as Eo, bW as Lo, bX as Fo, Z as ct, a0 as ee, aJ as P, bY as Mo, bZ as Oo, aK as lt, b_ as Ro, z as Q, b$ as _o, c0 as Bo, c1 as Wo, c2 as Go, c3 as Po, c4 as Uo, c5 as Vo, c6 as jo, B as Ko, c7 as Ho, c8 as qo, c9 as Zo, ai as Jo, ca as Xo, C as Yo, Y as he, cb as Qo, Q as tl, cc as el, cd as nl, ce as sl, am as il, cf as rl, a3 as al, an as ol, cg as ll, ap as ul, ch as cl, G as hl, ar as pl, ci as dl, cj as fl, ck as ml, cl as gl, at as bl, a4 as yl, cm as wl, cn as kl, co as xl, M as Nl, cp as vl, cq as Sl, cr as Al, X as Cl, _ as Il, ay as Dl, cs as zl, a6 as Tl, ct as $l, aw as El, P as Ll, cu as Fl, O as qn, az as Ml, cv as Ol, cw as Rl, N as _l, aC as Bl, aB as Wl, q as Gl, aS as Pl, cx as Ul, aT as Vl, cy as jl, aD as Kl, ak as Hl, af as ql, cz as Zl, T as Jl, ag as Xl, S as Yl, u as Ql, cA as tu, cB as eu, cC as nu, aF as su, cD as iu, y as ru, cE as au, a1 as ou, aH as lu, aG as uu, cF as Ie, cG as cu, g as hu, cH as Fs, F as Bt, $ as Fe, D as pu, w as du, aa as xe, cI as fu, cJ as mu, m as Ms, cK as gu, cL as Os, cM as bu } from "./index-XjBAhiFO.js";
-import { M as yu, s as wu, b as ku, g as xu, c as Nu, V as d, f as bn, N as G, h as yn, e as vu, l as Su, i as Zn, j as et, k as ye, m as Ge, n as Jn, o as Pi, p as Ui, t as Rt, R as $t, q as ht, A as Pt, u as K, v as le, x as Xn, y as pt, w as Ht, z as Pe, B as Ue, C as Yn, D as j, E as Ee, F as Gt, G as Au, H as be, I as Vi, J as en, K as Qn, L as ue, O as At, P as nt, Q as Ve, S as Cu, T as Iu, U as je, W as Ut, X as Et, Y as ts, Z as oe, _ as jt, $ as Xe, a0 as ce, a1 as Du, a2 as Rn, a3 as nn, a4 as wn, a5 as ji, a6 as Ct, a7 as Rs, a8 as zu, a9 as Ki, aa as Tu, ab as $u, ac as Eu, ad as Lu, ae as Fu, af as qt, ag as Mu, ah as es, ai as Dt, aj as Ye, ak as Ou, al as _t, am as ot, an as Hi, ao as gt, ap as ne, aq as _s, ar as Ne, d as Lt, as as Bs, at as Ke, au as qi, av as ns, aw as Ru, ax as _u, ay as Zi, az as Bu } from "./tfjs_backend-Bzl2SrRo.js";
-import { r as A } from "./reshape-DFzh97Sc.js";
-import { s as B } from "./sum-C6u3xMi3.js";
-import { m as Ot } from "./mat_mul-CKwFEV1Q.js";
-import { s as Kt } from "./split-CkbeVdF8.js";
-import { s as Wu, c as Ji } from "./sin-BYM-U4Ut.js";
-import { g as Xi, d as ss, e as Ws, c as Gu } from "./axis_util-DeydwOoC.js";
-import { a as Zt, e as Jt, l as Pu } from "./log_sum_exp-DJPkVZZn.js";
-import { s as kn } from "./stack-DaIMO5iX.js";
-import { o as pe } from "./ones-BO01zpJG.js";
-import { z as mt } from "./zeros-CmJFiC84.js";
-import { p as Uu } from "./slice_util-CnVNPQI-.js";
-import { c as is } from "./concat-DS_qH7MI.js";
-import { g as Vu } from "./gather-BUmJIS8n.js";
-import { m as at, a as rs } from "./moments-CrWRPcR3.js";
-import { s as Yi } from "./softmax-4DOn6cPq.js";
-import { m as ve } from "./max-DJvEiCAJ.js";
-import { t as ju } from "./tensor-Cu1fU7H7.js";
-import { r as Ku } from "./range-DQMNzBWs.js";
-import { m as Hu } from "./norm-BzY929B_.js";
-import { v as qu } from "./variable-BS4AKqNU.js";
+import { o as F, h as D, E as M, bb as fo, bc as mo, bd as mi, j as k, b8 as Mn, be as gi, x as L, bf as bi, bg as yi, bh as wi, bi as ki, bj as xi, bk as Ni, bl as vi, bm as go, bn as Si, bo, bp as Ai, bq as yo, br as Ci, p as Hn, al as wt, bs as wo, bt as Ii, bu as Di, bv as zi, c as On, s as V, b as w, bw as ko, bx as Ti, by as $i, bz as xo, bA as Ei, bB as Li, bC as Fi, bD as Mi, bE as Oi, bF as Ri, bG as _i, bH as Bi, k as No, aa as vo, bI as Wi, bJ as So, a5 as $, bK as Ls, bL as Ao, bM as Co, bN as Io, bO as Do, bP as zo, A as To, bQ as $o, bR as Eo, bS as Lo, bT as S, t as x, f as tt, n as Gi, bU as Be, bV as We, ab as Ft, a as Z, af as Fo, bW as Mo, bX as Oo, Z as ct, a0 as ee, ad as P, bY as Ro, bZ as _o, aN as lt, b_ as Bo, z as Q, b$ as Wo, c0 as Go, c1 as Po, c2 as Uo, c3 as Vo, c4 as jo, c5 as Ko, c6 as Ho, B as qo, c7 as Zo, c8 as Jo, c9 as Xo, aq as Yo, ca as Qo, C as tl, Y as he, cb as el, Q as nl, cc as sl, cd as il, ce as rl, at as al, cf as ol, a3 as ll, au as ul, cg as cl, a9 as hl, ch as pl, G as dl, aw as fl, ci as ml, cj as gl, ck as bl, cl as yl, ay as wl, a4 as kl, cm as xl, cn as Nl, co as vl, M as Sl, cp as Al, cq as Cl, cr as Il, X as Dl, _ as zl, aD as Tl, cs as $l, a6 as El, ct as Ll, aB as Fl, P as Ml, cu as Ol, O as qn, aE as Rl, cv as _l, cw as Bl, N as Wl, aH as Gl, aG as Pl, q as Ul, aV as Vl, cx as jl, aW as Kl, cy as Hl, aI as ql, ar as Zl, an as Jl, cz as Xl, T as Yl, ao as Ql, S as tu, u as eu, cA as nu, cB as su, cC as iu, aK as ru, cD as au, y as ou, cE as lu, a1 as uu, aM as cu, aL as hu, cF as Ie, cG as pu, g as du, cH as Fs, F as Bt, $ as Fe, D as fu, w as mu, ac as xe, cI as gu, cJ as bu, m as Ms, cK as yu, cL as Os, cM as wu } from "./index-C4JCoBvj.js";
+import { s as ku, a as xu, g as Nu, b as vu, V as d, N as G, r as bn, e as Su, l as Au, c as Zn, f as et, h as ye, i as Ge, j as Jn, k as Pi, m as Ui, t as Rt, R as Et, n as ht, A as Pt, o as K, p as le, q as Xn, u as pt, w as Ht, v as Pe, x as Ue, y as Yn, z as j, B as Ee, C as Gt, D as Cu, E as be, F as en, G as Qn, H as ue, I as Ct, J as nt, K as Ve, L as Iu, M as Du, O as je, P as Ut, Q as Lt, S as ts, T as oe, U as jt, W as Xe, X as ce, Y as zu, Z as Rn, _ as nn, $ as yn, a0 as Vi, a1 as It, a2 as Rs, a3 as Tu, a4 as ji, a5 as $u, a6 as Eu, a7 as Lu, a8 as Fu, a9 as Mu, aa as qt, ab as Ou, ac as es, ad as zt, ae as Ye, af as Ru, ag as _t, ah as ot, ai as Ki, aj as gt, ak as ne, al as _s, am as Ne, d as St, an as Bs, ao as Ke, ap as Hi, aq as ns, ar as _u, as as Bu, at as qi, au as Wu } from "./tfjs_backend-Cug-PH75.js";
+import { M as Gu, a as wn, f as Zi } from "./dropout-DfDdklfL.js";
+import { z as mt } from "./zeros-dnQxFgAD.js";
+import { o as pe } from "./ones-Bf3YR48P.js";
+import { v as Ji } from "./variable-LJT9Ld63.js";
+import { r as A } from "./reshape-Boe4DuIO.js";
+import { s as B } from "./sum-R28pucR5.js";
+import { m as Ot } from "./mat_mul-415y5Qn2.js";
+import { s as Kt } from "./split-DbcNm1-i.js";
+import { s as Pu, c as Xi } from "./sin-KmhiDuMa.js";
+import { g as Yi, d as ss, e as Ws, c as Uu } from "./axis_util-BgTGy5w8.js";
+import { a as Zt, e as Jt, l as Vu } from "./log_sum_exp-BswFnwOb.js";
+import { s as kn } from "./stack-D1YjmgKN.js";
+import { p as ju } from "./slice_util-19zDNNSn.js";
+import { c as is } from "./concat-CuRsVY-K.js";
+import { g as Qi } from "./gather-ZYRWhmXR.js";
+import { m as at, a as rs } from "./moments-CjeIaVdp.js";
+import { s as tr } from "./softmax-Cujsg4ay.js";
+import { m as ve } from "./max-CP_9O2Yd.js";
+import { t as Ku } from "./tensor-BVeHdl7V.js";
+import { r as Hu } from "./range-9AzeApCc.js";
+import { m as qu } from "./norm-CZM380I3.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.
@@ -38,7 +39,7 @@ import { v as qu } from "./variable-BS4AKqNU.js";
  */
 function Zu(s, t = null, e = !1) {
   const i = { x: D(s, "x", "all", "bool") }, r = { axis: t, keepDims: e };
-  return M.runKernel(ho, i, r);
+  return M.runKernel(fo, i, r);
 }
 const Ju = /* @__PURE__ */ F({ all_: Zu });
 /**
@@ -59,7 +60,7 @@ const Ju = /* @__PURE__ */ F({ all_: Zu });
  */
 function Xu(s, t = null, e = !1) {
   const i = { x: D(s, "x", "any", "bool") }, r = { axis: t, keepDims: e };
-  return M.runKernel(po, i, r);
+  return M.runKernel(mo, i, r);
 }
 const Gs = /* @__PURE__ */ F({ any_: Xu });
 /**
@@ -473,13 +474,13 @@ function Cc(s, t, e, n, i) {
   t.rank === 4 && (o = !0, a = A(t, [1, t.shape[0], t.shape[1], t.shape[2], t.shape[3]]), r = [1, s[0], s[1], s[2], s[3]]);
   const l = r[4], u = a.shape[4];
   k(r.length === 5, () => `Error in conv3dDerInput: inShape must be length 5, but got length ${r.length}.`), k(a.rank === 5, () => `Error in conv3dDerInput: dy must be rank 5, but got rank ${a.rank}`), k(e.rank === 5, () => `Error in conv3dDerInput: filter must be rank 5, but got rank ${e.rank}`), k(l === e.shape[3], () => `Error in conv3dDerInput: depth of input (${l}) must match input depth for filter ${e.shape[3]}.`), k(u === e.shape[4], () => `Error in conv3dDerInput: depth of output (${u}) must match output depth for filter ${e.shape[4]}.`);
-  const c = { dy: a, filter: e }, h = { pad: i, strides: n, inputShape: r }, p = M.runKernel(fo, c, h);
+  const c = { dy: a, filter: e }, h = { pad: i, strides: n, inputShape: r }, p = M.runKernel(go, c, h);
   return o ? A(p, [p.shape[1], p.shape[2], p.shape[3], p.shape[4]]) : p;
 }
-const Qi = /* @__PURE__ */ F({ conv3DBackpropInput_: Cc });
+const er = /* @__PURE__ */ F({ conv3DBackpropInput_: Cc });
 function Ic(s, t, e, n, i) {
   const r = D(s, "x", "conv3dTranspose"), a = D(t, "filter", "conv3dTranspose");
-  return Qi(e, r, a, n, i);
+  return er(e, r, a, n, i);
 }
 const Dc = /* @__PURE__ */ F({ conv3dTranspose_: Ic });
 /**
@@ -521,7 +522,7 @@ const Tc = /* @__PURE__ */ F({ cosh_: zc });
  */
 function $c(s, t = 0, e = !1, n = !1) {
   const r = { x: D(s, "x", "cumprod") }, a = { axis: t, exclusive: e, reverse: n };
-  return M.runKernel(mo, r, a);
+  return M.runKernel(bo, r, a);
 }
 const Ps = /* @__PURE__ */ F({ cumprod_: $c });
 /**
@@ -565,7 +566,7 @@ function Fc(s, t, e, n = !1) {
   const i = D(s, "x", "denseBincount"), r = D(t, "weights", "denseBincount");
   k(i.dtype === "int32", () => `Error in denseBincount: input dtype must be int32, but got ${i.dtype}`), k(i.rank <= 2, () => `Error in denseBincount: input must be at most rank 2, but got rank ${i.rank}.`), k(e >= 0, () => `size must be non-negative, but got ${e}.`), k(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 a = { x: i, weights: r }, o = { size: e, binaryOutput: n };
-  return M.runKernel(go, a, o);
+  return M.runKernel(yo, a, o);
 }
 const Us = /* @__PURE__ */ F({ denseBincount_: Fc });
 /**
@@ -593,7 +594,7 @@ function Mc(s, t, e, n, i = "NHWC", r = [1, 1], a) {
   const p = { x: u, filter: l }, f = { strides: e, pad: n, dataFormat: i, dilations: r, dimRoundingMode: a }, g = M.runKernel(Ci, p, f);
   return c ? A(g, [g.shape[1], g.shape[2], g.shape[3]]) : g;
 }
-const tr = /* @__PURE__ */ F({ depthwiseConv2d_: Mc });
+const nr = /* @__PURE__ */ F({ depthwiseConv2d_: Mc });
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.
@@ -614,7 +615,7 @@ function Oc(s, t) {
   let e = D(s, "a", "equal", "string_or_numeric"), n = D(t, "b", "equal", "string_or_numeric");
   [e, n] = Hn(e, n), wt(e.shape, n.shape);
   const i = { a: e, b: n };
-  return M.runKernel(bo, i);
+  return M.runKernel(wo, i);
 }
 const Xt = /* @__PURE__ */ F({ equal_: Oc });
 /**
@@ -729,7 +730,7 @@ const Uc = /* @__PURE__ */ F({ logSoftmax_: Pc });
  */
 function Vc(s) {
   const e = { x: D(s, "x", "logicalNot", "bool") };
-  return M.runKernel(yo, e);
+  return M.runKernel(ko, e);
 }
 const jc = /* @__PURE__ */ F({ logicalNot_: Vc });
 /**
@@ -800,7 +801,7 @@ function Jc(s, t) {
   let e = D(s, "a", "notEqual", "string_or_numeric"), n = D(t, "b", "notEqual", "string_or_numeric");
   [e, n] = Hn(e, n), wt(e.shape, n.shape);
   const i = { a: e, b: n };
-  return M.runKernel(wo, i);
+  return M.runKernel(xo, i);
 }
 const Bn = /* @__PURE__ */ F({ notEqual_: Jc });
 /**
@@ -846,7 +847,7 @@ function Qc(s) {
   const e = { x: D(s, "x", "onesLike") };
   return M.runKernel(Li, e);
 }
-const It = /* @__PURE__ */ F({ onesLike_: Qc });
+const Dt = /* @__PURE__ */ F({ onesLike_: Qc });
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.
@@ -870,7 +871,7 @@ function th(s, t, e = 0) {
   const i = { paddings: t, constantValue: e }, r = { x: n };
   return M.runKernel(Fi, r, i);
 }
-const er = /* @__PURE__ */ F({ pad_: th });
+const sr = /* @__PURE__ */ F({ pad_: th });
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.
@@ -965,7 +966,7 @@ function lh(s, t, e, n, i, r = [1, 1], a = "NHWC") {
   k(c.rank === 4, () => `Error in separableConv2d: input must be rank 4, but got rank ${c.rank}.`), k(l.rank === 4, () => `Error in separableConv2d: depthwise filter must be rank 4, but got rank ${l.rank}.`), k(u.rank === 4, () => `Error in separableConv2d: pointwise filter must be rank 4, but got rank ${l.rank}.`), k(u.shape[0] === 1, () => `Error in separableConv2d: the first dimension of pointwise filter  must be 1, but got ${u.shape[0]}.`), k(u.shape[1] === 1, () => `Error in separableConv2d: the second dimension of pointwise filter must be 1, but got ${u.shape[1]}.`);
   const p = l.shape[2], f = l.shape[3];
   k(u.shape[2] === p * f, () => `Error in separableConv2d: the third dimension of pointwise filter must be ${p * f}, but got ${u.shape[2]}.`);
-  const g = tr(c, l, n, i, a, r), m = Ce(g, u, 1, "valid", a);
+  const g = nr(c, l, n, i, a, r), m = Ce(g, u, 1, "valid", a);
   return h ? A(m, [m.shape[1], m.shape[2], m.shape[3]]) : m;
 }
 const uh = /* @__PURE__ */ F({ separableConv2d_: lh });
@@ -1007,14 +1008,14 @@ const hh = /* @__PURE__ */ F({ sinh_: ch });
  * =============================================================================
  */
 function ph(s, t = 0, e = 1, n, i) {
-  if (ko(s), n != null && n === "bool")
+  if (No(s), n != null && n === "bool")
     throw new Error("Unsupported data type $ { dtype }");
-  const r = new yu(t, e, n, !0, i), a = xo(s, n);
+  const r = new Gu(t, e, n, !0, i), a = vo(s, n);
   for (let o = 0; o < a.values.length; o++)
     a.values[o] = r.nextValue();
   return a.toTensor();
 }
-const nr = /* @__PURE__ */ F({ truncatedNormal_: ph });
+const ir = /* @__PURE__ */ F({ truncatedNormal_: ph });
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.
@@ -1062,7 +1063,7 @@ function mh(s, t, e, n, i, r = "NHWC", a) {
   const u = r === "NHWC" ? o.shape[3] : o.shape[1], c = r === "NHWC" ? l.shape[3] : l.shape[1];
   k(u === e[2], () => `Error in conv2dDerFilter: depth of input ${u}) must match input depth in filter (${e[2]}.`), k(c === e[3], () => `Error in conv2dDerFilter: depth of dy (${c}) must match output depth for filter (${e[3]}).`), ft("conv2dDerFilter", i, a);
   const h = { x: o, dy: l }, p = { strides: n, pad: i, dataFormat: r, dimRoundingMode: a, filterShape: e };
-  return M.runKernel(No, h, p);
+  return M.runKernel(So, h, p);
 }
 const cs = /* @__PURE__ */ F({ conv2DBackpropFilter_: mh });
 /**
@@ -1082,10 +1083,10 @@ const cs = /* @__PURE__ */ F({ conv2DBackpropFilter_: mh });
  * =============================================================================
  */
 function gh({ x: s, filter: t, strides: e, pad: n, dataFormat: i = "NHWC", dilations: r = [1, 1], dimRoundingMode: a, bias: o, activation: l = "linear", preluActivationWeights: u, leakyreluAlpha: c }) {
-  if (l = l || "linear", wu(M.state.gradientDepth, l) === !1) {
+  if (l = l || "linear", ku(M.state.gradientDepth, l) === !1) {
     k(i === "NHWC", () => `Error in fused conv2d: got dataFormat of ${i} but only NHWC is currently supported for the case of gradient depth is 0 and the activation is not linear.`);
     let z = Ce(s, t, e, n, i, r, a);
-    return o != null && (z = $(z, o)), ku(z, l, u, c);
+    return o != null && (z = $(z, o)), xu(z, l, u, c);
   }
   const h = D(s, "x", "conv2d", "float32"), p = D(t, "filter", "conv2d", "float32");
   let f = h, g = !1;
@@ -1111,12 +1112,12 @@ function gh({ x: s, filter: t, strides: e, pad: n, dataFormat: i = "NHWC", dilat
   }
   const C = (z, _) => {
     k(i === "NHWC", () => `Error in gradient of fused conv2D: got dataFormat of ${i} but only NHWC is currently supported.`);
-    const [T, E, R, q] = _, bt = xu(z, R, l);
+    const [T, E, R, q] = _, bt = Nu(z, R, l);
     k(Se(r), () => `Error in gradient of fused conv2D: dilation rates greater than 1 are not yet supported in gradients. Got dilations '${r}'`);
     const ie = ls(E.shape, bt, T, e, n), re = cs(E, bt, T.shape, e, n), xt = [ie, re];
     if (q != null) {
-      const zt = Nu(q, bt);
-      xt.push(zt);
+      const Tt = vu(q, bt);
+      xt.push(Tt);
     }
     return xt;
   }, N = {
@@ -1167,7 +1168,7 @@ function yh(s, t, e, n, i, r = [1, 1], a) {
   let l = t;
   l.rank === 3 && (l = A(t, [1, t.shape[0], t.shape[1], t.shape[2]]));
   const u = { x: o, dy: l }, c = { strides: n, pad: i, dimRoundingMode: a, dilations: r, filterShape: e };
-  return M.runKernel(vo, u, c);
+  return M.runKernel(Ao, u, c);
 }
 const wh = F({ depthwiseConv2dNativeBackpropFilter_: yh });
 /**
@@ -1191,7 +1192,7 @@ function kh(s, t, e, n, i, r = [1, 1], a) {
   t.rank === 3 && (l = !0, o = A(t, [1, t.shape[0], t.shape[1], t.shape[2]]));
   const u = { dy: o, filter: e }, c = { strides: n, pad: i, dimRoundingMode: a, dilations: r, inputShape: s }, h = (
     // tslint:disable-next-line: no-unnecessary-type-assertion
-    M.runKernel(So, u, c)
+    M.runKernel(Co, u, c)
   );
   return l ? A(h, [h.shape[1], h.shape[2], h.shape[3]]) : h;
 }
@@ -1251,7 +1252,7 @@ class Nh {
    * @doc {heading: 'Training', subheading: 'Optimizers', namespace: 'train'}
    */
   static sgd(t) {
-    return new Ao(t);
+    return new Io(t);
   }
   /**
    * Constructs a `tf.MomentumOptimizer` that uses momentum gradient
@@ -1269,7 +1270,7 @@ class Nh {
    * @doc {heading: 'Training', subheading: 'Optimizers', namespace: 'train'}
    */
   static momentum(t, e, n = !1) {
-    return new Co(t, e, n);
+    return new Do(t, e, n);
   }
   /**
    * Constructs a `tf.RMSPropOptimizer` that uses RMSProp gradient
@@ -1292,7 +1293,7 @@ class Nh {
    * @doc {heading: 'Training', subheading: 'Optimizers', namespace: 'train'}
    */
   static rmsprop(t, e = 0.9, n = 0, i = null, r = !1) {
-    return new Io(t, e, n, i, r);
+    return new zo(t, e, n, i, r);
   }
   /**
    * Constructs a `tf.AdamOptimizer` that uses the Adam algorithm.
@@ -1307,7 +1308,7 @@ class Nh {
    * @doc {heading: 'Training', subheading: 'Optimizers', namespace: 'train'}
    */
   static adam(t = 1e-3, e = 0.9, n = 0.999, i = null) {
-    return new Do(t, e, n, i);
+    return new To(t, e, n, i);
   }
   /**
    * Constructs a `tf.AdadeltaOptimizer` that uses the Adadelta algorithm.
@@ -1322,7 +1323,7 @@ class Nh {
    * @doc {heading: 'Training', subheading: 'Optimizers', namespace: 'train'}
    */
   static adadelta(t = 1e-3, e = 0.95, n = null) {
-    return new zo(t, e, n);
+    return new $o(t, e, n);
   }
   /**
    * Constructs a `tf.AdamaxOptimizer` that uses the Adamax algorithm.
@@ -1338,7 +1339,7 @@ class Nh {
    * @doc {heading: 'Training', subheading: 'Optimizers', namespace: 'train'}
    */
   static adamax(t = 2e-3, e = 0.9, n = 0.999, i = null, r = 0) {
-    return new To(t, e, n, i, r);
+    return new Eo(t, e, n, i, r);
   }
   /**
    * Constructs a `tf.AdagradOptimizer` that uses the Adagrad algorithm.
