@genai-fi/nanogpt 0.7.3 → 0.8.0

package/dist/{reshape-CdBq1WJ6.js → reshape-DEfQGSin.js}

@@ -1,4 +1,4 @@
-import { B as o, C as t, E as a, K as p } from "./index-BoWRt-10.js";
+import { B as o, C as t, E as a, U as p } from "./index-CUQrfsw_.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.

package/dist/{scatter_nd_util-DUstGbU1.js → scatter_nd_util-CUPPNLaA.js}

@@ -1,4 +1,4 @@
-import { j as p, a1 as w } from "./index-BoWRt-10.js";
+import { j as p, a2 as w } from "./index-CUQrfsw_.js";
 function k(o, t, r) {
   const n = t.rank > 1 ? t.shape[t.rank - 1] : 1, e = t.rank > 1 ? t.rank - 1 : 1, h = `Must have updates.shape = indices.shape[:batchDim] + shape[sliceDim:], got updates.shape: ${r.shape}, indices.shape: ${t.shape}, shape: ${o}, sliceDim: ${n}, and batchDim: ${e}.`;
   if (r.rank < e)

package/dist/{selu_util-BJEXVvjX.js → selu_util-8vv5JxQV.js}

@@ -1,6 +1,6 @@
-import { l as N, aq as H, B as _, C as S, E as L, ar as te, as as ne, an as se, ap as re, at as ie, au as oe, av as le, a as ue, aw as fe, ax as O } from "./index-BoWRt-10.js";
-import { r as ce } from "./reshape-CdBq1WJ6.js";
-import { s as ae } from "./sum-66ew2byf.js";
+import { n as N, as as H, B as _, C as S, E as L, at as te, au as ne, ap as se, ar as re, av as ie, aw as oe, ax as le, a as ue, ay as fe, az as O } from "./index-CUQrfsw_.js";
+import { r as ce } from "./reshape-DEfQGSin.js";
+import { s as ae } from "./sum-l_0SqM4h.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.

