@genai-fi/nanogpt 0.18.1 → 0.19.0

package/dist/Generator.js

@@ -1,41 +1,41 @@
 import { E as Ui } from "./index-DvYrXKkX.js";
-import { o as Hi, n as Xi, E as Ki, dl as Ss, a5 as pe, ab as _, as as oo, at as ao, e as Oe, aY as Dt, ay as ro, az as io, au as ji, av as Ft, aD as Ge, ae as co, aG as ws, aH as qi, N as G, af as _e, aI as Yi, A as Ns, aJ as Rs, R as Qi, aj as Ji, v as te, B as lo, _ as Ne, a9 as ee, bb as uo, c7 as Ts, c8 as Es, cO as po, bm as ho, ah as ue, L as Ye, bn as fo, bo as mo, c9 as go, ca as Ds, cb as Fs, cc as Ps, cd as Os, ce as As, bp as xo, ac as nt, cy as Co, cP as bo, cQ as Io, bs as yo, br as ko, bd as $o, de as vo, C as _s, cS as So, ao as wo, x as No, bt as Ro, c2 as $e, cD as To, bu as Eo, cz as Do, cT as Fo, cA as Po, bv as Ls, bw as Vs, bf as Oo, bx as Ao, am as Qe, Q as Zi, by as _o, cB as Lo, cf as Vo, bz as Wo, cF as Mo, cG as Bo, df as Go, cg as zo, bX as ns, bR as St, bU as Ws, cU as yn, dt as ut, du as Uo, cV as kn, dg as ec, K as tc, be as Ho, cW as Xo, bB as Ms, y as Ko, aX as jo, aV as mt, cp as qo, dc as Yo, dd as Qo, bg as Jo, cE as Zo, di as ea, al as ta, G as sa, cq as na, ch as Bs, ci as Gs, cj as zs, dj as oa, b3 as Us, b4 as Hs, bD as Xs, cl as Ks, ck as aa, cY as ra, aM as sc, bE as ia, cC as ca, cZ as la, c_ as ua, dk as da, aP as pa, aZ as ha, cm as fa, bQ as ma, M as js, D as ga, bi as xa, bk as Ca, bj as ba, bF as Ia, d3 as ya, bG as ka, P as $a, a6 as va, bH as Sa, c$ as qs, dv as wa, dw as Na, dx as Ra, X as Ta, cn as Ys, b9 as Ea, d0 as Da, ba as Fa, d1 as Pa, bJ as Oa, bh as Aa, b6 as Qs, ak as _a, dm as La, aL as Va, bL as Js, co as Zs, bM as en, bN as tn, bC as sn, bI as Wa, dy as Ma, dz as Ba, dn as Ga, dp as za, dq as Ua, H as Ha, d2 as Xa, aK as nn, cr as Ka, dr as ja, dA as qa, dB as Ya, cs as on, bq as an, ds as Qa, T as Ja, ct as Za, bl as er, dC as rn, cu as tr, b8 as sr, bO as nr, a as or, dD as ar, dE as $n, aw as vn, ax as nc, t as rr, b as oc, dF as ac, dG as rc, c0 as ic, ar as cc, bP as lc, bV as uc, S as dc, bW as pc, aQ as hc, aq as fc, bS as mc, bT as gc, bY as xc, aS as ir, bA as Cc, aN as bc, bZ as Ic, F as yc, b_ as kc, dh as $c, a$ as vc, b0 as Sc, b1 as wc, b2 as Nc, aO as Rc, b$ as Tc, b5 as Ec, c5 as Dc, ap as Fc, c1 as Pc, aW as cr, bK as Oc, aF as Ac, c3 as _c, b7 as Lc, c4 as Vc, k as Wc } from "./index-DSGwv2Yx.js";
-import { n as Mc } from "./random_width-B_fVXhGx.js";
-import { t as Bc } from "./zeros_like-rOHr54NY.js";
+import { o as Hi, q as Xi, E as Ki, dl as Ss, a5 as pe, ab as _, as as oo, at as ao, h as Oe, aY as Dt, ay as ro, az as io, au as ji, av as Ft, aD as Ge, ae as co, aG as ws, aH as qi, U as G, af as _e, aI as Yi, H as Ns, aJ as Rs, R as Qi, aj as Ji, x as te, D as lo, _ as Ne, a9 as ee, bb as uo, c7 as Ts, c8 as Es, cO as po, bm as ho, ah as ue, Q as Ye, bn as fo, bo as mo, c9 as go, ca as Ds, cb as Fs, cc as Ps, cd as Os, ce as As, bp as xo, ac as nt, cy as Co, cP as bo, cQ as Io, bs as yo, br as ko, bd as $o, de as vo, C as _s, cS as So, ao as wo, z as No, bt as Ro, c2 as $e, cD as To, bu as Eo, cz as Do, cT as Fo, cA as Po, bv as Ls, bw as Vs, bf as Oo, bx as Ao, am as Qe, V as Zi, by as _o, cB as Lo, cf as Vo, bz as Wo, cF as Mo, cG as Bo, df as Go, cg as zo, bX as ns, bR as St, bU as Ws, cU as yn, dt as ut, du as Uo, cV as kn, dg as ec, N as tc, be as Ho, cW as Xo, bB as Ms, A as Ko, aX as jo, aV as mt, cp as qo, dc as Yo, dd as Qo, bg as Jo, cE as Zo, di as ea, al as ta, G as sa, cq as na, ch as Bs, ci as Gs, cj as zs, dj as oa, b3 as Us, b4 as Hs, bD as Xs, cl as Ks, ck as aa, cY as ra, aM as sc, bE as ia, cC as ca, cZ as la, c_ as ua, dk as da, aP as pa, aZ as ha, cm as fa, bQ as ma, M as js, I as ga, bi as xa, bk as Ca, bj as ba, bF as Ia, d3 as ya, bG as ka, P as $a, a6 as va, bH as Sa, c$ as qs, dv as wa, dw as Na, dx as Ra, Z as Ta, cn as Ys, b9 as Ea, d0 as Da, ba as Fa, d1 as Pa, bJ as Oa, bh as Aa, b6 as Qs, ak as _a, dm as La, aL as Va, bL as Js, co as Zs, bM as en, bN as tn, bC as sn, bI as Wa, dy as Ma, dz as Ba, dn as Ga, dp as za, dq as Ua, J as Ha, d2 as Xa, aK as nn, cr as Ka, dr as ja, dA as qa, dB as Ya, cs as on, bq as an, ds as Qa, T as Ja, ct as Za, bl as er, dC as rn, cu as tr, b8 as sr, bO as nr, c as or, dD as ar, dE as $n, aw as vn, ax as nc, t as rr, a as oc, dF as ac, dG as rc, c0 as ic, ar as cc, bP as lc, bV as uc, S as dc, bW as pc, aQ as hc, aq as fc, bS as mc, bT as gc, bY as xc, aS as ir, bA as Cc, aN as bc, bZ as Ic, F as yc, b_ as kc, dh as $c, a$ as vc, b0 as Sc, b1 as wc, b2 as Nc, aO as Rc, b$ as Tc, b5 as Ec, c5 as Dc, ap as Fc, c1 as Pc, aW as cr, bK as Oc, aF as Ac, c3 as _c, b7 as Lc, c4 as Vc, k as Wc } from "./index-CUXkjxiT.js";