@@ -1357,7 +1358,7 @@ class Nh {
    * @doc {heading: 'Training', subheading: 'Optimizers', namespace: 'train'}
    */
   static adagrad(t, e = 0.1) {
-    return new $o(t, e);
+    return new Lo(t, e);
   }
 }
 /**
@@ -1423,7 +1424,7 @@ const Ah = 1.7580993408473768, Ch = 1.0507009873554805;
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-class sr {
+class rr {
   constructor(t) {
     this.maxEntries = t || 100, this.cache = /* @__PURE__ */ new Map();
   }
@@ -1478,7 +1479,7 @@ class sr {
  * =============================================================================
  */
 let Ih = 0;
-function ir() {
+function ar() {
   return Ih++;
 }
 const Qe = {};
@@ -1518,13 +1519,13 @@ class kt extends Be {
     return {};
   }
 }
-class rr extends kt {
+class or extends kt {
   apply(t, e) {
     return mt(t, e);
   }
 }
-rr.className = "Zeros";
-S(rr);
+or.className = "Zeros";
+S(or);
 class hs extends kt {
   apply(t, e) {
     return pe(t, e);
@@ -1532,7 +1533,7 @@ class hs extends kt {
 }
 hs.className = "Ones";
 S(hs);
-class ar extends kt {
+class lr extends kt {
   constructor(t) {
     if (super(), typeof t != "object")
       throw new d(`Expected argument of type ConstantConfig but got ${t}`);
@@ -1549,21 +1550,21 @@ class ar extends kt {
     };
   }
 }
-ar.className = "Constant";
-S(ar);
-class or extends kt {
+lr.className = "Constant";
+S(lr);
+class ur extends kt {
   constructor(t) {
     super(), this.DEFAULT_MINVAL = -0.05, this.DEFAULT_MAXVAL = 0.05, this.minval = t.minval || this.DEFAULT_MINVAL, this.maxval = t.maxval || this.DEFAULT_MAXVAL, this.seed = t.seed;
   }
   apply(t, e) {
-    return bn(t, this.minval, this.maxval, e, this.seed);
+    return wn(t, this.minval, this.maxval, e, this.seed);
   }
   getConfig() {
     return { minval: this.minval, maxval: this.maxval, seed: this.seed };
   }
 }
-or.className = "RandomUniform";
-S(or);
+ur.className = "RandomUniform";
+S(ur);
 class ps extends kt {
   constructor(t) {
     super(), this.DEFAULT_MEAN = 0, this.DEFAULT_STDDEV = 0.05, this.mean = t.mean || this.DEFAULT_MEAN, this.stddev = t.stddev || this.DEFAULT_STDDEV, this.seed = t.seed;
@@ -1571,7 +1572,7 @@ class ps extends kt {
   apply(t, e) {
     if (e = e || "float32", e !== "float32" && e !== "int32")
       throw new G(`randomNormal does not support dType ${e}.`);
-    return yn(t, this.mean, this.stddev, e, this.seed);
+    return bn(t, this.mean, this.stddev, e, this.seed);
   }
   getConfig() {
     return { mean: this.mean, stddev: this.stddev, seed: this.seed };
@@ -1579,22 +1580,22 @@ class ps extends kt {
 }
 ps.className = "RandomNormal";
 S(ps);
-class lr extends kt {
+class cr extends kt {
   constructor(t) {
     super(), this.DEFAULT_MEAN = 0, this.DEFAULT_STDDEV = 0.05, this.mean = t.mean || this.DEFAULT_MEAN, this.stddev = t.stddev || this.DEFAULT_STDDEV, this.seed = t.seed;
   }
   apply(t, e) {
     if (e = e || "float32", e !== "float32" && e !== "int32")
       throw new G(`truncatedNormal does not support dType ${e}.`);
-    return nr(t, this.mean, this.stddev, e, this.seed);
+    return ir(t, this.mean, this.stddev, e, this.seed);
   }
   getConfig() {
     return { mean: this.mean, stddev: this.stddev, seed: this.seed };
   }
 }
-lr.className = "TruncatedNormal";
-S(lr);
-class ur extends kt {
+cr.className = "TruncatedNormal";
+S(cr);
+class hr extends kt {
   constructor(t) {
     super(), this.gain = t.gain != null ? t.gain : 1;
   }
@@ -1602,15 +1603,15 @@ class ur extends kt {
     return x(() => {
       if (t.length !== 2 || t[0] !== t[1])
         throw new d("Identity matrix initializer can only be used for 2D square matrices.");
-      return w(this.gain, vu(t[0]));
+      return w(this.gain, Su(t[0]));
     });
   }
   getConfig() {
     return { gain: this.gain };
   }
 }
-ur.className = "Identity";
-S(ur);
+hr.className = "Identity";
+S(hr);
 function Eh(s, t = "channelsLast") {
   let e, n;
   if (et(t), s.length === 2)
@@ -1646,10 +1647,10 @@ class dt extends kt {
       const o = Math.sqrt(a);
       if (e = e || "float32", e !== "float32" && e !== "int32")
         throw new G(`${this.getClassName()} does not support dType ${e}.`);
-      return nr(t, 0, o, e, this.seed);
+      return ir(t, 0, o, e, this.seed);
     } else {
       const o = Math.sqrt(3 * a);
-      return bn(t, -o, o, e, this.seed);
+      return wn(t, -o, o, e, this.seed);
     }
   }
   getConfig() {
@@ -1767,7 +1768,7 @@ class ys extends dt {
 }
 ys.className = "LeCunUniform";
 S(ys);
-class cr extends kt {
+class pr extends kt {
   constructor(t) {
     super(), this.DEFAULT_GAIN = 1, this.ELEMENTS_WARN_SLOW = 2e3, this.gain = t.gain == null ? this.DEFAULT_GAIN : t.gain, this.seed = t.seed;
   }
@@ -1780,7 +1781,7 @@ class cr extends kt {
       e = e;
       const n = Gi(t.slice(0, -1)), i = t[t.length - 1], r = n * i;
       r > this.ELEMENTS_WARN_SLOW && console.warn(`Orthogonal initializer is being called on a matrix with more than ${this.ELEMENTS_WARN_SLOW} (${r}) elements: Slowness may result.`);
-      const a = [Math.max(i, n), Math.min(i, n)], o = yn(a, 0, 1, e, this.seed), l = Su.qr(o, !1);
+      const a = [Math.max(i, n), Math.min(i, n)], o = bn(a, 0, 1, e, this.seed), l = Au.qr(o, !1);
       let u = l[0];
       const h = l[1].flatten().stridedSlice([0], [Math.min(i, n) * Math.min(i, n)], [Math.min(i, n) + 1]);
       return u = w(u, h.sign()), n < i && (u = u.transpose()), w(tt(this.gain), u.reshape(t));
@@ -1793,8 +1794,8 @@ class cr extends kt {
     };
   }
 }
-cr.className = "Orthogonal";
-S(cr);
+pr.className = "Orthogonal";
+S(pr);
 const Vs = {
   constant: "Constant",
   glorotNormal: "GlorotNormal",
@@ -1913,7 +1914,7 @@ class Lh {
    * @throws ValueError if `name` is `null` or `undefined`.
    */
   constructor(t, e = "float32", n = Ks, i = !0, r = null) {
-    this.dtype = e ?? "float32", this.shape = t.shape, this.id = ir(), n = n ?? Ks, this.originalName = Pi(n), this.name = Ui(this.originalName), this.trainable_ = i, this.constraint = r, this.val = qu(t, this.trainable_, this.name, this.dtype);
+    this.dtype = e ?? "float32", this.shape = t.shape, this.id = ar(), n = n ?? Ks, this.originalName = Pi(n), this.name = Ui(this.originalName), this.trainable_ = i, this.constraint = r, this.val = Ji(t, this.trainable_, this.name, this.dtype);
   }
   /**
    * Get a snapshot of the Variable's value.
@@ -1993,7 +1994,7 @@ class Mt {
    *   returned by apply().
    */
   constructor(t, e, n, i, r, a, o) {
-    this.dtype = t, this.shape = e, this.sourceLayer = n, this.inputs = i, this.callArgs = r, this.outputTensorIndex = o, this.id = ir(), a != null && (this.originalName = Pi(a), this.name = Ui(this.originalName)), this.rank = e.length;
+    this.dtype = t, this.shape = e, this.sourceLayer = n, this.inputs = i, this.callArgs = r, this.outputTensorIndex = o, this.id = ar(), a != null && (this.originalName = Pi(a), this.name = Ui(this.originalName)), this.rank = e.length;
   }
 }
 let Mh = 0;
@@ -2060,7 +2061,7 @@ class W extends Be {
    */
   getNodeAtIndex(t, e) {
     if (this.inboundNodes.length === 0)
-      throw new $t(`The layer has never been called and thus has no defined ${e}.`);
+      throw new Et(`The layer has never been called and thus has no defined ${e}.`);
     if (this.inboundNodes.length <= t)
       throw new d(`Asked to get ${e} at node ${t}, but the layer has only ${this.inboundNodes.length} inbound nodes.`);
     return this.inboundNodes[t];
@@ -2419,7 +2420,7 @@ class W extends Be {
    */
   countParams() {
     if (!this.built)
-      throw new $t(`You tried to call countParams() on ${this.name}, but the layer is not built yet. Build it first by calling build(batchInputShape).`);
+      throw new Et(`You tried to call countParams() on ${this.name}, but the layer is not built yet. Build it first by calling build(batchInputShape).`);
     return un(this.weights);
   }
   /**
@@ -2688,7 +2689,7 @@ function Rh(s) {
 function _h(s) {
   return "float32";
 }
-function hr(s, t, e) {
+function dr(s, t, e) {
   if ((t == null || e != null && e > 0) && (t = s.sourceLayer, e = s.nodeIndex), t.inboundNodes.length === 0)
     return [s];
   {
@@ -2698,7 +2699,7 @@ function hr(s, t, e) {
     {
       const i = [];
       for (let r = 0; r < n.inboundLayers.length; r++) {
-        const a = n.inputTensors[r], o = n.inboundLayers[r], l = n.nodeIndices[r], u = hr(a, o, l);
+        const a = n.inputTensors[r], o = n.inboundLayers[r], l = n.nodeIndices[r], u = dr(a, o, l);
         for (const c of u)
           i.indexOf(c) === -1 && i.push(c);
       }
@@ -2911,7 +2912,7 @@ class Vt {
     this.id2Mask != null && Z(this.id2Mask);
   }
 }
-const cn = new sr(), hn = new sr();
+const cn = new rr(), hn = new rr();
 function Uh(s) {
   cn?.setMaxEntries(s), hn?.setMaxEntries(s);
 }
@@ -3032,7 +3033,7 @@ function Kh(s) {
  * limitations under the License.
  * =============================================================================
  */
-const Hh = Eo();
+const Hh = Fo();
 Hh.registerFlag("TOPOLOGICAL_SORT_CACHE_MAX_ENTRIES", () => 100, Uh);
 /**
  * @license
@@ -3050,8 +3051,8 @@ Hh.registerFlag("TOPOLOGICAL_SORT_CACHE_MAX_ENTRIES", () => 100, Uh);
  * limitations under the License.
  * =============================================================================
  */
-const pr = {
-  kernelName: Lo,
+const fr = {
+  kernelName: Mo,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -3075,7 +3076,7 @@ const pr = {
  * =============================================================================
  */
 const qh = {
-  kernelName: Fo,
+  kernelName: Oo,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -3104,7 +3105,7 @@ const qh = {
  * =============================================================================
  */
 const Zh = {
-  kernelName: Mo,
+  kernelName: Ro,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -3133,7 +3134,7 @@ const Zh = {
  * =============================================================================
  */
 const Jh = {
-  kernelName: Oo,
+  kernelName: _o,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t, i = wt(e.shape, n.shape);
@@ -3165,7 +3166,7 @@ const Jh = {
  * =============================================================================
  */
 const Xh = {
-  kernelName: Ro,
+  kernelName: Bo,
   saveAllInputs: !0,
   gradFunc: (s, t) => {
     const e = {};
@@ -3215,7 +3216,7 @@ const Yh = {
  * =============================================================================
  */
 const Qh = {
-  kernelName: _o,
+  kernelName: Wo,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -3239,7 +3240,7 @@ const Qh = {
  * =============================================================================
  */
 const tp = {
-  kernelName: Bo,
+  kernelName: Go,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -3263,7 +3264,7 @@ const tp = {
  * =============================================================================
  */
 const ep = {
-  kernelName: Wo,
+  kernelName: Po,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -3292,7 +3293,7 @@ const ep = {
  * =============================================================================
  */
 const np = {
-  kernelName: Go,
+  kernelName: Uo,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t, i = wt(e.shape, n.shape);
@@ -3326,7 +3327,7 @@ const np = {
  * =============================================================================
  */
 const sp = {
-  kernelName: Po,
+  kernelName: Vo,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -3350,7 +3351,7 @@ const sp = {
  * =============================================================================
  */
 const ip = {
-  kernelName: Uo,
+  kernelName: jo,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -3383,7 +3384,7 @@ function rp(s, t, e, n, i, r) {
     o.shape[2],
     o.shape[3]
   ])), k(l.rank === 5, () => `Error in avgPool3dGrad: dy must be rank 5 but got rank ${l.rank}.`), k(u.rank === 5, () => `Error in avgPool3dGrad: input must be rank 5 but got rank ${u.rank}.`), ft("avgPool3dGrad", i, r);
-  const h = { dy: l, input: u }, p = { filterSize: e, strides: n, pad: i, dimRoundingMode: r }, f = M.runKernel(Vo, h, p);
+  const h = { dy: l, input: u }, p = { filterSize: e, strides: n, pad: i, dimRoundingMode: r }, f = M.runKernel(Ko, h, p);
   return c ? A(f, [f.shape[1], f.shape[2], f.shape[3], f.shape[4]]) : f;
 }
 const ap = /* @__PURE__ */ F({ avgPool3dGrad_: rp });
@@ -3434,7 +3435,7 @@ function lp(s, t, e, n, i) {
   k(a.rank === r.rank, () => `Rank of input (${a.rank}) does not match rank of dy (${r.rank})`);
   let o = a, l = r, u = !1;
   a.rank === 3 && (u = !0, o = A(a, [1, a.shape[0], a.shape[1], a.shape[2]]), l = A(r, [1, r.shape[0], r.shape[1], r.shape[2]])), k(l.rank === 4, () => `Error in avgPoolGrad: dy must be rank 4 but got rank ${l.rank}.`), k(o.rank === 4, () => `Error in avgPoolGrad: input must be rank 4 but got rank ${o.rank}.`);
-  const c = { dy: l, input: o }, h = { filterSize: e, strides: n, pad: i }, p = M.runKernel(jo, c, h);
+  const c = { dy: l, input: o }, h = { filterSize: e, strides: n, pad: i }, p = M.runKernel(Ho, c, h);
   return u ? A(p, [p.shape[1], p.shape[2], p.shape[3]]) : p;
 }
 const up = /* @__PURE__ */ F({ avgPoolGrad_: lp });
@@ -3479,7 +3480,7 @@ const cp = {
  * =============================================================================
  */
 const hp = {
-  kernelName: Ko,
+  kernelName: qo,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t, e) => {
     const [n, i] = t, { transposeA: r, transposeB: a } = e;
@@ -3538,7 +3539,7 @@ const pp = {
  * =============================================================================
  */
 const dp = {
-  kernelName: Ho,
+  kernelName: Zo,
   gradFunc: (s, t, e) => {
     const n = e, i = n.inputShape, r = n.shape, a = Array.from(r);
     for (let l = i.length - 1; l >= 0; l--)
@@ -3574,7 +3575,7 @@ const dp = {
  * =============================================================================
  */
 const fp = {
-  kernelName: qo,
+  kernelName: Jo,
   gradFunc: (s) => ({ x: () => s.clone() })
 };
 /**
@@ -3594,7 +3595,7 @@ const fp = {
  * =============================================================================
  */
 const mp = {
-  kernelName: Zo,
+  kernelName: Xo,
   gradFunc: (s) => ({ x: () => Q(s) })
 };
 /**
@@ -3614,7 +3615,7 @@ const mp = {
  * =============================================================================
  */
 const gp = {
-  kernelName: Jo,
+  kernelName: Yo,
   inputsToSave: ["x"],
   gradFunc: (s, t, e) => {
     const [n] = t, { clipValueMin: i, clipValueMax: r } = e;
@@ -3640,9 +3641,9 @@ const gp = {
  * =============================================================================
  */
 const bp = {
-  kernelName: Xo,
+  kernelName: Qo,
   inputsToSave: ["x"],
-  gradFunc: pr.gradFunc
+  gradFunc: fr.gradFunc
 };
 /**
  * @license
@@ -3661,7 +3662,7 @@ const bp = {
  * =============================================================================
  */
 const yp = {
-  kernelName: Yo,
+  kernelName: tl,
   saveAllInputs: !0,
   gradFunc: (s, t, e) => {
     const n = t.map((l) => l.shape), { axis: i } = e, r = he(i, t[0].shape)[0], a = n.map((l) => l[r]);
@@ -3744,7 +3745,7 @@ function xp(s, t, e, n, i) {
   let a = t;
   a.rank === 4 && (a = A(t, [1, t.shape[0], t.shape[1], t.shape[2], t.shape[3]])), k(r.rank === 5, () => `Error in conv3dDerFilter: input must be rank 5, but got shape ${r.shape}.`), k(a.rank === 5, () => `Error in conv3dDerFilter: dy must be rank 5, but got shape ${a.shape}.`), k(e.length === 5, () => `Error in conv3dDerFilter: filterShape must be length 5, but got ${e}.`), k(r.shape[4] === e[3], () => `Error in conv3dDerFilter: depth of input ${r.shape[4]}) must match input depth in filter (${e[3]}.`), k(a.shape[4] === e[4], () => `Error in conv3dDerFilter: depth of dy (${a.shape[4]}) must match output depth for filter (${e[4]}).`);
   const o = { x: r, dy: a }, l = { strides: n, pad: i, filterShape: e };
-  return M.runKernel(Qo, o, l);
+  return M.runKernel(el, o, l);
 }
 const Np = /* @__PURE__ */ F({ conv3DBackpropFilter_: xp });
 /**
@@ -3771,7 +3772,7 @@ const vp = {
     k(Se(n), () => `Error in gradient of conv3D: dilation rates greater than 1 are not yet supported in gradients. Got dilations '${n}'`);
     const [a, o] = t;
     return {
-      x: () => Qi(a.shape, s, o, i, r),
+      x: () => er(a.shape, s, o, i, r),
       filter: () => Np(a, s, o.shape, i, r)
     };
   }
@@ -3793,11 +3794,11 @@ const vp = {
  * =============================================================================
  */
 const Sp = {
-  kernelName: tl,
+  kernelName: nl,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
-    return { x: () => w(pt(Wu(L(e, "float32"))), s) };
+    return { x: () => w(pt(Pu(L(e, "float32"))), s) };
   }
 };
 /**
@@ -3847,7 +3848,7 @@ const Cp = {
     const [n] = t, { axis: i, exclusive: r, reverse: a } = e;
     return {
       x: () => {
-        const o = Xi([i], n.rank);
+        const o = Yi([i], n.rank);
         let l = Lc(s, i, r, !a);
         return o != null && (l = j(l, o)), l;
       }
@@ -3900,13 +3901,13 @@ const Ip = {
  * =============================================================================
  */
 const Dp = {
-  kernelName: el,
+  kernelName: sl,
   inputsToSave: ["x", "filter"],
   gradFunc: (s, t, e) => {
     const [n, i] = t, r = { x: n, filter: i, dy: s }, a = { x: n, filter: i, dy: s };
     return {
-      x: () => M.runKernel(sl, r, e),
-      filter: () => M.runKernel(nl, a, e)
+      x: () => M.runKernel(rl, r, e),
+      filter: () => M.runKernel(il, a, e)
     };
   }
 };
@@ -3927,11 +3928,11 @@ const Dp = {
  * =============================================================================
  */
 const zp = {
-  kernelName: il,
+  kernelName: al,
   outputsToSave: [!0],
   gradFunc: (s, t) => {
     const [e] = t, n = { dy: s, y: e };
-    return { x: () => M.runKernel(rl, n) };
+    return { x: () => M.runKernel(ol, n) };
   }
 };
 /**
@@ -3975,7 +3976,7 @@ const Tp = {
  * =============================================================================
  */
 const $p = {
-  kernelName: al,
+  kernelName: ll,
   outputsToSave: [!0],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -3999,7 +4000,7 @@ const $p = {
  * =============================================================================
  */
 const Ep = {
-  kernelName: ol,
+  kernelName: ul,
   inputsToSave: ["input"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -4023,7 +4024,7 @@ const Ep = {
  * =============================================================================
  */
 const Lp = {
-  kernelName: ll,
+  kernelName: cl,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -4047,7 +4048,7 @@ const Lp = {
  * =============================================================================
  */
 const Fp = {
-  kernelName: ul,
+  kernelName: hl,
   gradFunc: (s) => ({ x: () => Q(s) })
 };
 /**
@@ -4067,7 +4068,7 @@ const Fp = {
  * =============================================================================
  */
 const Mp = {
-  kernelName: cl,
+  kernelName: pl,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t, i = wt(e.shape, n.shape);
@@ -4149,7 +4150,7 @@ const Op = {
  * =============================================================================
  */
 const Rp = {
-  kernelName: hl,
+  kernelName: dl,
   inputsToSave: ["x", "indices"],
   gradFunc: (s, t, e) => {
     const [n, i] = t, { axis: r, batchDims: a } = e, o = he(r, n.shape)[0], l = (u, c, h) => () => {
@@ -4199,7 +4200,7 @@ function Zs(s) {
  * =============================================================================
  */
 const _p = {
-  kernelName: pl,
+  kernelName: fl,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t;
@@ -4223,7 +4224,7 @@ const _p = {
  * =============================================================================
  */
 const Bp = {
-  kernelName: dl,
+  kernelName: ml,
   gradFunc: (s) => ({ x: () => L(s, "float32") })
 };
 /**
@@ -4243,7 +4244,7 @@ const Bp = {
  * =============================================================================
  */
 const Wp = {
-  kernelName: fl,
+  kernelName: gl,
   gradFunc: (s) => ({ x: () => Q(s) })
 };
 /**
@@ -4263,7 +4264,7 @@ const Wp = {
  * =============================================================================
  */
 const Gp = {
-  kernelName: ml,
+  kernelName: bl,
   gradFunc: (s) => ({ x: () => Q(s) })
 };
 /**
@@ -4283,7 +4284,7 @@ const Gp = {
  * =============================================================================
  */
 const Pp = {
-  kernelName: gl,
+  kernelName: yl,
   gradFunc: (s) => ({ x: () => Q(s) })
 };
 /**
@@ -4303,7 +4304,7 @@ const Pp = {
  * =============================================================================
  */
 const Up = {
-  kernelName: bl,
+  kernelName: wl,
   inputsToSave: ["x"],
   gradFunc: (s, t, e) => {
     const [n] = t, { alpha: i } = e, r = Gt(n, 0);
@@ -4351,7 +4352,7 @@ const Vp = {
  * =============================================================================
  */
 const jp = {
-  kernelName: yl,
+  kernelName: kl,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -4375,7 +4376,7 @@ const jp = {
  * =============================================================================
  */
 const Kp = {
-  kernelName: wl,
+  kernelName: xl,
   inputsToSave: [],
   outputsToSave: [!0],
   gradFunc: (s, t, e) => {
@@ -4406,7 +4407,7 @@ const Kp = {
  */
 function Hp(s, t, e, n = 5, i = 1, r = 1, a = 0.5) {
   const o = { x: s, y: t, dy: e }, l = { depthRadius: n, bias: i, alpha: r, beta: a };
-  return M.runKernel(kl, o, l);
+  return M.runKernel(Nl, o, l);
 }
 const qp = F({ localResponseNormalizationBackprop_: Hp });
 /**
@@ -4426,7 +4427,7 @@ const qp = F({ localResponseNormalizationBackprop_: Hp });
  * =============================================================================
  */
 const Zp = {
-  kernelName: xl,
+  kernelName: vl,
   inputsToSave: ["x"],
   outputsToSave: [!0],
   gradFunc: (s, t, e) => {
@@ -4452,7 +4453,7 @@ const Zp = {
  * limitations under the License.