package/dist/{shared-B8ztnyEk.js → shared-CkNorDcU.js}

@@ -1,9 +1,9 @@
-import { l as Lt, bf as Ce, j as V, k as Pt, a1 as H, bg as et, aa as ht, bh as dt, bi as at, a6 as De, a7 as st, am as We, b2 as $e, bj as Ue, bk as ze, bl as Be, bm as qt, al as K, bn as At, bo as W, bp as _t, bq as Vt, ae as Ct, br as Dt, ak as Wt, bs as $t, aW as Ut, aX as zt, aZ as Bt, a_ as jt, af as Gt, bt as Zt, b1 as Kt, bu as Ht, a$ as je, ao as Ge, bv as Xt, b5 as Ze, bw as Ke, p as Jt, u as He, t as Xe, bx as Qt, by as It, ar as Yt, W as Je, bz as te, bA as ee, bB as ne, aM as pt, bC as se, bD as ot } from "./index-BoWRt-10.js";
-import { f as it, m as kt, g as Qe, a as Ye, R as tn, v as en, d as nn, e as sn, h as on, i as rn, j as an, k as ln, l as cn, n as un, o as hn, p as fn, q as Nt, r as dn, s as gn, t as mn } from "./backend_util-TE7aTPhZ.js";
-import { g as pn, a as wn, e as In, c as bn } from "./axis_util-BzbKo31C.js";
-import { b as xn } from "./broadcast_to-CdbwV-Dj.js";
-import { r as En } from "./reshape-CdBq1WJ6.js";
-import { p as Fn, q as yn, r as kn, u as Nn } from "./selu_util-BJEXVvjX.js";
+import { n as Lt, bh as Ce, j as V, l as qt, a2 as H, bi as et, ad as ht, bj as dt, bk as at, a9 as De, aa as st, ao as We, b4 as $e, bl as Ue, bm as ze, bn as Be, bo as Pt, an as K, bp as At, bq as W, br as _t, bs as Vt, ag as Ct, bt as Dt, am as Wt, bu as $t, aY as Ut, aZ as zt, a$ as Bt, b0 as jt, ah as Gt, bv as Zt, b3 as Kt, bw as Ht, b1 as je, aq as Ge, bx as Xt, b7 as Ze, by as Ke, p as Yt, u as He, t as Xe, bz as Jt, bA as It, at as Qt, _ as Ye, bB as te, bC as ee, bD as ne, aO as pt, bE as se, bF as ot } from "./index-CUQrfsw_.js";
+import { f as it, m as kt, g as Je, a as Qe, R as tn, v as en, d as nn, e as sn, h as on, i as rn, j as an, k as ln, l as cn, n as un, o as hn, p as fn, q as Nt, r as dn, s as gn, t as mn } from "./backend_util-BJ-_jSeK.js";
+import { g as pn, a as wn, e as In, c as bn } from "./axis_util-DubwyOhW.js";
+import { b as xn } from "./broadcast_to-BYfCp5iL.js";
+import { r as En } from "./reshape-DEfQGSin.js";
+import { p as Fn, q as yn, r as kn, u as Nn } from "./selu_util-8vv5JxQV.js";
 import { g as vn } from "./_commonjsHelpers-ByX85dGu.js";
 function Sn(e, t) {
   for (var n = 0; n < t.length; n++) {
@@ -365,8 +365,8 @@ function Tn() {
       throw Error("interior hyphen");
     if (N === 0)
       return u(I.substring(1), c, E).neg();
-    for (var L = r(l(E, 8)), R = g, q = 0; q < I.length; q += 8) {
-      var U = Math.min(8, I.length - q), j = parseInt(I.substring(q, q + U), E);
+    for (var L = r(l(E, 8)), R = g, P = 0; P < I.length; P += 8) {
+      var U = Math.min(8, I.length - P), j = parseInt(I.substring(P, P + U), E);
       if (U < 8) {
         var C = r(l(E, U));
         R = R.mul(C).add(r(j));
@@ -412,13 +412,13 @@ function Tn() {
         return N.toString(c) + L.toInt().toString(c);
       } else
         return "-" + this.neg().toString(c);
-    for (var R = r(l(c, 6), this.unsigned), q = this, U = ""; ; ) {
-      var j = q.div(R), C = q.sub(j.mul(R)).toInt() >>> 0, P = C.toString(c);
-      if (q = j, q.isZero())
-        return P + U;
-      for (; P.length < 6; )
-        P = "0" + P;
-      U = "" + P + U;
+    for (var R = r(l(c, 6), this.unsigned), P = this, U = ""; ; ) {
+      var j = P.div(R), C = P.sub(j.mul(R)).toInt() >>> 0, q = C.toString(c);
+      if (P = j, P.isZero())
+        return q + U;
+      for (; q.length < 6; )
+        q = "0" + q;
+      U = "" + q + U;
     }
   }, x.getHighBits = function() {
     return this.high;
@@ -480,8 +480,8 @@ function Tn() {
     return !this.unsigned && this.eq(S) ? S : this.not().add(k);
   }, x.neg = x.negate, x.add = function(c) {
     n(c) || (c = h(c));
-    var E = this.high >>> 16, N = this.high & 65535, L = this.low >>> 16, R = this.low & 65535, q = c.high >>> 16, U = c.high & 65535, j = c.low >>> 16, C = c.low & 65535, P = 0, G = 0, _ = 0, B = 0;