+import { n as Mc } from "./random_width-DN5ZtQkM.js";
+import { t as Bc } from "./zeros_like-CWjDdwr-.js";
 import "./index-Cp39cXWe.js";
-import "./dataset-DlqAN81i.js";
-import { a as j, u as ae, c as ot, i as at, b as Gc, d as wt, t as Re, e as gt, f as dt, g as lr, r as Nt, h as Ae, j as zc, k as Uc, l as cn, z as Hc, m as ln, n as ur, o as Xc, p as Kc, q as jc, v as qc, w as Yc, x as Qc, y as Jc, A as Zc, B as el, C as tl, D as lt, E as sl, F as nl, G as dr, H as ol, I as al, J as rl, K as il, L as cl, M as ll, N as ul, O as dl, P as pl, Q as hl, R as fl, S as ml, T as gl, U as xl, V as Cl, W as bl, X as Il, Y as yl, Z as kl, _ as $l, $ as vl, a0 as Sl, a1 as wl, a2 as Nl, a3 as Rl, a4 as Tl, a5 as El, a6 as Dl, a7 as Fl, a8 as Pl, a9 as Ol, aa as Al, ab as _l, ac as Ll, ad as Vl, ae as Wl, af as Ml, ag as Bl, ah as Gl, ai as zl } from "./shared-DRWDyk9w.js";
+import "./dataset-CGGp1z9P.js";
+import { a as j, u as ae, c as ot, i as at, b as Gc, d as wt, t as Re, e as gt, f as dt, g as lr, r as Nt, h as Ae, j as zc, k as Uc, l as cn, z as Hc, m as ln, n as ur, o as Xc, p as Kc, q as jc, v as qc, w as Yc, x as Qc, y as Jc, A as Zc, B as el, C as tl, D as lt, E as sl, F as nl, G as dr, H as ol, I as al, J as rl, K as il, L as cl, M as ll, N as ul, O as dl, P as pl, Q as hl, R as fl, S as ml, T as gl, U as xl, V as Cl, W as bl, X as Il, Y as yl, Z as kl, _ as $l, $ as vl, a0 as Sl, a1 as wl, a2 as Nl, a3 as Rl, a4 as Tl, a5 as El, a6 as Dl, a7 as Fl, a8 as Pl, a9 as Ol, aa as Al, ab as _l, ac as Ll, ad as Vl, ae as Wl, af as Ml, ag as Bl, ah as Gl, ai as zl } from "./shared-CagdqkLh.js";
 import { m as pt, g as pr, s as Ul, c as Hl, b as Xl, d as Kl, a as jl, e as ql } from "./complex_util-Yc1A_gV1.js";
-import { a as ge, b as xe, d as ke, c as ve, e as Te, g as os } from "./axis_util-QWWgLjut.js";
-import { k as Je, h as Le, i as Ze, j as rt, b as Se, d as xt, g as as } from "./step-DQY6_ABw.js";
-import { z as rs, A as is, B as cs, C as hr, D as fr, F as mr, G as gr, H as xr, I as Cr, J as br, y as Ir, x as yr, w as kr, u as $r, t as vr, E as Sr, K as wr, L as Nr, M as Rr, c as Tr, f as Yl } from "./backend_util-qwSFfxYx.js";
-import { a as Er, c as Ue } from "./concat_util-C1Mxe27t.js";
-import { c as Dr, a as Ql, s as Jl } from "./segment_util-4zuHV5IG.js";
+import { a as ge, b as xe, d as ke, c as ve, e as Te, g as os } from "./axis_util-GTVlo58H.js";
+import { k as Je, h as Le, i as Ze, j as rt, b as Se, d as xt, g as as } from "./step-BTxPtq1r.js";
+import { z as rs, A as is, B as cs, C as hr, D as fr, F as mr, G as gr, H as xr, I as Cr, J as br, y as Ir, x as yr, w as kr, u as $r, t as vr, E as Sr, K as wr, L as Nr, M as Rr, c as Tr, f as Yl } from "./backend_util-GaFarB78.js";
+import { a as Er, c as Ue } from "./concat_util-D18dJ4fD.js";
+import { c as Dr, a as Ql, s as Jl } from "./segment_util-Dasb2Zaf.js";
 import { s as Zl } from "./index-CieiGp4Y.js";
 import { n as Fr, b as Pr, a as Or } from "./non_max_suppression_impl-B2W7YjZB.js";
-import { c as Ct } from "./scatter_nd_util-C-x73Cj6.js";
-import { S as Ar, a as _r } from "./selu_util-BXdhy_W6.js";
-import { s as Lr, b as Vr, p as eu, a as tu, i as su, c as nu } from "./slice_util-DPY56GzQ.js";
-import { h as Sn, j as ou, k as au, l as ru, m as iu, n as cu, o as lu, P as un, p as Ve, u as Pe, q as Wr, c as Mr, T as De, E as Br, g as Gr, a as zr, r as uu, s as du, t as Y, v as pu, w as wn, x as hu, y as fu, z as Pt, A as Ot, B as mu, C as gu, D as bs, F as Gt, G as zt, H as xu, I as Cu, J as Nn, K as bu, L as Iu, M as fs, N as yu, O as ku, Q as $u, R as Ut, S as ms, U as vu, f as he, V as be, W as Ht, X as Xt, Y as Su, d as Rn, e as Tn, i as Ur, Z as wu, _ as Nu, $ as Ru, a0 as Tu, a1 as Eu, a2 as Du, a3 as At } from "./gpgpu_math-3bCb5ooU.js";