  * =============================================================================
  */
-function dr(s, t, e, n) {
+function mr(s, t, e, n) {
   return t.rank < e.rank && (t = A(t, Ws(t.shape, n))), s.rank < e.rank && (s = A(s, Ws(s.shape, n))), {
     x: () => w(s, L(Xt(e, t), s.dtype))
   };
@@ -4474,11 +4475,11 @@ function dr(s, t, e, n) {
  * =============================================================================
  */
 const Js = {
-  kernelName: Nl,
+  kernelName: Sl,
   inputsToSave: ["x"],
   outputsToSave: [!0],
   gradFunc: (s, t, e) => {
-    const n = e, { reductionIndices: i } = n, r = t[0], a = t[1], o = he(i, r.shape), l = dr(s, a, r, o);
+    const n = e, { reductionIndices: i } = n, r = t[0], a = t[1], o = he(i, r.shape), l = mr(s, a, r, o);
     return {
       x: () => l.x()
     };
@@ -4501,11 +4502,11 @@ const Js = {
  * =============================================================================
  */
 const Jp = {
-  kernelName: vl,
+  kernelName: Al,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t;
-    return { a: () => w(s, L(Ue(e, n), "float32")), b: () => w(s, L(Au(e, n), "float32")) };
+    return { a: () => w(s, L(Ue(e, n), "float32")), b: () => w(s, L(Cu(e, n), "float32")) };
   }
 };
 /**
@@ -4540,7 +4541,7 @@ function Xp(s, t, e, n, i, r, a) {
     u.shape[2],
     u.shape[3]
   ])), k(c.rank === 5, () => `Error in maxPool3dGrad: dy must be rank 5 but got rank ${c.rank}.`), k(h.rank === 5, () => `Error in maxPool3dGrad: input must be rank 5 but got rank ${h.rank}.`), k(p.rank === 5, () => `Error in maxPool3dGrad: output must be rank 5 but got rank ${p.rank}.`), ft("maxPool3dGrad", r, a);
-  const g = { dy: c, input: h, output: p }, b = { filterSize: n, strides: i, pad: r, dimRoundingMode: a }, m = M.runKernel(Sl, g, b);
+  const g = { dy: c, input: h, output: p }, b = { filterSize: n, strides: i, pad: r, dimRoundingMode: a }, m = M.runKernel(Cl, g, b);
   return f ? A(m, [m.shape[1], m.shape[2], m.shape[3], m.shape[4]]) : m;
 }
 const Yp = /* @__PURE__ */ F({ maxPool3dGrad_: Xp });
@@ -4591,7 +4592,7 @@ function td(s, t, e, n, i, r, a) {
   const o = D(s, "dy", "maxPoolGrad"), l = D(t, "input", "maxPoolGrad"), u = D(e, "output", "maxPoolGrad");
   k(l.rank === o.rank, () => `Rank of input (${l.rank}) does not match rank of dy (${o.rank})`), k(o.rank === 4, () => `Error in maxPoolGrad: dy must be rank 4 but got rank ${o.rank}.`), k(l.rank === 4, () => `Error in maxPoolGrad: input must be rank 4 but got rank ${l.rank}.`), ft("maxPoolGrad", r, a);
   const c = { dy: o, input: l, output: u }, h = { filterSize: n, strides: i, pad: r, dimRoundingMode: a };
-  return M.runKernel(Al, c, h);
+  return M.runKernel(Il, c, h);
 }
 const ed = /* @__PURE__ */ F({ maxPoolGrad_: td });
 /**
@@ -4638,10 +4639,10 @@ const nd = {
  * =============================================================================
  */
 const sd = {
-  kernelName: Cl,
+  kernelName: Dl,
   inputsToSave: ["x"],
   gradFunc: (s, t, e) => {
-    const [n] = t, { axis: i } = e, r = he(i, n.shape), o = Gu(n.shape, r)[1], l = Gi(o);
+    const [n] = t, { axis: i } = e, r = he(i, n.shape), o = Uu(n.shape, r)[1], l = Gi(o);
     return { x: () => {
       const c = n.shape.slice();
       r.forEach((f) => {
@@ -4669,11 +4670,11 @@ const sd = {
  * =============================================================================
  */
 const id = {
-  kernelName: Il,
+  kernelName: zl,
   inputsToSave: ["x"],
   outputsToSave: [!0],
   gradFunc: (s, t, e) => {
-    const n = e, { axis: i } = n, [r, a] = t, o = he(i, r.shape), l = dr(s, a, r, o);
+    const n = e, { axis: i } = n, [r, a] = t, o = he(i, r.shape), l = mr(s, a, r, o);
     return {
       x: () => l.x()
     };
@@ -4696,7 +4697,7 @@ const id = {
  * =============================================================================
  */
 const rd = {
-  kernelName: Dl,
+  kernelName: Tl,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t;
@@ -4720,7 +4721,7 @@ const rd = {
  * =============================================================================
  */
 const ad = {
-  kernelName: zl,
+  kernelName: $l,
   inputsToSave: ["x"],
   gradFunc: (s, t, e) => {
     const n = t[0], { paddings: i } = e, r = i.map((a) => a[0]);
@@ -4744,7 +4745,7 @@ const ad = {
  * =============================================================================
  */
 const od = {
-  kernelName: Tl,
+  kernelName: El,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t, i = wt(e.shape, n.shape);
@@ -4752,7 +4753,7 @@ const od = {
       const o = lt(e.shape, i);
       return o.length > 0 ? A(B(s, o), e.shape) : s;
     }, b: () => {
-      const o = w(s, pt(Vi(P(e, n)))), l = lt(n.shape, i);
+      const o = w(s, pt(Zi(P(e, n)))), l = lt(n.shape, i);
       return l.length > 0 ? A(B(o, l), n.shape) : o;
     } };
   }
@@ -4774,7 +4775,7 @@ const od = {
  * =============================================================================
  */
 const ld = {
-  kernelName: $l,
+  kernelName: Ll,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t, i = wt(e.shape, n.shape);
@@ -4804,7 +4805,7 @@ const ld = {
  * =============================================================================
  */
 const ud = {
-  kernelName: El,
+  kernelName: Fl,
   gradFunc: (s) => ({ x: () => pt(s) })
 };
 /**
@@ -4868,7 +4869,7 @@ const hd = {
  * =============================================================================
  */
 const pd = {
-  kernelName: Ll,
+  kernelName: Ml,
   saveAllInputs: !0,
   gradFunc: (s, t, e) => {
     const { axis: n } = e;
@@ -4916,7 +4917,7 @@ const Xs = {
  * =============================================================================
  */
 const dd = {
-  kernelName: Fl,
+  kernelName: Ol,
   inputsToSave: ["a", "b"],
   outputsToSave: [!0],
   gradFunc: (s, t) => {
@@ -4951,7 +4952,7 @@ const dd = {
  * =============================================================================
  */
 const fd = {
-  kernelName: Ml,
+  kernelName: Rl,
   inputsToSave: ["x", "alpha"],
   gradFunc: (s, t) => {
     const [e, n] = t, i = Gt(e, 0);
@@ -4988,7 +4989,7 @@ function md(s, t, e) {
   return w(i, o);
 }
 function gd(s, t, e) {
-  const n = s.shape.length, i = n - e.length, r = Xi(e, n);
+  const n = s.shape.length, i = n - e.length, r = Yi(e, n);
   let a = s;
   r != null && (a = j(s, r));
   const o = a.shape.slice(), u = o.splice(n - e.length, e.length).reduce((p, f) => p * f, 1);
@@ -5002,7 +5003,7 @@ function gd(s, t, e) {
   return h;
 }
 const bd = {
-  kernelName: Ol,
+  kernelName: _l,
   inputsToSave: ["x"],
   gradFunc: (s, t, e) => {
     const [n] = t, { axis: i } = e;
@@ -5027,7 +5028,7 @@ const bd = {
  * =============================================================================
  */
 const yd = {
-  kernelName: Rl,
+  kernelName: Bl,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t, i = wt(e.shape, n.shape);
@@ -5060,7 +5061,7 @@ const yd = {
  * =============================================================================
  */
 const wd = {
-  kernelName: _l,
+  kernelName: Wl,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -5084,7 +5085,7 @@ const wd = {
  * =============================================================================
  */
 const kd = {
-  kernelName: Bl,
+  kernelName: Gl,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t, n = w(Yn(e, 6), Xn(e));
@@ -5108,7 +5109,7 @@ const kd = {
  * =============================================================================
  */
 const xd = {
-  kernelName: Wl,
+  kernelName: Pl,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -5132,7 +5133,7 @@ const xd = {
  * =============================================================================
  */
 const Nd = {
-  kernelName: Gl,
+  kernelName: Ul,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -5156,13 +5157,13 @@ const Nd = {
  * =============================================================================
  */
 const vd = {
-  kernelName: Pl,
+  kernelName: Vl,
   inputsToSave: ["images"],
   gradFunc: (s, t, e) => {
     const [n] = t, i = { dy: s, images: n };
     return { images: () => (
       // tslint:disable-next-line: no-unnecessary-type-assertion
-      M.runKernel(Ul, i, e)
+      M.runKernel(jl, i, e)
     ) };
   }
 };
@@ -5183,13 +5184,13 @@ const vd = {
  * =============================================================================
  */
 const Sd = {
-  kernelName: Vl,
+  kernelName: Kl,
   inputsToSave: ["images"],
   gradFunc: (s, t, e) => {
     const [n] = t, i = { dy: s, images: n };
     return { images: () => (
       // tslint:disable-next-line: no-unnecessary-type-assertion
-      M.runKernel(jl, i, e)
+      M.runKernel(Hl, i, e)
     ) };
   }
 };
@@ -5233,7 +5234,7 @@ const Ad = {
  * =============================================================================
  */
 const Cd = {
-  kernelName: Kl,
+  kernelName: ql,
   gradFunc: (s) => ({ x: () => Q(s) })
 };
 /**
@@ -5277,7 +5278,7 @@ const Id = {
  * =============================================================================
  */
 const Dd = {
-  kernelName: Hl,
+  kernelName: Zl,
   inputsToSave: ["condition"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -5336,7 +5337,7 @@ const zd = {
  * =============================================================================
  */
 const Td = {
-  kernelName: ql,
+  kernelName: Jl,
   outputsToSave: [!0],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -5360,7 +5361,7 @@ const Td = {
  * =============================================================================
  */
 const $d = {
-  kernelName: Zl,
+  kernelName: Xl,
   gradFunc: (s) => ({ x: () => Q(s) })
 };
 /**
@@ -5380,11 +5381,11 @@ const $d = {
  * =============================================================================
  */
 const Ed = {
-  kernelName: Jl,
+  kernelName: Yl,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
-    return { x: () => w(Ji(L(e, "float32")), s) };
+    return { x: () => w(Xi(L(e, "float32")), s) };
   }
 };
 /**
@@ -5428,13 +5429,13 @@ const Ld = {
  * =============================================================================
  */
 const Fd = {
-  kernelName: Xl,
+  kernelName: Ql,
   inputsToSave: ["x"],
   gradFunc: (s, t, e) => {
-    const [n] = t, { begin: i, size: r } = e, a = n.shape, [o, l] = Uu(n, i, r), u = [];
+    const [n] = t, { begin: i, size: r } = e, a = n.shape, [o, l] = ju(n, i, r), u = [];
     for (let c = 0; c < s.rank; c++)
       u.push([o[c], a[c] - o[c] - l[c]]);
-    return { x: () => er(s, u) };
+    return { x: () => sr(s, u) };
   }
 };
 /**
@@ -5454,7 +5455,7 @@ const Fd = {
  * =============================================================================
  */
 const Md = {
-  kernelName: Yl,
+  kernelName: tu,
   outputsToSave: [!0],
   gradFunc: (s, t, e) => {
     const [n] = t, { dim: i } = e, r = !0, a = w(s, n);
@@ -5527,7 +5528,7 @@ const Ys = {
  * =============================================================================
  */
 const Qs = {
-  kernelName: Ql,
+  kernelName: eu,
   gradFunc: (s, t, e) => {
     const { axis: n } = e;
     return { x: () => is(s, n) };
@@ -5550,7 +5551,7 @@ const Qs = {
  * =============================================================================
  */
 const Rd = {
-  kernelName: tu,
+  kernelName: nu,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -5574,7 +5575,7 @@ const Rd = {
  * =============================================================================
  */
 const _d = {
-  kernelName: eu,
+  kernelName: su,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
@@ -5598,7 +5599,7 @@ const _d = {
  * =============================================================================
  */
 const Bd = {
-  kernelName: nu,
+  kernelName: iu,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t, i = tt(2);
@@ -5622,7 +5623,7 @@ const Bd = {
  * =============================================================================
  */
 const Wd = {
-  kernelName: su,
+  kernelName: ru,
   gradFunc: (s) => ({ x: () => Q(s) })
 };
 /**
@@ -5642,7 +5643,7 @@ const Wd = {
  * =============================================================================
  */
 const Gd = {
-  kernelName: iu,
+  kernelName: au,
   inputsToSave: ["a", "b"],
   gradFunc: (s, t) => {
     const [e, n] = t, i = wt(e.shape, n.shape);
@@ -5674,7 +5675,7 @@ const Gd = {
  * =============================================================================
  */
 const Pd = {
-  kernelName: ru,
+  kernelName: ou,
   inputsToSave: ["x"],
   gradFunc: (s, t, e) => {
     const [n] = t, i = n.shape.slice(), { axis: r } = e;
@@ -5702,11 +5703,11 @@ const Pd = {
  * =============================================================================
  */
 const Ud = {
-  kernelName: au,
+  kernelName: lu,
   inputsToSave: ["x"],
   gradFunc: (s, t) => {
     const [e] = t;
-    return { x: () => P(s, ct(Ji(e))) };
+    return { x: () => P(s, ct(Xi(e))) };
   }
 };
 /**
@@ -5750,7 +5751,7 @@ const Vd = {
  * =============================================================================
  */
 const jd = {
-  kernelName: ou,
+  kernelName: uu,
   inputsToSave: ["x"],
   gradFunc: (s, t, e) => {
     const [n] = t, { reps: i } = e;
@@ -5805,7 +5806,7 @@ const jd = {
  * =============================================================================
  */
 const Kd = {
-  kernelName: lu,
+  kernelName: cu,
   gradFunc: (s, t, e) => {
     const n = e, { perm: i } = n, r = ss(i);
     return { x: () => j(s, r) };
@@ -5828,7 +5829,7 @@ const Kd = {
  * =============================================================================
  */
 const Hd = {
-  kernelName: uu,
+  kernelName: hu,
   gradFunc: (s, t, e) => {
     const n = e, { axis: i } = n;
     return { value: () => kn(s, i) };
@@ -5859,7 +5860,7 @@ const qd = {
   }
 };
 function Zd(s, t) {
-  const e = Ie(t, Q(t)), n = Vu(s, e);
+  const e = Ie(t, Q(t)), n = Qi(s, e);
   let i = Ue(t, tt(0, "int32"));
   const r = n.rank - i.rank;
   for (let o = 0; o < r; ++o)
@@ -5885,7 +5886,7 @@ function Zd(s, t) {
  * =============================================================================
  */
 const Jd = {
-  kernelName: cu,
+  kernelName: pu,
   gradFunc: (s) => ({ x: () => Q(s) })
 };
 /**
@@ -5905,7 +5906,7 @@ const Jd = {
  * =============================================================================
  */
 const Xd = [
-  pr,
+  fr,
   qh,
   Zh,
   Jh,
@@ -6012,7 +6013,7 @@ const Xd = [
   Jd
 ];
 for (const s of Xd)
-  hu(s);
+  du(s);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -6030,13 +6031,13 @@ class qe extends Be {
     return {};
   }
 }
-class fr extends qe {
+class gr extends qe {
   constructor(t) {
     super(), this.defaultMaxValue = 2, this.defaultAxis = 0, this.maxValue = t.maxValue != null ? t.maxValue : this.defaultMaxValue, this.axis = t.axis != null ? t.axis : this.defaultAxis;
   }
   apply(t) {
     return x(() => {
-      const e = ks(t, this.axis), n = At(e, 0, this.maxValue);
+      const e = ks(t, this.axis), n = Ct(e, 0, this.maxValue);
       return w(t, P(n, $(nt(), e)));
     });
   }
@@ -6044,9 +6045,9 @@ class fr extends qe {
     return { maxValue: this.maxValue, axis: this.axis };
   }
 }
-fr.className = "MaxNorm";
-S(fr);
-class mr extends qe {
+gr.className = "MaxNorm";
+S(gr);
+class br extends qe {
   constructor(t) {
     super(), this.defaultAxis = 0, this.axis = t.axis != null ? t.axis : this.defaultAxis;
   }
@@ -6057,22 +6058,22 @@ class mr extends qe {
     return { axis: this.axis };
   }
 }
-mr.className = "UnitNorm";
-S(mr);
-class gr extends qe {
+br.className = "UnitNorm";
+S(br);
+class yr extends qe {
   apply(t) {
     return Ve(t);
   }
 }
-gr.className = "NonNeg";
-S(gr);
-class br extends qe {
+yr.className = "NonNeg";
+S(yr);
+class wr extends qe {
   constructor(t) {
     super(), this.defaultMinValue = 0, this.defaultMaxValue = 1, this.defaultRate = 1, this.defaultAxis = 0, this.minValue = t.minValue != null ? t.minValue : this.defaultMinValue, this.maxValue = t.maxValue != null ? t.maxValue : this.defaultMaxValue, this.rate = t.rate != null ? t.rate : this.defaultRate, this.axis = t.axis != null ? t.axis : this.defaultAxis;
   }
   apply(t) {
     return x(() => {
-      const e = ks(t, this.axis), n = $(w(this.rate, At(e, this.minValue, this.maxValue)), w(1 - this.rate, e));
+      const e = ks(t, this.axis), n = $(w(this.rate, Ct(e, this.minValue, this.maxValue)), w(1 - this.rate, e));
       return w(t, P(n, $(nt(), e)));
     });
   }
@@ -6085,8 +6086,8 @@ class br extends qe {
     };
   }
 }
-br.className = "MinMaxNorm";
-S(br);
+wr.className = "MinMaxNorm";
+S(wr);
 const ti = {
   maxNorm: "MaxNorm",
   minMaxNorm: "MinMaxNorm",
@@ -6116,7 +6117,7 @@ function rt(s) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function Tm(s) {
+function Yd(s) {
   return new ps(s);
 }
 /**
@@ -6146,7 +6147,7 @@ async function ae(s) {
     Z(n);
   }
 }
-function yr(s) {
+function kr(s) {
   if (s != null)
     for (const t in s) {
       const e = s[t];
@@ -6166,7 +6167,7 @@ var ni;
 (function(s) {
   s[s.SILENT = 0] = "SILENT", s[s.VERBOSE = 1] = "VERBOSE";
 })(ni || (ni = {}));
-const Yd = 125;
+const Qd = 125;
 class Me {
   constructor() {
     this.validationData = null;
@@ -6196,7 +6197,7 @@ class Me {
   setModel(t) {
   }
 }
-class Qd {
+class tf {
   // TODO(cais): When the need arises, uncomment the following lines and
   // implement the queue for time values.
