@genai-fi/nanogpt 0.7.0 → 0.7.2

package/dist/{kernel_funcs_utils-CwRTFqrc.js → kernel_funcs_utils-BYKWV8Aa.js}

@@ -1,6 +1,6 @@
-import { k as B, j as G, ak as K, a1 as W, al as z, am as V, ac as N, an as F, u as S } from "./index-C0dhsYom.js";
-import { u as O, f as Y } from "./gpgpu_math-DJm3ZTAf.js";
-import { f as v } from "./backend_util-DWiwsi2N.js";
+import { k as B, j as G, am as K, a6 as W, an as z, ao as V, ac as N, ap as F, u as S } from "./index-BoWRt-10.js";
+import { u as O, f as Y } from "./gpgpu_math-DGNLNL4I.js";
+import { f as v } from "./backend_util-TE7aTPhZ.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/layers/BaseLayer.js

@@ -1,5 +1,5 @@
-import { T as g, q as p, e as o, w as v } from "../index-C0dhsYom.js";
-import { v as _ } from "../variable-DXEUOwew.js";
+import { T as g, y as p, e as o, A as v } from "../index-BoWRt-10.js";
+import { v as _ } from "../variable-BuddVFLa.js";
 class M {
   parent;
   config;

package/dist/layers/CausalSelfAttention.js

@@ -3,14 +3,14 @@ import O from "./BaseLayer.js";
 import { qkv as P } from "../ops/qkv.js";
 import { rope as v } from "../ops/rope.js";
 import { appendCache as V } from "../ops/appendCache.js";
-import { o as c, t as C } from "../index-C0dhsYom.js";
+import { w as c, t as C } from "../index-BoWRt-10.js";
 import { fusedSoftmax as T } from "../ops/fusedSoftmax.js";
-import { d as y } from "../random_width-DkYP8W8N.js";
-import { v as b } from "../variable-DXEUOwew.js";
-import { r as k, d as L } from "../dropout-hpDwECTe.js";
-import { r as N } from "../reshape-BMUzc1UY.js";
-import { m as R } from "../mat_mul-1nwdPkQ_.js";
-class $ extends O {
+import { d as y } from "../random_width-sZORGo5k.js";
+import { v as b } from "../variable-BuddVFLa.js";
+import { r as k, d as L } from "../dropout-DYs5QFGQ.js";
+import { r as N } from "../reshape-CdBq1WJ6.js";
+import { m as R } from "../mat_mul-8m8pfdcx.js";
+class W extends O {
   divisor;
   index;
   units;
@@ -92,5 +92,5 @@ class $ extends O {
   }
 }
 export {
-  $ as default
+  W as default
 };

package/dist/layers/MLP.js

@@ -1,10 +1,10 @@
-import { t as l } from "../index-C0dhsYom.js";
+import { t as l } from "../index-BoWRt-10.js";
 import u from "./BaseLayer.js";
 import { matMulGelu as M } from "../ops/matMulGelu.js";
-import { v as o } from "../variable-DXEUOwew.js";
-import { r as h, d as f } from "../dropout-hpDwECTe.js";
-import { r as d } from "../reshape-BMUzc1UY.js";
-import { m as c } from "../mat_mul-1nwdPkQ_.js";
+import { v as o } from "../variable-BuddVFLa.js";
+import { r as h, d as f } from "../dropout-DYs5QFGQ.js";
+import { r as d } from "../reshape-CdBq1WJ6.js";
+import { m as c } from "../mat_mul-8m8pfdcx.js";
 class V extends u {
   index;
   hiddenUnits;

package/dist/layers/RMSNorm.js

@@ -1,8 +1,8 @@
-import { t as s } from "../index-C0dhsYom.js";
+import { t as s } from "../index-BoWRt-10.js";
 import e from "./BaseLayer.js";
 import { normRMS as a } from "../ops/normRMS.js";
-import { v as i } from "../variable-DXEUOwew.js";
-import { o as m } from "../ones-D63HpSF_.js";
+import { v as i } from "../variable-BuddVFLa.js";
+import { o as m } from "../ones-Dj0SDhHf.js";
 class f extends e {
   GAMMA;
   constructor(r, t = "", o) {