-    return B += R + C, _ += B >>> 16, B &= 65535, _ += L + j, G += _ >>> 16, _ &= 65535, G += N + U, P += G >>> 16, G &= 65535, P += E + q, P &= 65535, i(_ << 16 | B, P << 16 | G, this.unsigned);
+    var E = this.high >>> 16, N = this.high & 65535, L = this.low >>> 16, R = this.low & 65535, P = c.high >>> 16, U = c.high & 65535, j = c.low >>> 16, C = c.low & 65535, q = 0, G = 0, _ = 0, B = 0;
+    return B += R + C, _ += B >>> 16, B &= 65535, _ += L + j, G += _ >>> 16, _ &= 65535, G += N + U, q += G >>> 16, G &= 65535, q += E + P, q &= 65535, i(_ << 16 | B, q << 16 | G, this.unsigned);
   }, x.subtract = function(c) {
     return n(c) || (c = h(c)), this.add(c.neg());
   }, x.sub = x.subtract, x.multiply = function(c) {
@@ -508,8 +508,8 @@ function Tn() {
       return this.mul(c.neg()).neg();
     if (this.lt(d) && c.lt(d))
       return r(this.toNumber() * c.toNumber(), this.unsigned);
-    var N = this.high >>> 16, L = this.high & 65535, R = this.low >>> 16, q = this.low & 65535, U = c.high >>> 16, j = c.high & 65535, C = c.low >>> 16, P = c.low & 65535, G = 0, _ = 0, B = 0, ct = 0;
-    return ct += q * P, B += ct >>> 16, ct &= 65535, B += R * P, _ += B >>> 16, B &= 65535, B += q * C, _ += B >>> 16, B &= 65535, _ += L * P, G += _ >>> 16, _ &= 65535, _ += R * C, G += _ >>> 16, _ &= 65535, _ += q * j, G += _ >>> 16, _ &= 65535, G += N * P + L * C + R * j + q * U, G &= 65535, i(B << 16 | ct, G << 16 | _, this.unsigned);
+    var N = this.high >>> 16, L = this.high & 65535, R = this.low >>> 16, P = this.low & 65535, U = c.high >>> 16, j = c.high & 65535, C = c.low >>> 16, q = c.low & 65535, G = 0, _ = 0, B = 0, ct = 0;
+    return ct += P * q, B += ct >>> 16, ct &= 65535, B += R * q, _ += B >>> 16, B &= 65535, B += P * C, _ += B >>> 16, B &= 65535, _ += L * q, G += _ >>> 16, _ &= 65535, _ += R * C, G += _ >>> 16, _ &= 65535, _ += P * j, G += _ >>> 16, _ &= 65535, G += N * q + L * C + R * j + P * U, G &= 65535, i(B << 16 | ct, G << 16 | _, this.unsigned);
   }, x.mul = x.multiply, x.divide = function(c) {
     if (n(c) || (c = h(c)), c.isZero())
       throw Error("division by zero");
@@ -539,8 +539,8 @@ function Tn() {
           return S;
         if (c.eq(S))
           return k;
-        var q = this.shr(1);
-        return N = q.div(c).shl(1), N.eq(g) ? c.isNegative() ? k : v : (L = this.sub(c.mul(N)), R = N.add(L.div(c)), R);
+        var P = this.shr(1);
+        return N = P.div(c).shl(1), N.eq(g) ? c.isNegative() ? k : v : (L = this.sub(c.mul(N)), R = N.add(L.div(c)), R);
       } else if (c.eq(S))
         return this.unsigned ? b : g;
       if (this.isNegative())
@@ -551,9 +551,9 @@ function Tn() {
     }
     for (L = this; L.gte(c); ) {
       N = Math.max(1, Math.floor(L.toNumber() / c.toNumber()));
-      for (var U = Math.ceil(Math.log(N) / Math.LN2), j = U <= 48 ? 1 : l(2, U - 48), C = r(N), P = C.mul(c); P.isNegative() || P.gt(L); )
-        N -= j, C = r(N, this.unsigned), P = C.mul(c);
-      C.isZero() && (C = k), R = R.add(C), L = L.sub(P);
+      for (var U = Math.ceil(Math.log(N) / Math.LN2), j = U <= 48 ? 1 : l(2, U - 48), C = r(N), q = C.mul(c); q.isNegative() || q.gt(L); )
+        N -= j, C = r(N, this.unsigned), q = C.mul(c);
+      C.isZero() && (C = k), R = R.add(C), L = L.sub(q);
     }
     return R;
   }, x.div = x.divide, x.modulo = function(c) {
@@ -661,7 +661,7 @@ const tt = (
 function gt(e) {
   return tt.fromString(e, !0, 16);
 }
-const ae = gt("c3a5c85c97cb3127"), Y = gt("b492b66fbe98f273"), D = gt("9ae16a3b2f90404f");
+const ae = gt("c3a5c85c97cb3127"), Q = gt("b492b66fbe98f273"), D = gt("9ae16a3b2f90404f");
 function bt(e) {
   return e.xor(e.shru(47));
 }
@@ -678,7 +678,7 @@ function St(e, t) {
 function A(e, t) {
   return t === 0 ? e : e.shru(t).or(e.shl(64 - t));
 }
-function J(e, t, n = gt("9ddfea08eb382d69")) {
+function Y(e, t, n = gt("9ddfea08eb382d69")) {
   let s = e.xor(t).mul(n);
   s = s.xor(s.shru(47));
   let o = t.xor(s).mul(n);
@@ -695,11 +695,11 @@ function ut(e, t, n, s) {
 function On(e, t = e.length) {
   if (t >= 8) {