   // private deltaTBatch: number;
@@ -6281,7 +6282,7 @@ class Qd {
       await e.onTrainEnd(t);
   }
 }
-class tf extends Me {
+class ef extends Me {
   constructor() {
     super();
   }
@@ -6313,7 +6314,7 @@ class tf extends Me {
         }));
   }
 }
-class ef extends Me {
+class nf extends Me {
   async onTrainBegin(t) {
     this.epoch = [], this.history = {};
   }
@@ -6340,11 +6341,11 @@ class ef extends Me {
       this.history[e[r]][n[r]].dispose(), this.history[e[r]][n[r]] = i[r][0];
   }
 }
-class nf extends Me {
+class sf extends Me {
   constructor(t, e) {
-    if (super(), this.currentEpoch = 0, this.nowFunc = t.nowFunc, this.nextFrameFunc = t.nextFrameFunc || Sh, this.yieldEvery = e || "auto", this.yieldEvery === "auto" && (this.yieldEvery = Yd), this.yieldEvery === "never" && t.onYield != null)
+    if (super(), this.currentEpoch = 0, this.nowFunc = t.nowFunc, this.nextFrameFunc = t.nextFrameFunc || Sh, this.yieldEvery = e || "auto", this.yieldEvery === "auto" && (this.yieldEvery = Qd), this.yieldEvery === "never" && t.onYield != null)
       throw new Error("yieldEvery is `never` but you provided an `onYield` callback. Either change `yieldEvery` or remove the callback");
-    Fs(this.yieldEvery) && (this.maybeWait = Cu(this.maybeWait.bind(this), this.yieldEvery, this.nowFunc)), this.trainBegin = t.onTrainBegin, this.trainEnd = t.onTrainEnd, this.epochBegin = t.onEpochBegin, this.epochEnd = t.onEpochEnd, this.batchBegin = t.onBatchBegin, this.batchEnd = t.onBatchEnd, this.yield = t.onYield;
+    Fs(this.yieldEvery) && (this.maybeWait = Iu(this.maybeWait.bind(this), this.yieldEvery, this.nowFunc)), this.trainBegin = t.onTrainBegin, this.trainEnd = t.onTrainEnd, this.epochBegin = t.onEpochBegin, this.epochEnd = t.onEpochEnd, this.batchBegin = t.onBatchBegin, this.batchEnd = t.onBatchEnd, this.yield = t.onYield;
   }
   async maybeWait(t, e, n) {
     const i = [];
@@ -6371,8 +6372,8 @@ class nf extends Me {
     this.trainEnd != null && (await ae(t), await this.trainEnd(t));
   }
 }
-function wr(s, t) {
-  return s == null && (s = {}), s instanceof Me ? [s] : Array.isArray(s) && s[0] instanceof Me ? s : K(s).map((n) => new nf(n, t));
+function xr(s, t) {
+  return s == null && (s = {}), s instanceof Me ? [s] : Array.isArray(s) && s[0] instanceof Me ? s : K(s).map((n) => new sf(n, t));
 }
 class yt {
   /**
@@ -6426,13 +6427,13 @@ class yt {
   }
 }
 yt.constructors = {};
-function kr(s, t, e, n, i, r, a, o, l) {
-  const u = new ef(), c = [
-    new tf(),
+function Nr(s, t, e, n, i, r, a, o, l) {
+  const u = new nf(), c = [
+    new ef(),
     ...yt.createCallbacks(t)
   ];
   s != null && c.push(...s), c.push(u);
-  const h = new Qd(c);
+  const h = new tf(c);
   return h.setParams({
     epochs: e,
     initialEpoch: n,
@@ -6468,7 +6469,7 @@ function Wt(s, t = {}, e = !1) {
 function pn(s, t) {
   return x(() => {
     s.dtype !== "float32" && (s = L(s, "float32"));
-    const e = B(je(s), t, !0), n = pu(e.shape, nt()), i = ee(Ie(e, n));
+    const e = B(je(s), t, !0), n = fu(e.shape, nt()), i = ee(Ie(e, n));
     return P(s, i);
   });
 }
@@ -6480,35 +6481,35 @@ function xs(s, t) {
 }
 function Ns(s, t) {
   return x(() => {
-    const e = V(s, t), n = At(Fe(s), nt(), Number.MAX_VALUE), i = Fe(P(e, n));
+    const e = V(s, t), n = Ct(Fe(s), nt(), Number.MAX_VALUE), i = Fe(P(e, n));
     return w(100, at(i, -1));
   });
 }
-function sf(s, t) {
+function rf(s, t) {
   return x(() => {
-    const e = At(t, nt(), Number.MAX_VALUE), n = Zt($(1, e)), i = At(s, nt(), Number.MAX_VALUE), r = Zt($(1, i));
+    const e = Ct(t, nt(), Number.MAX_VALUE), n = Zt($(1, e)), i = Ct(s, nt(), Number.MAX_VALUE), r = Zt($(1, i));
     return at(je(V(n, r)), -1);
   });
 }
-function rf(s, t) {
+function af(s, t) {
   return x(() => {
     const e = Ie(0, V(1, w(s, t)));
     return at(je(e), -1);
   });
 }
-function af(s, t) {
+function of(s, t) {
   return x(() => {
     const e = Ie(0, V(1, w(s, t)));
     return at(e, -1);
   });
 }
-function of(s, t) {
+function lf(s, t) {
   return x(() => {
     const e = B(w(s, t), -1), n = ve(w(V(1, s), t), -1);
     return Ie(0, $(1, V(n, e)));
   });
 }
-function lf(s, t) {
+function uf(s, t) {
   return x(() => {
     const e = Math.log(2), n = V(t, s), i = V($(n, us(w(-2, n))), e);
     return at(i, -1);
@@ -6517,23 +6518,23 @@ function lf(s, t) {
 function Oe(s, t, e = !1) {
   return x(() => {
     if (e)
-      t = Yi(t);
+      t = tr(t);
     else {
       const n = B(t, t.shape.length - 1, !0);
       t = P(t, n);
     }
-    return t = At(t, nt(), 1 - nt()), pt(B(w(L(s, "float32"), Zt(t)), t.shape.length - 1));
+    return t = Ct(t, nt(), 1 - nt()), pt(B(w(L(s, "float32"), Zt(t)), t.shape.length - 1));
   });
 }
 function dn(s, t, e = !1) {
   return x(() => {
-    const n = L(Vi(Iu(s)), "int32");
-    t = At(t, nt(), 1 - nt());
+    const n = L(Zi(Du(s)), "int32");
+    t = Ct(t, nt(), 1 - nt());
     const i = t.shape, r = A(Yc(n, i[i.length - 1]), i);
     return Oe(r, t, e);
   });
 }
-function uf(s, t) {
+function cf(s, t) {
   if (!Ft(s.shape, t.shape))
     throw new d(`logits and labels must have the same shape, but got shapes ${JSON.stringify(s.shape)} and ${JSON.stringify(t.shape)}`);
   return x(() => {
@@ -6544,22 +6545,22 @@ function uf(s, t) {
 function Sn(s, t) {
   return x(() => {
     let e;
-    return e = At(t, nt(), 1 - nt()), e = Zt(P(e, V(1, e))), at(uf(s, e), -1);
+    return e = Ct(t, nt(), 1 - nt()), e = Zt(P(e, V(1, e))), at(cf(s, e), -1);
   });
 }
-function cf(s, t) {
+function hf(s, t) {
   return x(() => {
-    const e = At(s, nt(), 1), n = At(t, nt(), 1);
+    const e = Ct(s, nt(), 1), n = Ct(t, nt(), 1);
     return B(w(s, Zt(P(e, n))), -1);
   });
 }
-function hf(s, t) {
+function pf(s, t) {
   return x(() => {
     const e = Zt($(nt(), t));
     return at(V(t, w(s, e)), -1);
   });
 }
-function xr(s, t) {
+function vr(s, t) {
   return x(() => {
     const e = pn(s, -1), n = pn(t, -1), i = w(e, n);
     return pt(B(i, -1));
@@ -6569,17 +6570,17 @@ const fn = {
   meanSquaredError: vn,
   meanAbsoluteError: xs,
   meanAbsolutePercentageError: Ns,
-  meanSquaredLogarithmicError: sf,
-  squaredHinge: rf,
-  hinge: af,
-  categoricalHinge: of,
-  logcosh: lf,
+  meanSquaredLogarithmicError: rf,
+  squaredHinge: af,
+  hinge: of,
+  categoricalHinge: lf,
+  logcosh: uf,
   categoricalCrossentropy: Oe,
   sparseCategoricalCrossentropy: dn,
   binaryCrossentropy: Sn,
-  kullbackLeiblerDivergence: cf,
-  poisson: hf,
-  cosineProximity: xr
+  kullbackLeiblerDivergence: hf,
+  poisson: pf,
+  cosineProximity: vr
 };
 function $n(s) {
   if (typeof s == "string") {
@@ -6599,48 +6600,48 @@ function $n(s) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function Nr(s, t) {
+function Sr(s, t) {
   return x(() => {
-    const e = w(0.5, It(t)), n = Et(Gt(t, e), s.dtype);
+    const e = w(0.5, Dt(t)), n = Lt(Gt(t, e), s.dtype);
     return at(Xt(s, n), -1);
   });
 }
-function vr(s, t) {
-  return x(() => Et(Xt(sn(s, -1), sn(t, -1)), "float32"));
+function Ar(s, t) {
+  return x(() => Lt(Xt(sn(s, -1), sn(t, -1)), "float32"));
 }
-function pf(s, t) {
+function df(s, t) {
   return x(() => L(B(Pe(Xt(s, 1), Xt(t, 1))), "float32"));
 }
-function df(s, t) {
+function ff(s, t) {
   return x(() => L(B(Pe(Xt(s, 0), Xt(t, 1))), "float32"));
 }
-function ff(s, t) {
+function mf(s, t) {
   return x(() => {
-    const e = pf(s, t), n = df(s, t), i = $(e, n);
+    const e = df(s, t), n = ff(s, t), i = $(e, n);
     return L(Ht(Gt(i, 0), P(e, i), 0), "float32");
   });
 }
-function mf(s, t) {
+function gf(s, t) {
   return Sn(s, t);
 }
-function gf(s, t) {
+function bf(s, t) {
   return s.rank === t.rank && (s = ts(s, [s.rank - 1])), t = sn(t, -1), t.dtype !== s.dtype && (t = L(t, s.dtype)), L(Xt(s, t), "float32");
 }
-const bf = vn, yf = vn, wf = xs, kf = xs, xf = Ns, Nf = Ns, Sr = Oe, vf = xr, Ar = dn, mn = {
-  binaryAccuracy: Nr,
-  categoricalAccuracy: vr,
-  precision: ff,
-  categoricalCrossentropy: Sr,
-  sparseCategoricalCrossentropy: Ar,
-  mse: bf,
-  MSE: yf,
-  mae: wf,
-  MAE: kf,
-  mape: xf,
-  MAPE: Nf,
-  cosine: vf
+const yf = vn, wf = vn, kf = xs, xf = xs, Nf = Ns, vf = Ns, Cr = Oe, Sf = vr, Ir = dn, mn = {
+  binaryAccuracy: Sr,
+  categoricalAccuracy: Ar,
+  precision: mf,
+  categoricalCrossentropy: Cr,
+  sparseCategoricalCrossentropy: Ir,
+  mse: yf,
+  MSE: wf,
+  mae: kf,
+  MAE: xf,
+  mape: Nf,
+  MAPE: vf,
+  cosine: Sf
 };
-function Sf(s) {
+function Af(s) {
   if (typeof s == "string" && s in mn)
     return mn[s];
   if (typeof s != "string" && s != null)
@@ -6676,7 +6677,7 @@ function tn(s) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function Af(s) {
+function Cf(s) {
   const t = {
     Adagrad: () => ge.adagrad(0.01),
     Adadelta: () => ge.adadelta(1, 0.95, nt()),
@@ -6738,8 +6739,8 @@ function Pn(s) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function Cf(s, t, e, n = console.log) {
-  const i = Df(s), r = ["Layer (type)", "Input Shape", "Output shape", "Param #"];
+function If(s, t, e, n = console.log) {
+  const i = zf(s), r = ["Layer (type)", "Input Shape", "Output shape", "Param #"];
   i ? (t = t || 90, e = e || [0.32, 0.61, 0.89, 1]) : (t = t || 115, e = e || [0.24, 0.48, 0.7, 0.8, 1]), e[e.length - 1] <= 1 && (e = e.map((c) => Math.floor(t * c)));
   let a;
   if (!i) {
@@ -6750,16 +6751,16 @@ function Cf(s, t, e, n = console.log) {
   n("_".repeat(t)), gn(r, e, n), n("=".repeat(t));
   const o = s.layers;
   for (let c = 0; c < o.length; ++c)
-    i ? zf(o[c], e, n) : Tf(o[c], e, a, n), n((c === o.length - 1 ? "=" : "_").repeat(t));
+    i ? Tf(o[c], e, n) : $f(o[c], e, a, n), n((c === o.length - 1 ? "=" : "_").repeat(t));
   s.checkTrainableWeightsConsistency();
-  const l = If(s), u = un(s.nonTrainableWeights);
+  const l = Df(s), u = un(s.nonTrainableWeights);
   n(`Total params: ${l + u}`), n(`Trainable params: ${l}`), n(`Non-trainable params: ${u}`), n("_".repeat(t));
 }
-function If(s) {
+function Df(s) {
   let t;
   return s.collectedTrainableWeights != null ? t = un(s.collectedTrainableWeights) : t = un(s.trainableWeights), t;
 }
-function Df(s) {
+function zf(s) {
   let t = !0;
   const e = [], n = [];
   for (const i in s.nodesByDepth)
@@ -6792,7 +6793,7 @@ function gn(s, t, e = console.log) {
     i > 0 && (n = n.slice(0, n.length - 1) + " "), n += s[i], n = n.slice(0, t[i]), n += " ".repeat(t[i] - n.length);
   e(n);
 }
-function zf(s, t, e) {
+function Tf(s, t, e) {
   let n, i;
   try {
     i = s.inboundNodes.map((l) => JSON.stringify(l.inputShapes)).join(",");
@@ -6812,7 +6813,7 @@ function zf(s, t, e) {
   ];
   gn(o, t, e);
 }
-function Tf(s, t, e, n) {
+function $f(s, t, e, n) {
   let i, r;
   try {
     r = s.inboundNodes.map((h) => JSON.stringify(h.inputShapes)).join(",");
@@ -6851,7 +6852,7 @@ function Tf(s, t, e, n) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function Cr(s, t, e) {
+function Dr(s, t, e) {
   return (s === "inboundNodes" || s === "outputLayers" || s === "inputLayers") && t === 0 && typeof e == "string";
 }
 function Un(s, t) {
@@ -6865,7 +6866,7 @@ function Un(s, t) {
     const e = [], n = s.length;
     for (let i = 0; i < n; ++i) {
       const r = s[i];
-      Cr(t, i, r) ? e.push(r) : e.push(Un(r, t));
+      Dr(t, i, r) ? e.push(r) : e.push(Un(r, t));
     }
     return e;
   } else {
@@ -6893,7 +6894,7 @@ function Vn(s, t) {
     const e = [], n = s.length;
     for (let i = 0; i < n; ++i) {
       const r = s[i];
-      Cr(t, i, r) ? e.push(r) : e.push(Vn(r, t));
+      Dr(t, i, r) ? e.push(r) : e.push(Vn(r, t));
     }
     return e;
   } else {
@@ -6906,7 +6907,7 @@ function Vn(s, t) {
   }
 }
 /** @license See the LICENSE file. */
-const Ir = "4.22.0";
+const zr = "4.22.0";
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -6916,7 +6917,7 @@ const Ir = "4.22.0";
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-const $f = (s) => {
+const Ef = (s) => {
   const t = Object.keys(s);
   if (t.length === 0)
     return !1;
@@ -6954,7 +6955,7 @@ class vt extends W {
       (I == null || z == null || _ == null) && (I = y.sourceLayer, z = y.nodeIndex, _ = y.tensorIndex);
       const T = I.inboundNodes[z];
       if (N.indexOf(T) !== -1)
-        throw new $t(`The tensor ${y.name} at layer "${I.name}" is part of a cycle.`);
+        throw new Et(`The tensor ${y.name} at layer "${I.name}" is part of a cycle.`);
       if (C.indexOf(T) !== -1)
         return;
       this.containerNodes.add(vt.nodeKey(I, z)), I.id in a || (a[I.id] = Object.keys(a).length), N.indexOf(T) === -1 && N.push(T);
@@ -7009,7 +7010,7 @@ class vt extends W {
         if (N != null) {
           for (const I of C.inputTensors)
             if (b.indexOf(I) === -1)
-              throw new $t(`Graph disconnected: cannot obtain value for tensor ${I} at layer "${N.name}". The following previous layers were accessed without issue: ${m}`);
+              throw new Et(`Graph disconnected: cannot obtain value for tensor ${I} at layer "${N.name}". The following previous layers were accessed without issue: ${m}`);
           for (const I of C.outputTensors)
             b.push(I);
           m.push(N.name);
@@ -7020,7 +7021,7 @@ class vt extends W {
     for (const y of v) {
       const C = v.filter((N) => N === y).length;
       if (C !== 1)
-        throw new $t(`The name "${y}" is used ${C} times in the model. All layer names should be unique. Layer names: ` + JSON.stringify(v));
+        throw new Et(`The name "${y}" is used ${C} times in the model. All layer names should be unique. Layer names: ` + JSON.stringify(v));
     }
     this.outboundNodes = [], this.inboundNodes = [], new Nn({
       outboundLayer: this,
@@ -7127,7 +7128,7 @@ class vt extends W {
   loadWeights(t, e = !0) {
     const n = {};
     let i = 0;
-    const r = $f(t);
+    const r = Ef(t);
     r && this.parseWeights(t);
     for (const o of this.layers)
       for (const [l, u] of o.weights.entries()) {
@@ -7170,7 +7171,7 @@ class vt extends W {
    */
   updatedConfig() {
     const t = this.getConfig(), e = {};
-    return e.className = this.getClassName(), e.config = t, e.kerasVersion = `tfjs-layers ${Ir}`, e.backend = "TensorFlow.js", e;
+    return e.className = this.getClassName(), e.config = t, e.kerasVersion = `tfjs-layers ${zr}`, e.backend = "TensorFlow.js", e;
   }
   /**
    * Returns a JSON string containing the network configuration.
@@ -7476,7 +7477,7 @@ class vt extends W {
     const c = e.name, h = e.layers;
     for (const m of h)
       u(m);
-    for (; !Du(a); )
+    for (; !zu(a); )
       for (const m of h) {
         const v = r[m.name];
         if (v.name in a) {
@@ -7539,7 +7540,7 @@ class vt extends W {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function Ef(s, t, e) {
+function Lf(s, t, e) {
   const n = t.length;
   if (s == null || Array.isArray(s) && s.length === 0)
     return t.map((i) => null);
@@ -7557,14 +7558,14 @@ function Ef(s, t, e) {
   } else
     throw new Error(`The model has multiple (${n}) outputs, so ${e} must be either an array with ${n} elements or an object with ${t} keys. Provided ${e} not understood: ${JSON.stringify(s)}`);
 }
-function Dr(s, t) {
-  return Ef(s, t, "classWeight");
+function Tr(s, t) {
+  return Lf(s, t, "classWeight");
 }
-async function zr(s, t, e, n) {
+async function $r(s, t, e, n) {
   if (e != null) {
     const i = x(() => {
       if (s.shape.length === 1)
-        return du(s);
+        return mu(s);
       if (s.shape.length === 2) {
         if (s.shape[1] > 1)
           return sn(s, 1);
@@ -7584,7 +7585,7 @@ async function zr(s, t, e, n) {
   } else
     return null;
 }
-function Lf(s, t) {
+function Ff(s, t) {
   return w(s, t);
 }
 /**
@@ -7596,8 +7597,8 @@ function Lf(s, t) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-const Ff = 32;
-function Tr(s, t) {
+const Mf = 32;
+function Er(s, t) {
   let e, n;
   const i = t;
   e = i.xs, n = i.ys, k(e != null && n != null, () => `A Dataset iterator for fitDataset() is expected to generate objects of the form \`{xs: xVal, ys: yVal}\`, where the two values may be \`tf.Tensor\`, an array of Tensors, or a map of string to Tensor.  The provided Dataset instead generates ${t}`);
@@ -7624,12 +7625,12 @@ function ri(s, t, e) {
     return n;
   }
 }
-function Mf(s) {
+function Of(s) {
   if (s.length === 3)
     throw new G("Validation with sample weights is not implemented yet.");
   return { xs: s[0], ys: s[1] };
 }
-async function Of(s, t, e) {
+async function Rf(s, t, e) {
   const n = e.batchesPerEpoch != null;
   if (k(s.optimizer != null, () => "You must compile a model before training/testing. Use LayersModel.compile(modelCompileConfig)."), k(e != null, () => "For fitDataset(), the 2nd argument (config) is required, but it is not provided in this call."), k(e.epochs != null && e.epochs > 0 && Number.isInteger(e.epochs), () => `For fitDataset(), config.epochs is expected to be a positive integer, but got ${e.epochs}`), k(!n || e.batchesPerEpoch > 0 && Number.isInteger(e.batchesPerEpoch), () => `For fitDataset(), config.batchesPerEpoch is expected to be a positive integer if specified, but got ${e.batchesPerEpoch}`), k(
     // tslint:disable-next-line:no-any
@@ -7645,19 +7646,19 @@ async function Of(s, t, e) {
       if (ai(e.validationData))
         k(e.validationBatches == null || e.validationBatches > 0 && Number.isInteger(e.validationBatches), () => `For fitDataset() with dataset-based validation, config.validationBatches is expected not to be provided, or to be a positive integer, but got ${e.validationBatches}`);
       else {
-        const m = Mf(e.validationData);
+        const m = Of(e.validationData);
         r = m.xs, a = m.ys;
       }
     const o = s.makeTrainFunction(), l = s.getDedupedMetricsNames();
     let u;
     i ? u = l.slice().concat(l.map((m) => "val_" + m)) : u = l.slice();
-    const c = wr(e.callbacks, e.yieldEvery), h = e.verbose == null ? 1 : e.verbose, { callbackList: p, history: f } = kr(
+    const c = xr(e.callbacks, e.yieldEvery), h = e.verbose == null ? 1 : e.verbose, { callbackList: p, history: f } = Nr(
       c,
       h,
       e.epochs,
       null,
       null,
-      Rf(t, e),
+      _f(t, e),
       null,
       // Batch size determined by the dataset itself.
       i,
@@ -7676,13 +7677,13 @@ async function Of(s, t, e) {
           break;
         }
         if (C.value != null) {
-          const { xs: N, ys: I } = Tr(s, C.value), z = {};
+          const { xs: N, ys: I } = Er(s, C.value), z = {};
           z.batch = y, z.size = N[0].shape[0], await p.onBatchBegin(y, z);
           const _ = [];
           if (e.classWeight != null) {
-            const R = Dr(e.classWeight, s.outputNames);
+            const R = Tr(e.classWeight, s.outputNames);
             for (let q = 0; q < R.length; ++q)
-              _.push(await zr(I[q], null, R[q]));
+              _.push(await $r(I[q], null, R[q]));
           }
           const T = N.concat(I).concat(_), E = o(T);
           Z(T);
@@ -7690,13 +7691,13 @@ async function Of(s, t, e) {
             const q = l[R], bt = E[R];
             z[q] = bt, Bt(bt);
           }
-          await p.onBatchEnd(y, z), yr(z), y++, v++;
+          await p.onBatchEnd(y, z), kr(z), y++, v++;
         }
         if (n ? v >= e.batchesPerEpoch : C.done) {
           if (i) {
             let N;
             ai(e.validationData) ? N = K(await s.evaluateDataset(e.validationData, { batches: e.validationBatches })) : N = K(s.evaluate(r, a, {
-              batchSize: e.validationBatchSize == null ? Ff : e.validationBatchSize,
+              batchSize: e.validationBatchSize == null ? Mf : e.validationBatchSize,
               verbose: 0
             }));
             for (let I = 0; I < s.metricsNames.length; ++I)
@@ -7715,30 +7716,30 @@ async function Of(s, t, e) {
     s.isTraining = !1;
   }
 }
-function Rf(s, t) {
+function _f(s, t) {
   let e = null;
   return t.batchesPerEpoch != null ? e = t.batchesPerEpoch : Number.isFinite(s.size) && (e = s.size), e;
 }
 function ai(s) {
   return typeof s.iterator == "function";
 }
-function _f(s) {
+function Bf(s) {
   return typeof s.next == "function";
 }
-async function Bf(s, t, e) {
+async function Wf(s, t, e) {
   e = e || {};
   const n = e.batches != null, i = s.testFunction;
   let r = [];
   if (e.verbose > 0)
     throw new G("Verbose mode is not implemented yet.");
   k(!n || e.batches > 0 && Number.isInteger(e.batches), () => `Test loop expects \`batches\` to be a positive integer, but received ${JSON.stringify(e.batches)}`);
-  const a = _f(t) ? t : await t.iterator();
+  const a = Bf(t) ? t : await t.iterator();
   let o = 0, l = 0;
   for (; !n || l < e.batches; ) {
     const u = await a.next();
     if (r = x(() => {
       if (u.value) {
-        const { xs: c, ys: h } = Tr(s, u.value), p = c.concat(h), f = x(() => i(p));
+        const { xs: c, ys: h } = Er(s, u.value), p = c.concat(h), f = x(() => i(p));
         if (Z(p), l === 0)
           for (let b = 0; b < f.length; ++b)
             r.push(tt(0));
@@ -7777,7 +7778,7 @@ function Te(s, t, e) {
   return s == null ? [null] : Array.isArray(s) ? s.map((n) => nn(n, t, e - t)) : nn(s, t, e - t);
 }
 function jn(s, t) {
-  return x(() => s == null ? null : Array.isArray(s) ? s.map((e) => jn(e, t)) : ji(s, t.dtype === "int32" ? t : L(t, "int32")));
+  return x(() => s == null ? null : Array.isArray(s) ? s.map((e) => jn(e, t)) : Vi(s, t.dtype === "int32" ? t : L(t, "int32")));
 }
 function Ln(s, t) {
   const e = [];
@@ -7786,13 +7787,13 @@ function Ln(s, t) {
     i = n + t, i >= s && (i = s), e.push([n, i]), n = i;
   return e;
 }
-function $r(s) {
+function Lr(s) {
   const t = [];
   s instanceof xe && (s = [s]);
   for (let e = 0; e < s.length; ++e) {
     const n = s[e];
     if (n.rank === 1)
-      t.push(wn(n, 1));
+      t.push(yn(n, 1));
     else {
       if (n.rank === 0)
         throw new Error("Expected tensor to be at least 1D, but received a 0D tensor (scalar).");
@@ -7839,14 +7840,14 @@ function Nt(s, t) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function Wf(s) {
+function Gf(s) {
   return s instanceof xe;
 }
 function Kn(s) {
   return Array.isArray(s);
 }
 function oi(s) {
-  return !Wf(s) && !Kn(s);
+  return !Gf(s) && !Kn(s);
 }
 function li(s, t, e, n = !0, i = "") {
   if (t == null || t.length === 0) {
@@ -7886,7 +7887,7 @@ function li(s, t, e, n = !0, i = "") {
       throw new d(`The model ${i} expects ${t.length} Tensor(s), but only received one Tensor. Found: Tensor with shape ${s.shape}`);
     r = [s];
   }
-  if (r = $r(r), e != null)
+  if (r = Lr(r), e != null)
     for (let a = 0; a < t.length; ++a) {
       if (e[a] == null)
         continue;
@@ -7903,7 +7904,7 @@ function li(s, t, e, n = !0, i = "") {
     }
   return r;
 }
-function Gf(s, t, e) {
+function Pf(s, t, e) {
   const n = jt(s.map((r) => r.shape[0]));
   n.sort();
   const i = jt(t.map((r) => r.shape[0]));
@@ -7914,7 +7915,7 @@ function Gf(s, t, e) {
   if (n.length > 0 && i.length > 0 && !Ft(n, i))
     throw new d(`Input Tensors should have the same number of samples as target Tensors. Found ${n[0]} input sample(s) and ${i[0]} target sample(s).`);
 }
-function Pf(s, t, e) {
+function Uf(s, t, e) {
   const n = [
     vn,
     Sn,
@@ -7963,7 +7964,7 @@ function ui(s, t, e, n = !0, i = "") {
       }
     }
 }
-function Uf(s, t) {
+function Vf(s, t) {
   if (s == null || Array.isArray(s) && s.length === 0)
     return t.map((n) => []);
   let e;
@@ -7984,7 +7985,7 @@ function Uf(s, t) {
     return n;
   }
 }
-const Vf = "layers-model";
+const jf = "layers-model";
 class we extends vt {
   constructor(t) {
     super(t), this.isTraining = !1;
@@ -8027,7 +8028,7 @@ class we extends vt {
   summary(t, e, n = console.log) {
     if (!this.built)
       throw new d("This model has never been called, thus its weights have not been created yet. So no summary can be displayed. Build the model first (e.g., by calling it on some test data).");
-    Cf(this, t, e, n);
+    If(this, t, e, n);
   }
   /**
    * Configures and prepares the model for training and evaluation.  Compiling
@@ -8041,9 +8042,9 @@ class we extends vt {
    */
   compile(t) {
     if (t.loss == null && (t.loss = []), this.loss = t.loss, typeof t.optimizer == "string")
-      this.optimizer_ = Af(t.optimizer), this.isOptimizerOwned = !0;
+      this.optimizer_ = Cf(t.optimizer), this.isOptimizerOwned = !0;
     else {
-      if (!(t.optimizer instanceof fu))
+      if (!(t.optimizer instanceof gu))
         throw new d("User-defined optimizer must be an instance of tf.Optimizer.");
       this.optimizer_ = t.optimizer, this.isOptimizerOwned = !1;
     }
@@ -8079,7 +8080,7 @@ class we extends vt {
         this.outputs.length > 1 && (this.metricsTensors.push([o, a]), this.metricsNames.push(this.outputNames[a] + "_loss"));
       }
     });
-    const i = Uf(t.metrics, this.outputNames), r = (a, o, l) => {
+    const i = Vf(t.metrics, this.outputNames), r = (a, o, l) => {
       this.outputNames.length > 1 && (o = this.outputNames[a] + "_" + o), this.metricsNames.push(o), this.metricsTensors.push([l, a]);
     };
     le("metric", () => {
@@ -8092,11 +8093,11 @@ class we extends vt {
           for (const g of u) {
             if (typeof g == "string" && ["accuracy", "acc", "crossentropy", "ce"].indexOf(g) !== -1) {
               const m = this.internalOutputShapes[a];
-              m[m.length - 1] === 1 || this.lossFunctions[a] === Sn ? ["accuracy", "acc"].indexOf(g) !== -1 ? p = Nr : ["crossentropy", "ce"].indexOf(g) !== -1 && (p = mf) : this.lossFunctions[a] === dn ? ["accuracy", "acc"].indexOf(g) !== -1 ? p = gf : ["crossentropy", "ce"].indexOf(g) !== -1 && (p = Ar) : ["accuracy", "acc"].indexOf(g) !== -1 ? p = vr : ["crossentropy", "ce"].indexOf(g) !== -1 && (p = Sr);
+              m[m.length - 1] === 1 || this.lossFunctions[a] === Sn ? ["accuracy", "acc"].indexOf(g) !== -1 ? p = Sr : ["crossentropy", "ce"].indexOf(g) !== -1 && (p = gf) : this.lossFunctions[a] === dn ? ["accuracy", "acc"].indexOf(g) !== -1 ? p = bf : ["crossentropy", "ce"].indexOf(g) !== -1 && (p = Ir) : ["accuracy", "acc"].indexOf(g) !== -1 ? p = Ar : ["crossentropy", "ce"].indexOf(g) !== -1 && (p = Cr);
               let v;
               ["accuracy", "acc"].indexOf(g) !== -1 ? v = "acc" : ["crossentropy", "ce"].indexOf(g) !== -1 && (v = "ce"), f = p, h = "" + v;
             } else
-              f = Sf(g), h = "" + tn(g);
+              f = Af(g), h = "" + tn(g);
             let b;
             le(h, () => {
               b = f;
@@ -8185,7 +8186,7 @@ class we extends vt {
    * @doc {heading: 'Models', subheading: 'Classes'}
    */
   async evaluateDataset(t, e) {
-    return this.makeTestFunction(), Bf(this, t, e);
+    return this.makeTestFunction(), Wf(this, t, e);
   }
   /**
    * Get number of samples provided for training, evaluation or prediction.