package/dist/layers/RoPECache.js

@@ -1,7 +1,7 @@
-import { b as t, n as h, t as n, o as p } from "../index-C0dhsYom.js";
-import { r as c } from "../reciprocal-_A9yv27J.js";
-import { c as f, s as m } from "../sin-Cocju-BY.js";
-import { r as a } from "../range-CYzpQY53.js";
+import { b as t, x as h, t as n, w as p } from "../index-BoWRt-10.js";
+import { r as c } from "../reciprocal-BvGAyKyu.js";
+import { c as f, s as m } from "../sin-BeA3tsEd.js";
+import { r as a } from "../range-CRuAh-gd.js";
 class D {
   constructor(o) {
     this.config = o;

package/dist/layers/TiedEmbedding.js

@@ -1,9 +1,9 @@
-import "../random_width-DkYP8W8N.js";
-import "../index-C0dhsYom.js";
-import { T as e } from "../TiedEmbedding-BLltddza.js";
+import "../random_width-sZORGo5k.js";
+import "../index-BoWRt-10.js";
+import { T as e } from "../TiedEmbedding-BxOerUmB.js";
 import "./BaseLayer.js";
-import "../variable-DXEUOwew.js";
-import "../gather-D0_gPiBz.js";
+import "../variable-BuddVFLa.js";
+import "../gather-CMMy2KEG.js";
 export {
   e as default
 };

package/dist/layers/TransformerBlock.js

@@ -2,7 +2,7 @@ import l from "./CausalSelfAttention.js";
 import r from "./MLP.js";
 import o from "./RMSNorm.js";
 import d from "./BaseLayer.js";
-import { t as p } from "../index-C0dhsYom.js";
+import { t as p } from "../index-BoWRt-10.js";
 class k extends d {
   ln1;
   attn;

package/dist/loader/loadTransformers.js

@@ -2,7 +2,7 @@ import b from "../NanoGPTModel.js";
 import c from "../tokeniser/CharTokeniser.js";
 import l from "../tokeniser/bpe.js";
 import { load_safetensors as u } from "../utilities/safetensors.js";
-import { Y as y } from "../index-C0dhsYom.js";
+import { a0 as y } from "../index-BoWRt-10.js";
 import { dummyPassAsync as h } from "../utilities/dummy.js";
 async function L(e, a, r, t) {
   const n = {

package/dist/loader/oldZipLoad.js

@@ -1,13 +1,17 @@
 import d from "../NanoGPTModel.js";
 import "../jszip.min-CjP2V1VV.js";
 import h from "../tokeniser/CharTokeniser.js";
-import { Y as k } from "../index-C0dhsYom.js";
+import { a0 as k } from "../index-BoWRt-10.js";
 import b from "../tokeniser/bpe.js";
 import { dummyPassAsync as u } from "../utilities/dummy.js";
 import "../Generator.js";
 import "../index-Dwqa6Zy2.js";
-import "../dataset-JDyjG3QR.js";
+import "../dataset-CtdBYwjo.js";
 import "../index-Tf7vU29b.js";
+import "../ops/cpu/adamAdjust.js";
+import "../ops/webgl/adamAdjust.js";
+import "../ops/cpu/adamMoments.js";
+import "../ops/webgl/adamMoments.js";
 import "../papaparse.min-C8l2Kvo1.js";
 import "../ops/cpu/scatterSub.js";
 import "../ops/webgl/scatterSub.js";
@@ -19,8 +23,8 @@ import "../ops/grads/attentionMask.js";
 import "../ops/cpu/qkv.js";
 import "../ops/webgl/qkv.js";
 import "../ops/grads/qkv.js";
-import "../random_width-DkYP8W8N.js";
-import "../register_all_kernels-guvSxp7M.js";
+import "../random_width-sZORGo5k.js";
+import "../register_all_kernels-BwDSRN-f.js";
 import "../ops/cpu/rope.js";
 import "../ops/webgl/rope.js";
 import "../ops/grads/rope.js";
@@ -34,13 +38,13 @@ import "../ops/webgl/matMulGelu.js";
 import "../ops/grads/matMulGelu.js";
 import "../ops/cpu/gelu.js";
 import "../ops/webgl/gelu.js";
-import "../gelu-uyHP1x1f.js";
+import "../gelu-C-dPj6Ku.js";
 import "../ops/cpu/normRMS.js";
 import "../ops/webgl/normRMS.js";
 import "../ops/grads/normRMS.js";
 import "../ops/webgl/log.js";
 import { importWeights as O } from "../utilities/weights.js";
-async function ft(o) {
+async function yt(o) {
   const n = /* @__PURE__ */ new Map(), s = await o.file("manifest.json")?.async("string");
   if (!s)
     throw new Error("Manifest file not found in the zip archive");
@@ -72,5 +76,5 @@ async function ft(o) {
   return { model: e, tokeniser: l };
 }
 export {
-  ft as default
+  yt as default
 };

