@genai-fi/nanogpt 0.5.6 → 0.6.0

package/dist/main.js

@@ -1,11 +1,11 @@
-import { default as E } from "./NanoGPTModel.js";
-import { default as G } from "./TeachableLLM.js";
-import { default as R } from "./tokeniser/CharTokeniser.js";
-import { default as q } from "./tokeniser/bpe.js";
-import { default as A } from "./utilities/waitForModel.js";
-import { default as I } from "./data/textLoader.js";
-import { estimateMemoryUsage as K, estimateParameterCount as O, estimateResources as Q, estimateTrainingMemoryUsage as S, validateConfig as V } from "./utilities/parameters.js";
-import "./index-Du-bmOP8.js";
+import { default as F } from "./NanoGPTModel.js";
+import { default as N } from "./TeachableLLM.js";
+import { default as j } from "./tokeniser/CharTokeniser.js";
+import { default as z } from "./tokeniser/bpe.js";
+import { default as H } from "./utilities/waitForModel.js";
+import { default as J } from "./data/textLoader.js";
+import { estimateMemoryUsage as O, estimateParameterCount as Q, estimateResources as S, estimateTrainingMemoryUsage as V, validateConfig as W } from "./utilities/parameters.js";
+import "./index-bMBtI-WR.js";
 import "./ops/cpu/scatterSub.js";
 import "./ops/webgl/scatterSub.js";
 import "./ops/cpu/gatherSub.js";
@@ -34,16 +34,17 @@ import "./ops/grads/gelu.js";
 import "./ops/cpu/normRMS.js";
 import "./ops/webgl/normRMS.js";
 import "./ops/grads/normRMS.js";
+import "./ops/webgl/log.js";
 export {
-  q as BPETokeniser,
-  R as CharTokeniser,
-  E as NanoGPT,
-  G as TeachableLLM,
-  K as estimateMemoryUsage,
-  O as estimateParameterCount,
-  Q as estimateResources,
-  S as estimateTrainingMemoryUsage,
-  I as loadTextData,
-  V as validateConfig,
-  A as waitForModel
+  z as BPETokeniser,
+  j as CharTokeniser,
+  F as NanoGPT,
+  N as TeachableLLM,
+  O as estimateMemoryUsage,
+  Q as estimateParameterCount,
+  S as estimateResources,
+  V as estimateTrainingMemoryUsage,
+  J as loadTextData,
+  W as validateConfig,
+  H as waitForModel
 };

package/dist/{mat_mul-CbiqIe2d.js → mat_mul-BsrLfy81.js}

@@ -1,4 +1,4 @@
-import { o as m, j as s, u as c, E as M, B as p } from "./index-Du-bmOP8.js";
+import { o as m, j as s, u as c, E as M, B as p } from "./index-bMBtI-WR.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.

package/dist/{max-0Xnlpv8k.js → max-DechV4Bc.js}

@@ -1,4 +1,4 @@
-import { o as r, j as e, E as x, M as c } from "./index-Du-bmOP8.js";
+import { o as r, j as e, E as x, M as c } from "./index-bMBtI-WR.js";
 /**
  * @license
  * Copyright 2020 Google LLC. All Rights Reserved.

package/dist/{norm-01kY9I2B.js → norm-B9hWHZH1.js}

@@ -1,8 +1,8 @@
-import { o as l, j as c, E as y, a0 as E, a1 as w, a2 as o, a3 as u, W as v, f as I, a4 as A } from "./index-Du-bmOP8.js";
-import { e as $ } from "./axis_util-BczFISHz.js";
-import { m as f } from "./max-0Xnlpv8k.js";
-import { r as h } from "./reshape-BH7eBpwq.js";
-import { s as t } from "./sum-Cvq06317.js";
+import { o as l, j as c, E as y, a0 as E, a1 as w, a2 as o, a3 as u, W as v, f as I, a4 as A } from "./index-bMBtI-WR.js";
+import { e as $ } from "./axis_util-D17qZRQm.js";
+import { m as f } from "./max-DechV4Bc.js";
+import { r as h } from "./reshape-T4yDEqoF.js";
+import { s as t } from "./sum-CutF5lj2.js";
 /**
  * @license
  * Copyright 2020 Google Inc. All Rights Reserved.