@@ -8319,7 +8320,7 @@ class we extends vt {
    * @doc {heading: 'Models', subheading: 'Classes'}
    */
   predict(t, e = {}) {
-    const n = $r(t);
+    const n = Lr(t);
     ui(n, this.inputNames, this.feedInputShapes, !1);
     try {
       const i = e.batchSize == null ? 32 : e.batchSize;
@@ -8350,13 +8351,13 @@ class we extends vt {
   }
   standardizeUserDataXY(t, e, n = !0, i) {
     if (this.optimizer_ == null)
-      throw new $t("You must compile a model before training/testing. Use LayersModel.compile(modelCompileArgs).");
+      throw new Et("You must compile a model before training/testing. Use LayersModel.compile(modelCompileArgs).");
     const r = [];
     for (let a = 0; a < this.feedOutputShapes.length; ++a) {
       const o = this.feedOutputShapes[a];
       this.feedLossFns[a] === dn ? r.push(o.slice(0, o.length - 1).concat([1])) : r.push(o);
     }
-    if (t = li(t, this.feedInputNames, this.feedInputShapes, !1, "input"), e = li(e, this.feedOutputNames, r, !1, "target"), Gf(t, e), Pf(e, this.feedLossFns, this.feedOutputShapes), this.stateful && i != null && i > 0 && t[0].shape[0] % i !== 0)
+    if (t = li(t, this.feedInputNames, this.feedInputShapes, !1, "input"), e = li(e, this.feedOutputNames, r, !1, "target"), Pf(t, e), Uf(e, this.feedLossFns, this.feedOutputShapes), this.stateful && i != null && i > 0 && t[0].shape[0] % i !== 0)
       throw new d(`In a stateful network, you should only pass inputs with a number of samples that is divisible by the batch size ${i}. Found: ${t[0].shape[0]} sample(s).`);
     return [t, e];
   }
@@ -8366,10 +8367,10 @@ class we extends vt {
       throw new Error("sample weight is not supported yet.");
     let u = null;
     if (i != null) {
-      const c = Dr(i, this.outputNames);
+      const c = Tr(i, this.outputNames);
       u = [];
       for (let h = 0; h < c.length; ++h)
-        u.push(await zr(l[h], null, c[h]));
+        u.push(await $r(l[h], null, c[h]));
     }
     return [o, l, u];
   }
@@ -8392,7 +8393,7 @@ class we extends vt {
       if (r != null)
         throw new G("steps mode in testLoop() is not implemented yet");
       {
-        const l = Ln(a, n), u = Rn(Ct(0, a));
+        const l = Ln(a, n), u = Rn(It(0, a));
         for (let c = 0; c < l.length; ++c) {
           const h = l[c][0], p = l[c][1], f = nn(u, h, p - h), g = jn(e, f), b = t(g);
           if (c === 0)
@@ -8443,7 +8444,7 @@ class we extends vt {
         for (let b = 0; b < this.lossFunctions.length; ++b) {
           const m = this.lossFunctions[b];
           let v = m(i[b], f[b]);
-          r[b] != null && (v = Lf(v, r[b]));
+          r[b] != null && (v = Ff(v, r[b]));
           const y = at(v);
           e.push(y), b === 0 ? g = v : g = $(g, v);
         }
@@ -8557,7 +8558,7 @@ class we extends vt {
       const C = this.makeTrainFunction(), N = this.getDedupedMetricsNames();
       let I, z;
       m ? (this.makeTestFunction(), I = this.testFunction, z = N.slice().concat(N.map((E) => "val_" + E))) : (I = null, v = [], z = N.slice());
-      const _ = wr(n.callbacks, n.yieldEvery);
+      const _ = xr(n.callbacks, n.yieldEvery);
       return await this.fitLoop(C, y, N, f, n.epochs, n.verbose, _, I, v, n.shuffle, z, n.initialEpoch, null, null);
     } finally {
       this.isTraining = !1, Nt(i, t), Nt(r, e), Nt(a, t), Nt(o, e), Nt(c, l), Nt(h, u), p != null && Z(p);
@@ -8597,8 +8598,8 @@ class we extends vt {
       throw new d("Can only use `validationSteps` when doing step-wise training, i.e., `stepsPerEpoch` must be set.");
     const m = this.checkNumSamples(e, i, f, "steps_per_epoch");
     let v;
-    m != null && (v = Ct(0, m)), a == null && (a = 1);
-    const { callbackList: y, history: C } = kr(o, a, r, p, m, f, i, b, h);
+    m != null && (v = It(0, m)), a == null && (a = 1);
+    const { callbackList: y, history: C } = Nr(o, a, r, p, m, f, i, b, h);
     y.setModel(this), this.history = C, await y.onTrainBegin(), this.stopTraining_ = !1;
     for (let N = p; N < r; ++N) {
       await y.onEpochBegin(N);
@@ -8608,7 +8609,7 @@ class we extends vt {
       {
         if (c === "batch")
           throw new G("batch shuffling is not implemneted yet");
-        c && mu(v);
+        c && bu(v);
         const z = Rn(v), _ = Ln(m, i);
         for (let T = 0; T < _.length; ++T) {
           const E = {};
@@ -8617,17 +8618,17 @@ class we extends vt {
             E.batch = T, E.size = q - R;
             const ie = jn(e, bt), re = t(ie);
             for (let xt = 0; xt < n.length; ++xt) {
-              const zt = n[xt], me = re[xt];
-              E[zt] = me, Bt(me);
+              const Tt = n[xt], me = re[xt];
+              E[Tt] = me, Bt(me);
             }
             if (T === _.length - 1 && b) {
               const xt = this.testLoop(l, u, i);
-              for (let zt = 0; zt < n.length; ++zt) {
-                const me = n[zt], ze = xt[zt];
+              for (let Tt = 0; Tt < n.length; ++Tt) {
+                const me = n[Tt], ze = xt[Tt];
                 Bt(ze), I["val_" + me] = ze;
               }
             }
-          }), await y.onBatchEnd(T, E), yr(E), this.stopTraining_)
+          }), await y.onBatchEnd(T, E), kr(E), this.stopTraining_)
             break;
         }
         z.dispose();
@@ -8661,7 +8662,7 @@ class we extends vt {
    * @doc {heading: 'Models', subheading: 'Classes'}
    */
   async fitDataset(t, e) {
-    return Of(this, t, e);
+    return Rf(this, t, e);
   }
   /**
    * Runs a single gradient update on a single batch of data.
@@ -8913,7 +8914,7 @@ class we extends vt {
    */
   async save(t, e) {
     if (typeof t == "string") {
-      const u = gu(t);
+      const u = yu(t);
       if (u.length === 0)
         throw new d(`Cannot find any save handlers for URL '${t}'`);
       if (u.length > 1)
@@ -8924,14 +8925,14 @@ class we extends vt {
       throw new d("LayersModel.save() cannot proceed because the IOHandler provided does not have the `save` attribute defined.");
     const n = await Os(this.getNamedWeights(e)), o = {
       modelTopology: this.toJSON(null, !1),
-      format: Vf,
-      generatedBy: `TensorFlow.js tfjs-layers v${Ir}`,
+      format: jf,
+      generatedBy: `TensorFlow.js tfjs-layers v${zr}`,
       convertedBy: null
     };
     if ((e == null ? !1 : e.includeOptimizer) && this.optimizer != null) {
       o.trainingConfig = this.getTrainingConfig();
       const u = "optimizer", { data: c, specs: h } = await Os(await this.optimizer.getWeights(), u);
-      n.specs.push(...h), n.data = bu([n.data, c]);
+      n.specs.push(...h), n.data = wu([n.data, c]);
     }
     return this.userDefinedMetadata != null && (ii(this.userDefinedMetadata, this.name, !0), o.userDefinedMetadata = this.userDefinedMetadata), o.weightData = n.data, o.weightSpecs = n.specs, t.save(o);
   }
@@ -8963,10 +8964,10 @@ class we extends vt {
 }
 we.className = "Model";
 S(we);
-class Er extends we {
+class Fr extends we {
 }
-Er.className = "Functional";
-S(Er);
+Fr.className = "Functional";
+S(Fr);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -9036,7 +9037,7 @@ class Re extends we {
           throw new d(`A layer added to a Sequential model must not already be connected somewhere else. LayersModel received layer ${t.name} which has ${t.inboundNodes.length} pre-existing inbound connections.`);
         if (t.inboundNodes[0].outputTensors.length !== 1)
           throw new d("All layers in a Sequential model should have a single output tensor. For multi-output layers, use the functional API.");
-        this.checkShape(t), this.outputs = [t.inboundNodes[0].outputTensors[0]], this.inputs = hr(this.outputs[0]);
+        this.checkShape(t), this.outputs = [t.inboundNodes[0].outputTensors[0]], this.inputs = dr(this.outputs[0]);
       }
       this.inboundNodes = [], new Nn({
         outboundLayer: this,
@@ -9165,7 +9166,7 @@ class Re extends we {
    */
   evaluate(t, e, n = {}) {
     if (!this.built)
-      throw new $t("The model needs to be compiled before being used.");
+      throw new Et("The model needs to be compiled before being used.");
     return this.model.evaluate(t, e, n);
   }
   // TODO(cais): Add code snippet below once real dataset objects are
@@ -9192,7 +9193,7 @@ class Re extends we {
    */
   async evaluateDataset(t, e) {
     if (!this.built)
-      throw new $t("The model needs to be compiled before being used.");
+      throw new Et("The model needs to be compiled before being used.");
     return this.model.evaluateDataset(t, e);
   }
   /**
@@ -9282,7 +9283,7 @@ class Re extends we {
    */
   async fit(t, e, n = {}) {
     if (!this.built)
-      throw new $t("The model needs to be compiled before being used.");
+      throw new Et("The model needs to be compiled before being used.");
     return this.model.fit(t, e, n);
   }
   /**
@@ -9372,7 +9373,7 @@ class Re extends we {
    */
   async fitDataset(t, e) {
     if (!this.built)
-      throw new $t("The model needs to be compiled before being used.");
+      throw new Et("The model needs to be compiled before being used.");
     return this.model.fitDataset(t, e);
   }
   /**
@@ -9485,7 +9486,7 @@ let ut = class extends Be {
     return {};
   }
 };
-class Lr extends ut {
+class Mr extends ut {
   /**
    * Calculate the activation function.
    *
@@ -9494,74 +9495,74 @@ class Lr extends ut {
    * @return Output of the ELU activation.
    */
   apply(t, e = 1) {
-    return zu(t, e);
+    return Tu(t, e);
   }
 }
-Lr.className = "elu";
-S(Lr);
-class Fr extends ut {
-  apply(t) {
-    return oh(t);
-  }
-}
-Fr.className = "selu";
-S(Fr);
-class Mr extends ut {
-  apply(t) {
-    return Ve(t);
-  }
-}
-Mr.className = "relu";
+Mr.className = "elu";
 S(Mr);
 class Or extends ut {
   apply(t) {
-    return x(() => Ki(6, Ve(t)));
+    return oh(t);
   }
 }
-Or.className = "relu6";
+Or.className = "selu";
 S(Or);
 class Rr extends ut {
   apply(t) {
-    return t;
+    return Ve(t);
   }
 }
-Rr.className = "linear";
+Rr.className = "relu";
 S(Rr);
 class _r extends ut {
   apply(t) {
-    return Qn(t);
+    return x(() => ji(6, Ve(t)));
   }
 }
-_r.className = "sigmoid";
+_r.className = "relu6";
 S(_r);
 class Br extends ut {
   apply(t) {
-    return Tu(t);
+    return t;
   }
 }
-Br.className = "hardSigmoid";
+Br.className = "linear";
 S(Br);
 class Wr extends ut {
   apply(t) {
-    return us(t);
+    return Qn(t);
   }
 }
-Wr.className = "softplus";
+Wr.className = "sigmoid";
 S(Wr);
 class Gr extends ut {
   apply(t) {
     return $u(t);
   }
 }
-Gr.className = "softsign";
+Gr.className = "hardSigmoid";
 S(Gr);
 class Pr extends ut {
   apply(t) {
-    return as(t);
+    return us(t);
   }
 }
-Pr.className = "tanh";
+Pr.className = "softplus";
 S(Pr);
+class Ur extends ut {
+  apply(t) {
+    return Eu(t);
+  }
+}
+Ur.className = "softsign";
+S(Ur);
+class Vr extends ut {
+  apply(t) {
+    return as(t);
+  }
+}
+Vr.className = "tanh";
+S(Vr);
 let vs = class extends ut {
   /**
    * Calculate the activation function.
@@ -9576,12 +9577,12 @@ let vs = class extends ut {
    * @throws ValueError: In case `dim(x) < 2`.
    */
   apply(t, e = -1) {
-    return Yi(t, e);
+    return tr(t, e);
   }
 };
 vs.className = "softmax";
 S(vs);
-class Ur extends ut {
+class jr extends ut {
   /**
    * Calculate the activation function of log softmax:
    * log( exp(x_i) / sum(exp(x)) )
@@ -9599,9 +9600,9 @@ class Ur extends ut {
     return Uc(t, e);
   }
 }
-Ur.className = "logSoftmax";
-S(Ur);
-class Vr extends ut {
+jr.className = "logSoftmax";
+S(jr);
+class Kr extends ut {
   /**
    * Calculate the activation function.
    *
@@ -9615,9 +9616,9 @@ class Vr extends ut {
     }));
   }
 }
-Vr.className = "gelu";
-S(Vr);
-class jr extends ut {
+Kr.className = "gelu";
+S(Kr);
+class Hr extends ut {
   /**
    * Calculate the activation function.
    *
@@ -9628,9 +9629,9 @@ class jr extends ut {
     return x(() => w(0.5, w(t, $(1, as(w(ee(P(2, Math.PI)), $(t, w(0.044715, qn(t, 3)))))))));
   }
 }
-jr.className = "gelu_new";
-S(jr);
-class Kr extends ut {
+Hr.className = "gelu_new";
+S(Hr);
+class qr extends ut {
   /**
    * Calculate the activation function.
    *
@@ -9641,9 +9642,9 @@ class Kr extends ut {
     return x(() => w(t, as(us(t))));
   }
 }
-Kr.className = "mish";
-S(Kr);
-class Hr extends ut {
+qr.className = "mish";
+S(qr);
+class Zr extends ut {
   /**
    * Calculate the activation function.
    *
@@ -9655,8 +9656,8 @@ class Hr extends ut {
     return x(() => w(Qn(w(t, e)), t));
   }
 }
-Hr.className = "swish";
-S(Hr);
+Zr.className = "swish";
+S(Zr);
 function Yt(s) {
   return s.getClassName();
 }
@@ -9682,15 +9683,15 @@ function Qt(s) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function jf(s) {
+function Kf(s) {
   if (s != null && typeof s != "object")
     throw new Error(`Argument to L1L2 regularizer's constructor is expected to be an object, but received: ${s}`);
 }
-class qr extends Be {
+class Jr extends Be {
 }
-class Zr extends qr {
+class Xr extends Jr {
   constructor(t) {
-    super(), jf(t), this.l1 = t == null || t.l1 == null ? 0.01 : t.l1, this.l2 = t == null || t.l2 == null ? 0.01 : t.l2, this.hasL1 = this.l1 !== 0, this.hasL2 = this.l2 !== 0;
+    super(), Kf(t), this.l1 = t == null || t.l1 == null ? 0.01 : t.l1, this.l2 = t == null || t.l2 == null ? 0.01 : t.l2, this.hasL1 = this.l1 !== 0, this.hasL2 = this.l2 !== 0;
   }
   /**
    * Porting note: Renamed from __call__.
@@ -9710,8 +9711,8 @@ class Zr extends qr {
     return new t({ l1: e.l1, l2: e.l2 });
   }
 }
-Zr.className = "L1L2";
-S(Zr);
+Xr.className = "L1L2";
+S(Xr);
 const ci = {
   l1l2: "L1L2"
 };
@@ -9727,7 +9728,7 @@ function X(s) {
   if (typeof s == "string") {
     const e = { className: s in ci ? ci[s] : s, config: {} };
     return hi(e);
-  } else return s instanceof qr ? s : hi(s);
+  } else return s instanceof Jr ? s : hi(s);
 }
 /**
  * @license
@@ -9738,14 +9739,14 @@ function X(s) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-class Jr extends W {
+class Yr extends W {
   constructor(t) {
     super(t ?? {}), this.supportsMasking = !0, t != null && (this.maxValue = t.maxValue);
   }
   call(t, e) {
     t = O(t);
     let n = Ve(t);
-    return this.maxValue != null && (n = At(n, 0, this.maxValue)), n;
+    return this.maxValue != null && (n = Ct(n, 0, this.maxValue)), n;
   }
   computeOutputShape(t) {
     return t;
@@ -9755,15 +9756,15 @@ class Jr extends W {
     return Object.assign(t, e), t;
   }
 }
-Jr.className = "ReLU";
-S(Jr);
-class Xr extends W {
+Yr.className = "ReLU";
+S(Yr);
+class Qr extends W {
   constructor(t) {
     super(t ?? {}), this.DEFAULT_ALPHA = 0.3, t == null && (t = {}), this.alpha = t.alpha == null ? this.DEFAULT_ALPHA : t.alpha;
   }
   call(t, e) {
     const n = O(t);
-    return Eu(n, this.alpha);
+    return Lu(n, this.alpha);
   }
   computeOutputShape(t) {
     return t;
@@ -9773,9 +9774,9 @@ class Xr extends W {
     return Object.assign(t, e), t;
   }
 }
-Xr.className = "LeakyReLU";
-S(Xr);
-class Yr extends W {
+Qr.className = "LeakyReLU";
+S(Qr);
+class ta extends W {
   constructor(t) {
     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 = rt(t.alphaConstraint), t.sharedAxes == null)
       this.sharedAxes = null;
@@ -9803,7 +9804,7 @@ class Yr extends W {
     })], this.built = !0;
   }
   call(t, e) {
-    return t = O(t), Lu(t, this.alpha.read());
+    return t = O(t), Fu(t, this.alpha.read());
   }
   getConfig() {
     const t = {
@@ -9815,9 +9816,9 @@ class Yr extends W {
     return Object.assign(t, e), t;
   }
 }
-Yr.className = "PReLU";
-S(Yr);
-class Qr extends W {
+ta.className = "PReLU";
+S(ta);
+class ea extends W {
   constructor(t) {
     if (super(t ?? {}), this.DEFAULT_ALPHA = 1, t == null && (t = {}), t.alpha != null && t.alpha !== this.DEFAULT_ALPHA)
       throw new G(`Non-default alpha value (${t.alpha}) is not supported by the ELU layer yet.`);
@@ -9825,7 +9826,7 @@ class Qr extends W {
   }
   call(t, e) {
     const n = O(t);
-    return Fu(n);
+    return Mu(n);
   }
   computeOutputShape(t) {
     return t;
@@ -9835,9 +9836,9 @@ class Qr extends W {
     return Object.assign(t, e), t;
   }
 }
-Qr.className = "ELU";
-S(Qr);
-class ta extends W {
+ea.className = "ELU";
+S(ea);
+class na extends W {
   constructor(t) {
     super(t ?? {}), this.DEFAULT_THETA = 1, t == null && (t = {}), this.theta = t.theta == null ? this.DEFAULT_THETA : t.theta;
   }
@@ -9853,9 +9854,9 @@ class ta extends W {
     return Object.assign(t, e), t;
   }
 }
-ta.className = "ThresholdedReLU";
-S(ta);
-class ea extends W {
+na.className = "ThresholdedReLU";
+S(na);
+class sa extends W {
   constructor(t) {
     super(t ?? {}), this.DEFAULT_AXIS = 1, t == null && (t = {}), this.softmax = new vs().apply, this.axis = t.axis == null ? this.DEFAULT_AXIS : t.axis;
   }
@@ -9867,7 +9868,7 @@ class ea extends W {
         const r = w(V(pe(n.shape), L(i, n.dtype)), tt(-1e9));
         n = $(n, r);
       }
-      return this.axis instanceof Array ? this.axis.length > 1 ? Jt(V(n, Pu(n, this.axis, !0))) : this.softmax(n, this.axis[0]) : this.softmax(n, this.axis);
+      return this.axis instanceof Array ? this.axis.length > 1 ? Jt(V(n, Vu(n, this.axis, !0))) : this.softmax(n, this.axis[0]) : this.softmax(n, this.axis);
     });
   }
   computeOutputShape(t) {
@@ -9878,8 +9879,8 @@ class ea extends W {
     return Object.assign(t, e), t;
   }
 }
-ea.className = "Softmax";
-S(ea);
+sa.className = "Softmax";
+S(sa);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -9896,19 +9897,19 @@ function ke(s, t, e) {
     throw new d(`The ${e} argument must be an integer or tuple of ${t} integers. Received: ${s.length} elements.`);
   for (let n = 0; n < t; ++n) {
     const i = s[n];
-    if (!Mu(i))
+    if (!Ou(i))
       throw new d(`The ${e} argument must be an integer or tuple of ${t} integers. Received: ${JSON.stringify(s)} including a non-integer number ${i}`);
   }
   return s;
 }
-function St(s, t, e, n, i = 1) {
+function At(s, t, e, n, i = 1) {
   if (s == null)
     return s;
   const r = t + (t - 1) * (i - 1);
   let a;
   return e === "same" ? a = s : a = s - r + 1, Math.floor((a + n - 1) / n);
 }
-function Tt(s, t, e, n) {
+function $t(s, t, e, n) {
   if (s == null)
     return null;
   if (n === "valid")
@@ -9931,10 +9932,10 @@ function Tt(s, t, e, n) {
 function Ss(s, t) {
   return x(() => (et(t), t === "channelsFirst" ? j(s, [0, 2, 3, 1]) : s));
 }
-function na(s, t) {
+function ia(s, t) {
   return x(() => (et(t), t === "channelsFirst" ? j(s, [0, 2, 3, 4, 1]) : s));
 }
-function Kf(s, t, e, n = 1, i = "valid", r, a = 1) {
+function Hf(s, t, e, n = 1, i = "valid", r, a = 1) {
   return x(() => {
     if (r == null && (r = ne()), et(r), s.shape.length !== 3)
       throw new d(`The input of a conv1dWithBias operation should be 3, but is ${s.shape.length} instead.`);
@@ -9945,7 +9946,7 @@ function Kf(s, t, e, n = 1, i = "valid", r, a = 1) {
     if (r === "channelsFirst" && (s = j(s, [0, 2, 1])), i === "causal")
       throw new G("The support for CAUSAL padding mode in conv1dWithBias is not implemented yet.");
     let o = kc(s, t, n, i === "same" ? "same" : "valid", "NWC", a);
-    return e != null && (o = Dt(o, e)), o;
+    return e != null && (o = zt(o, e)), o;
   });
 }
 function pi(s, t, e, n = [1, 1], i = "valid", r, a, o = null) {
@@ -9969,16 +9970,16 @@ function pi(s, t, e, n = [1, 1], i = "valid", r, a, o = null) {
     }), r === "channelsFirst" && (l = j(l, [0, 3, 1, 2])), l;
   });
 }
-function Hf(s, t, e, n = [1, 1, 1], i = "valid", r, a) {
+function qf(s, t, e, n = [1, 1, 1], i = "valid", r, a) {
   return x(() => {
     if (r == null && (r = ne()), et(r), s.rank !== 4 && s.rank !== 5)
       throw new d(`conv3dWithBias expects input to be of rank 4 or 5, but received ${s.rank}.`);
     if (t.rank !== 4 && t.rank !== 5)
       throw new d(`conv3dWithBias expects kernel to be of rank 4 or 5, but received ${s.rank}.`);
-    let o = na(s, r);
+    let o = ia(s, r);
     if (i === "causal")
       throw new G("The support for CAUSAL padding mode in conv3dWithBias is not implemented yet.");
-    return o = Ac(o, t, n, i === "same" ? "same" : "valid", "NDHWC", a), e != null && (o = Dt(o, e)), r === "channelsFirst" && (o = j(o, [0, 4, 1, 2, 3])), o;
+    return o = Ac(o, t, n, i === "same" ? "same" : "valid", "NDHWC", a), e != null && (o = zt(o, e)), r === "channelsFirst" && (o = j(o, [0, 4, 1, 2, 3])), o;
   });
 }
 class An extends W {
@@ -10036,16 +10037,16 @@ class De extends An {
     return x(() => {
       t = O(t);
       let n;
-      const i = this.bias == null ? null : this.bias.read(), r = Hi(this.activation.getClassName());
+      const i = this.bias == null ? null : this.bias.read(), r = Ki(this.activation.getClassName());
       if (r != null && this.rank === 2)
         n = pi(t, this.kernel.read(), i, this.strides, this.padding, this.dataFormat, this.dilationRate, r);
       else {
         if (this.rank === 1)
-          n = Kf(t, this.kernel.read(), i, this.strides[0], this.padding, this.dataFormat, this.dilationRate[0]);
+          n = Hf(t, this.kernel.read(), i, this.strides[0], this.padding, this.dataFormat, this.dilationRate[0]);
         else if (this.rank === 2)
           n = pi(t, this.kernel.read(), i, this.strides, this.padding, this.dataFormat, this.dilationRate);
         else if (this.rank === 3)
-          n = Hf(t, this.kernel.read(), i, this.strides, this.padding, this.dataFormat, this.dilationRate);
+          n = qf(t, this.kernel.read(), i, this.strides, this.padding, this.dataFormat, this.dilationRate);
         else
           throw new G("convolutions greater than 3D are not implemented yet.");
         this.activation != null && (n = this.activation.apply(n));
@@ -10057,7 +10058,7 @@ class De extends An {
     t = U(t);
     const e = [], n = this.dataFormat === "channelsLast" ? t.slice(1, t.length - 1) : t.slice(2);
     for (let r = 0; r < n.length; ++r) {
-      const a = St(n[r], this.kernelSize[r], this.padding, this.strides[r], typeof this.dilationRate == "number" ? this.dilationRate : this.dilationRate[r]);
+      const a = At(n[r], this.kernelSize[r], this.padding, this.strides[r], typeof this.dilationRate == "number" ? this.dilationRate : this.dilationRate[r]);
       e.push(a);
     }
     let i = [t[0]];
@@ -10107,7 +10108,7 @@ class Je extends De {
 }
 Je.className = "Conv3D";
 S(Je);
-class sa extends Ze {
+class ra extends Ze {
   constructor(t) {
     if (super(t), this.inputSpec = [new st({ ndim: 4 })], this.padding !== "same" && this.padding !== "valid")
       throw new d(`Conv2DTranspose currently supports only padding modes 'same' and 'valid', but received padding mode ${this.padding}`);
@@ -10129,10 +10130,10 @@ class sa extends Ze {
       const i = n.shape, r = i[0];
       let a, o;