package/dist/{log_sum_exp-D086OgZJ.js → log_sum_exp-DbjkV734.js}

@@ -1,8 +1,8 @@
-import { x as r, y as p, E as u, a7 as E, a8 as h, p as S, c as $, a6 as d } from "./index-C0dhsYom.js";
-import { e as K } from "./axis_util-DaAl5MER.js";
-import { m as T } from "./max-BQc2Aj-I.js";
-import { r as m } from "./reshape-BMUzc1UY.js";
-import { s as _ } from "./sum-B8wEpKsg.js";
+import { B as r, C as p, E as u, ae as E, af as h, p as S, c as $, o as d } from "./index-BoWRt-10.js";
+import { e as K } from "./axis_util-BzbKo31C.js";
+import { m as T } from "./max-Ddnnb5xe.js";
+import { r as m } from "./reshape-CdBq1WJ6.js";
+import { s as _ } from "./sum-66ew2byf.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.
@@ -61,7 +61,7 @@ const w = /* @__PURE__ */ r({ log_: v });
  * limitations under the License.
  * =============================================================================
  */
-function y(s, n = null, o = !1) {
+function A(s, n = null, o = !1) {
   const a = p(s, "x", "logSumExp"), t = S(n, a.shape), x = T(
     a,
     t,
@@ -74,9 +74,9 @@ function y(s, n = null, o = !1) {
   }
   return e;
 }
-const M = /* @__PURE__ */ r({ logSumExp_: y });
+const L = /* @__PURE__ */ r({ logSumExp_: A });
 export {
   w as a,
   N as e,
-  M as l
+  L as l
 };

package/dist/main.d.ts

@@ -1,3 +1,7 @@
+import { default as CausalSelfAttention } from './layers/CausalSelfAttention';
+import { default as MLP } from './layers/MLP';
+import { default as TransformerBlock } from './layers/TransformerBlock';
+import { default as RoPECache } from './layers/RoPECache';
 export { default as NanoGPT } from './NanoGPTModel';
 export { default as TeachableLLM } from './TeachableLLM';
 export { default as CharTokeniser } from './tokeniser/CharTokeniser';
@@ -13,3 +17,10 @@ export type { GPTConfig } from './config';
 export { estimateParameterCount, estimateMemoryUsage, estimateTrainingMemoryUsage, estimateResources, validateConfig, } from './utilities/parameters';
 export { selectBackend } from './backend';
 export { default as performanceTest } from './utilities/performance';
+export declare const layers: {
+    CausalSelfAttention: typeof CausalSelfAttention;
+    MLP: typeof MLP;
+    TransformerBlock: typeof TransformerBlock;
+    RoPECache: typeof RoPECache;
+};
+export { default as AdamExt } from './training/AdamExt';