     const n = D.add(t * 2), s = T(e, 0).add(D), o = T(e, t - 8), a = A(o, 37).mul(n).add(s), r = A(s, 25).add(o).mul(n);
-    return J(a, r, n);
+    return Y(a, r, n);
   }
   if (t >= 4) {
     const n = D.add(t * 2), s = St(e, 0);
-    return J(s.shl(3).add(t), St(e, t - 4), n);
+    return Y(s.shl(3).add(t), St(e, t - 4), n);
   }
   if (t > 0) {
     const n = e[0], s = e[t >> 1], o = e[t - 1], a = n + (s << 8), r = t + (o << 2);
@@ -708,28 +708,28 @@ function On(e, t = e.length) {
   return D;
 }
 function Ln(e, t = e.length) {
-  const n = D.add(t * 2), s = T(e, 0).mul(Y), o = T(e, 8), a = T(e, t - 8).mul(n), r = T(e, t - 16).mul(D);
-  return J(A(s.add(o), 43).add(A(a, 30)).add(r), s.add(A(o.add(D), 18)).add(a), n);
-}
-function Pn(e, t = e.length) {
-  const n = D.add(t * 2), s = T(e, 0).mul(D), o = T(e, 8), a = T(e, t - 8).mul(n), r = T(e, t - 16).mul(D), i = A(s.add(o), 43).add(A(a, 30)).add(r), l = J(i, s.add(A(o.add(D), 18)).add(a), n), u = T(e, 16).mul(n), h = T(e, 24), f = i.add(T(e, t - 32)).mul(n), p = l.add(T(e, t - 24)).mul(n);
-  return J(A(u.add(h), 43).add(A(f, 30)).add(p), u.add(A(h.add(s), 18)).add(f), n);
+  const n = D.add(t * 2), s = T(e, 0).mul(Q), o = T(e, 8), a = T(e, t - 8).mul(n), r = T(e, t - 16).mul(D);
+  return Y(A(s.add(o), 43).add(A(a, 30)).add(r), s.add(A(o.add(D), 18)).add(a), n);
 }
 function qn(e, t = e.length) {
+  const n = D.add(t * 2), s = T(e, 0).mul(D), o = T(e, 8), a = T(e, t - 8).mul(n), r = T(e, t - 16).mul(D), i = A(s.add(o), 43).add(A(a, 30)).add(r), l = Y(i, s.add(A(o.add(D), 18)).add(a), n), u = T(e, 16).mul(n), h = T(e, 24), f = i.add(T(e, t - 32)).mul(n), p = l.add(T(e, t - 24)).mul(n);
+  return Y(A(u.add(h), 43).add(A(f, 30)).add(p), u.add(A(h.add(s), 18)).add(f), n);
+}
+function Pn(e, t = e.length) {
   const n = tt.fromNumber(81, !0);
   if (t <= 32)
     return t <= 16 ? On(e, t) : Ln(e, t);
   if (t <= 64)
-    return Pn(e, t);
-  let s = n, o = n.mul(Y).add(113), a = bt(o.mul(D).add(113)).mul(D), r = [tt.UZERO, tt.UZERO], i = [tt.UZERO, tt.UZERO];
+    return qn(e, t);
+  let s = n, o = n.mul(Q).add(113), a = bt(o.mul(D).add(113)).mul(D), r = [tt.UZERO, tt.UZERO], i = [tt.UZERO, tt.UZERO];
   s = s.mul(D).add(T(e, 0));
   let l = 0;
   const u = (t - 1 >> 6) * 64, h = u + (t - 1 & 63) - 63;
   do
-    s = A(s.add(o).add(r[0]).add(T(e, l + 8)), 37).mul(Y), o = A(o.add(r[1]).add(T(e, l + 48)), 42).mul(Y), s = s.xor(i[1]), o = o.add(r[0]).add(T(e, l + 40)), a = A(a.add(i[0]), 33).mul(Y), r = ut(e, l, r[1].mul(Y), s.add(i[0])), i = ut(e, l + 32, a.add(i[1]), o.add(T(e, l + 16))), [a, s] = [s, a], l += 64;
+    s = A(s.add(o).add(r[0]).add(T(e, l + 8)), 37).mul(Q), o = A(o.add(r[1]).add(T(e, l + 48)), 42).mul(Q), s = s.xor(i[1]), o = o.add(r[0]).add(T(e, l + 40)), a = A(a.add(i[0]), 33).mul(Q), r = ut(e, l, r[1].mul(Q), s.add(i[0])), i = ut(e, l + 32, a.add(i[1]), o.add(T(e, l + 16))), [a, s] = [s, a], l += 64;
   while (l !== u);
-  const f = Y.add(a.and(255).shl(1));
-  return l = h, i[0] = i[0].add(t - 1 & 63), r[0] = r[0].add(i[0]), i[0] = i[0].add(r[0]), s = A(s.add(o).add(r[0]).add(T(e, l + 8)), 37).mul(f), o = A(o.add(r[1]).add(T(e, l + 48)), 42).mul(f), s = s.xor(i[1].mul(9)), o = o.add(r[0].mul(9).add(T(e, l + 40))), a = A(a.add(i[0]), 33).mul(f), r = ut(e, l, r[1].mul(f), s.add(i[0])), i = ut(e, l + 32, a.add(i[1]), o.add(T(e, l + 16))), [a, s] = [s, a], J(J(r[0], i[0], f).add(bt(o).mul(ae)).add(a), J(r[1], i[1], f).add(s), f);
+  const f = Q.add(a.and(255).shl(1));
+  return l = h, i[0] = i[0].add(t - 1 & 63), r[0] = r[0].add(i[0]), i[0] = i[0].add(r[0]), s = A(s.add(o).add(r[0]).add(T(e, l + 8)), 37).mul(f), o = A(o.add(r[1]).add(T(e, l + 48)), 42).mul(f), s = s.xor(i[1].mul(9)), o = o.add(r[0].mul(9).add(T(e, l + 40))), a = A(a.add(i[0]), 33).mul(f), r = ut(e, l, r[1].mul(f), s.add(i[0])), i = ut(e, l + 32, a.add(i[1]), o.add(T(e, l + 16))), [a, s] = [s, a], Y(Y(r[0], i[0], f).add(bt(o).mul(ae)).add(a), Y(r[1], i[1], f).add(s), f);
 }
 /**
  * @license
@@ -803,7 +803,7 @@ const An = (e) => {