       this.dataFormat === "channelsFirst" ? (a = 2, o = 3) : (a = 1, o = 2);
-      const l = i[a], u = i[o], c = this.kernelSize[0], h = this.kernelSize[1], p = this.strides[0], f = this.strides[1], g = Tt(l, p, c, this.padding), b = Tt(u, f, h, this.padding), m = [r, g, b, this.filters];
+      const l = i[a], u = i[o], c = this.kernelSize[0], h = this.kernelSize[1], p = this.strides[0], f = this.strides[1], g = $t(l, p, c, this.padding), b = $t(u, f, h, this.padding), m = [r, g, b, this.filters];
       this.dataFormat !== "channelsLast" && (n = j(n, [0, 2, 3, 1]));
       let v = vc(n, this.kernel.read(), m, this.strides, this.padding);
-      return this.dataFormat !== "channelsLast" && (v = j(v, [0, 3, 1, 2])), this.bias != null && (v = Dt(v, this.bias.read(), this.dataFormat)), this.activation != null && (v = this.activation.apply(v)), v;
+      return this.dataFormat !== "channelsLast" && (v = j(v, [0, 3, 1, 2])), this.bias != null && (v = zt(v, this.bias.read(), this.dataFormat)), this.activation != null && (v = this.activation.apply(v)), v;
     });
   }
   computeOutputShape(t) {
@@ -10141,16 +10142,16 @@ class sa extends Ze {
     let n, i, r;
     this.dataFormat === "channelsFirst" ? (n = 1, i = 2, r = 3) : (n = 3, i = 1, r = 2);
     const a = this.kernelSize[0], o = this.kernelSize[1], l = this.strides[0], u = this.strides[1];
-    return e[n] = this.filters, e[i] = Tt(e[i], l, a, this.padding), e[r] = Tt(e[r], u, o, this.padding), e;
+    return e[n] = this.filters, e[i] = $t(e[i], l, a, this.padding), e[r] = $t(e[r], u, o, this.padding), e;
   }
   getConfig() {
     const t = super.getConfig();
     return delete t.dilationRate, t;
   }
 }
-sa.className = "Conv2DTranspose";
-S(sa);
-class ia extends Je {
+ra.className = "Conv2DTranspose";
+S(ra);
+class aa extends Je {
   constructor(t) {
     if (super(t), this.inputSpec = [new st({ ndim: 5 })], this.padding !== "same" && this.padding !== "valid")
       throw new d(`Conv3DTranspose currently supports only padding modes 'same' and 'valid', but received padding mode ${this.padding}`);
@@ -10172,10 +10173,10 @@ class ia extends Je {
       const i = n.shape, r = i[0];
       let a, o, l;
       this.dataFormat === "channelsFirst" ? (l = 2, a = 3, o = 4) : (l = 1, a = 2, o = 3);
-      const u = i[l], c = i[a], h = i[o], p = this.kernelSize[0], f = this.kernelSize[1], g = this.kernelSize[2], b = this.strides[0], m = this.strides[1], v = this.strides[2], y = Tt(u, b, p, this.padding), C = Tt(c, m, f, this.padding), N = Tt(h, v, g, this.padding), I = [r, y, C, N, this.filters];
+      const u = i[l], c = i[a], h = i[o], p = this.kernelSize[0], f = this.kernelSize[1], g = this.kernelSize[2], b = this.strides[0], m = this.strides[1], v = this.strides[2], y = $t(u, b, p, this.padding), C = $t(c, m, f, this.padding), N = $t(h, v, g, this.padding), I = [r, y, C, N, this.filters];
       this.dataFormat !== "channelsLast" && (n = j(n, [0, 2, 3, 4, 1]));
       let z = Dc(n, this.kernel.read(), I, this.strides, this.padding);
-      return this.dataFormat !== "channelsLast" && (z = j(z, [0, 4, 1, 2, 3])), this.bias !== null && (z = Dt(z, this.bias.read(), this.dataFormat)), this.activation !== null && (z = this.activation.apply(z)), z;
+      return this.dataFormat !== "channelsLast" && (z = j(z, [0, 4, 1, 2, 3])), this.bias !== null && (z = zt(z, this.bias.read(), this.dataFormat)), this.activation !== null && (z = this.activation.apply(z)), z;
     });
   }
   computeOutputShape(t) {
@@ -10184,16 +10185,16 @@ class ia extends Je {
     let n, i, r, a;
     this.dataFormat === "channelsFirst" ? (n = 1, i = 2, r = 3, a = 4) : (n = 4, i = 1, r = 2, a = 3);
     const o = this.kernelSize[0], l = this.kernelSize[1], u = this.kernelSize[2], c = this.strides[0], h = this.strides[1], p = this.strides[2];
-    return e[n] = this.filters, e[i] = Tt(e[i], c, o, this.padding), e[r] = Tt(e[r], h, l, this.padding), e[a] = Tt(e[a], p, u, this.padding), e;
+    return e[n] = this.filters, e[i] = $t(e[i], c, o, this.padding), e[r] = $t(e[r], h, l, this.padding), e[a] = $t(e[a], p, u, this.padding), e;
   }
   getConfig() {
     const t = super.getConfig();
     return delete t.dilationRate, t;
   }
 }
-ia.className = "Conv3DTranspose";
-S(ia);
-class ra extends De {
+aa.className = "Conv3DTranspose";
+S(aa);
+class oa extends De {
   constructor(t, e) {
     if (super(t, e), this.DEFAULT_DEPTHWISE_INITIALIZER = "glorotUniform", this.DEFAULT_POINTWISE_INITIALIZER = "glorotUniform", this.depthwiseKernel = null, this.pointwiseKernel = null, e.filters == null)
       throw new d("The `filters` configuration field is required by SeparableConv, but is unspecified.");
@@ -10222,7 +10223,7 @@ class ra extends De {
       let n;
       if (this.rank === 1)
         throw new G("1D separable convolution is not implemented yet.");
-      return this.rank === 2 && (this.dataFormat === "channelsFirst" && (t = j(t, [0, 2, 3, 1])), n = uh(t, this.depthwiseKernel.read(), this.pointwiseKernel.read(), this.strides, this.padding, this.dilationRate, "NHWC")), this.useBias && (n = Dt(n, this.bias.read(), this.dataFormat)), this.activation != null && (n = this.activation.apply(n)), this.dataFormat === "channelsFirst" && (n = j(n, [0, 3, 1, 2])), n;
+      return this.rank === 2 && (this.dataFormat === "channelsFirst" && (t = j(t, [0, 2, 3, 1])), n = uh(t, this.depthwiseKernel.read(), this.pointwiseKernel.read(), this.strides, this.padding, this.dilationRate, "NHWC")), this.useBias && (n = zt(n, this.bias.read(), this.dataFormat)), this.activation != null && (n = this.activation.apply(n)), this.dataFormat === "channelsFirst" && (n = j(n, [0, 3, 1, 2])), n;
     });
   }
   getConfig() {
@@ -10230,14 +10231,14 @@ class ra extends De {
     return delete t.rank, delete t.kernelInitializer, delete t.kernelRegularizer, delete t.kernelConstraint, t.depthwiseInitializer = Y(this.depthwiseInitializer), t.pointwiseInitializer = Y(this.pointwiseInitializer), t.depthwiseRegularizer = H(this.depthwiseRegularizer), t.pointwiseRegularizer = H(this.pointwiseRegularizer), t.depthwiseConstraint = it(this.depthwiseConstraint), t.pointwiseConstraint = it(this.pointwiseConstraint), t;
   }
 }
-ra.className = "SeparableConv";
-class aa extends ra {
+oa.className = "SeparableConv";
+class la extends oa {
   constructor(t) {
     super(2, t);
   }
 }
-aa.className = "SeparableConv2D";
-S(aa);
+la.className = "SeparableConv2D";
+S(la);
 class Cn extends De {
   constructor(t) {
     super(1, t), Cn.verifyArgs(t), this.inputSpec = [{ ndim: 3 }];
@@ -10253,7 +10254,7 @@ class Cn extends De {
 }
 Cn.className = "Conv1D";
 S(Cn);
-class oa extends W {
+class ua extends W {
   constructor(t) {
     super(t), typeof t.cropping == "number" ? this.cropping = [[t.cropping, t.cropping], [t.cropping, t.cropping]] : typeof t.cropping[0] == "number" ? this.cropping = [
       [t.cropping[0], t.cropping[0]],
@@ -10289,11 +10290,11 @@ class oa extends W {
     return Object.assign(t, e), t;
   }
 }
-oa.className = "Cropping2D";
-S(oa);
-class la extends W {
+ua.className = "Cropping2D";
+S(ua);
+class ca extends W {
   constructor(t) {
-    super(t), this.DEFAULT_SIZE = [2, 2], this.inputSpec = [{ ndim: 4 }], this.size = t.size == null ? this.DEFAULT_SIZE : t.size, this.dataFormat = t.dataFormat == null ? "channelsLast" : t.dataFormat, et(this.dataFormat), this.interpolation = t.interpolation == null ? "nearest" : t.interpolation, Ou(this.interpolation);
+    super(t), this.DEFAULT_SIZE = [2, 2], this.inputSpec = [{ ndim: 4 }], this.size = t.size == null ? this.DEFAULT_SIZE : t.size, this.dataFormat = t.dataFormat == null ? "channelsLast" : t.dataFormat, et(this.dataFormat), this.interpolation = t.interpolation == null ? "nearest" : t.interpolation, Ru(this.interpolation);
   }
   computeOutputShape(t) {
     if (this.dataFormat === "channelsFirst") {
@@ -10327,8 +10328,8 @@ class la extends W {
     return Object.assign(t, e), t;
   }
 }
-la.className = "UpSampling2D";
-S(la);
+ca.className = "UpSampling2D";
+S(ca);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -10338,7 +10339,7 @@ S(la);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function qf(s, t, e = [1, 1], n = "valid", i, r) {
+function Zf(s, t, e = [1, 1], n = "valid", i, r) {
   return x(() => {
     i == null && (i = ne()), et(i);
     let a = Ss(s, i);
@@ -10346,10 +10347,10 @@ function qf(s, t, e = [1, 1], n = "valid", i, r) {
       throw new d(`Input for depthwiseConv2d is required to be 4-D, but is instead ${s.rank}-D`);
     if (t.rank !== 4)
       throw new d(`depthwiseKernel is required to be 4-D, but is instead ${t.rank}-D`);
-    return a = tr(a, t, e, n === "same" ? "same" : "valid", "NHWC", r), i === "channelsFirst" && (a = j(a, [0, 3, 1, 2])), a;
+    return a = nr(a, t, e, n === "same" ? "same" : "valid", "NHWC", r), i === "channelsFirst" && (a = j(a, [0, 3, 1, 2])), a;
   });
 }
-class ua extends An {
+class ha extends An {
   constructor(t) {
     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 = rt(t.depthwiseConstraint), this.depthwiseRegularizer = X(t.depthwiseRegularizer);
   }
@@ -10370,13 +10371,13 @@ class ua extends An {
   call(t, e) {
     return x(() => {
       t = O(t);
-      let n = qf(t, this.depthwiseKernel.read(), this.strides, this.padding, this.dataFormat, null);
-      return this.useBias && (n = Dt(n, this.bias.read(), this.dataFormat)), this.activation != null && (n = this.activation.apply(n)), n;
+      let n = Zf(t, this.depthwiseKernel.read(), this.strides, this.padding, this.dataFormat, null);
+      return this.useBias && (n = zt(n, this.bias.read(), this.dataFormat)), this.activation != null && (n = this.activation.apply(n)), n;
     });
   }
   computeOutputShape(t) {
     t = U(t);
-    const e = this.dataFormat === "channelsFirst" ? t[2] : t[1], n = this.dataFormat === "channelsFirst" ? t[3] : t[2], i = this.dataFormat === "channelsFirst" ? t[1] * this.depthMultiplier : t[3] * this.depthMultiplier, r = St(e, this.kernelSize[0], this.padding, this.strides[0]), a = St(n, this.kernelSize[1], this.padding, this.strides[1]);
+    const e = this.dataFormat === "channelsFirst" ? t[2] : t[1], n = this.dataFormat === "channelsFirst" ? t[3] : t[2], i = this.dataFormat === "channelsFirst" ? t[1] * this.depthMultiplier : t[3] * this.depthMultiplier, r = At(e, this.kernelSize[0], this.padding, this.strides[0]), a = At(n, this.kernelSize[1], this.padding, this.strides[1]);
     return this.dataFormat === "channelsFirst" ? [t[0], i, r, a] : [t[0], r, a, i];
   }
   getConfig() {
@@ -10384,8 +10385,8 @@ class ua extends An {
     return t.depthMultiplier = this.depthMultiplier, t.depthwiseInitializer = Y(this.depthwiseInitializer), t.depthwiseRegularizer = H(this.depthwiseRegularizer), t.depthwiseConstraint = it(this.depthwiseRegularizer), t;
   }
 }
-ua.className = "DepthwiseConv2D";
-S(ua);
+ha.className = "DepthwiseConv2D";
+S(ha);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -10395,7 +10396,7 @@ S(ua);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function ca(s, t, e, n) {
+function pa(s, t, e, n) {
   if (Array.isArray(s)) {
     if (t != null || e != null)
       throw new d("When inputs is an array, neither initialState or constants should be provided");
@@ -10406,12 +10407,12 @@ function ca(s, t, e, n) {
   }
   return t = i(t), e = i(e), { inputs: s, initialState: t, constants: e };
 }
-function ha(s, t, e, n = !1, i, r, a = !1, o = !1) {
+function da(s, t, e, n = !1, i, r, a = !1, o = !1) {
   return x(() => {
     const l = t.shape.length;
     if (l < 3)
       throw new d(`Input should be at least 3D, but is ${l}D.`);
-    const u = [1, 0].concat(Ct(2, l));
+    const u = [1, 0].concat(It(2, l));
     t = j(t, u), a && console.warn("Backend rnn(): the unroll = true option is not applicable to the imperative deeplearn.js backend."), i != null && (i = L(L(i, "bool"), "float32"), i.rank === l - 1 && (i = ue(i, -1)), i = j(i, u)), n && (t = on(t, 0), i != null && (i = on(i, 0)));
     const c = [];
     let h, p = e;
@@ -10424,7 +10425,7 @@ function ha(s, t, e, n = !1, i, r, a = !1, o = !1) {
         h = C[0], p = C[1];
       else {
         const N = x(() => {
-          const I = b[v], z = V(It(I), I), _ = $(w(C[0], I), w(p[0], z)), T = p.map((E, R) => $(w(C[1][R], I), w(E, z)));
+          const I = b[v], z = V(Dt(I), I), _ = $(w(C[0], I), w(p[0], z)), T = p.map((E, R) => $(w(C[1][R], I), w(E, z)));
           return { output: _, newStates: T };
         });
         h = N.output, p = N.newStates;
@@ -10450,7 +10451,7 @@ class se extends W {
   getStates() {
     if (this.states_ == null) {
       const t = Array.isArray(this.cell.stateSize) ? this.cell.stateSize.length : 1;
-      return Ct(0, t).map((e) => null);
+      return It(0, t).map((e) => null);
     } else
       return this.states_;
   }
@@ -10563,7 +10564,7 @@ class se extends W {
   apply(t, e) {
     let n = e == null ? null : e.initialState, i = e == null ? null : e.constants;
     e == null && (e = {});
-    const r = ca(t, n, i, this.numConstants);
+    const r = pa(t, n, i, this.numConstants);
     t = r.inputs, n = r.initialState, i = r.constants;
     let a = [], o = [];
     if (n != null) {
@@ -10590,7 +10591,7 @@ class se extends W {
       if (r.length !== a)
         throw new d(`RNN Layer has ${a} state(s) but was passed ${r.length} initial state(s).`);
       this.unroll && console.warn("Ignoring unroll = true for RNN layer, due to imperative backend.");
-      const o = { training: i }, u = ha((g, b) => {
+      const o = { training: i }, u = da((g, b) => {
         const m = this.cell.call([g].concat(b), o);
         return [m[0], m.slice(1)];
       }, t, r, this.goBackwards, n, null, this.unroll, this.returnSequences), c = u[0], h = u[1], p = u[2];
@@ -10602,7 +10603,7 @@ class se extends W {
   getInitialState(t) {
     return x(() => {
       let e = mt(t.shape);
-      return e = B(e, [1, 2]), e = wn(e), Array.isArray(this.cell.stateSize) ? this.cell.stateSize.map((n) => n > 1 ? _s(e, [1, n]) : e) : this.cell.stateSize > 1 ? [_s(e, [1, this.cell.stateSize])] : [e];
+      return e = B(e, [1, 2]), e = yn(e), Array.isArray(this.cell.stateSize) ? this.cell.stateSize.map((n) => n > 1 ? _s(e, [1, n]) : e) : this.cell.stateSize > 1 ? [_s(e, [1, this.cell.stateSize])] : [e];
     });
   }
   get trainableWeights() {
@@ -10663,20 +10664,20 @@ class As extends In {
       t = t[0];
       const i = e.training == null ? !1 : e.training;
       0 < this.dropout && this.dropout < 1 && this.dropoutMask == null && (this.dropoutMask = te({
-        ones: () => It(t),
+        ones: () => Dt(t),
         rate: this.dropout,
         training: i,
         dropoutFunc: this.dropoutFunc
       })), 0 < this.recurrentDropout && this.recurrentDropout < 1 && this.recurrentDropoutMask == null && (this.recurrentDropoutMask = te({
-        ones: () => It(n),
+        ones: () => Dt(n),
         rate: this.recurrentDropout,
         training: i,
         dropoutFunc: this.dropoutFunc
       }));
       let r;
       const a = this.dropoutMask, o = this.recurrentDropoutMask;
-      a != null ? r = Lt(w(t, a), this.kernel.read()) : r = Lt(t, this.kernel.read()), this.bias != null && (r = Dt(r, this.bias.read())), o != null && (n = w(n, o));
-      let l = $(r, Lt(n, this.recurrentKernel.read()));
+      a != null ? r = St(w(t, a), this.kernel.read()) : r = St(t, this.kernel.read()), this.bias != null && (r = zt(r, this.bias.read())), o != null && (n = w(n, o));
+      let l = $(r, St(n, this.recurrentKernel.read()));
       return this.activation != null && (l = this.activation.apply(l)), [l, l];
     });
   }
@@ -10703,7 +10704,7 @@ class As extends In {
 }
 As.className = "SimpleRNNCell";
 S(As);
-class pa extends se {
+class fa extends se {
   constructor(t) {
     t.cell = new As(t), super(t);
   }
@@ -10719,8 +10720,8 @@ class pa extends se {
     return new t(e);
   }
 }
-pa.className = "SimpleRNN";
-S(pa);
+fa.className = "SimpleRNN";
+S(fa);
 class Cs extends In {
   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)
@@ -10742,13 +10743,13 @@ class Cs extends In {
       const n = e.training == null ? !1 : e.training;
       let i = t[1];
       t = t[0], 0 < this.dropout && this.dropout < 1 && this.dropoutMask == null && (this.dropoutMask = te({
-        ones: () => It(t),
+        ones: () => Dt(t),
         rate: this.dropout,
         training: n,
         count: 3,
         dropoutFunc: this.dropoutFunc
       })), 0 < this.recurrentDropout && this.recurrentDropout < 1 && this.recurrentDropoutMask == null && (this.recurrentDropoutMask = te({
-        ones: () => It(i),
+        ones: () => Dt(i),
         rate: this.recurrentDropout,
         training: n,
         count: 3,
@@ -10757,11 +10758,11 @@ class Cs extends In {
       const r = this.dropoutMask, a = this.recurrentDropoutMask;
       let o, l, u;
       0 < this.dropout && this.dropout < 1 && (t = w(t, r[0]));
-      let c = Lt(t, this.kernel.read());
-      this.useBias && (c = Dt(c, this.bias.read())), 0 < this.recurrentDropout && this.recurrentDropout < 1 && (i = w(i, a[0]));
-      const h = this.recurrentKernel.read(), [p, f] = Kt(h, [2 * this.units, this.units], h.rank - 1), g = Lt(i, p), [b, m, v] = Kt(c, 3, c.rank - 1), [y, C] = Kt(g, 2, g.rank - 1);
+      let c = St(t, this.kernel.read());
+      this.useBias && (c = zt(c, this.bias.read())), 0 < this.recurrentDropout && this.recurrentDropout < 1 && (i = w(i, a[0]));
+      const h = this.recurrentKernel.read(), [p, f] = Kt(h, [2 * this.units, this.units], h.rank - 1), g = St(i, p), [b, m, v] = Kt(c, 3, c.rank - 1), [y, C] = Kt(g, 2, g.rank - 1);
       o = this.recurrentActivation.apply($(b, y)), l = this.recurrentActivation.apply($(m, C));
-      const N = Lt(w(l, i), f);
+      const N = St(w(l, i), f);
       u = this.activation.apply($(v, N));
       const I = $(w(o, i), w($(1, pt(o)), u));
       return [I, I];
@@ -10793,7 +10794,7 @@ class Cs extends In {
 }
 Cs.className = "GRUCell";
 S(Cs);
-class da extends se {
+class ma extends se {
   constructor(t) {
     t.implementation === 0 && console.warn("`implementation=0` has been deprecated, and now defaults to `implementation=1`. Please update your layer call."), t.cell = new Cs(t), super(t);
   }
@@ -10809,8 +10810,8 @@ class da extends se {
     return e.implmentation === 0 && (e.implementation = 1), new t(e);
   }
 }
-da.className = "GRU";
-S(da);
+ma.className = "GRU";
+S(ma);
 class Dn extends In {
   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, ot(this.units, "units"), this.activation = Qt(t.activation === void 0 ? this.DEFAULT_ACTIVATION : t.activation), this.recurrentActivation = Qt(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 = rt(t.kernelConstraint), this.recurrentConstraint = rt(t.recurrentConstraint), this.biasConstraint = rt(t.biasConstraint), this.dropout = Ne([1, qt([0, t.dropout == null ? 0 : t.dropout])]), this.recurrentDropout = Ne([
@@ -10849,13 +10850,13 @@ class Dn extends In {
       let i = t[1];
       const r = t[2];
       t = t[0], 0 < this.dropout && this.dropout < 1 && this.dropoutMask == null && (this.dropoutMask = te({
-        ones: () => It(t),
+        ones: () => Dt(t),
         rate: this.dropout,
         training: n,
         count: 4,
         dropoutFunc: this.dropoutFunc
       })), 0 < this.recurrentDropout && this.recurrentDropout < 1 && this.recurrentDropoutMask == null && (this.recurrentDropoutMask = te({
-        ones: () => It(i),
+        ones: () => Dt(i),
         rate: this.recurrentDropout,
         training: n,
         count: 4,
@@ -10864,8 +10865,8 @@ class Dn extends In {
       const a = this.dropoutMask, o = this.recurrentDropoutMask;
       let l, u, c, h;
       0 < this.dropout && this.dropout < 1 && (t = w(t, a[0]));
-      let p = Lt(t, this.kernel.read());
-      0 < this.recurrentDropout && this.recurrentDropout < 1 && (i = w(i, o[0])), p = $(p, Lt(i, this.recurrentKernel.read())), this.useBias && (p = Dt(p, this.bias.read()));
+      let p = St(t, this.kernel.read());
+      0 < this.recurrentDropout && this.recurrentDropout < 1 && (i = w(i, o[0])), p = $(p, St(i, this.recurrentKernel.read())), this.useBias && (p = zt(p, this.bias.read()));
       const [f, g, b, m] = Kt(p, 4, p.rank - 1);
       l = this.recurrentActivation.apply(f), u = this.recurrentActivation.apply(g), c = $(w(u, r), w(l, this.activation.apply(b))), h = this.recurrentActivation.apply(m);
       const v = w(h, this.activation.apply(c));
@@ -10898,7 +10899,7 @@ class Dn extends In {
 }
 Dn.className = "LSTMCell";
 S(Dn);
-class fa extends se {
+class ga extends se {
   constructor(t) {
     t.implementation === 0 && console.warn("`implementation=0` has been deprecated, and now defaults to `implementation=1`. Please update your layer call."), t.cell = new Dn(t), super(t);
   }
@@ -10914,8 +10915,8 @@ class fa extends se {
     return e.implmentation === 0 && (e.implementation = 1), new t(e);
   }
 }
-fa.className = "LSTM";
-S(fa);
+ga.className = "LSTM";
+S(ga);
 class Is extends In {
   constructor(t) {
     super(t), this.cells = t.cells;
@@ -11019,7 +11020,7 @@ class Is extends In {
 Is.className = "StackedRNNCells";
 S(Is);
 function te(s) {
-  const { ones: t, rate: e, training: n = !1, count: i = 1, dropoutFunc: r } = s, a = () => r != null ? r(t(), e) : qi(t(), e), o = () => Ke(a, t, n);
+  const { ones: t, rate: e, training: n = !1, count: i = 1, dropoutFunc: r } = s, a = () => r != null ? r(t(), e) : Hi(t(), e), o = () => Ke(a, t, n);
   return !i || i <= 1 ? Bt(o().clone()) : Array(i).fill(void 0).map(o).map((u) => Bt(u.clone()));
 }
 /**
@@ -11031,7 +11032,7 @@ function te(s) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-var Zf = function(s, t) {
+var Jf = function(s, t) {
   var e = {};
   for (var n in s) Object.prototype.hasOwnProperty.call(s, n) && t.indexOf(n) < 0 && (e[n] = s[n]);
   if (s != null && typeof Object.getOwnPropertySymbols == "function")
@@ -11039,7 +11040,7 @@ var Zf = function(s, t) {
       t.indexOf(n[i]) < 0 && Object.prototype.propertyIsEnumerable.call(s, n[i]) && (e[n[i]] = s[n[i]]);
   return e;
 };
-class ma extends se {
+class ba extends se {
   constructor(t) {
     if (t.unroll)
       throw new G("Unrolling is not possible with convolutional RNNs.");
@@ -11091,14 +11092,14 @@ class ma extends se {
     });
   }
   computeSingleOutputShape(t) {
-    const { dataFormat: e, filters: n, kernelSize: i, padding: r, strides: a, dilationRate: o } = this.cell, l = e === "channelsFirst", u = t[l ? 3 : 2], c = t[l ? 4 : 3], h = St(u, i[0], r, a[0], o[0]), p = St(c, i[1], r, a[1], o[1]);
+    const { dataFormat: e, filters: n, kernelSize: i, padding: r, strides: a, dilationRate: o } = this.cell, l = e === "channelsFirst", u = t[l ? 3 : 2], c = t[l ? 4 : 3], h = At(u, i[0], r, a[0], o[0]), p = At(c, i[1], r, a[1], o[1]);
