craa commited on
Commit
1967bee
·
verified ·
1 Parent(s): a724b7e

Training in progress, step 10000, checkpoint

Browse files
checkpoint-10000/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9f9ecb7d2cdd0667c4486020af5ba7e4fa897d83b09ba91b38fb038a95b63412
3
  size 503128704
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03ac1bf824faf271bae58dce673b9d7952ef68a6eccdb4f5c5dd65b8f9aa84a7
3
  size 503128704
checkpoint-10000/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:82cb4fcc8c34b3cd40b17526742c583c9bd86a97648c7965ff9715953b8f8c71
3
  size 1006351290
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:72ab1143702e11fdaaf19b5fff3b4c60a793fc84d845bcd5583d8263cc76e0cb
3
  size 1006351290
checkpoint-10000/rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f1a0d474d254d7072642df0d6e147a602036c487497f36ec913c1166677568b0
3
  size 14244
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1abdd4c7c045dcb1f8a6404bddc83e4806c8d3e3173eb1017f9b082b3c161e85
3
  size 14244
checkpoint-10000/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e5a954fd0891b9a63bb5018718359621b31d6887975b43dbbefda70f75ac15a5
3
  size 1064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:baaddc0dc177f093cb7927eae851400a36fde6f45129bcaf2df443bdde5269f4
3
  size 1064
checkpoint-10000/trainer_state.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "best_metric": 3.7458271980285645,
3
- "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_495/checkpoint-10000",
4
- "epoch": 1.0763104079216446,
5
  "eval_steps": 1000,
6
  "global_step": 10000,
7
  "is_hyper_param_search": false,
@@ -9,1498 +9,1498 @@
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.005381552039608223,
13
- "grad_norm": 2.1159660816192627,
14
- "learning_rate": 0.00028199999999999997,
15
- "loss": 8.9316,
16
  "step": 50
17
  },
18
  {
19
- "epoch": 0.010763104079216447,
20
- "grad_norm": 1.9895623922348022,
21
- "learning_rate": 0.0005819999999999999,
22
- "loss": 6.9096,
23
  "step": 100
24
  },
25
  {
26
- "epoch": 0.01614465611882467,
27
- "grad_norm": 1.8703136444091797,
28
- "learning_rate": 0.0005996961534317422,
29
- "loss": 6.4463,
30
  "step": 150
31
  },
32
  {
33
- "epoch": 0.021526208158432893,
34
- "grad_norm": 1.2748037576675415,
35
- "learning_rate": 0.0005993729124016807,
36
- "loss": 6.2017,
37
  "step": 200
38
  },
39
  {
40
- "epoch": 0.026907760198041114,
41
- "grad_norm": 1.3479403257369995,
42
- "learning_rate": 0.0005990496713716194,
43
  "loss": 6.1031,
44
  "step": 250
45
  },
46
  {
47
- "epoch": 0.03228931223764934,
48
- "grad_norm": 1.0388602018356323,
49
- "learning_rate": 0.000598726430341558,
50
- "loss": 5.9726,
51
  "step": 300
52
  },
53
  {
54
- "epoch": 0.03767086427725756,
55
- "grad_norm": 1.1332666873931885,
56
- "learning_rate": 0.0005984031893114966,
57
- "loss": 5.886,
58
  "step": 350
59
  },
60
  {
61
- "epoch": 0.04305241631686579,
62
- "grad_norm": 1.398230791091919,
63
- "learning_rate": 0.0005980799482814351,
64
- "loss": 5.8019,
65
  "step": 400
66
  },
67
  {
68
- "epoch": 0.048433968356474004,
69
- "grad_norm": 1.5357556343078613,
70
- "learning_rate": 0.0005977567072513737,
71
- "loss": 5.7437,
72
  "step": 450
73
  },
74
  {
75
- "epoch": 0.05381552039608223,
76
- "grad_norm": 1.175200343132019,
77
- "learning_rate": 0.0005974334662213123,
78
- "loss": 5.6616,
79
  "step": 500
80
  },
81
  {
82
- "epoch": 0.05919707243569045,
83
- "grad_norm": 1.386881709098816,
84
- "learning_rate": 0.0005971102251912509,
85
- "loss": 5.5895,
86
  "step": 550
87
  },
88
  {
89
- "epoch": 0.06457862447529868,
90
- "grad_norm": 1.3578399419784546,
91
- "learning_rate": 0.0005967869841611895,
92
- "loss": 5.4912,
93
  "step": 600
94
  },
95
  {
96
- "epoch": 0.0699601765149069,
97
- "grad_norm": 1.2151751518249512,
98
- "learning_rate": 0.000596463743131128,
99
- "loss": 5.4275,
100
  "step": 650
101
  },
102
  {
103
- "epoch": 0.07534172855451512,
104
- "grad_norm": 1.3390365839004517,
105
- "learning_rate": 0.0005961405021010667,
106
- "loss": 5.3731,
107
  "step": 700
108
  },
109
  {
110
- "epoch": 0.08072328059412334,
111
- "grad_norm": 1.103965401649475,
112
- "learning_rate": 0.0005958172610710052,
113
- "loss": 5.3186,
114
  "step": 750
115
  },
116
  {
117
- "epoch": 0.08610483263373157,
118
- "grad_norm": 1.1866347789764404,
119
- "learning_rate": 0.0005954940200409439,
120
- "loss": 5.2687,
121
  "step": 800
122
  },
123
  {
124
- "epoch": 0.09148638467333979,
125
- "grad_norm": 1.0097291469573975,
126
- "learning_rate": 0.0005951707790108824,
127
- "loss": 5.2072,
128
  "step": 850
129
  },
130
  {
131
- "epoch": 0.09686793671294801,
132
- "grad_norm": 1.3843412399291992,
133
- "learning_rate": 0.0005948475379808209,
134
- "loss": 5.1549,
135
  "step": 900
136
  },
137
  {
138
- "epoch": 0.10224948875255624,
139
- "grad_norm": 1.1859126091003418,
140
- "learning_rate": 0.0005945242969507596,
141
- "loss": 5.0977,
142
  "step": 950
143
  },
144
  {
145
- "epoch": 0.10763104079216446,
146
- "grad_norm": 1.1825942993164062,
147
- "learning_rate": 0.0005942010559206981,
148
- "loss": 5.0644,
149
  "step": 1000
150
  },
151
  {
152
- "epoch": 0.10763104079216446,
153
- "eval_accuracy": 0.22798678911267642,
154
- "eval_loss": 5.015804290771484,
155
- "eval_runtime": 186.987,
156
- "eval_samples_per_second": 96.322,
157
- "eval_steps_per_second": 6.022,
158
  "step": 1000
159
  },
160
  {
161
- "epoch": 0.11301259283177269,
162
- "grad_norm": 1.036710262298584,
163
- "learning_rate": 0.0005938778148906367,
164
- "loss": 5.045,
165
  "step": 1050
166
  },
167
  {
168
- "epoch": 0.1183941448713809,
169
- "grad_norm": 1.1089870929718018,
170
- "learning_rate": 0.0005935545738605753,
171
- "loss": 5.0081,
172
  "step": 1100
173
  },
174
  {
175
- "epoch": 0.12377569691098914,
176
- "grad_norm": 0.9135867357254028,
177
- "learning_rate": 0.000593231332830514,
178
- "loss": 4.9925,
179
  "step": 1150
180
  },
181
  {
182
- "epoch": 0.12915724895059735,
183
- "grad_norm": 0.8327316641807556,
184
- "learning_rate": 0.0005929080918004525,
185
- "loss": 4.9437,
186
  "step": 1200
187
  },
188
  {
189
- "epoch": 0.13453880099020557,
190