  */
 function $(e) {
   return (t, n, s, o, a) => {
-    const r = Pt(t, n), i = r.length, l = H(r), u = V(r), h = et(a, u), f = t.length, p = n.length, w = H(t), m = H(n), F = ht(t, r), d = ht(n, r);
+    const r = qt(t, n), i = r.length, l = H(r), u = V(r), h = et(a, u), f = t.length, p = n.length, w = H(t), m = H(n), F = ht(t, r), d = ht(n, r);
     if (F.length + d.length === 0)
       for (let g = 0; g < h.length; ++g)
         h[g] = e(s[g % s.length], o[g % o.length]);
@@ -842,7 +842,7 @@ function mt(e) {
     imag: n.makeTensorInfo(o.shape, "float32", r)
   }, i;
 }
-const Js = {
+const Ys = {
   kernelName: De,
   backendName: "cpu",
   kernelFunc: mt
@@ -891,7 +891,7 @@ function Et(e) {
   const { inputs: t, backend: n } = e, { x: s } = t;
   return n.incRef(s.dataId), { dataId: s.dataId, shape: s.shape, dtype: s.dtype };
 }
-const Qs = {
+const Js = {
   kernelName: We,
   backendName: "cpu",
   kernelFunc: Et
@@ -916,7 +916,7 @@ function ce(e) {
   const { inputs: t, backend: n } = e, { input: s } = t, o = n.data.get(s.dataId).complexTensorInfos.real, a = n.data.get(o.dataId).values;
   return n.makeTensorInfo(o.shape, o.dtype, a);
 }
-const Ys = {
+const Qs = {
   kernelName: $e,
   backendName: "cpu",
   kernelFunc: ce
@@ -1013,7 +1013,7 @@ function z(e, t, n, s) {
 }
 function Ft(e) {
   return (t, n, s, o, a, r) => {
-    const i = Pt(t, n), l = V(i), u = i.length, h = H(i), f = et("float32", l), p = et("float32", l), w = ht(t, i), m = ht(n, i), F = kt(s, o), d = kt(a, r), g = t.length, b = H(t), k = n.length, y = H(n);
+    const i = qt(t, n), l = V(i), u = i.length, h = H(i), f = et("float32", l), p = et("float32", l), w = ht(t, i), m = ht(n, i), F = kt(s, o), d = kt(a, r), g = t.length, b = H(t), k = n.length, y = H(n);
     if (w.length + m.length === 0)
       for (let v = 0; v < f.length; v++) {
         const M = v % F.length, O = v % d.length, S = e(F[M * 2], F[M * 2 + 1], d[O * 2], d[O * 2 + 1]);
@@ -1047,8 +1047,8 @@ function Ft(e) {
  * limitations under the License.
  * =============================================================================
  */
-const he = $((e, t) => e + t), _n = Ft((e, t, n, s) => ({ real: e + n, imag: t + s })), Vn = z(qt, he, _n), eo = {
-  kernelName: qt,
+const he = $((e, t) => e + t), _n = Ft((e, t, n, s) => ({ real: e + n, imag: t + s })), Vn = z(Pt, he, _n), eo = {
+  kernelName: Pt,
   backendName: "cpu",
   kernelFunc: Vn
 };
@@ -1152,9 +1152,9 @@ function X(e) {
  */
 function de(e, t, n) {
   const s = X(t);
-  return Q(e, s, n);
+  return J(e, s, n);
 }
-function Q(e, t, n) {
+function J(e, t, n) {
   return ({ inputs: s, attrs: o, backend: a }) => {
     const { x: r } = s;
     nt(r, e);
@@ -1186,7 +1186,7 @@ function Q(e, t, n) {
  * limitations under the License.
  * =============================================================================
  */
-const ge = X((e) => Math.ceil(e)), $n = Q(_t, ge), so = {
+const ge = X((e) => Math.ceil(e)), $n = J(_t, ge), so = {
   kernelName: _t,
   backendName: "cpu",
   kernelFunc: $n
@@ -1267,7 +1267,7 @@ const me = $((e, t) => e === t ? 1 : 0), zn = z(Vt, me, null, "bool"), oo = {
  * limitations under the License.
  * =============================================================================
  */
-const pe = X((e) => Math.exp(e)), Bn = Q(Ct, pe, "float32"), ro = {
+const pe = X((e) => Math.exp(e)), Bn = J(Ct, pe, "float32"), ro = {
   kernelName: Ct,
   backendName: "cpu",
   kernelFunc: Bn
@@ -1288,7 +1288,7 @@ const pe = X((e) => Math.exp(e)), Bn = Q(Ct, pe, "float32"), ro = {
  * limitations under the License.
  * =============================================================================
  */
-const we = X((e) => Math.expm1(e)), jn = Q(Dt, we), ao = {
+const we = X((e) => Math.expm1(e)), jn = J(Dt, we), ao = {
   kernelName: Dt,
   backendName: "cpu",
   kernelFunc: jn
@@ -1309,7 +1309,7 @@ const we = X((e) => Math.expm1(e)), jn = Q(Dt, we), ao = {
  * limitations under the License.