package/dist/{ones-CrutWGas.js → ones-g0K8jVwm.js}

@@ -1,5 +1,5 @@
-import { n, p as t, q as m, E as i } from "./index-Du-bmOP8.js";
-import { z as c, c as f } from "./zeros-BaHhQTWf.js";
+import { n, p as t, q as m, E as i } from "./index-bMBtI-WR.js";
+import { z as c, c as f } from "./zeros-CAbHfODe.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/ops/appendCache.js

@@ -1,8 +1,8 @@
-import { e as a } from "../index-Du-bmOP8.js";
+import { e as a } from "../index-bMBtI-WR.js";
 import "./cpu/appendCache.js";
 import "./webgl/appendCache.js";
-import { c as s } from "../concat-DdKPyAtw.js";
-import { z as c } from "../zeros-BaHhQTWf.js";
+import { c as s } from "../concat-DhZfF1GY.js";
+import { z as c } from "../zeros-CAbHfODe.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-Du-bmOP8.js";
+import { e as o } from "../index-bMBtI-WR.js";
 import "./cpu/attentionMask.js";
 import "./webgl/attentionMask.js";
 import "./grads/attentionMask.js";

package/dist/ops/cpu/appendCache.js

@@ -1,5 +1,5 @@
-import { r as d } from "../../index-Du-bmOP8.js";
-import { c as h } from "../../concat-DdKPyAtw.js";
+import { r as d } from "../../index-bMBtI-WR.js";
+import { c as h } from "../../concat-DhZfF1GY.js";
 function u(p) {
   const { cache: n, item: s } = p.inputs, { maxSize: r, pastLen: c } = p.attrs, t = n.shape[0], o = n.shape[1], a = n.shape[3], e = s.shape[2];
   if (c + e <= r) {

package/dist/ops/cpu/attentionMask.js

@@ -1,8 +1,8 @@
-import { r as a, g as p, f as u } from "../../index-Du-bmOP8.js";
-import { l as N, w as b } from "../../ops-CJNniCAV.js";
-import { o as g } from "../../ones-CrutWGas.js";
-import { z as A } from "../../zeros-BaHhQTWf.js";
-import { m as I } from "../../mat_mul-CbiqIe2d.js";
+import { r as a, g as p, f as u } from "../../index-bMBtI-WR.js";
+import { l as N, w as b } from "../../ops-Mv7Ta72x.js";
+import { o as g } from "../../ones-g0K8jVwm.js";
+import { z as A } from "../../zeros-CAbHfODe.js";
+import { m as I } from "../../mat_mul-BsrLfy81.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(g([t, t]), -1, 0).cast("bool"), l = A([t, t]), i = p([t, t], Number.NEGATIVE_INFINITY), f = b(m, l, i), k = I(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 { r as n } from "../../index-Du-bmOP8.js";
-import { s as f } from "../../softmax-DhWoBa7r.js";
+import { r as n } from "../../index-bMBtI-WR.js";
+import { s as f } from "../../softmax-B2_IKPDR.js";
 function r(t) {
   const { inputs: s, attrs: i } = t, { logits: o } = s, { dim: a, dropoutRate: e } = i;
   if (!o)

package/dist/ops/cpu/gatherSub.js

@@ -1,6 +1,6 @@
-import { o as u, j as c, E as g, O as h, r as m, s as p } from "../../index-Du-bmOP8.js";
-import { r as l } from "../../range-Bvs1hidm.js";
-import { s as N } from "../../stack-BV1v7l3S.js";
+import { o as u, j as c, E as g, O as h, r as m, s as p } from "../../index-bMBtI-WR.js";
+import { r as l } from "../../range-DMaG9A3G.js";
+import { s as N } from "../../stack-lpJ5kYvE.js";
 /**
  * @license
  * Copyright 2018 Google LLC. All Rights Reserved.