- "grad_norm": 1.0137102603912354,
191
- "learning_rate": 0.000592584850770391,
192
- "loss": 4.9288,
193
  "step": 1250
194
  },
195
  {
196
- "epoch": 0.1399203530298138,
197
- "grad_norm": 1.620171070098877,
198
- "learning_rate": 0.0005922616097403296,
199
- "loss": 4.8813,
200
  "step": 1300
201
  },
202
  {
203
- "epoch": 0.14530190506942203,
204
- "grad_norm": 1.1357035636901855,
205
- "learning_rate": 0.0005919383687102682,
206
- "loss": 4.8713,
207
  "step": 1350
208
  },
209
  {
210
- "epoch": 0.15068345710903025,
211
- "grad_norm": 0.7975968718528748,
212
- "learning_rate": 0.0005916151276802069,
213
- "loss": 4.835,
214
  "step": 1400
215
  },
216
  {
217
- "epoch": 0.15606500914863847,
218
- "grad_norm": 0.9254979491233826,
219
- "learning_rate": 0.0005912918866501454,
220
- "loss": 4.806,
221
  "step": 1450
222
  },
223
  {
224
- "epoch": 0.16144656118824668,
225
- "grad_norm": 0.8283946514129639,
226
- "learning_rate": 0.000590968645620084,
227
- "loss": 4.8075,
228
  "step": 1500
229
  },
230
  {
231
- "epoch": 0.1668281132278549,
232
- "grad_norm": 1.1234453916549683,
233
- "learning_rate": 0.0005906454045900226,
234
- "loss": 4.7791,
235
  "step": 1550
236
  },
237
  {
238
- "epoch": 0.17220966526746315,
239
- "grad_norm": 0.9318310618400574,
240
- "learning_rate": 0.0005903221635599612,
241
- "loss": 4.7354,
242
  "step": 1600
243
  },
244
  {
245
- "epoch": 0.17759121730707136,
246
- "grad_norm": 0.8706691861152649,
247
- "learning_rate": 0.0005899989225298998,
248
- "loss": 4.7278,
249
  "step": 1650
250
  },
251
  {
252
- "epoch": 0.18297276934667958,
253
- "grad_norm": 1.0983690023422241,
254
- "learning_rate": 0.0005896756814998383,
255
- "loss": 4.7116,
256
  "step": 1700
257
  },
258
  {
259
- "epoch": 0.1883543213862878,
260
- "grad_norm": 1.1487178802490234,
261
- "learning_rate": 0.0005893524404697769,
262
- "loss": 4.7026,
263
  "step": 1750
264
  },
265
  {
266
- "epoch": 0.19373587342589602,
267
- "grad_norm": 0.8954223394393921,
268
- "learning_rate": 0.0005890291994397155,
269
- "loss": 4.6491,
270
  "step": 1800
271
  },
272
  {
273
- "epoch": 0.19911742546550426,
274
- "grad_norm": 1.1864057779312134,
275
- "learning_rate": 0.0005887059584096541,
276
- "loss": 4.6526,
277
  "step": 1850
278
  },
279
  {
280
- "epoch": 0.20449897750511248,
281
- "grad_norm": 0.945884108543396,
282
- "learning_rate": 0.0005883827173795926,
283
- "loss": 4.6132,
284
  "step": 1900
285
  },
286
  {
287
- "epoch": 0.2098805295447207,
288
- "grad_norm": 1.2601373195648193,
289
- "learning_rate": 0.0005880594763495313,
290
- "loss": 4.6044,
291
  "step": 1950
292
  },
293
  {
294
- "epoch": 0.2152620815843289,
295
- "grad_norm": 0.9447725415229797,
296
- "learning_rate": 0.0005877362353194698,
297
- "loss": 4.5877,
298
  "step": 2000
299
  },
300
  {
301
- "epoch": 0.2152620815843289,
302
- "eval_accuracy": 0.27040933128385014,
303
- "eval_loss": 4.508606910705566,
304
- "eval_runtime": 184.8398,
305
- "eval_samples_per_second": 97.441,
306
- "eval_steps_per_second": 6.092,
307
  "step": 2000
308
  },
309
  {
310
- "epoch": 0.22064363362393713,
311
- "grad_norm": 0.7851864695549011,
312
- "learning_rate": 0.0005874129942894084,
313
- "loss": 4.5643,
314
  "step": 2050
315
  },
316
  {
317
- "epoch": 0.22602518566354537,
318
- "grad_norm": 0.862130880355835,
319
- "learning_rate": 0.000587089753259347,
320
- "loss": 4.5481,
321
  "step": 2100
322
  },
323
  {
324
- "epoch": 0.2314067377031536,
325
- "grad_norm": 0.8470994830131531,
326
- "learning_rate": 0.0005867665122292855,
327
- "loss": 4.5219,
328
  "step": 2150
329
  },
330
  {
331
- "epoch": 0.2367882897427618,
332
- "grad_norm": 0.8713746070861816,
333
- "learning_rate": 0.0005864432711992242,
334
- "loss": 4.511,
335
  "step": 2200
336
  },
337
  {
338
- "epoch": 0.24216984178237003,
339
- "grad_norm": 0.8879591226577759,
340
- "learning_rate": 0.0005861200301691628,
341
- "loss": 4.5126,
342
  "step": 2250
343
  },
344
  {
345
- "epoch": 0.24755139382197827,
346
- "grad_norm": 0.9361649751663208,
347
- "learning_rate": 0.0005857967891391014,
348
- "loss": 4.4803,
349
  "step": 2300
350
  },
351
  {
352
- "epoch": 0.2529329458615865,
353
- "grad_norm": 0.8480615019798279,
354
- "learning_rate": 0.0005854735481090399,
355
- "loss": 4.4446,
356
  "step": 2350
357
  },
358
  {
359
- "epoch": 0.2583144979011947,
360
- "grad_norm": 1.0228937864303589,
361
- "learning_rate": 0.0005851503070789784,
362
- "loss": 4.4508,
363
  "step": 2400
364
  },
365
  {
366
- "epoch": 0.2636960499408029,
367
- "grad_norm": 0.9693055152893066,
368
- "learning_rate": 0.0005848270660489171,
369
- "loss": 4.4283,
370
  "step": 2450
371
  },
372
  {
373
- "epoch": 0.26907760198041114,
374
- "grad_norm": 0.7247486114501953,
375
- "learning_rate": 0.0005845038250188556,
376
- "loss": 4.4015,
377
  "step": 2500
378
  },
379
  {
380
- "epoch": 0.27445915402001936,
381
- "grad_norm": 0.7183290123939514,
382
- "learning_rate": 0.0005841805839887943,
383
- "loss": 4.3918,
384
  "step": 2550
385
  },
386
  {
387
- "epoch": 0.2798407060596276,
388
- "grad_norm": 0.900417149066925,
389
- "learning_rate": 0.0005838573429587328,
390
- "loss": 4.3701,
391
  "step": 2600
392
  },
393
  {
394
- "epoch": 0.2852222580992358,
395
- "grad_norm": 0.8292120099067688,
396
- "learning_rate": 0.0005835341019286715,
397
- "loss": 4.3891,
398
  "step": 2650
399
  },
400
  {
401
- "epoch": 0.29060381013884407,
402
- "grad_norm": 0.798037052154541,
403
- "learning_rate": 0.00058321086089861,
404
- "loss": 4.36,
405
  "step": 2700
406
  },
407
  {
408
- "epoch": 0.2959853621784523,
409
- "grad_norm": 0.8176091313362122,
410
- "learning_rate": 0.0005828876198685486,
411
- "loss": 4.3491,
412
  "step": 2750
413
  },
414
  {
415
- "epoch": 0.3013669142180605,
416
- "grad_norm": 0.8014965057373047,
417
- "learning_rate": 0.0005825643788384872,
418
- "loss": 4.3642,
419
  "step": 2800
420
  },
421
  {
422
- "epoch": 0.3067484662576687,
423
- "grad_norm": 1.2503150701522827,
424
- "learning_rate": 0.0005822411378084257,
425
- "loss": 4.3251,
426
  "step": 2850
427
  },
428
  {
429
- "epoch": 0.31213001829727693,
430
- "grad_norm": 0.7830126285552979,