package/dist/main.js

@@ -1,11 +1,11 @@
-import { default as R } from "./NanoGPTModel.js";
-import { default as q } from "./TeachableLLM.js";
-import { default as A } from "./tokeniser/CharTokeniser.js";
-import { default as I } from "./tokeniser/bpe.js";
-import { default as K } from "./utilities/waitForModel.js";
-import { default as Q } from "./data/textLoader.js";
-import { estimateMemoryUsage as V, estimateParameterCount as W, estimateResources as X, estimateTrainingMemoryUsage as Y, validateConfig as Z } from "./utilities/parameters.js";
-import "./index-C0dhsYom.js";
+import { default as W } from "./NanoGPTModel.js";
+import { default as Y } from "./TeachableLLM.js";
+import { default as _ } from "./tokeniser/CharTokeniser.js";
+import { default as oo } from "./tokeniser/bpe.js";
+import { default as to } from "./utilities/waitForModel.js";
+import { default as eo } from "./data/textLoader.js";
+import { estimateMemoryUsage as po, estimateParameterCount as ao, estimateResources as fo, estimateTrainingMemoryUsage as so, validateConfig as lo } from "./utilities/parameters.js";
+import "./index-BoWRt-10.js";
 import "./ops/cpu/scatterSub.js";
 import "./ops/webgl/scatterSub.js";
 import "./ops/cpu/gatherSub.js";
@@ -16,10 +16,10 @@ import "./ops/grads/attentionMask.js";
 import "./ops/cpu/qkv.js";
 import "./ops/webgl/qkv.js";
 import "./ops/grads/qkv.js";
-import "./random_width-DkYP8W8N.js";
-import "./register_all_kernels-guvSxp7M.js";
+import "./random_width-sZORGo5k.js";
+import "./register_all_kernels-BwDSRN-f.js";
 import "./index-Tf7vU29b.js";
-import "./dataset-JDyjG3QR.js";
+import "./dataset-CtdBYwjo.js";
 import "./ops/cpu/rope.js";
 import "./ops/webgl/rope.js";
 import "./ops/grads/rope.js";
@@ -33,25 +33,42 @@ import "./ops/webgl/matMulGelu.js";
 import "./ops/grads/matMulGelu.js";
 import "./ops/cpu/gelu.js";
 import "./ops/webgl/gelu.js";
-import "./gelu-uyHP1x1f.js";
+import "./gelu-C-dPj6Ku.js";
 import "./ops/cpu/normRMS.js";
 import "./ops/webgl/normRMS.js";
 import "./ops/grads/normRMS.js";
 import "./ops/webgl/log.js";
-import { selectBackend as $ } from "./backend.js";
-import { default as ot } from "./utilities/performance.js";
+import "./ops/cpu/adamMoments.js";
+import "./ops/webgl/adamMoments.js";
+import "./ops/cpu/adamAdjust.js";
+import "./ops/webgl/adamAdjust.js";
+import { selectBackend as xo } from "./backend.js";
+import { default as co } from "./utilities/performance.js";
+import o from "./layers/CausalSelfAttention.js";
+import r from "./layers/MLP.js";
+import t from "./layers/TransformerBlock.js";
+import m from "./layers/RoPECache.js";
+import { default as ko } from "./training/AdamExt.js";
+const O = {
+  CausalSelfAttention: o,
+  MLP: r,
+  TransformerBlock: t,
+  RoPECache: m
+};
 export {
-  I as BPETokeniser,
-  A as CharTokeniser,
-  R as NanoGPT,
-  q as TeachableLLM,
-  V as estimateMemoryUsage,
-  W as estimateParameterCount,
-  X as estimateResources,
-  Y as estimateTrainingMemoryUsage,
-  Q as loadTextData,
-  ot as performanceTest,
-  $ as selectBackend,
-  Z as validateConfig,
-  K as waitForModel
+  ko as AdamExt,
+  oo as BPETokeniser,
+  _ as CharTokeniser,
+  W as NanoGPT,
+  Y as TeachableLLM,
+  po as estimateMemoryUsage,
+  ao as estimateParameterCount,
+  fo as estimateResources,
+  so as estimateTrainingMemoryUsage,
+  O as layers,
+  eo as loadTextData,
+  co as performanceTest,
+  xo as selectBackend,
+  lo as validateConfig,
+  to as waitForModel
 };

package/dist/{mat_mul-1nwdPkQ_.js → mat_mul-8m8pfdcx.js}

@@ -1,4 +1,4 @@
-import { x as m, y as s, F as c, E as M, H as p } from "./index-C0dhsYom.js";
+import { B as m, C as s, I as c, E as M, J as p } from "./index-BoWRt-10.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.

