@genai-fi/nanogpt 0.18.1 → 0.19.1

package/dist/ops/webgpu/normRMSGrad.js

@@ -1,8 +1,8 @@
-import { a as $, a6 as y, e as L } from "../../index-DSGwv2Yx.js";
+import { c as $, a6 as y, h as L } from "../../index-CUXkjxiT.js";
 import { createReduceInfo as M } from "./utils/reductions.js";
-import { f as w } from "../../webgpu_util-DuofJBMo.js";
-import { e as _ } from "../../webgpu_program-CbjdYLYk.js";
-import { p as x, u as R } from "../../pack16-WlOSOuZA.js";
+import { f as w } from "../../webgpu_util-DYlGSwOJ.js";
+import { e as _ } from "../../webgpu_program-B4HmApL1.js";
+import { p as x, u as R } from "../../pack16-bqltoUlR.js";
 import { isPackedTensor as h } from "../../utilities/packed.js";
 import { reshape16 as z } from "../reshape16.js";
 import { sum16 as N } from "../sum16.js";

package/dist/ops/webgpu/pack16.js

@@ -1,14 +1,14 @@
-import { a as i } from "../../index-DSGwv2Yx.js";
+import { c as i } from "../../index-CUXkjxiT.js";
 import p from "./pack16_program.js";
 function m(e) {
-  const { x: n } = e.inputs, { scaling: a, padding: r } = e.attrs, s = e.backend;
+  const { x: n } = e.inputs, { scaling: a, padding: r } = e.attrs, c = e.backend;
   if (n.shape[n.shape.length - 1] % 2 !== 0)
     throw new Error("Last dimension of input tensor must be even to use Pack16.");
   e.attrs && (e.attrs.originalShape = n.shape);
   const t = new p(n.shape, r), o = a !== 1;
   o && t.useScaling();
-  const c = [{ type: "float32", data: [a] }];
-  return s.runWebGPUProgram(t, [n], "packedF16", o ? c : void 0);
+  const s = [{ type: "float32", data: [a] }];
+  return c.runWebGPUProgram(t, [n], "packedF16", o ? s : void 0);
 }
 const k = {
   kernelName: "Pack16",

package/dist/ops/webgpu/pack16_program.js

@@ -1,5 +1,5 @@
-import { f as o, c as a } from "../../webgpu_util-DuofJBMo.js";
-import { e as s } from "../../webgpu_program-CbjdYLYk.js";
+import { f as o, c as a } from "../../webgpu_util-DYlGSwOJ.js";
+import { e as s } from "../../webgpu_program-B4HmApL1.js";
 class h {
   outputShape;
   shaderKey = "Pack16";

package/dist/ops/webgpu/qkv.js

@@ -1,5 +1,5 @@
-import { a as h, a6 as l } from "../../index-DSGwv2Yx.js";
-import { b as f } from "../../matMul16-BIT70Vya.js";
+import { c as h, a6 as l } from "../../index-CUXkjxiT.js";
+import { b as f } from "../../matMul16-BcVC_E62.js";
 import { slice16 as a } from "../slice16.js";
 import { isPackedTensor as u } from "../../utilities/packed.js";
 function k(i) {

package/dist/ops/webgpu/rope.js

@@ -1,7 +1,7 @@
 import { isPackedTensor as y } from "../../utilities/packed.js";
-import { e as c } from "../../webgpu_program-CbjdYLYk.js";
-import { f as x, c as l } from "../../webgpu_util-DuofJBMo.js";
-import { a as w, a6 as b } from "../../index-DSGwv2Yx.js";
+import { e as c } from "../../webgpu_program-B4HmApL1.js";
+import { f as x, c as l } from "../../webgpu_util-DYlGSwOJ.js";
+import { c as w, a6 as b } from "../../index-CUXkjxiT.js";
 class v {
   variableNames = ["x", "sin", "cos"];
   outputShape;