431
- "learning_rate": 0.0005819178967783644,
432
- "loss": 4.3251,
433
  "step": 2900
434
  },
435
  {
436
- "epoch": 0.31751157033688515,
437
- "grad_norm": 0.9035398364067078,
438
- "learning_rate": 0.0005815946557483029,
439
- "loss": 4.3313,
440
  "step": 2950
441
  },
442
  {
443
- "epoch": 0.32289312237649337,
444
- "grad_norm": 0.6857808828353882,
445
- "learning_rate": 0.0005812714147182415,
446
- "loss": 4.2989,
447
  "step": 3000
448
  },
449
  {
450
- "epoch": 0.32289312237649337,
451
- "eval_accuracy": 0.2990024252411089,
452
- "eval_loss": 4.2293620109558105,
453
- "eval_runtime": 185.031,
454
- "eval_samples_per_second": 97.34,
455
- "eval_steps_per_second": 6.085,
456
  "step": 3000
457
  },
458
  {
459
- "epoch": 0.3282746744161016,
460
- "grad_norm": 0.7547754049301147,
461
- "learning_rate": 0.0005809481736881801,
462
- "loss": 4.296,
463
  "step": 3050
464
  },
465
  {
466
- "epoch": 0.3336562264557098,
467
- "grad_norm": 0.7557756304740906,
468
- "learning_rate": 0.0005806249326581187,
469
- "loss": 4.2866,
470
  "step": 3100
471
  },
472
  {
473
- "epoch": 0.3390377784953181,
474
- "grad_norm": 0.7948616147041321,
475
- "learning_rate": 0.0005803016916280573,
476
- "loss": 4.2736,
477
  "step": 3150
478
  },
479
  {
480
- "epoch": 0.3444193305349263,
481
- "grad_norm": 0.864937424659729,
482
- "learning_rate": 0.0005799784505979959,
483
- "loss": 4.2902,
484
  "step": 3200
485
  },
486
  {
487
- "epoch": 0.3498008825745345,
488
- "grad_norm": 0.8035250902175903,
489
- "learning_rate": 0.0005796552095679344,
490
- "loss": 4.2524,
491
  "step": 3250
492
  },
493
  {
494
- "epoch": 0.35518243461414273,
495
- "grad_norm": 0.7827973365783691,
496
- "learning_rate": 0.000579331968537873,
497
- "loss": 4.2573,
498
  "step": 3300
499
  },
500
  {
501
- "epoch": 0.36056398665375095,
502
- "grad_norm": 0.7145817875862122,
503
- "learning_rate": 0.0005790087275078116,
504
- "loss": 4.233,
505
  "step": 3350
506
  },
507
  {
508
- "epoch": 0.36594553869335916,
509
- "grad_norm": 0.7260717749595642,
510
- "learning_rate": 0.0005786854864777502,
511
- "loss": 4.2351,
512
  "step": 3400
513
  },
514
  {
515
- "epoch": 0.3713270907329674,
516
- "grad_norm": 0.846215009689331,
517
- "learning_rate": 0.0005783622454476888,
518
- "loss": 4.236,
519
  "step": 3450
520
  },
521
  {
522
- "epoch": 0.3767086427725756,
523
- "grad_norm": 0.6381611227989197,
524
- "learning_rate": 0.0005780390044176273,
525
- "loss": 4.2186,
526
  "step": 3500
527
  },
528
  {
529
- "epoch": 0.3820901948121838,
530
- "grad_norm": 0.7176727652549744,
531
- "learning_rate": 0.000577715763387566,
532
- "loss": 4.2159,
533
  "step": 3550
534
  },
535
  {
536
- "epoch": 0.38747174685179203,
537
- "grad_norm": 0.7199215888977051,
538
- "learning_rate": 0.0005773925223575045,
539
- "loss": 4.2187,
540
  "step": 3600
541
  },
542
  {
543
- "epoch": 0.3928532988914003,
544
- "grad_norm": 0.7120049595832825,
545
- "learning_rate": 0.0005770692813274432,
546
- "loss": 4.1858,
547
  "step": 3650
548
  },
549
  {
550
- "epoch": 0.3982348509310085,
551
- "grad_norm": 0.682905912399292,
552
- "learning_rate": 0.0005767460402973817,
553
- "loss": 4.2067,
554
  "step": 3700
555
  },
556
  {
557
- "epoch": 0.40361640297061674,
558
- "grad_norm": 0.8106050491333008,
559
- "learning_rate": 0.0005764227992673203,
560
- "loss": 4.1901,
561
  "step": 3750
562
  },
563
  {
564
- "epoch": 0.40899795501022496,
565
- "grad_norm": 0.6139677166938782,
566
- "learning_rate": 0.0005760995582372589,
567
- "loss": 4.1922,
568
  "step": 3800
569
  },
570
  {
571
- "epoch": 0.4143795070498332,
572
- "grad_norm": 0.6996195912361145,
573
- "learning_rate": 0.0005757763172071974,
574
- "loss": 4.1884,
575
  "step": 3850
576
  },
577
  {
578
- "epoch": 0.4197610590894414,
579
- "grad_norm": 0.6972913146018982,
580
- "learning_rate": 0.000575453076177136,
581
- "loss": 4.1469,
582
  "step": 3900
583
  },
584
  {
585
- "epoch": 0.4251426111290496,
586
- "grad_norm": 0.7805142998695374,
587
- "learning_rate": 0.0005751298351470746,
588
- "loss": 4.1681,
589
  "step": 3950
590
  },
591
  {
592
- "epoch": 0.4305241631686578,
593
- "grad_norm": 0.6750819683074951,
594
- "learning_rate": 0.0005748065941170133,
595
- "loss": 4.1677,
596
  "step": 4000
597
  },
598
  {
599
- "epoch": 0.4305241631686578,
600
- "eval_accuracy": 0.3131229545414788,
601
- "eval_loss": 4.085065841674805,
602
- "eval_runtime": 184.7939,
603
- "eval_samples_per_second": 97.465,
604
- "eval_steps_per_second": 6.093,
605
  "step": 4000
606
  },
607
  {
608
- "epoch": 0.43590571520826604,
609
- "grad_norm": 0.743445098400116,
610
- "learning_rate": 0.0005744833530869518,
611
- "loss": 4.1605,
612
  "step": 4050
613
  },
614
  {
615
- "epoch": 0.44128726724787426,
616
- "grad_norm": 0.6816270351409912,
617
- "learning_rate": 0.0005741601120568903,
618
- "loss": 4.1494,
619
  "step": 4100
620
  },
621
  {
622
- "epoch": 0.44666881928748253,
623
- "grad_norm": 0.6031078696250916,
624
- "learning_rate": 0.0005738368710268289,
625
- "loss": 4.1373,
626
  "step": 4150
627
  },
628
  {
629
- "epoch": 0.45205037132709075,
630
- "grad_norm": 0.869440495967865,
631
- "learning_rate": 0.0005735136299967675,
632
- "loss": 4.1274,
633
  "step": 4200
634
  },
635
  {
636
- "epoch": 0.45743192336669897,
637
- "grad_norm": 0.6965010762214661,
638
- "learning_rate": 0.0005731903889667062,
639
- "loss": 4.1363,
640
  "step": 4250
641
  },
642
  {
643
- "epoch": 0.4628134754063072,
644
- "grad_norm": 0.6586082577705383,
645
- "learning_rate": 0.0005728671479366447,
646
- "loss": 4.1073,
647
  "step": 4300
648
  },
649
  {
650
- "epoch": 0.4681950274459154,
651
- "grad_norm": 0.6008425951004028,
652
- "learning_rate": 0.0005725439069065833,
653
- "loss": 4.1131,
654
  "step": 4350
655
  },
656
  {
657
- "epoch": 0.4735765794855236,
658
- "grad_norm": 0.6233063340187073,
659
- "learning_rate": 0.0005722206658765219,
660
- "loss": 4.1266,
661
  "step": 4400
662
  },
663
  {
664
- "epoch": 0.47895813152513184,
665
- "grad_norm": 0.7806559205055237,
666
- "learning_rate": 0.0005718974248464605,
667
- "loss": 4.1052,
668
  "step": 4450
669