-import { s as Hr, a as Fu, t as Xr, b as Pu, c as Ou, d as Kr, e as Au, n as _u, f as Lu, g as Vu, h as Wu, i as Mu, j as Bu, k as Gu, l as zu, o as Uu, p as Hu, q as Xu, r as Ku, u as ju, v as qu, w as Yu, x as Qu, y as Ju, z as Zu, A as ed, B as td, C as sd, D as nd, E as od, F as ad, G as rd, H as id, I as cd, J as ld, K as ud, L as dd, M as jr, N as pd, O as hd, P as fd, Q as md, R as gd, S as xd, T as Cd, U as bd, V as Id, W as yd } from "./shared-zTaJ5siv.js";
-import { a as ye, c as kd, U as st, d as qe, e as ze, A as En, f as bt, B as dn, h as pn, m as Rt, u as se, C as We, b as Ce, i as Fe, j as hn, k as it, l as It, n as $d, o as vd, p as Sd, q as wd } from "./kernel_funcs_utils-DGqzNlHT.js";
-import { R as Nd, r as U, a as Rd } from "./Reshape-BZC-ebeR.js";
-import { M as qr } from "./matMulGelu-CsZnh18H.js";
-import { t as Yr, s as fn, a as _t, m as Td, r as Ed, b as Dd, c as Fd, d as Pd } from "./RealDiv-ioj6Z-ox.js";
-import { z as Od } from "./zeros-Bw0puq_w.js";
+import { c as Ct } from "./scatter_nd_util-ByNJaL6I.js";
+import { S as Ar, a as _r } from "./selu_util-BLhIqRkw.js";
+import { s as Lr, b as Vr, p as eu, a as tu, i as su, c as nu } from "./slice_util-CC35pLmT.js";
+import { h as Sn, j as ou, k as au, l as ru, m as iu, n as cu, o as lu, P as un, p as Ve, u as Pe, q as Wr, c as Mr, T as De, E as Br, g as Gr, a as zr, r as uu, s as du, t as Y, v as pu, w as wn, x as hu, y as fu, z as Pt, A as Ot, B as mu, C as gu, D as bs, F as Gt, G as zt, H as xu, I as Cu, J as Nn, K as bu, L as Iu, M as fs, N as yu, O as ku, Q as $u, R as Ut, S as ms, U as vu, f as he, V as be, W as Ht, X as Xt, Y as Su, d as Rn, e as Tn, i as Ur, Z as wu, _ as Nu, $ as Ru, a0 as Tu, a1 as Eu, a2 as Du, a3 as At } from "./gpgpu_math-BwvV12df.js";
+import { s as Hr, a as Fu, t as Xr, b as Pu, c as Ou, d as Kr, e as Au, n as _u, f as Lu, g as Vu, h as Wu, i as Mu, j as Bu, k as Gu, l as zu, o as Uu, p as Hu, q as Xu, r as Ku, u as ju, v as qu, w as Yu, x as Qu, y as Ju, z as Zu, A as ed, B as td, C as sd, D as nd, E as od, F as ad, G as rd, H as id, I as cd, J as ld, K as ud, L as dd, M as jr, N as pd, O as hd, P as fd, Q as md, R as gd, S as xd, T as Cd, U as bd, V as Id, W as yd } from "./shared-3agzAqQ_.js";
+import { a as ye, c as kd, U as st, d as qe, e as ze, A as En, f as bt, B as dn, h as pn, m as Rt, u as se, C as We, b as Ce, i as Fe, j as hn, k as it, l as It, n as $d, o as vd, p as Sd, q as wd } from "./kernel_funcs_utils-pq0CK9co.js";
+import { R as Nd, r as U, a as Rd } from "./Reshape-BW__R4mZ.js";
+import { M as qr } from "./matMulGelu-JNLZqKQp.js";
+import { t as Yr, s as fn, a as _t, m as Td, r as Ed, b as Dd, c as Fd, d as Pd } from "./RealDiv-CGwv0liw.js";
+import { z as Od } from "./zeros-DvZpK8s6.js";
 import "./ops/cpu/attentionMask.js";
 import "./ops/webgl/attentionMask.js";
 import "./ops/grads/attentionMask.js";
 import "./ops/cpu/rope.js";
 import "./ops/webgl/rope.js";
-import "./rope-CC5RjmKU.js";
+import "./rope-DcrZM_e6.js";
 import "./ops/cpu/appendCache.js";
 import "./ops/webgl/appendCache.js";
 import "./ops/grads/softmax16.js";
-import "./matMul16-BIT70Vya.js";
+import "./matMul16-BcVC_E62.js";
 import "./ops/webgl/matMul16.js";
 import "./ops/cpu/matMul16.js";
-import "./pack16-WlOSOuZA.js";
+import "./pack16-bqltoUlR.js";
 import "./ops/transpose16.js";
 import "./ops/reshape16.js";
 import "./ops/cpu/qkv.js";
@@ -65,17 +65,17 @@ import "./ops/cpu/matMulGelu.js";
 import "./ops/grads/matMulGelu.js";
 import "./ops/cpu/gelu.js";
 import "./ops/webgl/gelu.js";
-import "./gelu-DpTCC3eB.js";
+import "./gelu-B220X1Go.js";
 import "./ops/webgl/log.js";
 import "./checks/normRMS.js";
 import "./checks/normRMSGrad.js";
 import Wd from "./utilities/multinomialCPU.js";
-import { r as Dn } from "./reshape-BIN71H3p.js";
-import { t as Kt } from "./tensor2d-DTtQ1QcT.js";
-import { z as Md } from "./unsorted_segment_sum-BfFVV9Zm.js";
-import { s as gs } from "./softmax-BLGJqdwx.js";
-import { g as Bd } from "./gather-DEyjXNb1.js";
-import { c as Gd } from "./concat-Dqk7Xk7h.js";
+import { r as Dn } from "./reshape-BDOuCSNW.js";
+import { t as Kt } from "./tensor2d-BnXMKScO.js";
+import { z as Md } from "./unsorted_segment_sum-DJvk5xnh.js";
+import { s as gs } from "./softmax-D4q1LJN7.js";
+import { g as Bd } from "./gather-qIqEqaGn.js";
+import { c as Gd } from "./concat-CAQpCret.js";
 function zd(a, t, e, n = !1) {
   const s = Xi(a, "logits", "multinomial"), o = s.size, r = s.rank;
   if (o < 2)