package/dist/ops/webgpu/scatterSub.js

@@ -1,6 +1,6 @@
-import { e as p } from "../../webgpu_program-CbjdYLYk.js";
-import { f as u, c as d } from "../../webgpu_util-DuofJBMo.js";
-import { a as h, a6 as o } from "../../index-DSGwv2Yx.js";
+import { e as p } from "../../webgpu_program-B4HmApL1.js";
+import { f as u, c as d } from "../../webgpu_util-DYlGSwOJ.js";
+import { c as h, a6 as o } from "../../index-CUXkjxiT.js";
 class b {
   variableNames = ["labels", "softmaxProbs", "dy"];
   outputShape;

package/dist/ops/webgpu/slice16.js

@@ -1,7 +1,7 @@
-import { b as u, c as m, e as l } from "../../webgpu_program-CbjdYLYk.js";
-import { f, c as g } from "../../webgpu_util-DuofJBMo.js";
-import { a as S, N as k } from "../../index-DSGwv2Yx.js";
-import { p as y, a as $ } from "../../slice_util-DPY56GzQ.js";
+import { b as u, c as m, e as l } from "../../webgpu_program-B4HmApL1.js";
+import { f, c as g } from "../../webgpu_util-DYlGSwOJ.js";
+import { c as S, U as k } from "../../index-CUXkjxiT.js";
+import { p as y, a as $ } from "../../slice_util-CC35pLmT.js";
 function x(o) {
   switch (o) {
     case 1:

package/dist/ops/webgpu/softmax16.js

@@ -1,12 +1,12 @@
-import { a as S, N as h, e as b } from "../../index-DSGwv2Yx.js";
+import { c as h, U as S, h as b } from "../../index-CUXkjxiT.js";
 import { reshape16 as d } from "../reshape16.js";
 import x from "./softmax16_program.js";
 import k from "./softmax16_subgroup_program.js";
 import l from "./utils/deviceInfo.js";
-import { r as z } from "../../reshape-BIN71H3p.js";
+import { r as z } from "../../reshape-BDOuCSNW.js";
 function F(a) {
   const { inputs: t, backend: o, attrs: p } = a, { logits: e } = t, { dim: r } = p, m = o.subgroupMinSize, i = o.subgroupMaxSize, c = l(o).subgroupsSupported, s = z(e, [
-    h(e.shape) / e.shape[r],
+    S(e.shape) / e.shape[r],
     e.shape[r]
   ]), u = c ? new k(s.shape, m, i) : new x(s.shape), f = o.runWebGPUProgram(u, [s], "packedF16");
   s.dispose();
@@ -18,4 +18,4 @@ const I = {
   backendName: "webgpu",
   kernelFunc: F
 };
-S(I);
+h(I);

package/dist/ops/webgpu/softmax16_program.js

@@ -1,5 +1,5 @@
-import { e } from "../../webgpu_program-CbjdYLYk.js";
-import { f as o } from "../../webgpu_util-DuofJBMo.js";
+import { e } from "../../webgpu_program-B4HmApL1.js";
+import { f as o } from "../../webgpu_util-DYlGSwOJ.js";
 class i {
   variableNames = ["logits"];
   outputShape;

package/dist/ops/webgpu/softmax16_subgroup_program.js

@@ -1,5 +1,5 @@
-import { e as o } from "../../webgpu_program-CbjdYLYk.js";
-import { f as u } from "../../webgpu_util-DuofJBMo.js";
+import { e as o } from "../../webgpu_program-B4HmApL1.js";
+import { f as u } from "../../webgpu_util-DYlGSwOJ.js";
 class i {
   variableNames = ["logits"];
   outputShape;

package/dist/ops/webgpu/softmax16grad.js

@@ -1,4 +1,4 @@
-import { a as p } from "../../index-DSGwv2Yx.js";
+import { c as p } from "../../index-CUXkjxiT.js";
 import { createReduceInfo as m, reduce as l, ReduceProgram as i } from "./utils/reductions.js";
 import { isPackedTensor as n } from "../../utilities/packed.js";
 import k from "./utils/deviceInfo.js";
@@ -23,11 +23,11 @@ class x extends i {
   }
 }
 function b(o) {
-  const { dy: e, softmaxOutput: t } = o.inputs, a = o.backend, s = k(a), u = n(e), c = n(t);
+  const { dy: e, softmaxOutput: t } = o.inputs, r = o.backend, s = k(r), u = n(e), c = n(t);
   if (!(u && c))
     throw new Error("softmaxGradGPU: dy and softmaxOutput must be packed tensors");
-  const r = [e, t], f = m(r, -1), d = new x(s, f);
-  return l(d, r, a);
+  const a = [e, t], f = m(a, -1), d = new x(s, f);
+  return l(d, a, r);
 }
 const v = {
   kernelName: "Softmax16Grad",