     return [
       ...t.slice(0, 2),
       ...l ? [n, h, p] : [h, p, n]
     ];
   }
 }
-ma.className = "ConvRNN2D";
+ba.className = "ConvRNN2D";
 class Ds extends Dn {
   constructor(t) {
     const { filters: e, kernelSize: n, strides: i, padding: r, dataFormat: a, dilationRate: o } = t;
@@ -11136,7 +11137,7 @@ class Ds extends Dn {
         throw new d(`ConvLSTM2DCell expects 3 input Tensors (inputs, h, c), got ${t.length}.`);
       const n = e.training || !1, i = t[0], r = t[1], a = t[2], o = 4;
       0 < this.dropout && this.dropout < 1 && this.dropoutMask == null && (this.dropoutMask = te({
-        ones: () => It(i),
+        ones: () => Dt(i),
         rate: this.dropout,
         training: n,
         count: o,
@@ -11145,7 +11146,7 @@ class Ds extends Dn {
       const l = this.dropoutMask, u = ($s, Tn, Es) => !Tn || !Tn[Es] ? $s : w(Tn[Es], $s);
       let c = u(i, l, 0), h = u(i, l, 1), p = u(i, l, 2), f = u(i, l, 3);
       0 < this.recurrentDropout && this.recurrentDropout < 1 && this.recurrentDropoutMask == null && (this.recurrentDropoutMask = te({
-        ones: () => It(r),
+        ones: () => Dt(r),
         rate: this.recurrentDropout,
         training: n,
         count: o,
@@ -11157,12 +11158,12 @@ class Ds extends Dn {
       c = this.inputConv(c, N, T, this.padding), h = this.inputConv(h, I, E, this.padding), p = this.inputConv(p, z, R, this.padding), f = this.inputConv(f, _, q, this.padding);
       const [bt, ie, re, xt] = Kt(this.recurrentKernel.read(), o, C);
       b = this.recurrentConv(b, bt), m = this.recurrentConv(m, ie), v = this.recurrentConv(v, re), y = this.recurrentConv(y, xt);
-      const zt = this.recurrentActivation.apply($(c, b)), me = this.recurrentActivation.apply($(h, m)), ze = $(w(me, a), w(zt, this.activation.apply($(p, v)))), Ts = w(this.recurrentActivation.apply($(f, y)), this.activation.apply(ze));
+      const Tt = this.recurrentActivation.apply($(c, b)), me = this.recurrentActivation.apply($(h, m)), ze = $(w(me, a), w(Tt, this.activation.apply($(p, v)))), Ts = w(this.recurrentActivation.apply($(f, y)), this.activation.apply(ze));
       return [Ts, Ts, ze];
     });
   }
   getConfig() {
-    const t = super.getConfig(), { units: e } = t, n = Zf(t, ["units"]), i = {
+    const t = super.getConfig(), { units: e } = t, n = Jf(t, ["units"]), i = {
       filters: this.filters,
       kernelSize: this.kernelSize,
       padding: this.padding,
@@ -11174,7 +11175,7 @@ class Ds extends Dn {
   }
   inputConv(t, e, n, i) {
     const r = Ce(t, e, this.strides, i || "valid", this.dataFormat === "channelsFirst" ? "NCHW" : "NHWC", this.dilationRate);
-    return n ? Dt(r, n, this.dataFormat) : r;
+    return n ? zt(r, n, this.dataFormat) : r;
   }
   recurrentConv(t, e) {
     return Ce(t, e, 1, "same", this.dataFormat === "channelsFirst" ? "NCHW" : "NHWC");
@@ -11182,7 +11183,7 @@ class Ds extends Dn {
 }
 Ds.className = "ConvLSTM2DCell";
 S(Ds);
-class ga extends ma {
+class ya extends ba {
   constructor(t) {
     const e = new Ds(t);
     super(Object.assign(Object.assign({}, t), { cell: e }));
@@ -11192,8 +11193,8 @@ class ga extends ma {
     return new t(e);
   }
 }
-ga.className = "ConvLSTM2D";
-S(ga);
+ya.className = "ConvLSTM2D";
+S(ya);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -11221,7 +11222,7 @@ class zs extends W {
       const n = O(t);
       if (0 < this.rate && this.rate < 1) {
         const i = e.training == null ? !1 : e.training, r = this.getNoiseShape(n);
-        return Ke(() => qi(n, this.rate, r, this.seed), () => n, i);
+        return Ke(() => Hi(n, this.rate, r, this.seed), () => n, i);
       }
       return t;
     });
@@ -11240,7 +11241,7 @@ class zs extends W {
 }
 zs.className = "Dropout";
 S(zs);
-class ba extends zs {
+class wa extends zs {
   constructor(t) {
     super(t), this.inputSpec = [{ ndim: 3 }];
   }
@@ -11249,9 +11250,9 @@ class ba extends zs {
     return [e[0], 1, e[2]];
   }
 }
-ba.className = "SpatialDropout1D";
-S(ba);
-class ya extends W {
+wa.className = "SpatialDropout1D";
+S(wa);
+class ka extends W {
   constructor(t) {
     if (super(t), this.activation = null, this.useBias = !0, this.kernel = null, this.bias = null, this.DEFAULT_KERNEL_INITIALIZER = "glorotNormal", this.DEFAULT_BIAS_INITIALIZER = "zeros", t.batchInputShape == null && t.inputShape == null && t.inputDim != null) {
       let e = null;
@@ -11272,9 +11273,9 @@ class ya extends W {
   call(t, e) {
     return x(() => {
       this.invokeCallHook(t, e);
-      const n = O(t), i = Hi(this.activation.getClassName());
+      const n = O(t), i = Ki(this.activation.getClassName());
       let r;
-      return i != null ? r = Lt(n, this.kernel.read(), i, this.bias ? this.bias.read() : null) : (r = Lt(n, this.kernel.read()), this.bias != null && (r = Dt(r, this.bias.read())), this.activation != null && (r = this.activation.apply(r))), r;
+      return i != null ? r = St(n, this.kernel.read(), i, this.bias ? this.bias.read() : null) : (r = St(n, this.kernel.read()), this.bias != null && (r = zt(r, this.bias.read())), this.activation != null && (r = this.activation.apply(r))), r;
     });
   }
   getConfig() {
@@ -11293,9 +11294,9 @@ class ya extends W {
     return Object.assign(t, e), t;
   }
 }
-ya.className = "Dense";
-S(ya);
-class wa extends W {
+ka.className = "Dense";
+S(ka);
+class xa extends W {
   constructor(t) {
     t = t || {}, super(t), this.inputSpec = [{ minNDim: 3 }], this.dataFormat = t.dataFormat;
   }
@@ -11316,7 +11317,7 @@ class wa extends W {
           i.push(r);
         i.push(1), n = j(n, i);
       }
-      return Ru(n);
+      return _u(n);
     });
   }
   getConfig() {
@@ -11326,9 +11327,9 @@ class wa extends W {
     return Object.assign(t, e), t;
   }
 }
-wa.className = "Flatten";
-S(wa);
-class ka extends W {
+xa.className = "Flatten";
+S(xa);
+class Na extends W {
   constructor(t) {
     super(t), this.supportsMasking = !0, this.activation = Qt(t.activation);
   }
@@ -11344,9 +11345,9 @@ class ka extends W {
     return Object.assign(t, e), t;
   }
 }
-ka.className = "Activation";
-S(ka);
-class xa extends W {
+Na.className = "Activation";
+S(Na);
+class va extends W {
   constructor(t) {
     super(t), this.n = t.n, this.inputSpec = [{ ndim: 2 }];
   }
@@ -11354,7 +11355,7 @@ class xa extends W {
     return [t[0], this.n, t[1]];
   }
   call(t, e) {
-    return x(() => (t = O(t), _u(t, this.n)));
+    return x(() => (t = O(t), Bu(t, this.n)));
   }
   getConfig() {
     const t = {
@@ -11363,9 +11364,9 @@ class xa extends W {
     return Object.assign(t, e), t;
   }
 }
-xa.className = "RepeatVector";
-S(xa);
-class Na extends W {
+va.className = "RepeatVector";
+S(va);
+class Sa extends W {
   constructor(t) {
     super(t), this.targetShape = t.targetShape;
     for (let e = 0; e < this.targetShape.length; ++e)
@@ -11433,15 +11434,15 @@ class Na extends W {
     return Object.assign(t, e), t;
   }
 }
-Na.className = "Reshape";
-S(Na);
-class va extends W {
+Sa.className = "Reshape";
+S(Sa);
+class Aa extends W {
   constructor(t) {
     if (super(t), t.dims == null)
       throw new Error("Required configuration field `dims` is missing during Permute constructor call.");
     if (!Array.isArray(t.dims))
       throw new Error(`Permute constructor requires \`dims\` to be an Array, but received ${t.dims} instead.`);
-    const e = Ct(1, t.dims.length + 1);
+    const e = It(1, t.dims.length + 1);
     if (!Ft(t.dims.slice().sort(), e))
       throw new Error("Invalid permutation `dims`: " + JSON.stringify(t.dims) + " `dims` must contain consecutive integers starting from 1.");
     this.dims = t.dims, this.dimsIncludingBatch = [0].concat(this.dims), this.inputSpec = [new st({ ndim: this.dims.length + 1 })];
@@ -11463,9 +11464,9 @@ class va extends W {
     return Object.assign(t, e), t;
   }
 }
-va.className = "Permute";
-S(va);
-class Sa extends W {
+Aa.className = "Permute";
+S(Aa);
+class Ca extends W {
   constructor(t) {
     super(t ?? {}), this.supportsMasking = !0, t != null ? this.maskValue = t.maskValue == null ? 0 : t.maskValue : this.maskValue = 0;
   }
@@ -11488,8 +11489,8 @@ class Sa extends W {
     });
   }
 }
-Sa.className = "Masking";
-S(Sa);
+Ca.className = "Masking";
+S(Ca);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -11499,7 +11500,7 @@ S(Sa);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-class Aa extends W {
+class Ia extends W {
   constructor(t) {
     if (super(t), this.embeddings = null, this.DEFAULT_EMBEDDINGS_INITIALIZER = "randomUniform", t.batchInputShape == null && t.inputShape == null) {
       let e = null;
@@ -11538,8 +11539,8 @@ class Aa extends W {
     return x(() => {
       this.invokeCallHook(t, e);
       let n = O(t);
-      n.dtype !== "int32" && (n = Et(n, "int32"));
-      const i = ji(this.embeddings.read(), A(n, [n.size]));
+      n.dtype !== "int32" && (n = Lt(n, "int32"));
+      const i = Vi(this.embeddings.read(), A(n, [n.size]));
       return A(i, U(this.computeOutputShape(n.shape)));
     });
   }
@@ -11557,8 +11558,8 @@ class Aa extends W {
     return Object.assign(t, e), t;
   }
 }
-Aa.className = "Embedding";
-S(Aa);
+Ia.className = "Embedding";
+S(Ia);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -11638,7 +11639,7 @@ class fe extends W {
           for (let a of t) {
             const o = a.rank;
             for (let l = 0; l < r - o; ++l)
-              a = wn(a, 1);
+              a = yn(a, 1);
             n.push(a);
           }
           return this.mergeFunction(n);
@@ -11651,7 +11652,7 @@ class fe extends W {
               let f = A(l, [h].concat(ye(c.slice(1))));
               f = j(f, [1, 0]), f = A(f, p), n.push(f), r = !0;
             } else if (u > 1) {
-              const c = Ct(1, u).concat([0]);
+              const c = It(1, u).concat([0]);
               n.push(j(l, c)), r = !0;
             } else
               n.push(l);
@@ -11663,7 +11664,7 @@ class fe extends W {
               const l = a.shape, u = l.length, c = l[u - 1], h = [c].concat(l.slice(0, l.length - 1));
               a = A(j(A(a, [-1, c]), [1, 0]), h);
             } else if (o > 1) {
-              const l = [o - 1].concat(Ct(0, o - 1));
+              const l = [o - 1].concat(It(0, o - 1));
               a = j(a, l);
             }
           }
@@ -11706,7 +11707,7 @@ class fe extends W {
     });
   }
 }
-class Ca extends fe {
+class Da extends fe {
   constructor(t) {
     super(t);
   }
@@ -11719,9 +11720,9 @@ class Ca extends fe {
     });
   }
 }
-Ca.className = "Add";
-S(Ca);
-class Ia extends fe {
+Da.className = "Add";
+S(Da);
+class za extends fe {
   constructor(t) {
     super(t);
   }
@@ -11734,9 +11735,9 @@ class Ia extends fe {
     });
   }
 }
-Ia.className = "Multiply";
-S(Ia);
-class Da extends fe {
+za.className = "Multiply";
+S(za);
+class Ta extends fe {
   constructor(t) {
     super(t);
   }
@@ -11749,9 +11750,9 @@ class Da extends fe {
     });
   }
 }
-Da.className = "Average";
-S(Da);
-class za extends fe {
+Ta.className = "Average";
+S(Ta);
+class $a extends fe {
   constructor(t) {
     super(t);
   }
@@ -11764,9 +11765,9 @@ class za extends fe {
     });
   }
 }
-za.className = "Maximum";
-S(za);
-class Ta extends fe {
+$a.className = "Maximum";
+S($a);
+class Ea extends fe {
   constructor(t) {
     super(t);
   }
@@ -11774,14 +11775,14 @@ class Ta extends fe {
     return x(() => {
       let e = t[0];
       for (let n = 1; n < t.length; ++n)
-        e = Ki(e, t[n]);
+        e = ji(e, t[n]);
       return e;
     });
   }
 }
-Ta.className = "Minimum";
-S(Ta);
-class $a extends fe {
+Ea.className = "Minimum";
+S(Ea);
+class La extends fe {
   constructor(t) {
     super(t), this.DEFAULT_AXIS = -1, t == null && (t = {}), this.axis = t.axis == null ? this.DEFAULT_AXIS : t.axis, this.supportsMasking = !0, this.reshapeRequired = !1;
   }
@@ -11848,7 +11849,7 @@ class $a extends fe {
         return null;
       const i = [];
       for (let a = 0; a < t.length; ++a)
-        e[a] == null ? i.push(L(It(t[a]), "bool")) : e[a].rank < t[a].rank ? i.push(ue(e[a], -1)) : i.push(e[a]);
+        e[a] == null ? i.push(L(Dt(t[a]), "bool")) : e[a].rank < t[a].rank ? i.push(ue(e[a], -1)) : i.push(e[a]);
       const r = is(i, this.axis);
       return Ju(r, -1, !1);
     });
@@ -11860,14 +11861,14 @@ class $a extends fe {
     return Object.assign(t, e), t;
   }
 }
-$a.className = "Concatenate";
-S($a);
+La.className = "Concatenate";
+S(La);
 function $e(s, t) {
   for (; s < 0; )
     s += t;
   return s;
 }
-function Jf(s, t, e) {
+function Xf(s, t, e) {
   if (s.shape.length > 3 || t.shape.length > 3)
     throw new G("batchDot is not implemented for tensors of 4D or higher rank yet");
   if (k(s.shape.length >= 2, () => `batchDot requires the rank of x to be >= 2, but got ${s.shape.length}`), k(s.shape.length >= 2, () => `batchDot requires the rank of y to be >= 2, but got ${t.shape.length}`), typeof e == "number" && (e = [e, e]), s.dtype === "complex64" || t.dtype === "complex64")
@@ -11909,7 +11910,7 @@ function Jf(s, t, e) {
     return o.shape.length === 1 && (o = ue(o, 1)), o;
   });
 }
-class Ea extends fe {
+class Fa extends fe {
   constructor(t) {
     super(t), this.axes = t.axes, this.normalize = t.normalize == null ? !1 : t.normalize, this.supportsMasking = !0, this.reshapeRequired = !1;
   }
@@ -11929,7 +11930,7 @@ class Ea extends fe {
     return Array.isArray(this.axes) ? i = this.axes.map((r, a) => $e(r, t[a].shape.length)) : i = [
       $e(this.axes, e.shape.length),
       $e(this.axes, n.shape.length)
-    ], this.normalize && (e = pn(e, i[0]), n = pn(n, i[1])), Jf(e, n, i);
+    ], this.normalize && (e = pn(e, i[0]), n = pn(n, i[1])), Xf(e, n, i);
   }
   interpretAxes(t, e) {
     let n;
@@ -11959,8 +11960,8 @@ class Ea extends fe {
     return Object.assign(t, e), t;
   }
 }
-Ea.className = "Dot";
-S(Ea);
+Fa.className = "Dot";
+S(Fa);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -11970,7 +11971,7 @@ S(Ea);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-class La extends W {
+class Ma extends W {
   constructor(t) {
     super(t), this.supportsMasking = !0, this.stddev = t.stddev;
   }
@@ -11985,13 +11986,13 @@ class La extends W {
     return x(() => {
       this.invokeCallHook(t, e);
       const n = O(t);
-      return Ke(() => $(yn(n.shape, 0, this.stddev), n), () => n, e.training || !1);
+      return Ke(() => $(bn(n.shape, 0, this.stddev), n), () => n, e.training || !1);
     });
   }
 }
-La.className = "GaussianNoise";
-S(La);
-class Fa extends W {
+Ma.className = "GaussianNoise";
+S(Ma);
+class Oa extends W {
   constructor(t) {
     super(t), this.supportsMasking = !0, this.rate = t.rate;
   }
@@ -12008,14 +12009,14 @@ class Fa extends W {
       const n = O(t);
       return this.rate > 0 && this.rate < 1 ? Ke(() => {
         const r = Math.sqrt(this.rate / (1 - this.rate));
-        return w(n, yn(n.shape, 1, r));
+        return w(n, bn(n.shape, 1, r));
       }, () => n, e.training || !1) : n;
     });
   }
 }
-Fa.className = "GaussianDropout";
-S(Fa);
-class Ma extends W {
+Oa.className = "GaussianDropout";
+S(Oa);
+class Ra extends W {
   constructor(t) {
     super(t), this.supportsMasking = !0, this.rate = t.rate, this.noiseShape = t.noiseShape;
   }
@@ -12035,8 +12036,8 @@ class Ma extends W {
         const n = this._getNoiseShape(t);
         return Ke(() => {
           const r = O(t), o = -1.6732632423543772 * 1.0507009873554805;
-          let l = Ue(bn(n), this.rate);
-          l = Et(l, "float32");
+          let l = Ue(wn(n), this.rate);
+          l = Lt(l, "float32");
           const u = ((1 - this.rate) * (1 + this.rate * o ** 2)) ** -0.5, c = -u * o * this.rate, h = $(w(r, l), w($(l, -1), o));
           return $(w(h, u), c);
         }, () => O(t), e.training || !1);
@@ -12045,8 +12046,8 @@ class Ma extends W {
     });
   }
 }
-Ma.className = "AlphaDropout";
-S(Ma);
+Ra.className = "AlphaDropout";
+S(Ra);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -12068,25 +12069,25 @@ function _e(s, t, e, n, i, r = 1e-3) {
     throw new G(`batchNormalization is not implemented for array of rank ${s.rank} yet`);
   return a;
 }
-function Xf(s, t, e, n, i = 1e-3) {
+function Yf(s, t, e, n, i = 1e-3) {
   return x(() => {
     const r = rs(s, n), a = r.mean, o = r.variance;
     return [_e(s, a, o, e, t, i), a, o];
   });
 }
-function Yf(s, t, e, n, i = 1e-3) {
+function Qf(s, t, e, n, i = 1e-3) {
   return x(() => {
     const r = rs(s, n), a = r.mean, o = r.variance, l = [];
-    for (const g of Ct(0, s.rank))
+    for (const g of It(0, s.rank))
       n.indexOf(g) !== -1 ? l.push(1) : l.push(s.shape[g]);
     const u = A(a, l), c = A(o, l), h = t == null ? null : A(t, l), p = e == null ? null : A(e, l);
     return [_e(s, u, c, p, h, i), a, o];
   });
 }
-function Qf(s, t, e, n, i = 1e-3) {
-  return Ft(n.slice().sort(), Ct(0, s.rank - 1)) ? Xf(s, t, e, n, i) : Yf(s, t, e, n, i);
+function tm(s, t, e, n, i = 1e-3) {
+  return Ft(n.slice().sort(), It(0, s.rank - 1)) ? Yf(s, t, e, n, i) : Qf(s, t, e, n, i);
 }
-class Oa extends W {
+class _a extends W {
   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 = J(t.betaInitializer || "zeros"), this.gammaInitializer = J(t.gammaInitializer || "ones"), this.movingMeanInitializer = J(t.movingMeanInitializer || "zeros"), this.movingVarianceInitializer = J(t.movingVarianceInitializer || "ones"), this.betaConstraint = rt(t.betaConstraint), this.gammaConstraint = rt(t.gammaConstraint), this.betaRegularizer = X(t.betaRegularizer), this.gammaRegularizer = X(t.gammaRegularizer);
   }
@@ -12101,13 +12102,13 @@ class Oa extends W {
   }
   call(t, e) {
     return x(() => {
-      const n = e.training == null ? !1 : e.training, i = O(t), r = i.shape, a = r.length, o = Ct(0, a), l = this.axis >= 0 ? this.axis : this.axis + a;
+      const n = e.training == null ? !1 : e.training, i = O(t), r = i.shape, a = r.length, o = It(0, a), l = this.axis >= 0 ? this.axis : this.axis + a;
       o.splice(l, 1);
       const u = ce(1, a);
       u[l] = r[l];
       const c = o.slice();
       c.sort();
-      const h = !Ft(c, Ct(0, a).slice(0, a - 1)), p = () => {
+      const h = !Ft(c, It(0, a).slice(0, a - 1)), p = () => {
         if (h) {
           const y = A(this.movingMean.read(), u), C = A(this.movingVariance.read(), u), N = this.center ? A(this.beta.read(), u) : null, I = this.scale ? A(this.gamma.read(), u) : null;
           return _e(i, y, C, N, I, this.epsilon);
@@ -12116,7 +12117,7 @@ class Oa extends W {
       };
       if (!n)
         return p();
-      const [f, g, b] = Qf(i, this.gamma.read(), this.beta.read(), o, this.epsilon), m = (y, C, N) => {
+      const [f, g, b] = tm(i, this.gamma.read(), this.beta.read(), o, this.epsilon), m = (y, C, N) => {
         x(() => {
           const I = 1 - N, z = y.read(), _ = w(V(z, C), I);
           y.write(V(z, _));
@@ -12146,9 +12147,9 @@ class Oa extends W {
     return Object.assign(t, e), t;
   }
 }
-Oa.className = "BatchNormalization";
-S(Oa);
-class Ra extends W {
+_a.className = "BatchNormalization";
+S(_a);
+class Ba extends W {
   constructor(t) {
     if (t == null && (t = {}), super(t), this.axis = t.axis == null ? -1 : t.axis, typeof this.axis == "number") {
       if (!Number.isInteger(this.axis))
@@ -12204,8 +12205,8 @@ class Ra extends W {
     return Object.assign(t, e), t;
   }
 }
-Ra.className = "LayerNormalization";
-S(Ra);
+Ba.className = "LayerNormalization";
+S(Ba);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -12215,7 +12216,7 @@ S(Ra);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function tm(s, t, e) {
+function em(s, t, e) {
   return x(() => {
     if (s.rank !== 4)
       throw new d(`temporalPadding expects input tensor to be 4-D, but received a ${s.rank}-D tensor.`);
@@ -12224,10 +12225,10 @@ function tm(s, t, e) {
     if (e == null && (e = ne()), e !== "channelsLast" && e !== "channelsFirst")
       throw new d(`Unknown data format: ${e}. Supported data formats are 'channelsLast' and 'channelsFirst.`);
     let n;
-    return e === "channelsFirst" ? n = [[0, 0], [0, 0], t[0], t[1]] : n = [[0, 0], t[0], t[1], [0, 0]], er(s, n);
+    return e === "channelsFirst" ? n = [[0, 0], [0, 0], t[0], t[1]] : n = [[0, 0], t[0], t[1], [0, 0]], sr(s, n);
   });
 }
-class _a extends W {
+class Wa extends W {
   constructor(t) {
     if (t == null && (t = {}), super(t), this.dataFormat = t.dataFormat == null ? ne() : t.dataFormat, t.padding == null)
       this.padding = [[1, 1], [1, 1]];
@@ -12256,7 +12257,7 @@ class _a extends W {
     return this.dataFormat === "channelsFirst" ? (t[2] != null && t[2] >= 0 ? e = t[2] + this.padding[0][0] + this.padding[0][1] : e = null, t[3] != null && t[3] >= 0 ? n = t[3] + this.padding[1][0] + this.padding[1][1] : n = null, [t[0], t[1], e, n]) : (t[1] != null && t[1] >= 0 ? e = t[1] + this.padding[0][0] + this.padding[0][1] : e = null, t[2] != null && t[2] >= 0 ? n = t[2] + this.padding[1][0] + this.padding[1][1] : n = null, [t[0], e, n, t[3]]);
   }
   call(t, e) {
-    return x(() => tm(O(t), this.padding, this.dataFormat));
+    return x(() => em(O(t), this.padding, this.dataFormat));
   }
   getConfig() {
     const t = {
@@ -12266,8 +12267,8 @@ class _a extends W {
     return Object.assign(t, e), t;
   }
 }
-_a.className = "ZeroPadding2D";
-S(_a);
+Wa.className = "ZeroPadding2D";
+S(Wa);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -12279,7 +12280,7 @@ S(_a);
  */
 function zn(s, t, e, n, i, r) {
   return x(() => {
-    et(i), Zi(r), gt(n), e == null && (e = [1, 1]), n == null && (n = "valid"), i == null && (i = ne()), r == null && (r = "max"), s = Ss(s, i);
+    et(i), qi(r), gt(n), e == null && (e = [1, 1]), n == null && (n = "valid"), i == null && (i = ne()), r == null && (r = "max"), s = Ss(s, i);
     let a;
     const o = n === "same" ? "same" : "valid";
     return r === "max" ? a = Hc(s, t, e, o) : a = ic(
@@ -12291,15 +12292,15 @@ function zn(s, t, e, n, i, r) {
     ), i === "channelsFirst" && (a = j(a, [0, 3, 1, 2])), a;
   });
 }
-function Ba(s, t, e, n, i, r) {
+function Ga(s, t, e, n, i, r) {
   return x(() => {
-    et(i), Zi(r), gt(n), e == null && (e = [1, 1, 1]), n == null && (n = "valid"), i == null && (i = ne()), r == null && (r = "max"), s = na(s, i);
+    et(i), qi(r), gt(n), e == null && (e = [1, 1, 1]), n == null && (n = "valid"), i == null && (i = ne()), r == null && (r = "max"), s = ia(s, i);
     let a;
     const o = n === "same" ? "same" : "valid";
     return r === "max" ? a = Zc(s, t, e, o) : a = ac(s, t, e, o), i === "channelsFirst" && (a = j(a, [0, 4, 1, 2, 3])), a;
   });
 }
-class Wa extends W {
+class Pa extends W {
   /**
    *
    * @param args Parameters for the Pooling layer.