  },
670
  {
671
- "epoch": 0.48433968356474005,
672
- "grad_norm": 0.7604005336761475,
673
- "learning_rate": 0.000571574183816399,
674
- "loss": 4.0944,
675
  "step": 4500
676
  },
677
  {
678
- "epoch": 0.48972123560434827,
679
- "grad_norm": 0.8495274782180786,
680
- "learning_rate": 0.0005712509427863376,
681
- "loss": 4.1176,
682
  "step": 4550
683
  },
684
  {
685
- "epoch": 0.49510278764395654,
686
- "grad_norm": 0.6198326349258423,
687
- "learning_rate": 0.0005709277017562762,
688
- "loss": 4.1016,
689
  "step": 4600
690
  },
691
  {
692
- "epoch": 0.5004843396835648,
693
- "grad_norm": 0.6641437411308289,
694
- "learning_rate": 0.0005706044607262148,
695
- "loss": 4.0835,
696
  "step": 4650
697
  },
698
  {
699
- "epoch": 0.505865891723173,
700
- "grad_norm": 0.7691634297370911,
701
- "learning_rate": 0.0005702812196961534,
702
- "loss": 4.0815,
703
  "step": 4700
704
  },
705
  {
706
- "epoch": 0.5112474437627812,
707
- "grad_norm": 0.7156651020050049,
708
- "learning_rate": 0.0005699579786660919,
709
- "loss": 4.0818,
710
  "step": 4750
711
  },
712
  {
713
- "epoch": 0.5166289958023894,
714
- "grad_norm": 0.6472979784011841,
715
- "learning_rate": 0.0005696347376360306,
716
- "loss": 4.067,
717
  "step": 4800
718
  },
719
  {
720
- "epoch": 0.5220105478419976,
721
- "grad_norm": 0.724949300289154,
722
- "learning_rate": 0.0005693114966059691,
723
- "loss": 4.0615,
724
  "step": 4850
725
  },
726
  {
727
- "epoch": 0.5273920998816058,
728
- "grad_norm": 0.7559602856636047,
729
- "learning_rate": 0.0005689882555759077,
730
- "loss": 4.0676,
731
  "step": 4900
732
  },
733
  {
734
- "epoch": 0.5327736519212141,
735
- "grad_norm": 0.6394193768501282,
736
- "learning_rate": 0.0005686650145458463,
737
- "loss": 4.0461,
738
  "step": 4950
739
  },
740
  {
741
- "epoch": 0.5381552039608223,
742
- "grad_norm": 0.7001257538795471,
743
- "learning_rate": 0.0005683417735157848,
744
- "loss": 4.0729,
745
  "step": 5000
746
  },
747
  {
748
- "epoch": 0.5381552039608223,
749
- "eval_accuracy": 0.320991922635667,
750
- "eval_loss": 3.990946054458618,
751
- "eval_runtime": 184.8283,
752
- "eval_samples_per_second": 97.447,
753
- "eval_steps_per_second": 6.092,
754
  "step": 5000
755
  },
756
  {
757
- "epoch": 0.5435367560004305,
758
- "grad_norm": 0.7765357494354248,
759
- "learning_rate": 0.0005680185324857235,
760
- "loss": 4.0425,
761
  "step": 5050
762
  },
763
  {
764
- "epoch": 0.5489183080400387,
765
- "grad_norm": 0.6510383486747742,
766
- "learning_rate": 0.0005676952914556621,
767
- "loss": 4.0239,
768
  "step": 5100
769
  },
770
  {
771
- "epoch": 0.5542998600796469,
772
- "grad_norm": 0.6882658004760742,
773
- "learning_rate": 0.0005673720504256007,
774
- "loss": 4.0242,
775
  "step": 5150
776
  },
777
  {
778
- "epoch": 0.5596814121192552,
779
- "grad_norm": 0.7188264727592468,
780
- "learning_rate": 0.0005670488093955392,
781
- "loss": 4.0442,
782
  "step": 5200
783
  },
784
  {
785
- "epoch": 0.5650629641588634,
786
- "grad_norm": 0.6599631905555725,
787
- "learning_rate": 0.0005667255683654777,
788
- "loss": 4.0497,
789
  "step": 5250
790
  },
791
  {
792
- "epoch": 0.5704445161984716,
793
- "grad_norm": 0.5674473643302917,
794
- "learning_rate": 0.0005664023273354164,
795
- "loss": 4.0271,
796
  "step": 5300
797
  },
798
  {
799
- "epoch": 0.5758260682380799,
800
- "grad_norm": 0.6670402884483337,
801
- "learning_rate": 0.0005660790863053549,
802
- "loss": 4.0284,
803
  "step": 5350
804
  },
805
  {
806
- "epoch": 0.5812076202776881,
807
- "grad_norm": 0.7389664649963379,
808
- "learning_rate": 0.0005657558452752936,
809
- "loss": 4.0449,
810
  "step": 5400
811
  },
812
  {
813
- "epoch": 0.5865891723172963,
814
- "grad_norm": 0.6841723322868347,
815
- "learning_rate": 0.0005654326042452321,
816
- "loss": 4.051,
817
  "step": 5450
818
  },
819
  {
820
- "epoch": 0.5919707243569046,
821
- "grad_norm": 0.6119431853294373,
822
- "learning_rate": 0.0005651093632151708,
823
- "loss": 4.0158,
824
  "step": 5500
825
  },
826
  {
827
- "epoch": 0.5973522763965128,
828
- "grad_norm": 0.8850950598716736,
829
- "learning_rate": 0.0005647861221851093,
830
- "loss": 4.0183,
831
  "step": 5550
832
  },
833
  {
834
- "epoch": 0.602733828436121,
835
- "grad_norm": 0.6376111507415771,
836
- "learning_rate": 0.0005644628811550479,
837
- "loss": 4.0076,
838
  "step": 5600
839
  },
840
  {
841
- "epoch": 0.6081153804757292,
842
- "grad_norm": 0.5115376114845276,
843
- "learning_rate": 0.0005641396401249865,
844
- "loss": 4.0006,
845
  "step": 5650
846
  },
847
  {
848
- "epoch": 0.6134969325153374,
849
- "grad_norm": 0.6021561026573181,
850
- "learning_rate": 0.000563816399094925,
851
- "loss": 4.0056,
852
  "step": 5700
853
  },
854
  {
855
- "epoch": 0.6188784845549457,
856
- "grad_norm": 0.5795146226882935,
857
- "learning_rate": 0.0005634931580648637,
858
- "loss": 4.0159,
859
  "step": 5750
860
  },
861
  {
862
- "epoch": 0.6242600365945539,
863
- "grad_norm": 0.6918138861656189,
864
- "learning_rate": 0.0005631699170348022,
865
- "loss": 3.991,
866
  "step": 5800
867
  },
868
  {
869
- "epoch": 0.6296415886341621,
870
- "grad_norm": 0.6135892271995544,
871
- "learning_rate": 0.0005628466760047408,
872
- "loss": 4.0127,
873
  "step": 5850
874
  },
875
  {
876
- "epoch": 0.6350231406737703,
877
- "grad_norm": 0.6184191703796387,
878
- "learning_rate": 0.0005625234349746794,
879
- "loss": 3.9862,
880
  "step": 5900
881
  },
882
  {
883
- "epoch": 0.6404046927133785,
884
- "grad_norm": 0.6259024739265442,
885
- "learning_rate": 0.000562200193944618,
886
- "loss": 3.9922,
887
  "step": 5950
888
  },
889
  {
890
- "epoch": 0.6457862447529867,
891
- "grad_norm": 0.5602654814720154,
892
- "learning_rate": 0.0005618769529145566,
893
- "loss": 3.995,
894
  "step": 6000
895
  },
896
  {
897
- "epoch": 0.6457862447529867,
898
- "eval_accuracy": 0.32913100180896193,
899
- "eval_loss": 3.9143569469451904,
900
- "eval_runtime": 184.5106,
901
- "eval_samples_per_second": 97.615,
902
- "eval_steps_per_second": 6.103,
903
  "step": 6000
904
  },
905
  {
906
- "epoch": 0.651167796792595,
907
- "grad_norm": 0.7381386160850525,
908
- "learning_rate": 0.0005615537118844952,