package/dist/ops/cpu/gelu.js

@@ -1,4 +1,4 @@
-import { r as t, t as d } from "../../index-Du-bmOP8.js";
+import { r as t, t as d } from "../../index-bMBtI-WR.js";
 const o = 0.7978845608028654, c = 0.044715;
 function m(u) {
   const { inputs: l } = u, { x: e } = l, n = e;

package/dist/ops/cpu/matMulGelu.js

@@ -1,4 +1,4 @@
-import { r as a, t as i } from "../../index-Du-bmOP8.js";
+import { r as a, t as i } from "../../index-bMBtI-WR.js";
 const c = 0.7978845608028654, m = 0.044715;
 function M(o) {
   const { inputs: s } = o, { x: t, kernel: l } = s, e = t, u = l;

package/dist/ops/cpu/matMulMul.js

@@ -1,4 +1,4 @@
-import { r as n, t as M } from "../../index-Du-bmOP8.js";
+import { r as n, t as M } from "../../index-bMBtI-WR.js";
 function e(t) {
   const { inputs: r, attrs: o } = t, { transposeA: s, transposeB: l } = o, { x: c, kernel: u, y: a } = r, m = c, i = u, k = a;
   return M(() => m.matMul(i, s, l).mul(k));

package/dist/ops/cpu/mulDropout.js

@@ -1,4 +1,4 @@
-import { r as e, b as u } from "../../index-Du-bmOP8.js";
+import { r as e, b as u } from "../../index-bMBtI-WR.js";
 function n(o) {
   const { inputs: r } = o, { a: l, b: t } = r;
   return console.warn("Using fallback mulDrop implementation without dropout."), u(l, t);

package/dist/ops/cpu/normRMS.js

@@ -1,4 +1,4 @@
-import { r as o, t as d } from "../../index-Du-bmOP8.js";
+import { r as o, t as d } from "../../index-bMBtI-WR.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 { r as q } from "../../index-Du-bmOP8.js";
-import { r as o } from "../../reshape-BH7eBpwq.js";
-import { s as x } from "../../split-BCUhuU7B.js";
+import { r as q } from "../../index-bMBtI-WR.js";
+import { r as o } from "../../reshape-T4yDEqoF.js";
+import { s as x } from "../../split-dcks18H1.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 { r as S } from "../../index-Du-bmOP8.js";
-import { r as F } from "../../range-Bvs1hidm.js";
-import { g as I } from "../../gather-D6MsdXqc.js";
-import { s as E } from "../../stack-BV1v7l3S.js";
-import { c as T } from "../../concat-DdKPyAtw.js";
+import { r as S } from "../../index-bMBtI-WR.js";
+import { r as F } from "../../range-DMaG9A3G.js";
+import { g as I } from "../../gather-DZCMHZuN.js";
+import { s as E } from "../../stack-lpJ5kYvE.js";
+import { c as T } from "../../concat-DhZfF1GY.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

@@ -1,7 +1,7 @@
-import { o as l, n as g, j as h, E as k, a5 as w, r as $, s as d, b as m } from "../../index-Du-bmOP8.js";
-import { r as b } from "../../range-Bvs1hidm.js";
-import { s as E } from "../../stack-BV1v7l3S.js";
-import { o as D } from "../../ones-CrutWGas.js";
+import { o as l, n as g, j as h, E as k, a5 as w, r as $, s as d, b as m } from "../../index-bMBtI-WR.js";
+import { r as b } from "../../range-DMaG9A3G.js";
+import { s as E } from "../../stack-lpJ5kYvE.js";
+import { o as D } from "../../ones-g0K8jVwm.js";
 function N(a, r, t) {
   const s = r.rank > 1 ? r.shape[r.rank - 1] : 1, e = r.rank > 1 ? r.rank - 1 : 1, o = `Must have updates.shape = indices.shape[:batchDim] + shape[sliceDim:], got updates.shape: ${t.shape}, indices.shape: ${r.shape}, shape: ${a}, sliceDim: ${s}, and batchDim: ${e}.`;
   if (t.rank < e)

package/dist/ops/fusedSoftmax.js

@@ -1,4 +1,4 @@
-import { e as t } from "../index-Du-bmOP8.js";
+import { e as t } from "../index-bMBtI-WR.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-Du-bmOP8.js";
+import { e as n } from "../index-bMBtI-WR.js";
 import "./cpu/gatherSub.js";
 import "./webgl/gatherSub.js";
 function f(r, e, t) {