package/dist/ops/webgpu/sub16.js

@@ -1,13 +1,13 @@
-import { a as t } from "../../index-DSGwv2Yx.js";
+import { c as t } from "../../index-CUXkjxiT.js";
 import { BinaryOpProgram as s } from "./utils/binary_op.js";
-import { B as p } from "../../binary_op_util-pKXltfxI.js";
-function c(e) {
-  const { a: r, b: n } = e.inputs, a = e.backend, o = new s(p.SUB, r.shape, n.shape);
-  return a.runWebGPUProgram(o, [r, n], "packedF16");
+import { B as c } from "../../binary_op_util-pKXltfxI.js";
+function p(e) {
+  const { a: r, b: n } = e.inputs, o = e.backend, a = new s(c.SUB, r.shape, n.shape);
+  return o.runWebGPUProgram(a, [r, n], "packedF16");
 }
 const m = {
   kernelName: "Sub16",
   backendName: "webgpu",
-  kernelFunc: c
+  kernelFunc: p
 };
 t(m);

package/dist/ops/webgpu/sum16.js

@@ -1,10 +1,10 @@
 import { createReduceInfo as h, reduce as g, ReduceProgram as x } from "./utils/reductions.js";
-import { a as k, ah as A } from "../../index-DSGwv2Yx.js";
+import { c as k, ah as A } from "../../index-CUXkjxiT.js";
 import { isPackedTensor as P } from "../../utilities/packed.js";
 import { transpose16 as b } from "../transpose16.js";
 import I from "./utils/deviceInfo.js";
-import { s as w } from "../../sum-D39FeU5h.js";
-import { a as D, b as K } from "../../axis_util-QWWgLjut.js";
+import { s as w } from "../../sum-BpiwSWvg.js";
+import { a as D, b as K } from "../../axis_util-GTVlo58H.js";
 class v extends x {
   shaderKey = "sum16";
   constructor(e, t, o) {

package/dist/ops/webgpu/transpose16.js

@@ -1,13 +1,13 @@
 import { isPackedTensor as d } from "../../utilities/packed.js";
-import { a as k, e as l } from "../../index-DSGwv2Yx.js";
+import { c as k, h as l } from "../../index-CUXkjxiT.js";
 import { reshape16 as u } from "../reshape16.js";
 import P from "./transpose16_shared_program.js";
 import b from "./transpose16_program.js";
-import { t as T } from "../../transpose-Qxz-4os3.js";
-function w(a) {
-  const { inputs: h, attrs: i } = a, { x: e } = h, { perm: r } = i, p = a.backend, m = d(e);
-  if (m && r[r.length - 1] !== e.shape.length - 1) {
-    const o = e.shape.length, c = o === 4 ? r.map((s) => s - 1).filter((s) => s >= 0) : r, t = o === 4 ? u(e, [e.shape[0] * e.shape[1], e.shape[2], e.shape[3]]) : e, f = new P(t.shape, c), n = p.runWebGPUProgram(f, [t], "packedF16");
+import { t as T } from "../../transpose-9kRxIXWR.js";
+function w(p) {
+  const { inputs: h, attrs: i } = p, { x: e } = h, { perm: r } = i, a = p.backend, c = d(e);
+  if (c && r[r.length - 1] !== e.shape.length - 1) {
+    const o = e.shape.length, m = o === 4 ? r.map((s) => s - 1).filter((s) => s >= 0) : r, t = o === 4 ? u(e, [e.shape[0] * e.shape[1], e.shape[2], e.shape[3]]) : e, f = new P(t.shape, m), n = a.runWebGPUProgram(f, [t], "packedF16");
     if (o === 4) {
       t.dispose();
       const s = l().makeTensorFromTensorInfo(n), g = u(s, [
@@ -20,9 +20,9 @@ function w(a) {
     }
     return n;
   }
-  if (m) {
+  if (c) {
     const o = new b(e.shape, r);
-    return p.runWebGPUProgram(o, [e], "packedF16");
+    return a.runWebGPUProgram(o, [e], "packedF16");
   } else
     return T(e, r);
 }