@@ -12325,12 +12326,12 @@ class Wa extends W {
   }
   computeOutputShape(t) {
     t = U(t);
-    const e = St(t[1], this.poolSize[0], this.padding, this.strides[0]);
+    const e = At(t[1], this.poolSize[0], this.padding, this.strides[0]);
     return [t[0], e, t[2]];
   }
   call(t, e) {
     return x(() => {
-      this.invokeCallHook(t, e), t = wn(O(t), 2);
+      this.invokeCallHook(t, e), t = yn(O(t), 2);
       const n = this.poolingFunction(O(t), [this.poolSize[0], 1], [this.strides[0], 1], this.padding, "channelsLast");
       return ts(n, [2]);
     });
@@ -12344,7 +12345,7 @@ class Wa extends W {
     return Object.assign(t, e), t;
   }
 }
-class Ga extends Wa {
+class Ua extends Pa {
   constructor(t) {
     super(t);
   }
@@ -12352,9 +12353,9 @@ class Ga extends Wa {
     return et(r), gt(i), zn(t, e, n, i, r, "max");
   }
 }
-Ga.className = "MaxPooling1D";
-S(Ga);
-class Pa extends Wa {
+Ua.className = "MaxPooling1D";
+S(Ua);
+class Va extends Pa {
   constructor(t) {
     super(t);
   }
@@ -12362,9 +12363,9 @@ class Pa extends Wa {
     return et(r), gt(i), zn(t, e, n, i, r, "avg");
   }
 }
-Pa.className = "AveragePooling1D";
-S(Pa);
-class Ua extends W {
+Va.className = "AveragePooling1D";
+S(Va);
+class ja extends W {
   constructor(t) {
     if (t.poolSize == null && (t.poolSize = [2, 2]), super(t), this.poolSize = Array.isArray(t.poolSize) ? t.poolSize : [t.poolSize, t.poolSize], t.strides == null)
       this.strides = this.poolSize;
@@ -12379,7 +12380,7 @@ class Ua extends W {
   computeOutputShape(t) {
     t = U(t);
     let e = this.dataFormat === "channelsFirst" ? t[2] : t[1], n = this.dataFormat === "channelsFirst" ? t[3] : t[2];
-    return e = St(e, this.poolSize[0], this.padding, this.strides[0]), n = St(n, this.poolSize[1], this.padding, this.strides[1]), this.dataFormat === "channelsFirst" ? [t[0], t[1], e, n] : [t[0], e, n, t[3]];
+    return e = At(e, this.poolSize[0], this.padding, this.strides[0]), n = At(n, this.poolSize[1], this.padding, this.strides[1]), this.dataFormat === "channelsFirst" ? [t[0], t[1], e, n] : [t[0], e, n, t[3]];
   }
   call(t, e) {
     return x(() => (this.invokeCallHook(t, e), this.poolingFunction(O(t), this.poolSize, this.strides, this.padding, this.dataFormat)));
@@ -12394,7 +12395,7 @@ class Ua extends W {
     return Object.assign(t, e), t;
   }
 }
-class Va extends Ua {
+class Ka extends ja {
   constructor(t) {
     super(t);
   }
@@ -12402,9 +12403,9 @@ class Va extends Ua {
     return et(r), gt(i), zn(t, e, n, i, r, "max");
   }
 }
-Va.className = "MaxPooling2D";
-S(Va);
-class ja extends Ua {
+Ka.className = "MaxPooling2D";
+S(Ka);
+class Ha extends ja {
   constructor(t) {
     super(t);
   }
@@ -12412,9 +12413,9 @@ class ja extends Ua {
     return et(r), gt(i), zn(t, e, n, i, r, "avg");
   }
 }
-ja.className = "AveragePooling2D";
-S(ja);
-class Ka extends W {
+Ha.className = "AveragePooling2D";
+S(Ha);
+class qa extends W {
   constructor(t) {
     if (t.poolSize == null && (t.poolSize = [2, 2, 2]), super(t), this.poolSize = Array.isArray(t.poolSize) ? t.poolSize : [t.poolSize, t.poolSize, t.poolSize], t.strides == null)
       this.strides = this.poolSize;
@@ -12429,7 +12430,7 @@ class Ka extends W {
   computeOutputShape(t) {
     t = U(t);
     let e = this.dataFormat === "channelsFirst" ? t[2] : t[1], n = this.dataFormat === "channelsFirst" ? t[3] : t[2], i = this.dataFormat === "channelsFirst" ? t[4] : t[3];
-    return e = St(e, this.poolSize[0], this.padding, this.strides[0]), n = St(n, this.poolSize[1], this.padding, this.strides[1]), i = St(i, this.poolSize[2], this.padding, this.strides[2]), this.dataFormat === "channelsFirst" ? [t[0], t[1], e, n, i] : [t[0], e, n, i, t[4]];
+    return e = At(e, this.poolSize[0], this.padding, this.strides[0]), n = At(n, this.poolSize[1], this.padding, this.strides[1]), i = At(i, this.poolSize[2], this.padding, this.strides[2]), this.dataFormat === "channelsFirst" ? [t[0], t[1], e, n, i] : [t[0], e, n, i, t[4]];
   }
   call(t, e) {
     return x(() => (this.invokeCallHook(t, e), this.poolingFunction(O(t), this.poolSize, this.strides, this.padding, this.dataFormat)));
@@ -12444,27 +12445,27 @@ class Ka extends W {
     return Object.assign(t, e), t;
   }
 }
-class Ha extends Ka {
+class Za extends qa {
   constructor(t) {
     super(t);
   }
   poolingFunction(t, e, n, i, r) {
-    return et(r), gt(i), Ba(t, e, n, i, r, "max");
+    return et(r), gt(i), Ga(t, e, n, i, r, "max");
   }
 }
-Ha.className = "MaxPooling3D";
-S(Ha);
-class qa extends Ka {
+Za.className = "MaxPooling3D";
+S(Za);
+class Ja extends qa {
   constructor(t) {
     super(t);
   }
   poolingFunction(t, e, n, i, r) {
-    return et(r), gt(i), Ba(t, e, n, i, r, "avg");
+    return et(r), gt(i), Ga(t, e, n, i, r, "avg");
   }
 }
-qa.className = "AveragePooling3D";
-S(qa);
-class Za extends W {
+Ja.className = "AveragePooling3D";
+S(Ja);
+class Xa extends W {
   constructor(t) {
     super(t), this.inputSpec = [new st({ ndim: 3 })];
   }
@@ -12475,7 +12476,7 @@ class Za extends W {
     throw new G();
   }
 }
-class Ja extends Za {
+class Ya extends Xa {
   constructor(t) {
     super(t || {});
   }
@@ -12486,9 +12487,9 @@ class Ja extends Za {
     });
   }
 }
-Ja.className = "GlobalAveragePooling1D";
-S(Ja);
-class Xa extends Za {
+Ya.className = "GlobalAveragePooling1D";
+S(Ya);
+class Qa extends Xa {
   constructor(t) {
     super(t || {});
   }
@@ -12499,9 +12500,9 @@ class Xa extends Za {
     });
   }
 }
-Xa.className = "GlobalMaxPooling1D";
-S(Xa);
-class Ya extends W {
+Qa.className = "GlobalMaxPooling1D";
+S(Qa);
+class to extends W {
   constructor(t) {
     super(t), this.dataFormat = t.dataFormat == null ? "channelsLast" : t.dataFormat, et(this.dataFormat), this.inputSpec = [new st({ ndim: 4 })];
   }
@@ -12516,7 +12517,7 @@ class Ya extends W {
     return Object.assign(t, e), t;
   }
 }
-class Qa extends Ya {
+class eo extends to {
   call(t, e) {
     return x(() => {
       const n = O(t);
@@ -12524,9 +12525,9 @@ class Qa extends Ya {
     });
   }
 }
-Qa.className = "GlobalAveragePooling2D";
-S(Qa);
-class to extends Ya {
+eo.className = "GlobalAveragePooling2D";
+S(eo);
+class no extends to {
   call(t, e) {
     return x(() => {
       const n = O(t);
@@ -12534,8 +12535,8 @@ class to extends Ya {
     });
   }
 }
-to.className = "GlobalMaxPooling2D";
-S(to);
+no.className = "GlobalMaxPooling2D";
+S(no);
 /**
  * @license
  * Copyright 2018 Google LLC
@@ -12545,7 +12546,7 @@ S(to);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-class eo extends W {
+class so extends W {
   constructor(t) {
     super(t), this.layer = t.layer;
   }
@@ -12601,7 +12602,7 @@ class eo extends W {
     return Object.assign(a, e), new t(a);
   }
 }
-class no extends eo {
+class io extends so {
   constructor(t) {
     super(t), this.supportsMasking = !0;
   }
@@ -12618,7 +12619,7 @@ class no extends eo {
     return [n[0], i].concat(n.slice(1));
   }
   call(t, e) {
-    return x(() => (t = O(t), ha(
+    return x(() => (t = O(t), da(
       (a, o) => [O(this.layer.call(a, e)), []],
       t,
       [],
@@ -12631,19 +12632,19 @@ class no extends eo {
     )[1]));
   }
 }
-no.className = "TimeDistributed";
-S(no);
-function em(s) {
-  Zn(Bu, "BidirectionalMergeMode", s);
+io.className = "TimeDistributed";
+S(io);
+function nm(s) {
+  Zn(Wu, "BidirectionalMergeMode", s);
 }
-const nm = "concat";
-class so extends eo {
+const sm = "concat";
+class ro extends so {
   constructor(t) {
     super(t);
     const e = t.layer.getConfig(), n = {};
     n.className = t.layer.getClassName(), n.config = e, this.forwardLayer = Wt(n), e.goBackwards = e.goBackwards !== !0;
     const i = {};
-    if (i.className = t.layer.getClassName(), i.config = e, this.backwardLayer = Wt(i), this.forwardLayer.name = "forward_" + this.forwardLayer.name, this.backwardLayer.name = "backward_" + this.backwardLayer.name, this.mergeMode = t.mergeMode === void 0 ? nm : t.mergeMode, em(this.mergeMode), t.weights)
+    if (i.className = t.layer.getClassName(), i.config = e, this.backwardLayer = Wt(i), this.forwardLayer.name = "forward_" + this.forwardLayer.name, this.backwardLayer.name = "backward_" + this.backwardLayer.name, this.mergeMode = t.mergeMode === void 0 ? sm : t.mergeMode, nm(this.mergeMode), t.weights)
       throw new G("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;
   }
@@ -12669,7 +12670,7 @@ class so extends eo {
   apply(t, e) {
     let n = e == null ? null : e.initialState, i = e == null ? null : e.constants;
     e == null && (e = {});
-    const r = ca(t, n, i, this.numConstants);
+    const r = pa(t, n, i, this.numConstants);
     if (t = r.inputs, n = r.initialState, i = r.constants, Array.isArray(t) && (n = t.slice(1), t = t[0]), (n == null || n.length === 0) && i == null)
       return super.apply(t, e);
     const a = [], o = [];
@@ -12755,8 +12756,8 @@ class so extends eo {
     return i.layer = n, new t(i);
   }
 }
-so.className = "Bidirectional";
-S(so);
+ro.className = "Bidirectional";
+S(ro);
 /**
  * @license
  * Copyright 2022 CodeSmith LLC
@@ -12766,7 +12767,7 @@ S(so);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-class io extends W {
+class ao extends W {
   constructor(t) {
     super(t), this.scale = t.scale, t.offset ? this.offset = t.offset : this.offset = 0;
   }
@@ -12778,11 +12779,11 @@ class io extends W {
     return Object.assign(t, e), t;
   }
   call(t, e) {
-    return x(() => (t = O(t), t.dtype !== "float32" && (t = Et(t, "float32")), $(w(t, this.scale), this.offset)));
+    return x(() => (t = O(t), t.dtype !== "float32" && (t = Lt(t, "float32")), $(w(t, this.scale), this.offset)));
   }
 }
-io.className = "Rescaling";
-S(io);
+ao.className = "Rescaling";
+S(ao);
 /**
  * @license
  * Copyright 2022 CodeSmith LLC
@@ -12792,8 +12793,8 @@ S(io);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-const { resizeBilinear: sm, cropAndResize: im } = _t;
-class ro extends W {
+const { resizeBilinear: im, cropAndResize: rm } = _t;
+class oo extends W {
   constructor(t) {
     super(t), this.height = t.height, this.width = t.width;
   }
@@ -12804,14 +12805,14 @@ class ro extends W {
       t.rank === 3 ? (c = !0, u = kn([t])) : u = t;
       for (let I = 0; I < u.shape[0]; I++)
         m.push(b);
-      const v = ju(m, [m.length, 4]), y = Ku(0, m.length, 1, "int32"), N = im(u, v, y, [i, r], "nearest");
-      return c ? Et(O(en(N)), l) : Et(N, l);
+      const v = Ku(m, [m.length, 4]), y = Hu(0, m.length, 1, "int32"), N = rm(u, v, y, [i, r], "nearest");
+      return c ? Lt(O(en(N)), l) : Lt(N, l);
     });
   }
   upsize(t, e, n, i) {
     return x(() => {
-      const r = sm(t, [e, n]);
-      return Et(r, i);
+      const r = im(t, [e, n]);
+      return Lt(r, i);
     });
   }
   call(t, e) {
@@ -12836,8 +12837,8 @@ class ro extends W {
     return t[e] = this.height, t[n] = this.width, t;
   }
 }
-ro.className = "CenterCrop";
-S(ro);
+oo.className = "CenterCrop";
+S(oo);
 /**
  * @license
  * Copyright 2022 CodeSmith LLC
@@ -12847,9 +12848,9 @@ S(ro);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-function rm(s, t, e, n) {
+function am(s, t, e, n) {
   let i = O(s);
-  if (i.dtype !== "int32" && (i = Et(i, "int32")), t === "int")
+  if (i.dtype !== "int32" && (i = Lt(i, "int32")), t === "int")
     return i;
   const r = i.shape;
   if (i.rank === 0 && (i = ue(i, -1)), t === "oneHot" && i.shape[i.shape.length - 1] !== 1 && (i = ue(i, -1)), i.rank > 2)
@@ -12871,7 +12872,7 @@ function rm(s, t, e, n) {
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-class ao extends W {
+class lo extends W {
   constructor(t) {
     super(t), this.numTokens = t.numTokens, t.outputMode ? this.outputMode = t.outputMode : this.outputMode = "multiHot";
   }
@@ -12887,7 +12888,7 @@ class ao extends W {
   }
   call(t, e) {
     return x(() => {
-      t = O(t), t.dtype !== "int32" && (t = Et(t, "int32"));
+      t = O(t), t.dtype !== "int32" && (t = Lt(t, "int32"));
       let n;
       if (typeof e.countWeights < "u") {
         if (this.outputMode !== "count")
@@ -12895,15 +12896,15 @@ class ao extends W {
               Received countWeights=${e.countWeights}`);
         n = O(e.countWeights);
       }
-      const i = ve(t), r = Hu(t), a = Gt(this.numTokens, i).bufferSync().get(0), o = Ue(r, 0).bufferSync().get(0);
+      const i = ve(t), r = qu(t), a = Gt(this.numTokens, i).bufferSync().get(0), o = Ue(r, 0).bufferSync().get(0);
       if (!(a && o))
         throw new d(`Input values must be between 0 < values <= numTokens with numTokens=${this.numTokens}`);
-      return rm(t, this.outputMode, this.numTokens, n);
+      return am(t, this.outputMode, this.numTokens, n);
     });
   }
 }
-ao.className = "CategoryEncoding";
-S(ao);
+lo.className = "CategoryEncoding";
+S(lo);
 /**
  * @license
  * Copyright 2022 CodeSmith LLC
@@ -12913,8 +12914,8 @@ S(ao);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-const am = ["bilinear", "nearest"], di = new Set(am);
-class oo extends W {
+const om = ["bilinear", "nearest"], di = new Set(om);
+class uo extends W {
   constructor(t) {
     if (super(t), this.height = t.height, this.width = t.width, t.interpolation)
       if (di.has(t.interpolation))
@@ -12950,8 +12951,8 @@ class oo extends W {
     });
   }
 }
-oo.className = "Resizing";
-S(oo);
+uo.className = "Resizing";
+S(uo);
 /**
  * @license
  * Copyright 2023 CodeSmith LLC
@@ -12961,7 +12962,7 @@ S(oo);
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-class lo {
+class co {
   constructor(t) {
     this.seed = t;
   }
@@ -12970,7 +12971,7 @@ class lo {
       return this.seed++;
   }
 }
-lo.className = "RandomSeed";
+co.className = "RandomSeed";
 /**
  * @license
  * Copyright 2023 CodeSmith LLC
@@ -12980,9 +12981,9 @@ lo.className = "RandomSeed";
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-class uo extends W {
+class ho extends W {
   constructor(t) {
-    super(t), this.randomGenerator = new lo(t.seed);
+    super(t), this.randomGenerator = new co(t.seed);
   }
   getConfig() {
     const t = {
@@ -12991,7 +12992,7 @@ class uo extends W {
     return Object.assign(t, e), t;
   }
 }
-uo.className = "BaseRandomLayer";
+ho.className = "BaseRandomLayer";
 /**
  * @license
  * Copyright 2023 CodeSmith LLC
@@ -13001,8 +13002,8 @@ uo.className = "BaseRandomLayer";
  * https://opensource.org/licenses/MIT.
  * =============================================================================
  */
-const om = ["bilinear", "nearest"], fi = new Set(om);
-class co extends uo {
+const lm = ["bilinear", "nearest"], fi = new Set(lm);
+class po extends ho {
   constructor(t) {
     super(t);
     const { factor: e, interpolation: n = "bilinear" } = t;
@@ -13042,7 +13043,7 @@ class co extends uo {
       const n = O(t);
       this.imgHeight = n.shape[n.shape.length - 3];
       const i = n.shape[n.shape.length - 2];
-      this.widthFactor = bn([1], 1 + this.widthLower, 1 + this.widthUpper, "float32", this.randomGenerator.next());
+      this.widthFactor = wn([1], 1 + this.widthLower, 1 + this.widthUpper, "float32", this.randomGenerator.next());
       let r = this.widthFactor.dataSync()[0] * i;
       r = Math.round(r);
       const a = [this.imgHeight, r];
@@ -13058,12 +13059,52 @@ class co extends uo {
     });
   }
 }
-co.className = "RandomWidth";
-S(co);
+po.className = "RandomWidth";
+S(po);
+class Om {
+  vocabSize;
+  embedDim;
+  tiedWeights;
+  initializer;
+  constructor(t, e) {
+    this.vocabSize = t.vocabSize, this.embedDim = t.embedDim, this.initializer = Yd({
+      mean: 0,
+      stddev: 0.02
+    }), this.tiedWeights = Ji(
+      this.initializer.apply([this.vocabSize, this.embedDim]),
+      !0,
+      e || "tied_embedding"
+    );
+  }
+  get variables() {
+    return [this.tiedWeights];
+  }
+  embed(t) {
+    return Qi(this.tiedWeights, t, 0);
+  }
+  project(t) {
+    return St(t, this.tiedWeights.transpose());
+  }
+  getWeights() {
+    return [this.tiedWeights];
+  }
+  setWeights(t) {
+    this.tiedWeights.assign(t[0]);
+  }
+  getConfig() {
+    return {
+      vocabSize: this.vocabSize,
+      embedDim: this.embedDim
+    };
+  }
+  dispose() {
+    this.tiedWeights.dispose();
+  }
+}
 export {
   zs as D,
-  Aa as E,
-  ya as a,
-  er as p,
-  Tm as r
+  Ia as E,
+  Om as T,
+  sr as p,
+  Yd as r
 };