package/dist/{max-BQc2Aj-I.js → max-Ddnnb5xe.js}

@@ -1,4 +1,4 @@
-import { x as r, y as x, E as e, N as c } from "./index-C0dhsYom.js";
+import { B as r, C as e, E as x, U as c } from "./index-BoWRt-10.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.
@@ -16,8 +16,8 @@ import { x as r, y as x, E as e, N as c } from "./index-C0dhsYom.js";
  * =============================================================================
  */
 function m(n, s = null, o = !1) {
-  const t = { x: x(n, "x", "max") }, a = { reductionIndices: s, keepDims: o };
-  return e.runKernel(c, t, a);
+  const t = { x: e(n, "x", "max") }, a = { reductionIndices: s, keepDims: o };
+  return x.runKernel(c, t, a);
 }
 const l = /* @__PURE__ */ r({ max_: m });
 export {

package/dist/{mulmat_packed_gpu-Gzf3I9UV.js → mulmat_packed_gpu-VSekgsNv.js}

@@ -1,4 +1,4 @@
-import { u as z } from "./gpgpu_math-DJm3ZTAf.js";
+import { u as z } from "./gpgpu_math-DGNLNL4I.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/{ones-D63HpSF_.js → ones-Dj0SDhHf.js}

@@ -1,5 +1,5 @@
-import { C as n, D as t, j as m, E as i } from "./index-C0dhsYom.js";
-import { z as c, c as f } from "./zeros-DCPCdFGq.js";
+import { F as n, H as t, j as m, E as i } from "./index-BoWRt-10.js";
+import { z as c, c as f } from "./zeros--BdLQ3oG.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/ops/adamAdjust.d.ts

@@ -0,0 +1,2 @@
+import { Tensor } from '@tensorflow/tfjs-core';
+export declare function adamAdjust(moments: Tensor, value: Tensor, beta1: number, beta2: number, epsilon: number, learningRate: number): Tensor;

package/dist/ops/adamAdjust.js

@@ -0,0 +1,9 @@
+import { e as i } from "../index-BoWRt-10.js";
+import "./cpu/adamAdjust.js";
+import "./webgl/adamAdjust.js";
+function p(r, t, e, n, m, o) {
+  return i().runKernel("AdamAdjust", { moments: r, value: t }, { beta1: e, beta2: n, epsilon: m, learningRate: o });
+}
+export {
+  p as adamAdjust
+};

package/dist/ops/adamMoments.d.ts

@@ -0,0 +1,2 @@
+import { Tensor } from '@tensorflow/tfjs-core';
+export declare function adamMoments(moments: Tensor, gradient: Tensor, beta1: number, beta2: number): Tensor;

package/dist/ops/adamMoments.js

@@ -0,0 +1,9 @@
+import { e as o } from "../index-BoWRt-10.js";
+import "./cpu/adamMoments.js";
+import "./webgl/adamMoments.js";
+function p(e, n, r, m) {
+  return o().runKernel("AdamMoments", { moments: e, gradient: n }, { beta1: r, beta2: m });
+}
+export {
+  p as adamMoments
+};

package/dist/ops/appendCache.js

@@ -1,8 +1,8 @@
-import { e as a } from "../index-C0dhsYom.js";
+import { e as a } from "../index-BoWRt-10.js";
 import "./cpu/appendCache.js";
 import "./webgl/appendCache.js";
-import { c as s } from "../concat-CsHeR4zV.js";
-import { z as c } from "../zeros-DCPCdFGq.js";
+import { c as s } from "../concat-CsxrgovM.js";
+import { z as c } from "../zeros--BdLQ3oG.js";
 function i(r, p, n, o) {
   if (!o) {
     const e = r.shape[2];