909
- "loss": 3.9749,
910
  "step": 6050
911
  },
912
  {
913
- "epoch": 0.6565493488322032,
914
- "grad_norm": 0.6130694150924683,
915
- "learning_rate": 0.0005612304708544337,
916
- "loss": 3.9791,
917
  "step": 6100
918
  },
919
  {
920
- "epoch": 0.6619309008718114,
921
- "grad_norm": 0.6282258629798889,
922
- "learning_rate": 0.0005609072298243723,
923
- "loss": 3.9766,
924
  "step": 6150
925
  },
926
  {
927
- "epoch": 0.6673124529114196,
928
- "grad_norm": 0.7320833802223206,
929
- "learning_rate": 0.0005605839887943109,
930
- "loss": 3.9744,
931
  "step": 6200
932
  },
933
  {
934
- "epoch": 0.6726940049510278,
935
- "grad_norm": 0.7067676782608032,
936
- "learning_rate": 0.0005602607477642495,
937
- "loss": 3.9755,
938
  "step": 6250
939
  },
940
  {
941
- "epoch": 0.6780755569906362,
942
- "grad_norm": 0.6241488456726074,
943
- "learning_rate": 0.0005599375067341881,
944
- "loss": 3.9639,
945
  "step": 6300
946
  },
947
  {
948
- "epoch": 0.6834571090302444,
949
- "grad_norm": 0.6031643152236938,
950
- "learning_rate": 0.0005596142657041266,
951
- "loss": 3.9535,
952
  "step": 6350
953
  },
954
  {
955
- "epoch": 0.6888386610698526,
956
- "grad_norm": 0.5485600829124451,
957
- "learning_rate": 0.0005592910246740653,
958
- "loss": 3.9616,
959
  "step": 6400
960
  },
961
  {
962
- "epoch": 0.6942202131094608,
963
- "grad_norm": 0.6337101459503174,
964
- "learning_rate": 0.0005589677836440038,
965
- "loss": 3.9509,
966
  "step": 6450
967
  },
968
  {
969
- "epoch": 0.699601765149069,
970
- "grad_norm": 0.546737790107727,
971
- "learning_rate": 0.0005586445426139425,
972
- "loss": 3.9416,
973
  "step": 6500
974
  },
975
  {
976
- "epoch": 0.7049833171886772,
977
- "grad_norm": 0.6616541147232056,
978
- "learning_rate": 0.000558321301583881,
979
- "loss": 3.9618,
980
  "step": 6550
981
  },
982
  {
983
- "epoch": 0.7103648692282855,
984
- "grad_norm": 0.5739938020706177,
985
- "learning_rate": 0.0005579980605538196,
986
- "loss": 3.949,
987
  "step": 6600
988
  },
989
  {
990
- "epoch": 0.7157464212678937,
991
- "grad_norm": 0.5481446981430054,
992
- "learning_rate": 0.0005576748195237582,
993
- "loss": 3.961,
994
  "step": 6650
995
  },
996
  {
997
- "epoch": 0.7211279733075019,
998
- "grad_norm": 0.6824391484260559,
999
- "learning_rate": 0.0005573515784936967,
1000
- "loss": 3.9376,
1001
  "step": 6700
1002
  },
1003
  {
1004
- "epoch": 0.7265095253471101,
1005
- "grad_norm": 0.6663408279418945,
1006
- "learning_rate": 0.0005570283374636353,
1007
- "loss": 3.9525,
1008
  "step": 6750
1009
  },
1010
  {
1011
- "epoch": 0.7318910773867183,
1012
- "grad_norm": 0.5435226559638977,
1013
- "learning_rate": 0.0005567050964335739,
1014
- "loss": 3.9249,
1015
  "step": 6800
1016
  },
1017
  {
1018
- "epoch": 0.7372726294263265,
1019
- "grad_norm": 0.6284464001655579,
1020
- "learning_rate": 0.0005563818554035126,
1021
- "loss": 3.9307,
1022
  "step": 6850
1023
  },
1024
  {
1025
- "epoch": 0.7426541814659348,
1026
- "grad_norm": 0.5826367139816284,
1027
- "learning_rate": 0.0005560586143734511,
1028
- "loss": 3.9251,
1029
  "step": 6900
1030
  },
1031
  {
1032
- "epoch": 0.748035733505543,
1033
- "grad_norm": 0.5755831003189087,
1034
- "learning_rate": 0.0005557353733433896,
1035
- "loss": 3.9271,
1036
  "step": 6950
1037
  },
1038
  {
1039
- "epoch": 0.7534172855451512,
1040
- "grad_norm": 0.6272289156913757,
1041
- "learning_rate": 0.0005554121323133283,
1042
- "loss": 3.9517,
1043
  "step": 7000
1044
  },
1045
  {
1046
- "epoch": 0.7534172855451512,
1047
- "eval_accuracy": 0.33305206722441094,
1048
- "eval_loss": 3.8637194633483887,
1049
- "eval_runtime": 184.9315,
1050
- "eval_samples_per_second": 97.393,
1051
- "eval_steps_per_second": 6.089,
1052
  "step": 7000
1053
  },
1054
  {
1055
- "epoch": 0.7587988375847594,
1056
- "grad_norm": 0.6227930188179016,
1057
- "learning_rate": 0.0005550888912832668,
1058
- "loss": 3.916,
1059
  "step": 7050
1060
  },
1061
  {
1062
- "epoch": 0.7641803896243676,
1063
- "grad_norm": 0.6726216077804565,
1064
- "learning_rate": 0.0005547656502532055,
1065
- "loss": 3.9217,
1066
  "step": 7100
1067
  },
1068
  {
1069
- "epoch": 0.7695619416639758,
1070
- "grad_norm": 0.5420123338699341,
1071
- "learning_rate": 0.000554442409223144,
1072
- "loss": 3.9238,
1073
  "step": 7150
1074
  },
1075
  {
1076
- "epoch": 0.7749434937035841,
1077
- "grad_norm": 0.5879136919975281,
1078
- "learning_rate": 0.0005541191681930826,
1079
- "loss": 3.9191,
1080
  "step": 7200
1081
  },
1082
  {
1083
- "epoch": 0.7803250457431924,
1084
- "grad_norm": 0.5138105750083923,
1085
- "learning_rate": 0.0005537959271630212,
1086
- "loss": 3.9145,
1087
  "step": 7250
1088
  },
1089
  {
1090
- "epoch": 0.7857065977828006,
1091
- "grad_norm": 0.5751107931137085,
1092
- "learning_rate": 0.0005534726861329598,
1093
- "loss": 3.9067,
1094
  "step": 7300
1095
  },
1096
  {
1097
- "epoch": 0.7910881498224088,
1098
- "grad_norm": 0.6342102289199829,
1099
- "learning_rate": 0.0005531494451028983,
1100
- "loss": 3.9038,
1101
  "step": 7350
1102
  },
1103
  {
1104
- "epoch": 0.796469701862017,
1105
- "grad_norm": 0.5363113284111023,
1106
- "learning_rate": 0.0005528262040728369,
1107
- "loss": 3.9121,
1108
  "step": 7400
1109
  },
1110
  {
1111
- "epoch": 0.8018512539016253,
1112
- "grad_norm": 0.6412177681922913,
1113
- "learning_rate": 0.0005525029630427755,
1114
- "loss": 3.894,
1115
  "step": 7450
1116
  },
1117
  {
1118
- "epoch": 0.8072328059412335,
1119
- "grad_norm": 0.6429057121276855,
1120
- "learning_rate": 0.0005521797220127141,
1121
- "loss": 3.904,
1122
  "step": 7500
1123
  },
1124
  {
1125
- "epoch": 0.8126143579808417,
1126
- "grad_norm": 0.6362873315811157,
1127
- "learning_rate": 0.0005518564809826527,
1128
- "loss": 3.9328,
1129
  "step": 7550
1130
  },
1131
  {
1132
- "epoch": 0.8179959100204499,
1133
- "grad_norm": 0.6349565386772156,
1134
- "learning_rate": 0.0005515332399525912,
1135
- "loss": 3.8807,
1136
  "step": 7600
1137
  },
1138
  {
1139
- "epoch": 0.8233774620600581,
1140
- "grad_norm": 0.6769669651985168,
1141
- "learning_rate": 0.0005512099989225299,
1142
- "loss": 3.8932,
1143
  "step": 7650