package/dist/ops/gelu.js

@@ -1,4 +1,4 @@
-import "../index-Du-bmOP8.js";
+import "../index-bMBtI-WR.js";
 import "./cpu/gelu.js";
 import "./webgl/gelu.js";
 import { d as e, g as i } from "./grads/gelu.js";

package/dist/ops/grads/attentionMask.js

@@ -1,25 +1,29 @@
-import { h as l, f as a } from "../../index-Du-bmOP8.js";
-import { matMulMul as i } from "../matMulMul.js";
-const m = {
+import { h as m, f as i } from "../../index-bMBtI-WR.js";
+import { matMulMul as u } from "../matMulMul.js";
+const p = {
   kernelName: "AttentionMask",
   inputsToSave: ["q", "k"],
   outputsToSave: [],
-  gradFunc: (t, u, c) => {
+  gradFunc: (t, c, l) => {
     if (Array.isArray(t))
       throw new Error("Expected dy to be a single Tensor");
-    const [e, o] = u, { divisor: n } = c;
+    const [e, n] = c, { divisor: a } = l;
     return {
-      q: () => i(t, o, a(n)),
+      q: () => u(t, n, i(a)),
       k: () => {
-        const s = e.transpose([0, 1, 3, 2]), r = i(s, t, a(n));
-        return s.dispose(), r.transpose([0, 1, 3, 2]);
+        const s = e.transpose([0, 1, 3, 2]), r = u(s, t, i(a));
+        s.dispose();
+        const o = r.transpose([0, 1, 3, 2]);
+        return r.dispose(), o;
       },
       mask: () => t,
       divisor: () => {
-        const s = e.matMul(o, !1, !0), r = t.mul(s);
-        return s.dispose(), r.sum();
+        const s = e.matMul(n, !1, !0), r = t.mul(s);
+        s.dispose();
+        const o = r.sum();
+        return r.dispose(), o;
       }
     };
   }
 };
-l(m);
+m(p);

package/dist/ops/grads/fusedSoftmax.js

@@ -1,20 +1,22 @@
-import { h as d, b as u, s as f } from "../../index-Du-bmOP8.js";
-import { mulDrop as l } from "../mulDrop.js";
-import { s as g } from "../../sum-Cvq06317.js";
-const T = {
+import { h as f, b as i, s as l } from "../../index-bMBtI-WR.js";
+import { mulDrop as g } from "../mulDrop.js";
+import { s as T } from "../../sum-CutF5lj2.js";
+const Y = {
   kernelName: "FusedSoftmax",
   outputsToSave: [!0],
-  gradFunc: (o, i, n) => {
-    const [s] = i, { dim: a, dropoutRate: t, seed: e } = n, p = !0, r = t && e ? l(o, s, t, e) : u(o, s);
+  gradFunc: (o, n, a) => {
+    const [s] = n, { dim: p, dropoutRate: t, seed: e } = a, c = !0, r = t && e ? g(o, s, t, e) : i(o, s);
     return {
       logits: () => {
-        const m = g(r, [a], p), c = u(m, s);
-        return m.dispose(), f(r, c);
+        const m = T(r, [p], c), u = i(m, s);
+        m.dispose();
+        const d = l(r, u);
+        return u.dispose(), d;
       }
     };
   }
 };
-d(T);
+f(Y);
 export {
-  T as softmaxGradConfig
+  Y as softmaxGradConfig
 };

package/dist/ops/grads/gelu.js

@@ -1,4 +1,4 @@
-import { h as t, e as n } from "../../index-Du-bmOP8.js";
+import { h as t, e as n } from "../../index-bMBtI-WR.js";
 import "../cpu/gelu.js";
 import "../webgl/gelu.js";
 const o = {

package/dist/ops/grads/matMulGelu.js

@@ -1,4 +1,4 @@
-import { h as a, e as o } from "../../index-Du-bmOP8.js";
+import { h as a, e as o } from "../../index-bMBtI-WR.js";
 function s(e, n, r) {
   return o().runKernel("MatMulGeluGrad", { dy: e, x: n, kernel: r });
 }