package/dist/ops/attentionMask.js

@@ -1,4 +1,4 @@
-import { e as o } from "../index-C0dhsYom.js";
+import { e as o } from "../index-BoWRt-10.js";
 import "./cpu/attentionMask.js";
 import "./webgl/attentionMask.js";
 import "./grads/attentionMask.js";

package/dist/ops/cpu/adamAdjust.d.ts

@@ -0,0 +1 @@
+export {};

package/dist/ops/cpu/adamAdjust.js

@@ -0,0 +1,18 @@
+import { f as k, n as t, o as i, a as q, q as z } from "../../index-BoWRt-10.js";
+function A(a) {
+  const { moments: n, value: r } = a.inputs, { beta1: l, beta2: u, epsilon: m, learningRate: d } = a.attrs, e = n.shape.length, c = new Array(e).fill(0), s = n.shape.slice();
+  s[e - 1] = 1;
+  const o = c.slice();
+  o[e - 1] = 1;
+  const b = s.slice(), p = n.slice(c, s).squeeze([e - 1]), M = n.slice(o, b).squeeze([e - 1]), f = t(p, l), g = t(M, u);
+  return i(
+    q(t(f, i(z(g), m ?? 1e-8)), -d),
+    r
+  );
+}
+const C = {
+  kernelName: "AdamAdjust",
+  backendName: "cpu",
+  kernelFunc: A
+};
+k(C);

package/dist/ops/cpu/adamMoments.d.ts

@@ -0,0 +1 @@
+export {};

package/dist/ops/cpu/adamMoments.js

@@ -0,0 +1,16 @@
+import { f as p } from "../../index-BoWRt-10.js";
+import { s as f } from "../../stack-B17UN2nn.js";
+function b(t) {
+  const { moments: n, gradient: c } = t.inputs, { beta1: o, beta2: m } = t.attrs, e = n.shape.length, a = new Array(e).fill(0), s = n.shape.slice();
+  s[e - 1] = 1;
+  const i = a.slice();
+  i[e - 1] = 1;
+  const r = s.slice(), l = n.slice(a, s).squeeze([e - 1]), u = n.slice(i, r).squeeze([e - 1]), M = l.mul(o).add(c.mul(1 - o)), d = u.mul(m).add(c.square().mul(1 - m));
+  return f([M, d], -1);
+}
+const g = {
+  kernelName: "AdamMoments",
+  backendName: "cpu",
+  kernelFunc: b
+};
+p(g);

package/dist/ops/cpu/appendCache.js

@@ -1,5 +1,5 @@
-import { f as d } from "../../index-C0dhsYom.js";
-import { c as h } from "../../concat-CsHeR4zV.js";
+import { f as d } from "../../index-BoWRt-10.js";
+import { c as h } from "../../concat-CsxrgovM.js";
 function u(p) {
   const { cache: n, item: s } = p.inputs, { maxSize: i, pastLen: c } = p.attrs, t = n.shape[0], o = n.shape[1], a = n.shape[3], e = s.shape[2];
   if (c + e <= i) {

package/dist/ops/cpu/attentionMask.js

@@ -1,8 +1,8 @@
-import { f as a, h as p, b as u } from "../../index-C0dhsYom.js";
-import { l as N, w as b } from "../../ops-CIQLNshk.js";
-import { o as A } from "../../ones-D63HpSF_.js";
-import { z as I } from "../../zeros-DCPCdFGq.js";
-import { m as g } from "../../mat_mul-1nwdPkQ_.js";
+import { f as a, h as p, b as u } from "../../index-BoWRt-10.js";
+import { l as N, w as b } from "../../ops-BFGCx8Ri.js";
+import { o as A } from "../../ones-Dj0SDhHf.js";
+import { z as I } from "../../zeros--BdLQ3oG.js";
+import { m as g } from "../../mat_mul-8m8pfdcx.js";
 function o(n) {
   const { q: s, k: e } = n.inputs, { divisor: r } = n.attrs, c = s.shape[2], t = e.shape[2], m = N.bandPart(A([t, t]), -1, 0).cast("bool"), l = I([t, t]), i = p([t, t], Number.NEGATIVE_INFINITY), f = b(m, l, i), k = g(s, e, !1, !0).mul(u(r)), d = f.slice([0, 0], [c, t]).expandDims(0).expandDims(0);
   return k.add(d);