package/dist/{RealDiv-ioj6Z-ox.js → RealDiv-CGwv0liw.js}

@@ -1,10 +1,10 @@
-import { aE as E, ab as T, ah as O, N as V, aW as B, K as F, aM as K, aX as W } from "./index-DSGwv2Yx.js";
-import { r as $ } from "./Reshape-BZC-ebeR.js";
-import { a as A, b as k, d as C, c as N, e as R } from "./axis_util-QWWgLjut.js";
-import { t as U, m as _ } from "./shared-zTaJ5siv.js";
-import { c as j } from "./backend_util-qwSFfxYx.js";
-import { f as y } from "./gpgpu_math-3bCb5ooU.js";
-import { g as G, b as L } from "./kernel_funcs_utils-DGqzNlHT.js";
+import { aE as E, ab as T, ah as O, U as V, aW as B, N as F, aM as U, aX as W } from "./index-CUXkjxiT.js";
+import { r as $ } from "./Reshape-BW__R4mZ.js";
+import { a as A, b as k, d as C, c as N, e as R } from "./axis_util-GTVlo58H.js";
+import { t as K, m as _ } from "./shared-3agzAqQ_.js";
+import { c as j } from "./backend_util-GaFarB78.js";
+import { f as y } from "./gpgpu_math-BwvV12df.js";
+import { g as G, b as L } from "./kernel_funcs_utils-pq0CK9co.js";
 class w {
   constructor(s, e) {
     this.variableNames = ["x"];
@@ -299,7 +299,7 @@ function te(a) {
       const I = e.texData.get(d.dataId).values, m = new Array(i);
       for (let v = 0; v < m.length; v++)
         m[v] = n.shape[u[v]];
-      const z = U(I, n.shape, n.dtype, u, m);
+      const z = K(I, n.shape, n.dtype, u, m);
       d = e.makeTensorInfo(m, n.dtype);
       const D = e.texData.get(d.dataId);
       D.values = z;
@@ -322,7 +322,7 @@ function te(a) {
   return p && e.disposeIntermediateTensorInfo(d), x;
 }
 const he = {
-  kernelName: K,
+  kernelName: U,
   backendName: "webgl",
   kernelFunc: te
 };

package/dist/{Reshape-BZC-ebeR.js → Reshape-BW__R4mZ.js}

@@ -1,6 +1,6 @@
-import { R as C, N as c, aj as R, v as f } from "./index-DSGwv2Yx.js";
-import { u as g, g as I, a as x, b as F, c as $, d as u, e as m, i as l } from "./gpgpu_math-3bCb5ooU.js";
-class v {
+import { R as C, U as c, aj as R, x as f } from "./index-CUXkjxiT.js";
+import { u as g, g as I, a as x, b as F, c as $, d as u, e as m, i as l } from "./gpgpu_math-BwvV12df.js";
+class S {
   constructor(t, i) {
     this.variableNames = ["A"], this.packedInputs = !0, this.packedOutput = !0, this.customUniforms = [{ name: "inputShape", type: "ivec3" }], this.outputShape = t, this.enableShapeUniforms = g(this.outputShape.length);
     let a = "";
@@ -20,7 +20,7 @@ class v {
       `;
     }
     this.userCode = `
-      ${S(i, this.enableShapeUniforms)}
+      ${v(i, this.enableShapeUniforms)}
       ${this.enableShapeUniforms ? I() : x(t)}
 
       void main() {
@@ -39,7 +39,7 @@ class v {
     `;
   }
 }
-function S(s, t) {
+function v(s, t) {
   return `
     ivec3 inputCoordsFromReshapedOutCoords(int index) {
       ${t ? F(["r", "c", "d"], "inputShape") : $(["r", "c", "d"], s)}
@@ -58,7 +58,7 @@ function b(s, t, i) {
   }, o = [
     u(t),
     ...m(t)
-  ], r = new v(o, a), p = !0, n = [a], h = i.runWebGLProgram(r, [e], s.dtype, n, p);
+  ], r = new S(o, a), p = !0, n = [a], h = i.runWebGLProgram(r, [e], s.dtype, n, p);
   return { dataId: h.dataId, shape: t, dtype: h.dtype };
 }
 function y(s) {
@@ -73,7 +73,7 @@ const U = {
   kernelFunc: y
 };
 export {
-  v as R,
+  S as R,
   U as a,
   y as r
 };