  * =============================================================================
  */
-const Ie = X((e) => Math.floor(e)), Gn = Q(Wt, Ie), io = {
+const Ie = X((e) => Math.floor(e)), Gn = J(Wt, Ie), io = {
   kernelName: Wt,
   backendName: "cpu",
   kernelFunc: Gn
@@ -1430,10 +1430,10 @@ const xe = $((e, t) => e > t ? 1 : 0), Xn = z(Ut, xe, null, "bool"), co = {
  * limitations under the License.
  * =============================================================================
  */
-const Ee = $((e, t) => e >= t ? 1 : 0), Jn = z(zt, Ee, null, "bool"), uo = {
+const Ee = $((e, t) => e >= t ? 1 : 0), Yn = z(zt, Ee, null, "bool"), uo = {
   kernelName: zt,
   backendName: "cpu",
-  kernelFunc: Jn
+  kernelFunc: Yn
 };
 /**
  * @license
@@ -1451,10 +1451,10 @@ const Ee = $((e, t) => e >= t ? 1 : 0), Jn = z(zt, Ee, null, "bool"), uo = {
  * limitations under the License.
  * =============================================================================
  */
-const Fe = $((e, t) => e < t ? 1 : 0), Qn = z(Bt, Fe, null, "bool"), ho = {
+const Fe = $((e, t) => e < t ? 1 : 0), Jn = z(Bt, Fe, null, "bool"), ho = {
   kernelName: Bt,
   backendName: "cpu",
-  kernelFunc: Qn
+  kernelFunc: Jn
 };
 /**
  * @license
@@ -1472,10 +1472,10 @@ const Fe = $((e, t) => e < t ? 1 : 0), Qn = z(Bt, Fe, null, "bool"), ho = {
  * limitations under the License.
  * =============================================================================
  */
-const ye = $((e, t) => e <= t ? 1 : 0), Yn = z(jt, ye, null, "bool"), fo = {
+const ye = $((e, t) => e <= t ? 1 : 0), Qn = z(jt, ye, null, "bool"), fo = {
   kernelName: jt,
   backendName: "cpu",
-  kernelFunc: Yn
+  kernelFunc: Qn
 };
 /**
  * @license
@@ -1516,7 +1516,7 @@ function ts(e, t, n) {
  * limitations under the License.
  * =============================================================================
  */
-const ke = X((e) => Math.log(e)), es = Q(Gt, ke), go = {
+const ke = X((e) => Math.log(e)), es = J(Gt, ke), go = {
   kernelName: Gt,
   backendName: "cpu",
   kernelFunc: es
@@ -1755,7 +1755,7 @@ function Oe(e, t, n, s) {
 function cs(e) {
   const { inputs: t, backend: n, attrs: s } = e, { x: o } = t, { axis: a, keepDims: r } = s;
   nt(o, "prod");
-  const i = o.shape.length, l = Jt(a, o.shape), u = pn(l, i);
+  const i = o.shape.length, l = Yt(a, o.shape), u = pn(l, i);
   let h = l, f = o;
   const p = [];
   u != null && (f = Me({ inputs: { x: o }, backend: n, attrs: { perm: u } }), p.push(f), h = wn(h.length, i));
@@ -1948,7 +1948,7 @@ function ws(e, t, n, s, o, a, r) {
 var Z = tn;
 class ft {
   constructor(t, n, s, o, a, r, i, l, u, h) {
-    this.shape = t, this.shapeShape = n, this.values = s, this.valuesShape = o, this.valuesDType = a, this.defaultValue = r, this.defaultValueShape = i, this.rowPartitionValues = l, this.rowPartitionValuesShapes = u, this.rowPartitionTypes = Qe(h), this.raggedRank = Ye(this.rowPartitionTypes);
+    this.shape = t, this.shapeShape = n, this.values = s, this.valuesShape = o, this.valuesDType = a, this.defaultValue = r, this.defaultValueShape = i, this.rowPartitionValues = l, this.rowPartitionValuesShapes = u, this.rowPartitionTypes = Je(h), this.raggedRank = Qe(this.rowPartitionTypes);
   }
   getRowPartitionTypeByDimension(t) {
     return this.rowPartitionTypes[0] === Z.FIRST_DIM_SIZE ? this.rowPartitionTypes[t + 1] : this.rowPartitionTypes[t];
@@ -2236,8 +2236,8 @@ function bs(e, t, n, s) {
  * limitations under the License.
  * =============================================================================
  */
-const Le = X((e) => 1 / Math.sqrt(e)), xs = Q(Qt, Le), Fo = {
-  kernelName: Qt,
+const Le = X((e) => 1 / Math.sqrt(e)), xs = J(Jt, Le), Fo = {
+  kernelName: Jt,
   backendName: "cpu",
   kernelFunc: xs
 };
@@ -2293,8 +2293,8 @@ function Es(e, t, n, s, o, a, r, i, l, u) {
  * limitations under the License.
  * =============================================================================
  */
-const Fs = X((e) => 1 / (1 + Math.exp(-e))), ys = de(Yt, (e) => 1 / (1 + Math.exp(-e))), yo = {