package/dist/ops/cpu/fusedSoftmax.js

@@ -1,5 +1,5 @@
-import { f as e } from "../../index-C0dhsYom.js";
-import { s as f } from "../../softmax-GPNK3o-U.js";
+import { f as e } from "../../index-BoWRt-10.js";
+import { s as f } from "../../softmax-Bv_6lyMX.js";
 function n(t) {
   const { inputs: s, attrs: a } = t, { logits: o } = s, { dim: i, dropoutRate: r } = a;
   if (!o)

package/dist/ops/cpu/gatherSub.js

@@ -1,6 +1,6 @@
-import { x as u, y as c, E as g, $ as h, f as m, c as p } from "../../index-C0dhsYom.js";
-import { r as f } from "../../range-CYzpQY53.js";
-import { s as l } from "../../stack-Dpgg_1W1.js";
+import { B as u, C as c, E as g, a5 as h, f as m, c as p } from "../../index-BoWRt-10.js";
+import { r as f } from "../../range-CRuAh-gd.js";
+import { s as l } from "../../stack-B17UN2nn.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/ops/cpu/gelu.js

@@ -1,4 +1,4 @@
-import { f as t, t as d } from "../../index-C0dhsYom.js";
+import { f as t, t as d } from "../../index-BoWRt-10.js";
 const o = 0.7978845608028654, c = 0.044715;
 function m(r) {
   const { inputs: u } = r, { x: n } = u, e = n;

package/dist/ops/cpu/matMulGelu.js

@@ -1,5 +1,5 @@
-import { f as e, t as m } from "../../index-C0dhsYom.js";
-import { g as d, d as M } from "../../gelu-uyHP1x1f.js";
+import { f as e, t as m } from "../../index-BoWRt-10.js";
+import { g as d, d as M } from "../../gelu-C-dPj6Ku.js";
 function c(t) {
   const { inputs: u } = t, { x: n, kernel: l } = u, a = n, r = l;
   return m(() => {

package/dist/ops/cpu/matMulMul.js

@@ -1,4 +1,4 @@
-import { f as e, t as i } from "../../index-C0dhsYom.js";
+import { f as e, t as i } from "../../index-BoWRt-10.js";
 function n(t) {
   const { inputs: r, attrs: l } = t, { transposeA: o, transposeB: s } = l, { x: u, kernel: a, y: c } = r, m = u, k = a, M = c;
   return i(() => m.matMul(k, o, s).mul(M));

package/dist/ops/cpu/mulDropout.js

@@ -1,4 +1,4 @@
-import { f as e, a as t } from "../../index-C0dhsYom.js";
+import { f as e, a as t } from "../../index-BoWRt-10.js";
 function n(o) {
   const { inputs: r } = o, { a: l, b: u } = r;
   return console.warn("Using fallback mulDrop implementation without dropout."), t(l, u);

package/dist/ops/cpu/normRMS.js

@@ -1,4 +1,4 @@
-import { f as o, t as d } from "../../index-C0dhsYom.js";
+import { f as o, t as d } from "../../index-BoWRt-10.js";
 function i(t) {
   const { inputs: e } = t, { x: n, gamma: s } = e, r = n, a = s;
   return d(() => {