package/dist/{Reshape-pwprEaej.js → Reshape-CPBkTIH2.js}

@@ -1,4 +1,4 @@
-import { N as h, aj as d, v as c, R as m } from "./index-DSGwv2Yx.js";
+import { U as h, aj as d, x as c, R as m } from "./index-CUXkjxiT.js";
 function i(n) {
   const { inputs: p, attrs: o } = n, { x: e } = p, { shape: r } = o, a = h(e.shape), s = d(r, a), t = h(s);
   return c(a === t, () => `The new shape (${s}) has ${t} elements and the old shape (${e.shape}) has ${a} elements. The new shape and old shape must have the same number of elements.`), n.backend.incRef(e.dataId), { dataId: e.dataId, shape: s, dtype: e.dtype };

package/dist/TeachableLLM.js

@@ -1,28 +1,28 @@
-import { validateConfig as l } from "./models/config.js";
+import { validateConfig as _ } from "./models/config.js";
 import { saveModel as p } from "./loader/save.js";
 import { loadModel as u } from "./loader/load.js";
 import f from "./Generator.js";
 import h from "./Trainer.js";
 import { E as c } from "./index-DvYrXKkX.js";
 import { dummyPassTrainAsync as d } from "./utilities/dummy.js";
-import "./index-DSGwv2Yx.js";
-import "./random_width-B_fVXhGx.js";
-import "./zeros_like-rOHr54NY.js";
+import "./index-CUXkjxiT.js";
+import "./random_width-DN5ZtQkM.js";
+import "./zeros_like-CWjDdwr-.js";
 import "./index-Cp39cXWe.js";
-import "./dataset-DlqAN81i.js";
+import "./dataset-CGGp1z9P.js";
 import "./ops/cpu/attentionMask.js";
 import "./ops/webgl/attentionMask.js";
 import "./ops/grads/attentionMask.js";
 import "./ops/cpu/rope.js";
 import "./ops/webgl/rope.js";
-import "./rope-CC5RjmKU.js";
+import "./rope-DcrZM_e6.js";
 import "./ops/cpu/appendCache.js";
 import "./ops/webgl/appendCache.js";
 import "./ops/grads/softmax16.js";
-import "./matMul16-BIT70Vya.js";
+import "./matMul16-BcVC_E62.js";
 import "./ops/webgl/matMul16.js";
 import "./ops/cpu/matMul16.js";
-import "./pack16-WlOSOuZA.js";
+import "./pack16-bqltoUlR.js";
 import "./ops/transpose16.js";
 import "./ops/reshape16.js";
 import "./ops/cpu/qkv.js";
@@ -43,11 +43,11 @@ import "./ops/webgl/gatherSub.js";
 import "./ops/cpu/scatterSub.js";
 import "./ops/webgl/scatterSub.js";
 import "./ops/cpu/matMulGelu.js";
-import "./matMulGelu-CsZnh18H.js";
+import "./matMulGelu-JNLZqKQp.js";
 import "./ops/grads/matMulGelu.js";
 import "./ops/cpu/gelu.js";
 import "./ops/webgl/gelu.js";
-import "./gelu-DpTCC3eB.js";
+import "./gelu-B220X1Go.js";
 import "./ops/webgl/log.js";
 import "./ops/cpu/adamMoments.js";
 import "./ops/webgl/adamMoments.js";
@@ -57,7 +57,7 @@ import "./checks/normRMS.js";
 import "./checks/normRMSGrad.js";
 import k from "./utilities/profile.js";
 import y from "./models/factory.js";
-class _ {
+class l {
   ee = new c();
   _config;
   _model;
@@ -133,7 +133,7 @@ class _ {
   attachLoRA(t) {
     if (!this._model)
       throw new Error("model_not_initialized.");
-    this._model.attachLoRA(t), this.ee.emit("changeLoRA");
+    this.model.lora?.name !== t && (this._trainer && (this._trainer.dispose(), this._trainer = null), this._model.attachLoRA(t), this.ee.emit("changeLoRA"));
   }
   detachLoRA() {
     if (!this._model)
@@ -171,9 +171,9 @@ class _ {
     );
   }
   static loadModel(t, i) {
-    const e = new _();
+    const e = new l();
     return u(t, i).then(({ model: o, tokeniser: s, metaData: r, optimizer: n, log: a }) => {
-      l(o.config), e._model = o, e._tokeniser = s, e._config = o.config, r && (e.meta = r), e.setStatus("warmup"), d(o).then((m) => {
+      _(o.config), e._model = o, e._tokeniser = s, e._config = o.config, r && (e.meta = r), e.setStatus("warmup"), d(o).then((m) => {
         e._memoryRequirements = m, n && (e._trainer = new h(
           o,
           s,
@@ -189,8 +189,8 @@ class _ {
     }), e;
   }
   static create(t, i) {
-    l(i);
-    const e = i, o = t === "char" ? new g(e.vocabSize) : new w(e.vocabSize), s = y(e), r = new _(o, s);
+    _(i);
+    const e = i, o = t === "char" ? new g(e.vocabSize) : new w(e.vocabSize), s = y(e), r = new l(o, s);
     return r.setStatus("warmup"), d(s).then((n) => {
       r._memoryRequirements = n, r.tokeniser.trained ? (r.setStatus("ready"), r.ee.emit("loaded"), r.ee.emit("phase", r.phase)) : (r.setStatus("awaitingTokens"), r.ee.emit("loaded"), r.ee.emit("phase", r.phase), r.tokeniser.once("trainStatus", (a) => {
         a === "trained" && r.setStatus("ready");
@@ -269,5 +269,5 @@ class _ {
   }
 }
 export {
-  _ as default
+  l as default
 };