package/dist/ops/grads/normRMS.js

@@ -1,4 +1,4 @@
-import { h as t, e as g } from "../../index-Du-bmOP8.js";
+import { h as t, e as g } from "../../index-bMBtI-WR.js";
 function s(r, a, n) {
   return g().runKernel("RMSNormGrad", { dy: r, x: a, gamma: n });
 }

package/dist/ops/grads/qkv.js

@@ -1,4 +1,4 @@
-import { h as Q } from "../../index-Du-bmOP8.js";
+import { h as Q } from "../../index-bMBtI-WR.js";
 const V = {
   kernelName: "QKV",
   inputsToSave: ["x", "kernel"],

package/dist/ops/grads/rope.js

@@ -1,4 +1,4 @@
-import { h as a, e as i } from "../../index-Du-bmOP8.js";
+import { h as a, e as i } from "../../index-bMBtI-WR.js";
 function p(n, e, s, o) {
   return i().runKernel("Rope", { x: n, sin: e, cos: s }, { pastLen: o });
 }

package/dist/ops/log.js

@@ -0,0 +1 @@
+import "./webgl/log.js";

package/dist/ops/matMulGelu.js

@@ -1,4 +1,4 @@
-import { e as u } from "../index-Du-bmOP8.js";
+import { e as u } from "../index-bMBtI-WR.js";
 import "./cpu/matMulGelu.js";
 import "./webgl/matMulGelu.js";
 import "./grads/matMulGelu.js";

package/dist/ops/matMulMul.js

@@ -1,4 +1,4 @@
-import { e as u } from "../index-Du-bmOP8.js";
+import { e as u } from "../index-bMBtI-WR.js";
 import "./cpu/matMulMul.js";
 import "./webgl/matMulMul.js";
 function m(e, r, t, l = !1, n = !1) {

package/dist/ops/mulDrop.js

@@ -1,4 +1,4 @@
-import { e as t } from "../index-Du-bmOP8.js";
+import { e as t } from "../index-bMBtI-WR.js";
 import "./cpu/mulDropout.js";
 import "./webgl/mulDropout.js";
 function m(r, o, e, n) {

package/dist/ops/node/sparseCrossEntropy.js

@@ -1,4 +1,4 @@
-import { r as o } from "../../index-Du-bmOP8.js";
+import { r as o } from "../../index-bMBtI-WR.js";
 function r(e) {
   const { logits: t, labels: n } = e.inputs;
   return e.backend.executeMultipleOutputs("SparseSoftmaxCrossEntropyWithLogits", [], [t, n], 2);

package/dist/ops/normRMS.js

@@ -1,4 +1,4 @@
-import { e as n } from "../index-Du-bmOP8.js";
+import { e as n } from "../index-bMBtI-WR.js";
 import "./cpu/normRMS.js";
 import "./webgl/normRMS.js";
 import "./grads/normRMS.js";

package/dist/ops/qkv.js

@@ -1,4 +1,4 @@
-import { e as o } from "../index-Du-bmOP8.js";
+import { e as o } from "../index-bMBtI-WR.js";
 import "./cpu/qkv.js";
 import "./webgl/qkv.js";
 import "./grads/qkv.js";

package/dist/ops/scatterSub.js

@@ -1,4 +1,4 @@
-import { e as i } from "../index-Du-bmOP8.js";
+import { e as i } from "../index-bMBtI-WR.js";
 import "./cpu/scatterSub.js";
 import "./webgl/scatterSub.js";
 function c(t, r, e) {

package/dist/ops/webgl/appendCache.js

@@ -1,4 +1,4 @@
-import { r as p } from "../../index-Du-bmOP8.js";
+import { r as p } from "../../index-bMBtI-WR.js";
 class m {
   variableNames = ["cache", "item"];
   outputShape;

package/dist/ops/webgl/attentionMask.js

@@ -1,4 +1,4 @@
-import { r as m } from "../../index-Du-bmOP8.js";
+import { r as m } from "../../index-bMBtI-WR.js";
 class h {
   variableNames = ["q", "k"];
   outputShape;