package/dist/ops/webgpu/transpose16_program.js

@@ -1,5 +1,5 @@
-import { f as a, c as i } from "../../webgpu_util-DuofJBMo.js";
-import { b as h, e as d, c as n } from "../../webgpu_program-CbjdYLYk.js";
+import { f as a, c as i } from "../../webgpu_util-DYlGSwOJ.js";
+import { b as h, e as d, c as n } from "../../webgpu_program-B4HmApL1.js";
 function p(r) {
   const e = r.length;
   if (e > 6)

package/dist/ops/webgpu/transpose16_shared_program.js

@@ -1,6 +1,6 @@
-import { c as a } from "../../webgpu_util-DuofJBMo.js";
-import { e as l } from "../../webgpu_program-CbjdYLYk.js";
-import { v as p } from "../../index-DSGwv2Yx.js";
+import { c as a } from "../../webgpu_util-DYlGSwOJ.js";
+import { e as l } from "../../webgpu_program-B4HmApL1.js";
+import { x as p } from "../../index-CUXkjxiT.js";
 class n {
   variableNames = ["A"];
   outputShape;

package/dist/ops/webgpu/unpack16.js

@@ -1,6 +1,6 @@
-import { f as c, c as r } from "../../webgpu_util-DuofJBMo.js";
-import { e as u } from "../../webgpu_program-CbjdYLYk.js";
-import { a as p } from "../../index-DSGwv2Yx.js";
+import { f as c, c as r } from "../../webgpu_util-DYlGSwOJ.js";
+import { e as u } from "../../webgpu_program-B4HmApL1.js";
+import { c as p } from "../../index-CUXkjxiT.js";
 class l {
   outputShape;
   shaderKey = "Unpack16";

package/dist/ops/webgpu/utils/binary_op.js

@@ -1,8 +1,8 @@
-import { f as r, c as n } from "../../../webgpu_util-DuofJBMo.js";
+import { f as r, c as n } from "../../../webgpu_util-DYlGSwOJ.js";
 import { g as p } from "../../../binary_op_util-pKXltfxI.js";
 import { B as k } from "../../../binary_op_util-pKXltfxI.js";
-import { _ as c, ai as s } from "../../../index-DSGwv2Yx.js";
-import { e as u } from "../../../webgpu_program-CbjdYLYk.js";
+import { _ as c, ai as s } from "../../../index-CUXkjxiT.js";
+import { e as u } from "../../../webgpu_program-B4HmApL1.js";
 class x {
   dispatch;
   dispatchLayout;

package/dist/ops/webgpu/utils/reductions.js

@@ -1,9 +1,9 @@
-import { ah as h, N as S, e as f } from "../../../index-DSGwv2Yx.js";
-import { e as d } from "../../../webgpu_program-CbjdYLYk.js";
+import { ah as h, U as S, h as f } from "../../../index-CUXkjxiT.js";
+import { e as d } from "../../../webgpu_program-B4HmApL1.js";
 import { reshape16 as g } from "../../reshape16.js";
-import { f as z } from "../../../webgpu_util-DuofJBMo.js";
-import { c as k } from "../../../axis_util-QWWgLjut.js";
-import { z as x } from "../../../zeros-Bw0puq_w.js";
+import { f as z } from "../../../webgpu_util-DYlGSwOJ.js";
+import { c as k } from "../../../axis_util-GTVlo58H.js";
+import { z as x } from "../../../zeros-DvZpK8s6.js";
 function c(e, u, t, i) {
   return e && !u ? `
             bestValue = subgroupAdd(bestValue);