package/dist/ops/cpu/qkv.js

@@ -1,6 +1,6 @@
-import { f as q } from "../../index-C0dhsYom.js";
-import { r as o } from "../../reshape-BMUzc1UY.js";
-import { s as x } from "../../split-CHzJjxDv.js";
+import { f as q } from "../../index-BoWRt-10.js";
+import { r as o } from "../../reshape-CdBq1WJ6.js";
+import { s as x } from "../../split-B-dikLRw.js";
 function v(p) {
   const { x: c, kernel: K } = p.inputs, { heads: n } = p.attrs, [s, e, t] = c.shape, a = o(c, [s * e, t]), i = a.dot(K);
   a.dispose();

package/dist/ops/cpu/rope.js

@@ -1,8 +1,8 @@
-import { f as S } from "../../index-C0dhsYom.js";
-import { r as F } from "../../range-CYzpQY53.js";
-import { g as I } from "../../gather-D0_gPiBz.js";
-import { s as E } from "../../stack-Dpgg_1W1.js";
-import { c as T } from "../../concat-CsHeR4zV.js";
+import { f as S } from "../../index-BoWRt-10.js";
+import { r as F } from "../../range-CRuAh-gd.js";
+import { g as I } from "../../gather-CMMy2KEG.js";
+import { s as E } from "../../stack-B17UN2nn.js";
+import { c as T } from "../../concat-CsxrgovM.js";
 function U(t, c, p, o, r) {
   const n = o.shape[3], s = p;
   if (s > n) return o;

package/dist/ops/cpu/scatterSub.js

@@ -1,8 +1,8 @@
-import { x as f, C as g, y as r, E as l, _ as N, f as b, c as S, a as h } from "../../index-C0dhsYom.js";
-import { v as D } from "../../scatter_nd_util-IRBqKz_b.js";
-import { r as k } from "../../range-CYzpQY53.js";
-import { s as v } from "../../stack-Dpgg_1W1.js";
-import { o as E } from "../../ones-D63HpSF_.js";
+import { B as f, F as g, C as r, E as l, a4 as N, f as b, c as S, a as h } from "../../index-BoWRt-10.js";
+import { v as D } from "../../scatter_nd_util-DUstGbU1.js";
+import { r as k } from "../../range-CRuAh-gd.js";
+import { s as v } from "../../stack-B17UN2nn.js";
+import { o as E } from "../../ones-Dj0SDhHf.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.
@@ -26,14 +26,14 @@ function I(a, e, s) {
   const c = { indices: n, updates: t }, o = { shape: s };
   return l.runKernel(N, c, o);
 }
-const y = /* @__PURE__ */ f({ scatterND_: I });
-function C(a) {
-  const { logits: e, labels: s, dy: n } = a.inputs, t = s.shape[0], c = e.shape[1], o = k(0, t, 1, "int32"), i = v([o, s], 1), d = E([t]), u = y(i, d, [t, c]), p = S(e, u), m = n.reshape([t, 1]);
+const C = /* @__PURE__ */ f({ scatterND_: I });
+function K(a) {
+  const { logits: e, labels: s, dy: n } = a.inputs, t = s.shape[0], c = e.shape[1], o = k(0, t, 1, "int32"), i = v([o, s], 1), d = E([t]), u = C(i, d, [t, c]), p = S(e, u), m = n.reshape([t, 1]);
   return h(p, m);
 }
-const K = {
+const L = {
   kernelName: "EfficientScatterSub",
   backendName: "cpu",
-  kernelFunc: C
+  kernelFunc: K
 };
-b(K);
+b(L);

package/dist/ops/fusedSoftmax.js

@@ -1,4 +1,4 @@
-import { e as t } from "../index-C0dhsYom.js";
+import { e as t } from "../index-BoWRt-10.js";
 import "./cpu/fusedSoftmax.js";
 import "./webgl/fusedSoftmax.js";
 import "./grads/fusedSoftmax.js";

package/dist/ops/gatherSub.js

@@ -1,4 +1,4 @@
-import { e as n } from "../index-C0dhsYom.js";
+import { e as n } from "../index-BoWRt-10.js";
 import "./cpu/gatherSub.js";
 import "./webgl/gatherSub.js";
 function f(r, e, t) {