-  kernelName: Yt,
+const Fs = X((e) => 1 / (1 + Math.exp(-e))), ys = de(Qt, (e) => 1 / (1 + Math.exp(-e))), yo = {
+  kernelName: Qt,
   backendName: "cpu",
   kernelFunc: ys
 };
@@ -2314,7 +2314,7 @@ const Fs = X((e) => 1 / (1 + Math.exp(-e))), ys = de(Yt, (e) => 1 / (1 + Math.ex
  * limitations under the License.
  * =============================================================================
  */
-function Pe(e, t, n, s, o) {
+function qe(e, t, n, s, o) {
   const a = kn(s, t, n), r = V(n), i = H(s);
   if (a) {
     const f = Nn(t, i);
@@ -2332,11 +2332,11 @@ function ks(e) {
   nt(o, "slice");
   const [i, l] = Fn(o, a, r);
   yn(o, i, l);
-  const u = n.data.get(o.dataId).values, h = Pe(u, i, l, o.shape, o.dtype);
+  const u = n.data.get(o.dataId).values, h = qe(u, i, l, o.shape, o.dtype);
   return n.makeTensorInfo(l, o.dtype, h);
 }
 const ko = {
-  kernelName: Je,
+  kernelName: Ye,
   backendName: "cpu",
   kernelFunc: ks
 };
@@ -2578,10 +2578,10 @@ const Ts = X((e) => Math.sqrt(e)), Rs = de(te, (e) => Math.sqrt(e)), No = {
  * limitations under the License.
  * =============================================================================
  */
-const qe = $((e, t) => {
+const Pe = $((e, t) => {
   const n = e - t;
   return n * n;
-}), Ms = z(ee, qe), vo = {
+}), Ms = z(ee, Pe), vo = {
   kernelName: ee,
   backendName: "cpu",
   kernelFunc: Ms
@@ -2605,7 +2605,7 @@ const qe = $((e, t) => {
 const Ae = X((e, t) => {
   const { pattern: n, replaceGlobal: s, rewrite: o } = t;
   return e.replace(new RegExp(n, s ? "g" : ""), o);
-}), Os = Q(ne, Ae), So = {
+}), Os = J(ne, Ae), So = {
   kernelName: ne,
   backendName: "cpu",
   kernelFunc: Os
@@ -2652,7 +2652,7 @@ function Ls(e, t, n, s) {
  * limitations under the License.
  * =============================================================================
  */
-class Ps {
+class qs {
   constructor(t, n, s, o, a, r) {
     this.separator = pt(t), this.nGramWidths = n, this.leftPad = pt(s), this.rightPad = pt(o), this.padWidth = a, this.preserveShort = r;
   }
@@ -2742,8 +2742,8 @@ class Ps {
     return [i, r];
   }
 }
-function qs(e, t, n, s, o, a, r, i) {
-  return new Ps(n, s, o, a, r, i).compute(e, t);
+function Ps(e, t, n, s, o, a, r, i) {
+  return new qs(n, s, o, a, r, i).compute(e, t);
 }
 /**
  * @license
@@ -2822,7 +2822,7 @@ function _s(e, t, n) {
 function Vs(e, t) {
   const n = W("int32", e.length);
   for (let s = 0; s < e.length; ++s)
-    n[s] = qn(e[s]).modulo(t).getLowBitsUnsigned();
+    n[s] = Pn(e[s]).modulo(t).getLowBitsUnsigned();
   return n;
 }
 /**
@@ -2946,7 +2946,7 @@ function $s(e, t, n, s, o) {
  * =============================================================================
  */
 function Us(e, t, n, s) {
-  const o = Jt(t, n)[0], a = [1, n[0], 1];
+  const o = Yt(t, n)[0], a = [1, n[0], 1];
   for (let m = 0; m < o; m++)
     a[0] *= n[m];
   a[1] = n[o];
@@ -3040,15 +3040,15 @@ const Ro = /* @__PURE__ */ Object.freeze(/* @__PURE__ */ Object.defineProperty({
   scatterImpl: Es,
   sigmoidImpl: Fs,
   simpleAbsImpl: le,
-  sliceImpl: Pe,
+  sliceImpl: qe,
   sparseFillEmptyRowsImpl: Ns,
   sparseReshapeImpl: vs,
   sparseSegmentReductionImpl: Ss,
   sqrtImpl: Ts,
-  squaredDifferenceImpl: qe,
+  squaredDifferenceImpl: Pe,
   staticRegexReplaceImpl: Ae,
   stridedSliceImpl: Ls,
-  stringNGramsImpl: qs,
+  stringNGramsImpl: Ps,
   stringSplitImpl: _s,
   stringToHashBucketFastImpl: Vs,
   subImpl: _e,
@@ -3067,7 +3067,7 @@ export {
   vs as F,
   Ss as G,
   Ls as H,
-  qs as I,
+  Ps as I,
   _s as J,
   Vs as K,
   Ws as L,
@@ -3079,7 +3079,7 @@ export {
   no as R,
   to as S,
   so as T,
-  Js as U,
+  Ys as U,
   oo as V,
   ro as W,
   ao as X,
@@ -3087,7 +3087,7 @@ export {
   lo as Z,
   co as _,
   nt as a,
-  Qs as a0,
+  Js as a0,
   ho as a1,
   fo as a2,
   go as a3,
@@ -3097,7 +3097,7 @@ export {
   Io as a7,
   bo as a8,
   Eo as a9,
-  Ys as aa,
+  Qs as aa,
   Fo as ab,
   yo as ac,
   ko as ad,