1144
  },
1145
  {
1146
- "epoch": 0.8287590140996663,
1147
- "grad_norm": 0.601761519908905,
1148
- "learning_rate": 0.0005508867578924685,
1149
- "loss": 3.897,
1150
  "step": 7700
1151
  },
1152
  {
1153
- "epoch": 0.8341405661392746,
1154
- "grad_norm": 0.6315547823905945,
1155
- "learning_rate": 0.000550563516862407,
1156
- "loss": 3.8814,
1157
  "step": 7750
1158
  },
1159
  {
1160
- "epoch": 0.8395221181788828,
1161
- "grad_norm": 0.5754820108413696,
1162
- "learning_rate": 0.0005502402758323456,
1163
- "loss": 3.9101,
1164
  "step": 7800
1165
  },
1166
  {
1167
- "epoch": 0.844903670218491,
1168
- "grad_norm": 0.627307653427124,
1169
- "learning_rate": 0.0005499170348022841,
1170
- "loss": 3.9035,
1171
  "step": 7850
1172
  },
1173
  {
1174
- "epoch": 0.8502852222580992,
1175
- "grad_norm": 0.6498885154724121,
1176
- "learning_rate": 0.0005495937937722228,
1177
- "loss": 3.9012,
1178
  "step": 7900
1179
  },
1180
  {
1181
- "epoch": 0.8556667742977074,
1182
- "grad_norm": 0.5503144264221191,
1183
- "learning_rate": 0.0005492705527421614,
1184
- "loss": 3.8942,
1185
  "step": 7950
1186
  },
1187
  {
1188
- "epoch": 0.8610483263373157,
1189
- "grad_norm": 0.6200529336929321,
1190
- "learning_rate": 0.0005489473117121,
1191
- "loss": 3.8717,
1192
  "step": 8000
1193
  },
1194
  {
1195
- "epoch": 0.8610483263373157,
1196
- "eval_accuracy": 0.3380203291726159,
1197
- "eval_loss": 3.8143210411071777,
1198
- "eval_runtime": 184.8318,
1199
- "eval_samples_per_second": 97.445,
1200
- "eval_steps_per_second": 6.092,
1201
  "step": 8000
1202
  },
1203
  {
1204
- "epoch": 0.8664298783769239,
1205
- "grad_norm": 0.7157406210899353,
1206
- "learning_rate": 0.0005486240706820385,
1207
- "loss": 3.8746,
1208
  "step": 8050
1209
  },
1210
  {
1211
- "epoch": 0.8718114304165321,
1212
- "grad_norm": 0.643976628780365,
1213
- "learning_rate": 0.0005483008296519772,
1214
- "loss": 3.8859,
1215
  "step": 8100
1216
  },
1217
  {
1218
- "epoch": 0.8771929824561403,
1219
- "grad_norm": 0.5374308824539185,
1220
- "learning_rate": 0.0005479775886219157,
1221
- "loss": 3.8586,
1222
  "step": 8150
1223
  },
1224
  {
1225
- "epoch": 0.8825745344957485,
1226
- "grad_norm": 0.6169512867927551,
1227
- "learning_rate": 0.0005476543475918542,
1228
- "loss": 3.8734,
1229
  "step": 8200
1230
  },
1231
  {
1232
- "epoch": 0.8879560865353568,
1233
- "grad_norm": 0.6148831248283386,
1234
- "learning_rate": 0.0005473311065617929,
1235
- "loss": 3.8731,
1236
  "step": 8250
1237
  },
1238
  {
1239
- "epoch": 0.8933376385749651,
1240
- "grad_norm": 0.5680239796638489,
1241
- "learning_rate": 0.0005470078655317314,
1242
- "loss": 3.8674,
1243
  "step": 8300
1244
  },
1245
  {
1246
- "epoch": 0.8987191906145733,
1247
- "grad_norm": 0.5679119825363159,
1248
- "learning_rate": 0.0005466846245016701,
1249
- "loss": 3.8556,
1250
  "step": 8350
1251
  },
1252
  {
1253
- "epoch": 0.9041007426541815,
1254
- "grad_norm": 0.5395457744598389,
1255
- "learning_rate": 0.0005463613834716086,
1256
- "loss": 3.8536,
1257
  "step": 8400
1258
  },
1259
  {
1260
- "epoch": 0.9094822946937897,
1261
- "grad_norm": 0.5603843331336975,
1262
- "learning_rate": 0.0005460381424415472,
1263
- "loss": 3.8579,
1264
  "step": 8450
1265
  },
1266
  {
1267
- "epoch": 0.9148638467333979,
1268
- "grad_norm": 0.5550517439842224,
1269
- "learning_rate": 0.0005457149014114858,
1270
- "loss": 3.858,
1271
  "step": 8500
1272
  },
1273
  {
1274
- "epoch": 0.9202453987730062,
1275
- "grad_norm": 0.6112827062606812,
1276
- "learning_rate": 0.0005453916603814243,
1277
- "loss": 3.8563,
1278
  "step": 8550
1279
  },
1280
  {
1281
- "epoch": 0.9256269508126144,
1282
- "grad_norm": 0.6317242383956909,
1283
- "learning_rate": 0.000545068419351363,
1284
- "loss": 3.854,
1285
  "step": 8600
1286
  },
1287
  {
1288
- "epoch": 0.9310085028522226,
1289
- "grad_norm": 0.574309766292572,
1290
- "learning_rate": 0.0005447451783213015,
1291
- "loss": 3.8455,
1292
  "step": 8650
1293
  },
1294
  {
1295
- "epoch": 0.9363900548918308,
1296
- "grad_norm": 0.5467745065689087,
1297
- "learning_rate": 0.0005444219372912401,
1298
- "loss": 3.858,
1299
  "step": 8700
1300
  },
1301
  {
1302
- "epoch": 0.941771606931439,
1303
- "grad_norm": 0.5530813932418823,
1304
- "learning_rate": 0.0005440986962611787,
1305
- "loss": 3.8518,
1306
  "step": 8750
1307
  },
1308
  {
1309
- "epoch": 0.9471531589710472,
1310
- "grad_norm": 0.7071423530578613,
1311
- "learning_rate": 0.0005437754552311173,
1312
- "loss": 3.8473,
1313
  "step": 8800
1314
  },
1315
  {
1316
- "epoch": 0.9525347110106555,
1317
- "grad_norm": 0.5127449631690979,
1318
- "learning_rate": 0.0005434522142010559,
1319
- "loss": 3.8471,
1320
  "step": 8850
1321
  },
1322
  {
1323
- "epoch": 0.9579162630502637,
1324
- "grad_norm": 0.5855455994606018,
1325
- "learning_rate": 0.0005431289731709945,
1326
- "loss": 3.8589,
1327
  "step": 8900
1328
  },
1329
  {
1330
- "epoch": 0.9632978150898719,
1331
- "grad_norm": 0.6145156025886536,
1332
- "learning_rate": 0.000542805732140933,
1333
- "loss": 3.8295,
1334
  "step": 8950
1335
  },
1336
  {
1337
- "epoch": 0.9686793671294801,
1338
- "grad_norm": 0.5550306439399719,
1339
- "learning_rate": 0.0005424824911108716,
1340
- "loss": 3.8378,
1341
  "step": 9000
1342
  },
1343
  {
1344
- "epoch": 0.9686793671294801,
1345
- "eval_accuracy": 0.3415591537287335,
1346
- "eval_loss": 3.7806262969970703,
1347
- "eval_runtime": 184.435,
1348
- "eval_samples_per_second": 97.655,
1349
- "eval_steps_per_second": 6.105,
1350
  "step": 9000
1351
  },
1352
  {
1353
- "epoch": 0.9740609191690883,
1354
- "grad_norm": 0.5801270008087158,
1355
- "learning_rate": 0.0005421592500808102,
1356
- "loss": 3.8502,
1357
  "step": 9050
1358
  },
1359
  {
1360
- "epoch": 0.9794424712086965,
1361
- "grad_norm": 0.5607963800430298,
1362
- "learning_rate": 0.0005418360090507488,
1363
- "loss": 3.825,
1364
  "step": 9100
1365
  },
1366
  {
1367
- "epoch": 0.9848240232483048,
1368
- "grad_norm": 0.5534636974334717,
1369
- "learning_rate": 0.0005415127680206874,
1370
- "loss": 3.8361,
1371
  "step": 9150
1372
  },
1373
  {
1374
- "epoch": 0.9902055752879131,
1375