package/dist/Trainer.js

@@ -1,29 +1,28 @@
-import { E as m } from "./index-DvYrXKkX.js";
-import g from "./training/PreTrainer.js";
-import { createTrainValidationSplit as u } from "./training/validation.js";
-import c from "./training/SFTTrainer.js";
-import p from "./training/tasks/splitter.js";
-const r = [];
+import { E as w } from "./index-DvYrXKkX.js";
+import c from "./training/PreTrainer.js";
+import { createTrainValidationSplit as f } from "./training/validation.js";
+import g from "./training/SFTTrainer.js";
+const s = [];
 for (let n = 0; n < 256; ++n)
-  r.push((n + 256).toString(16).slice(1));
-function w(n, t = 0) {
-  return (r[n[t + 0]] + r[n[t + 1]] + r[n[t + 2]] + r[n[t + 3]] + "-" + r[n[t + 4]] + r[n[t + 5]] + "-" + r[n[t + 6]] + r[n[t + 7]] + "-" + r[n[t + 8]] + r[n[t + 9]] + "-" + r[n[t + 10]] + r[n[t + 11]] + r[n[t + 12]] + r[n[t + 13]] + r[n[t + 14]] + r[n[t + 15]]).toLowerCase();
+  s.push((n + 256).toString(16).slice(1));
+function T(n, t = 0) {
+  return (s[n[t + 0]] + s[n[t + 1]] + s[n[t + 2]] + s[n[t + 3]] + "-" + s[n[t + 4]] + s[n[t + 5]] + "-" + s[n[t + 6]] + s[n[t + 7]] + "-" + s[n[t + 8]] + s[n[t + 9]] + "-" + s[n[t + 10]] + s[n[t + 11]] + s[n[t + 12]] + s[n[t + 13]] + s[n[t + 14]] + s[n[t + 15]]).toLowerCase();
 }
-const T = new Uint8Array(16);
-function D() {
-  return crypto.getRandomValues(T);
+const p = new Uint8Array(16);
+function k() {
+  return crypto.getRandomValues(p);
 }
-function d(n, t, a) {
-  return crypto.randomUUID ? crypto.randomUUID() : k(n);
+function m(n, t, e) {
+  return crypto.randomUUID ? crypto.randomUUID() : L(n);
 }
-function k(n, t, a) {
+function L(n, t, e) {
   n = n || {};
-  const i = n.random ?? n.rng?.() ?? D();
+  const i = n.random ?? n.rng?.() ?? k();
   if (i.length < 16)
     throw new Error("Random bytes length must be >= 16");
-  return i[6] = i[6] & 15 | 64, i[8] = i[8] & 63 | 128, w(i);
+  return i[6] = i[6] & 15 | 64, i[8] = i[8] & 63 | 128, T(i);
 }
-class f extends m {
+class u extends w {
   trainer;
   trainingType = "pretraining";
   hasTrained = !1;
@@ -39,21 +38,34 @@ class f extends m {
     logInterval: 10
   };
   tokenizer;
-  constructor(t, a, i = "pretraining", e, s) {
-    if (super(), t instanceof f) {
-      const o = a || t.options, h = t.options;
+  constructor(t, e, i, a, o) {
+    if (super(), t instanceof u) {
+      const r = e || t.options, h = t.options;
       let l = !1;
-      t.trainingType === "sft" && o.sftMode !== h.sftMode && (l = !0), i !== t.trainingType && (l = !0), l ? (t.trainingType === "sft" ? this.trainer = new c(t.model, t.tokenizer, o) : this.trainer = new g(t.model, t.tokenizer, o), this.trainingType = i, this.options = o, this.tokenizer = t.tokenizer) : (this.trainer = t.trainer, this.trainingType = i, this.options = o, this.trainer.updateOptimizer(this.options), this.log = t.log, this.progress = t.progress, this.totalTokens = t.totalTokens, this.tokenizer = t.tokenizer, o.batchSize === h.batchSize && (this.trainDataset = t.trainDataset, this.validationDataset = t.validationDataset));
+      if (t.trainingType === "sft" && r.sftMode !== h.sftMode && (l = !0), t.trainer instanceof g && t.trainer.loraName && r.loraName !== t.trainer.loraName && (l = !0), i !== void 0 && i !== t.trainingType && (l = !0), l) {
+        if (t.trainingType === "sft") {
+          const d = new g(t.model, t.tokenizer, r);
+          this.trainer = d, d.loraName = r.loraName;
+        } else
+          this.trainer = new c(t.model, t.tokenizer, r);
+        this.trainingType = i || t.trainingType, this.options = r, this.tokenizer = t.tokenizer;
+      } else
+        this.trainer = t.trainer, this.trainingType = i || t.trainingType, this.options = r, this.trainer.updateOptimizer(this.options), this.log = t.log, this.progress = t.progress, this.totalTokens = t.totalTokens, this.tokenizer = t.tokenizer, r.batchSize === h.batchSize && (this.trainDataset = t.trainDataset, this.validationDataset = t.validationDataset);
       return;
     }
-    if (!a)
+    if (!e)
       throw new Error("Tokeniser must be provided when initializing Trainer with a model");
     if (!t)
       throw new Error("Model must be provided when initializing Trainer");
-    this.options = e || {
+    if (this.options = a || {
       batchSize: 32,
       sftMode: "full"
-    }, i === "sft" ? this.trainer = new c(t, a, e, s) : this.trainer = new g(t, a, e, s), this.trainingType = i, this.tokenizer = a;
+    }, i === "sft") {
+      const r = new g(t, e, a, o);
+      this.trainer = r, r.loraName = a?.loraName;
+    } else
+      this.trainer = new c(t, e, a, o);
+    this.trainingType = i || "pretraining", this.tokenizer = e;
   }
   get model() {
     return this.trainer.model;
@@ -74,85 +86,75 @@ class f extends m {
     return this.totalTokens;
   }
   setOptions(t) {
-    const a = new Set(
+    const e = new Set(
       Object.keys(t).filter(
         (i) => t[i] !== this.options[i]
       )
     );
     if (this.trainer.isRunning) {
-      if (a.has("batchSize"))
+      if (e.has("batchSize"))
         throw new Error("Cannot change batch size during training");
-      if (a.has("sftMode"))
+      if (e.has("sftMode"))
         throw new Error("Cannot change SFT mode during training");
-      if (a.has("loraConfig"))
+      if (e.has("loraConfig"))