package/dist/{shared-wS99K7_n.js → shared-D1elLckx.js}

@@ -1,4 +1,4 @@
-import { s as l } from "./shared-B8ztnyEk.js";
+import { s as l } from "./shared-CkNorDcU.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.

package/dist/{sin-BeA3tsEd.js → sin-D2CKKmyR.js}

@@ -1,4 +1,4 @@
-import { B as o, C as t, E as c, a8 as a, a9 as e } from "./index-BoWRt-10.js";
+import { B as o, C as t, E as c, ab as a, ac as e } from "./index-CUQrfsw_.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/{slice-BiOsknYS.js → slice-BnyE-M_7.js}

@@ -1,4 +1,4 @@
-import { B as c, C as e, E as a, W as l } from "./index-BoWRt-10.js";
+import { B as c, C as e, E as a, _ as l } from "./index-CUQrfsw_.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/{softmax-Bv_6lyMX.js → softmax-DLoZWYBx.js}

@@ -1,4 +1,4 @@
-import { B as r, C as f, E as e, S as i } from "./index-BoWRt-10.js";
+import { B as r, C as f, E as e, S as i } from "./index-CUQrfsw_.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/{split-B-dikLRw.js → split-By_n4TKP.js}

@@ -1,4 +1,4 @@
-import { B as p, C as i, E as a, L as c } from "./index-BoWRt-10.js";
+import { B as p, C as i, E as a, V as c } from "./index-CUQrfsw_.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.

package/dist/{stack-B17UN2nn.js → stack-DkdFLq37.js}

@@ -1,4 +1,4 @@
-import { B as e, D as c, l as n, E as k, P as i } from "./index-BoWRt-10.js";
+import { B as e, D as c, n, E as k, P as i } from "./index-CUQrfsw_.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.

package/dist/{sum-66ew2byf.js → sum-l_0SqM4h.js}

@@ -1,4 +1,4 @@
-import { B as e, C as u, X as c, E as l, Y as m } from "./index-BoWRt-10.js";
+import { B as e, C as u, $ as c, E as l, a0 as m } from "./index-CUQrfsw_.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.
@@ -18,8 +18,8 @@ import { B as e, C as u, X as c, E as l, Y as m } from "./index-BoWRt-10.js";
 function i(t, n = null, o = !1) {
   let s = u(t, "x", "sum");
   s.dtype === "bool" && (s = c(s, "int32"));
-  const r = { x: s }, a = { axis: n, keepDims: o };
-  return l.runKernel(m, r, a);
+  const a = { x: s }, r = { axis: n, keepDims: o };
+  return l.runKernel(m, a, r);
 }
 const f = /* @__PURE__ */ e({ sum_: i });
 export {

package/dist/{tensor-JwS7ZYY6.js → tensor-BAQdLqoU.js}

@@ -1,4 +1,4 @@
-import { Z as t, _ as a } from "./index-BoWRt-10.js";
+import { z as t, A as a } from "./index-CUQrfsw_.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/{tensor2d-wxPAnDQy.js → tensor2d-BHy261cI.js}

@@ -1,4 +1,4 @@
-import { $ as t, Z as s, _ as a } from "./index-BoWRt-10.js";
+import { y as t, z as s, A as a } from "./index-CUQrfsw_.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/training/Adam.js

@@ -1,7 +1,7 @@
 import { adamAdjust as b } from "../ops/adamAdjust.js";
 import { adamMoments as d } from "../ops/adamMoments.js";
-import { O as g, e as h, t as o, d as B } from "../index-BoWRt-10.js";
-import { z as M } from "../zeros--BdLQ3oG.js";
+import { O as g, e as h, t as o, d as B } from "../index-CUQrfsw_.js";
+import { z as M } from "../zeros-VZ72lWXM.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/training/AdamExt.js

@@ -1,4 +1,4 @@
-import { a as r, b as c, c as h, e as o } from "../index-BoWRt-10.js";
+import { a as r, b as c, c as h, e as o } from "../index-CUQrfsw_.js";
 import { AdamOptimizer as g } from "./Adam.js";
 class y extends g {
   constructor(t, e, s, i, a) {

package/dist/training/DatasetBuilder.js

@@ -1,5 +1,5 @@
-import { t as g } from "../index-BoWRt-10.js";
-import { d as u, i as d } from "../dataset-CtdBYwjo.js";
+import { t as g } from "../index-CUQrfsw_.js";
+import { d as u, i as d } from "../dataset-CJmEGu6D.js";
 import "../index-Tf7vU29b.js";
 /**
  * @license

package/dist/training/Evaluator.d.ts

@@ -1,9 +1,9 @@
-import { default as NanoGPT } from '../NanoGPTModel';
 import { Dataset } from '@tensorflow/tfjs-data';
 import { TensorContainer } from '@tensorflow/tfjs-core';
+import { default as Model, ModelForwardAttributes } from '../models/model';
 export default class Evaluator {
     private model;
     private iterator;
-    constructor(model: NanoGPT, dataset: Dataset<TensorContainer>);
+    constructor(model: Model<ModelForwardAttributes>, dataset: Dataset<TensorContainer>);
     evaluate(maxBatches?: number): Promise<number>;
 }

package/dist/training/FullTrainer.d.ts

@@ -1,10 +1,10 @@
 import { ITokeniser } from '../tokeniser/type';
-import { default as NanoGPT, TrainingLogEntry } from '../NanoGPTModel';
-import { default as GPTTrainer, TrainingOptions, TrainingProgress } from './Trainer';
+import { default as GPTTrainer, TrainingLogEntry, TrainingOptions, TrainingProgress } from './Trainer';
 import { Tensor } from '@tensorflow/tfjs-core';
 import { Dataset } from '@tensorflow/tfjs-data';
+import { default as Model, ModelForwardAttributes } from '../models/model';
 export default class FullTrainer extends GPTTrainer {
-    constructor(model: NanoGPT, tokenizer: ITokeniser, learningRate?: number);
+    constructor(model: Model<ModelForwardAttributes>, tokenizer: ITokeniser, learningRate?: number);
     private createEmptyState;
     private createLogEntry;
     private createProgress;