- "grad_norm": 0.692423939704895,
1376
- "learning_rate": 0.0005411895269906259,
1377
- "loss": 3.8318,
1378
  "step": 9200
1379
  },
1380
  {
1381
- "epoch": 0.9955871273275213,
1382
- "grad_norm": 0.5722038745880127,
1383
- "learning_rate": 0.0005408662859605646,
1384
- "loss": 3.8376,
1385
  "step": 9250
1386
  },
1387
  {
1388
- "epoch": 1.0009686793671295,
1389
- "grad_norm": 0.6226180195808411,
1390
- "learning_rate": 0.0005405430449305031,
1391
- "loss": 3.8294,
1392
  "step": 9300
1393
  },
1394
  {
1395
- "epoch": 1.0063502314067376,
1396
- "grad_norm": 0.6096346974372864,
1397
- "learning_rate": 0.0005402198039004416,
1398
- "loss": 3.7878,
1399
  "step": 9350
1400
  },
1401
  {
1402
- "epoch": 1.011731783446346,
1403
- "grad_norm": 0.5371617078781128,
1404
- "learning_rate": 0.0005398965628703803,
1405
- "loss": 3.7627,
1406
  "step": 9400
1407
  },
1408
  {
1409
- "epoch": 1.017113335485954,
1410
- "grad_norm": 0.6130861639976501,
1411
- "learning_rate": 0.0005395733218403189,
1412
- "loss": 3.769,
1413
  "step": 9450
1414
  },
1415
  {
1416
- "epoch": 1.0224948875255624,
1417
- "grad_norm": 0.5652937889099121,
1418
- "learning_rate": 0.0005392500808102575,
1419
- "loss": 3.7714,
1420
  "step": 9500
1421
  },
1422
  {
1423
- "epoch": 1.0278764395651705,
1424
- "grad_norm": 0.5810158252716064,
1425
- "learning_rate": 0.000538926839780196,
1426
- "loss": 3.7566,
1427
  "step": 9550
1428
  },
1429
  {
1430
- "epoch": 1.0332579916047788,
1431
- "grad_norm": 0.6366006731987,
1432
- "learning_rate": 0.0005386035987501346,
1433
- "loss": 3.7647,
1434
  "step": 9600
1435
  },
1436
  {
1437
- "epoch": 1.0386395436443872,
1438
- "grad_norm": 0.5298298597335815,
1439
- "learning_rate": 0.0005382803577200732,
1440
- "loss": 3.7488,
1441
  "step": 9650
1442
  },
1443
  {
1444
- "epoch": 1.0440210956839953,
1445
- "grad_norm": 1.2535181045532227,
1446
- "learning_rate": 0.0005379571166900119,
1447
- "loss": 3.7539,
1448
  "step": 9700
1449
  },
1450
  {
1451
- "epoch": 1.0494026477236036,
1452
- "grad_norm": 0.601189374923706,
1453
- "learning_rate": 0.0005376338756599504,
1454
- "loss": 3.7654,
1455
  "step": 9750
1456
  },
1457
  {
1458
- "epoch": 1.0547841997632117,
1459
- "grad_norm": 0.6602137684822083,
1460
- "learning_rate": 0.0005373106346298889,
1461
- "loss": 3.7686,
1462
  "step": 9800
1463
  },
1464
  {
1465
- "epoch": 1.06016575180282,
1466
- "grad_norm": 0.6176168918609619,
1467
- "learning_rate": 0.0005369873935998276,
1468
- "loss": 3.7589,
1469
  "step": 9850
1470
  },
1471
  {
1472
- "epoch": 1.0655473038424281,
1473
- "grad_norm": 0.6260287165641785,
1474
- "learning_rate": 0.0005366641525697661,
1475
- "loss": 3.7655,
1476
  "step": 9900
1477
  },
1478
  {
1479
- "epoch": 1.0709288558820365,
1480
- "grad_norm": 0.5720372796058655,
1481
- "learning_rate": 0.0005363409115397048,
1482
- "loss": 3.7559,
1483
  "step": 9950
1484
  },
1485
  {
1486
- "epoch": 1.0763104079216446,
1487
- "grad_norm": 0.5716636180877686,
1488
- "learning_rate": 0.0005360176705096433,
1489
- "loss": 3.7573,
1490
  "step": 10000
1491
  },
1492
  {
1493
- "epoch": 1.0763104079216446,
1494
- "eval_accuracy": 0.3451164492757796,
1495
- "eval_loss": 3.7458271980285645,
1496
- "eval_runtime": 185.1344,
1497
- "eval_samples_per_second": 97.286,
1498
- "eval_steps_per_second": 6.082,
1499
  "step": 10000
1500
  }
1501
  ],
1502
  "logging_steps": 50,
1503
- "max_steps": 92910,
1504
  "num_input_tokens_seen": 0,
1505
  "num_train_epochs": 10,
1506
  "save_steps": 10000,
@@ -1516,7 +1516,7 @@
1516
  "attributes": {}
1517
  }
1518
  },
1519
- "total_flos": 8.3610314735616e+16,
1520
  "train_batch_size": 32,
1521
  "trial_name": null,
1522
  "trial_params": null
 
1
  {
2
+ "best_metric": 3.753229856491089,
3
+ "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_high_2000_495/checkpoint-10000",
4
+ "epoch": 1.0781671159029649,
5
  "eval_steps": 1000,
6
  "global_step": 10000,
7
  "is_hyper_param_search": false,
 
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.005390835579514825,
13
+ "grad_norm": 1.4431322813034058,
14
+ "learning_rate": 0.000294,
15
+ "loss": 8.5391,
16
  "step": 50
17
  },
18
  {
19
+ "epoch": 0.01078167115902965,
20
+ "grad_norm": 3.01206111907959,
21
+ "learning_rate": 0.0005939999999999999,
22
+ "loss": 6.9498,
23
  "step": 100
24
  },
25
  {
26
+ "epoch": 0.016172506738544475,
27
+ "grad_norm": 2.0440971851348877,
28
+ "learning_rate": 0.0005996826767404209,
29
+ "loss": 6.5454,
30
  "step": 150
31
  },
32
  {
33
+ "epoch": 0.0215633423180593,
34
+ "grad_norm": 1.441593885421753,
35
+ "learning_rate": 0.0005993588774959525,
36
+ "loss": 6.29,
37
  "step": 200
38
  },
39
  {
40
+ "epoch": 0.026954177897574125,
41
+ "grad_norm": 1.1196218729019165,
42
+ "learning_rate": 0.000599035078251484,
43
  "loss": 6.1031,
44
  "step": 250
45
  },
46
  {
47
+ "epoch": 0.03234501347708895,
48
+ "grad_norm": 1.7114529609680176,
49
+ "learning_rate": 0.0005987112790070156,
50
+ "loss": 5.9896,
51
  "step": 300
52
  },
53
  {
54
+ "epoch": 0.03773584905660377,
55
+ "grad_norm": 1.0614324808120728,
56
+ "learning_rate": 0.0005983874797625472,
57
+ "loss": 5.8898,
58
  "step": 350
59
  },
60
  {
61
+ "epoch": 0.0431266846361186,
62
+ "grad_norm": 0.9476490616798401,
63
+ "learning_rate": 0.0005980636805180788,
64
+ "loss": 5.8227,
65
  "step": 400
66
  },
67
  {
68
+ "epoch": 0.04851752021563342,
69
+ "grad_norm": 1.198143482208252,
70
+ "learning_rate": 0.0005977398812736103,
71
+ "loss": 5.7384,
72
  "step": 450
73
  },
74
  {
75
+ "epoch": 0.05390835579514825,
76
+ "grad_norm": 1.5752428770065308,
77
+ "learning_rate": 0.0005974160820291419,
78
+ "loss": 5.6852,
79
  "step": 500
80
  },
81
  {
82
+ "epoch": 0.05929919137466307,
83
+ "grad_norm": 2.0579349994659424,
84
+ "learning_rate": 0.0005970922827846734,
85
+ "loss": 5.5695,
86
  "step": 550
87
  },
88
  {
89
+ "epoch": 0.0646900269541779,
90
+ "grad_norm": 1.3935744762420654,
91
+ "learning_rate": 0.0005967684835402051,
92
+ "loss": 5.535,
93
  "step": 600
94
  },
95
  {
96
+ "epoch": 0.07008086253369272,
97
+ "grad_norm": 1.303139090538025,
98
+ "learning_rate": 0.0005964446842957366,
99