         throw new Error("Cannot change LoRA configuration during training");
-      if (a.has("validationSplit"))
+      if (e.has("validationSplit"))
         throw new Error("Cannot change validation split during training");
-      if (a.has("trainableWeights"))
+      if (e.has("trainableWeights"))
         throw new Error("Cannot change trainable weights during training");
-      if (a.has("mixedPrecision"))
+      if (e.has("mixedPrecision"))
         throw new Error("Cannot change mixed precision setting during training");
-      if (a.has("gradientCheckpointing"))
+      if (e.has("gradientCheckpointing"))
         throw new Error("Cannot change gradient checkpointing setting during training");
     }
     this.options = {
       ...this.options,
       ...t
-    }, this.trainer.updateOptimizer(this.options), a.has("metrics") && this.trainer.setMetrics(t.metrics || []);
+    }, this.trainer.updateOptimizer(this.options), e.has("metrics") && this.trainer.setMetrics(t.metrics || []);
   }
-  async prepare(t = [], a) {
+  async prepare(t = [], e) {
     const i = this.options;
-    if (a && (this.model.metaData.pretrainingData = a.map((e) => ({
-      id: e.id,
-      name: e.name
-    }))), this.trainingType === "pretraining" && this.trainer instanceof g) {
-      const { trainDataset: e, validationDataset: s, size: o } = await u(
+    if (e && (this.model.metaData.pretrainingData = e.map((a) => ({
+      id: a.id,
+      name: a.name
+    }))), this.trainingType === "pretraining" && this.trainer instanceof c) {
+      const { trainDataset: a, validationDataset: o, size: r } = await f(
         t,
         this.trainer.tokenizer,
         this.trainer.datasetBuilder,
         i?.batchSize || 32,
-        i?.validationSplit || 0.1
-      ), h = o * (1 - (i?.validationSplit || 0));
-      this.trainDataset = e, this.validationDataset = s, this.totalTokens = h, this.options.epochSteps = Math.ceil(
+        i?.validationSplit || 0.1,
+        i?.maskedLoss ?? !1
+      ), h = r * (1 - (i?.validationSplit || 0));
+      this.trainDataset = a, this.validationDataset = o, this.totalTokens = h, this.options.epochSteps = Math.ceil(
         this.totalTokens / ((i?.batchSize || 32) * this.model.config.blockSize)
       ), this.trainer.updateOptimizer(this.options);
-    } else if (this.trainingType === "sft" && this.trainer instanceof c) {
+    } else if (this.trainingType === "sft" && this.trainer instanceof g) {
       if (t instanceof Uint16Array)
         throw new Error("SFT training requires Task[] input");
-      if (i?.validationSplit && i.validationSplit > 0) {
-        const e = p(t, i?.validationSplit), s = await this.trainer.datasetBuilder.createSFTDataset(
-          [e.training],
-          i?.batchSize || 32,
-          -100
-        ), o = await this.trainer.datasetBuilder.createSFTDataset(
-          [e.validation],
-          i?.batchSize || 32,
-          -100
-        );
-        this.validationDataset = o, this.trainDataset = s;
-      } else {
-        const e = await this.trainer.datasetBuilder.createSFTDataset(
-          t,
-          i?.batchSize || 32,
-          -100
-        );
-        this.trainDataset = e;
-      }
-      this.totalTokens = t.reduce((e, s) => e + s.length, 0), this.options.epochSteps = Math.ceil(
+      const { trainDataset: a, validationDataset: o, size: r } = await f(
+        t,
+        this.trainer.tokenizer,
+        this.trainer.datasetBuilder,
+        i?.batchSize || 32,
+        i?.validationSplit || 0.1,
+        i?.maskedLoss ?? !0
+      ), h = r * (1 - (i?.validationSplit || 0));
+      this.trainDataset = a, this.validationDataset = o, this.totalTokens = h, this.options.epochSteps = Math.ceil(
         this.totalTokens / ((i?.batchSize || 32) * this.model.config.blockSize)
       ), this.trainer.updateOptimizer(this.options);
     }
   }
   configureModel(t) {
-    const a = t?.sftMode || "full";
+    const e = t?.sftMode || "full";
     if (this.trainingType === "pretraining" && (this.trainer.model.hasLoRA() && this.trainer.model.detachLoRA(), this.trainer.model.weightStore.setTrainable(["*"])), this.trainingType === "sft") {
-      if (a === "lora") {
+      if (e === "lora") {
         const i = this.trainer.model;
-        if (t?.loraName)
+        if (console.log("Configuring model for LoRA fine-tuning with options:", t), t?.loraName)
           if (i.hasLoRA(t.loraName)) {
             if (i.attachLoRA(t.loraName), t.loraConfig) {
-              const e = i.lora;
-              (e.alpha !== t.loraConfig.alpha || e.rank !== t.loraConfig.rank) && (i.detachLoRA(), i.deleteLoRA(t.loraName), i.createLoRA(t.loraName, t.loraConfig), i.attachLoRA(t.loraName), console.warn("Resetting LoRA with new configuration."));
+              const a = i.lora;
+              (a.alpha !== t.loraConfig.alpha || a.rank !== t.loraConfig.rank) && (i.detachLoRA(), i.deleteLoRA(t.loraName), i.createLoRA(t.loraName, t.loraConfig), i.attachLoRA(t.loraName), console.warn("Resetting LoRA with new configuration."));
             }
           } else if (t.loraConfig)
             i.createLoRA(t.loraName, t.loraConfig), i.attachLoRA(t.loraName);
@@ -162,23 +164,23 @@ class f extends m {
             );
         else if (t?.loraConfig)
           if (i.hasLoRA()) {
-            const e = i.lora;
-            if (e.alpha !== t.loraConfig.alpha || e.rank !== t.loraConfig.rank) {
+            const a = i.lora;
+            if (a.alpha !== t.loraConfig.alpha || a.rank !== t.loraConfig.rank) {
               i.detachLoRA();
-              const s = t.loraName || d();
-              i.createLoRA(s, t.loraConfig), i.attachLoRA(s);
+              const o = t.loraName || m();
+              i.createLoRA(o, t.loraConfig), i.attachLoRA(o);
             }
           } else {
-            const e = t.loraName || d();
-            i.createLoRA(e, t.loraConfig), i.attachLoRA(e);
+            const a = t.loraName || m();
+            i.createLoRA(a, t.loraConfig), i.attachLoRA(a);
           }
         else if (!i.hasLoRA()) throw new Error("LoRA configuration must be provided for lora SFT mode");
       } else
         this.trainer.model.hasLoRA() && this.trainer.model.detachLoRA();
-      a === "last-layer" ? this.trainer.model.weightStore.setTrainable([
+      e === "last-layer" ? this.trainer.model.weightStore.setTrainable([
         `block_${this.trainer.model.config.nLayer - 1}_*`,
         "token_embedding"
-      ]) : a === "full" && this.trainer.model.weightStore.setTrainable(["*"]);
+      ]) : e === "full" && this.trainer.model.weightStore.setTrainable(["*"]);
     }
     t?.trainableWeights && this.trainer.model.weightStore.setTrainable(t.trainableWeights);
   }
@@ -187,20 +189,20 @@ class f extends m {
     if (!this.trainDataset)
       throw new Error("Dataset not prepared");
     this.hasTrained || this.trainer.setLearningRate(t?.learningRate || 1e-3), this.hasTrained = !0, this.emit("start"), this.model.metaData.pretrainingSettings = t;
-    const a = Date.now();
+    const e = Date.now();
     this.log.length > 0 && this.trainer.resumeFromLog(this.log[this.log.length - 1]), this.trainer.setGradientCheckpointing(t?.gradientCheckpointing || !1), this.trainer.setMixedPrecision(t?.mixedPrecision || !1), this.trainer.setLabelSmoothing(t?.labelSmoothing || 0), this.trainer.setDropout(t?.dropout || 0), this.trainer.setLayerDrop(t?.layerDrop || 0), this.configureModel(t), await this.trainer.trainOnDataset(
       this.trainDataset,
       {
         ...t,
-        onStep: async (e) => {
-          this.log.push(e), this.progress = {
-            lastLog: e,
-            progress: e.totalTokens / this.totalTokens,
-            remaining: Math.max(0, (this.totalTokens - e.totalTokens) / e.totalTokens * e.duration)
-          }, this.tokensProcessed = e.totalTokens;
-          const s = this.listeners("log");
-          for (const o of s)
-            await o(e, this.progress);
+        onStep: async (a) => {
+          this.log.push(a), this.progress = {
+            lastLog: a,
+            progress: a.totalTokens / this.totalTokens,
+            remaining: Math.max(0, (this.totalTokens - a.totalTokens) / a.totalTokens * a.duration)
+          }, this.tokensProcessed = a.totalTokens;
+          const o = this.listeners("log");
+          for (const r of o)
+            await r(a, this.progress);
         }
       },
       this.validationDataset
@@ -209,7 +211,7 @@ class f extends m {
     this.model.metaData.actionLog.push({
       action: "pretrain",
       timestamp: i,
-      duration: i - a,
+      duration: i - e,
       tokensProcessed: this.tokensProcessed,
       options: t
     }), this.emit("stop");
@@ -218,12 +220,12 @@ class f extends m {
     if (!this.trainDataset)
       throw new Error("Dataset not prepared");
     this.hasTrained || this.trainer.setLearningRate(t?.learningRate || 1e-3), this.hasTrained = !0, this.emit("start");
-    const { log: a } = await this.trainer.stepDataset(this.trainDataset, t || {}, this.validationDataset), i = this.listeners("log");
-    for (const e of i)
-      await e(a, {
-        lastLog: a,
-        progress: a.totalTokens / this.totalTokens,
-        remaining: Math.max(0, (this.totalTokens - a.totalTokens) / a.totalTokens * a.duration)
+    const { log: e } = await this.trainer.stepDataset(this.trainDataset, t || {}, this.validationDataset), i = this.listeners("log");
+    for (const a of i)
+      await a(e, {
+        lastLog: e,
+        progress: e.totalTokens / this.totalTokens,
+        remaining: Math.max(0, (this.totalTokens - e.totalTokens) / e.totalTokens * e.duration)
       });
     this.emit("stop");
   }
@@ -238,5 +240,5 @@ class f extends m {
   }
 }
 export {
-  f as default
+  u as default
 };

package/dist/{axis_util-QWWgLjut.js → axis_util-GTVlo58H.js}

@@ -1,4 +1,4 @@
-import { v as c } from "./index-DSGwv2Yx.js";
+import { x as c } from "./index-CUXkjxiT.js";
 function i(e, n) {
   for (let t = 0; t < e.length; ++t)
     if (e[e.length - t - 1] !== n - 1 - t)

package/dist/backend.js

@@ -1,9 +1,9 @@
-import { g as o, s as e, r as s } from "./index-DSGwv2Yx.js";
+import { g as o, s as e, r as s } from "./index-CUXkjxiT.js";
 async function c(t, a) {
   if (o() !== t) {
     if (t === "webgpu") {
       const { registerWebGPUBackend: i } = await import("./patches/webgpu_base.js");
-      i(a), await import("./index-BQvB7LCC.js"), await import("./ops/webgpu/index.js");
+      i(a), await import("./index-CjOWnMXP.js"), await import("./ops/webgpu/index.js");
     }
     await e(t), await s(), console.log(`Backend set to ${t}`);
   }