+ "loss": 5.4307,
100
  "step": 650
101
  },
102
  {
103
+ "epoch": 0.07547169811320754,
104
+ "grad_norm": 0.969357967376709,
105
+ "learning_rate": 0.0005961208850512682,
106
+ "loss": 5.4075,
107
  "step": 700
108
  },
109
  {
110
+ "epoch": 0.08086253369272237,
111
+ "grad_norm": 1.0970923900604248,
112
+ "learning_rate": 0.0005957970858067997,
113
+ "loss": 5.3339,
114
  "step": 750
115
  },
116
  {
117
+ "epoch": 0.0862533692722372,
118
+ "grad_norm": 1.444526195526123,
119
+ "learning_rate": 0.0005954732865623314,
120
+ "loss": 5.2734,
121
  "step": 800
122
  },
123
  {
124
+ "epoch": 0.09164420485175202,
125
+ "grad_norm": 1.2398624420166016,
126
+ "learning_rate": 0.0005951494873178628,
127
+ "loss": 5.2435,
128
  "step": 850
129
  },
130
  {
131
+ "epoch": 0.09703504043126684,
132
+ "grad_norm": 1.0231698751449585,
133
+ "learning_rate": 0.0005948256880733944,
134
+ "loss": 5.1615,
135
  "step": 900
136
  },
137
  {
138
+ "epoch": 0.10242587601078167,
139
+ "grad_norm": 0.8052211403846741,
140
+ "learning_rate": 0.000594501888828926,
141
+ "loss": 5.1565,
142
  "step": 950
143
  },
144
  {
145
+ "epoch": 0.1078167115902965,
146
+ "grad_norm": 0.9758635759353638,
147
+ "learning_rate": 0.0005941780895844576,
148
+ "loss": 5.0973,
149
  "step": 1000
150
  },
151
  {
152
+ "epoch": 0.1078167115902965,
153
+ "eval_accuracy": 0.22682387725439804,
154
+ "eval_loss": 5.0240583419799805,
155
+ "eval_runtime": 187.7541,
156
+ "eval_samples_per_second": 95.929,
157
+ "eval_steps_per_second": 5.997,
158
  "step": 1000
159
  },
160
  {
161
+ "epoch": 0.11320754716981132,
162
+ "grad_norm": 1.1241644620895386,
163
+ "learning_rate": 0.0005938542903399891,
164
+ "loss": 5.0594,
165
  "step": 1050
166
  },
167
  {
168
+ "epoch": 0.11859838274932614,
169
+ "grad_norm": 1.3577865362167358,
170
+ "learning_rate": 0.0005935304910955207,
171
+ "loss": 5.0184,
172
  "step": 1100
173
  },
174
  {
175
+ "epoch": 0.12398921832884097,
176
+ "grad_norm": 1.2023968696594238,
177
+ "learning_rate": 0.0005932066918510523,
178
+ "loss": 4.9997,
179
  "step": 1150
180
  },
181
  {
182
+ "epoch": 0.1293800539083558,
183
+ "grad_norm": 1.0888073444366455,
184
+ "learning_rate": 0.0005928828926065839,
185
+ "loss": 4.9799,
186
  "step": 1200
187
  },
188
  {
189
+ "epoch": 0.1347708894878706,
190
+ "grad_norm": 1.1558417081832886,
191
+ "learning_rate": 0.0005925590933621154,
192
+ "loss": 4.9525,
193
  "step": 1250
194
  },
195
  {
196
+ "epoch": 0.14016172506738545,
197
+ "grad_norm": 0.8390737771987915,
198
+ "learning_rate": 0.000592235294117647,
199
+ "loss": 4.8991,
200
  "step": 1300
201
  },
202
  {
203
+ "epoch": 0.14555256064690028,
204
+ "grad_norm": 1.2147760391235352,
205
+ "learning_rate": 0.0005919114948731785,
206
+ "loss": 4.8807,
207
  "step": 1350
208
  },
209
  {
210
+ "epoch": 0.1509433962264151,
211
+ "grad_norm": 0.8272018432617188,
212
+ "learning_rate": 0.0005915876956287102,
213
+ "loss": 4.8533,
214
  "step": 1400
215
  },
216
  {
217
+ "epoch": 0.15633423180592992,
218
+ "grad_norm": 0.9127699732780457,
219
+ "learning_rate": 0.0005912638963842416,
220
+ "loss": 4.8361,
221
  "step": 1450
222
  },
223
  {
224
+ "epoch": 0.16172506738544473,
225
+ "grad_norm": 1.053229808807373,
226
+ "learning_rate": 0.0005909400971397733,
227
+ "loss": 4.822,
228
  "step": 1500
229
  },
230
  {
231
+ "epoch": 0.16711590296495957,
232
+ "grad_norm": 0.7915006279945374,
233
+ "learning_rate": 0.0005906162978953049,
234
+ "loss": 4.7808,
235
  "step": 1550
236
  },
237
  {
238
+ "epoch": 0.1725067385444744,
239
+ "grad_norm": 1.0942384004592896,
240
+ "learning_rate": 0.0005902924986508364,
241
+ "loss": 4.7932,
242
  "step": 1600
243
  },
244
  {
245
+ "epoch": 0.1778975741239892,
246
+ "grad_norm": 1.114819884300232,
247
+ "learning_rate": 0.000589968699406368,
248
+ "loss": 4.7716,
249
  "step": 1650
250
  },
251
  {
252
+ "epoch": 0.18328840970350405,
253
+ "grad_norm": 0.8059048652648926,
254
+ "learning_rate": 0.0005896449001618996,
255
+ "loss": 4.6935,
256
  "step": 1700
257
  },
258
  {
259
+ "epoch": 0.18867924528301888,
260
+ "grad_norm": 1.0432419776916504,
261
+ "learning_rate": 0.0005893211009174312,
262
+ "loss": 4.7057,
263
  "step": 1750
264
  },
265
  {
266
+ "epoch": 0.1940700808625337,
267
+ "grad_norm": 1.0190030336380005,
268
+ "learning_rate": 0.0005889973016729627,
269
+ "loss": 4.6937,
270
  "step": 1800
271
  },
272
  {
273
+ "epoch": 0.19946091644204852,
274
+ "grad_norm": 0.9340100884437561,
275
+ "learning_rate": 0.0005886735024284943,
276
+ "loss": 4.6527,
277
  "step": 1850
278
  },
279
  {
280
+ "epoch": 0.20485175202156333,
281
+ "grad_norm": 1.0046193599700928,
282
+ "learning_rate": 0.0005883497031840258,
283
+ "loss": 4.65,
284
  "step": 1900
285
  },
286
  {
287
+ "epoch": 0.21024258760107817,
288
+ "grad_norm": 1.0945899486541748,
289
+ "learning_rate": 0.0005880259039395575,
290
+ "loss": 4.6138,
291
  "step": 1950
292
  },
293
  {
294
+ "epoch": 0.215633423180593,
295
+ "grad_norm": 1.0984909534454346,
296
+ "learning_rate": 0.000587702104695089,
297
+ "loss": 4.5905,
298
  "step": 2000
299
  },
300
  {
301
+ "epoch": 0.215633423180593,
302
+ "eval_accuracy": 0.26831504687122604,
303
+ "eval_loss": 4.5331034660339355,
304
+ "eval_runtime": 188.7106,
305
+ "eval_samples_per_second": 95.442,
306
+ "eval_steps_per_second": 5.967,
307
  "step": 2000
308
  },
309
  {
310
+ "epoch": 0.2210242587601078,
311
+ "grad_norm": 0.9245219230651855,
312
+ "learning_rate": 0.0005873783054506206,
313
+ "loss": 4.6042,
314
  "step": 2050
315
  },
316
  {
317
+ "epoch": 0.22641509433962265,
318
+ "grad_norm": 1.0915874242782593,
319
+ "learning_rate": 0.0005870545062061521,
320
+ "loss": 4.5669,
321
  "step": 2100
322
  },
323
  {
324
+ "epoch": 0.23180592991913745,
325
+ "grad_norm": 1.1114705801010132,
326
+ "learning_rate": 0.0005867307069616837,
327
+ "loss": 4.513,
328
  "step": 2150
329
  },
330
  {
331
+ "epoch": 0.2371967654986523,
332
+ "grad_norm": 1.0507372617721558,
333
+ "learning_rate": 0.0005864069077172152,
334
+ "loss": 4.555,
335
  "step": 2200
336
  },
337
  {
338
+ "epoch": 0.24258760107816713,
339
+ "grad_norm": 0.7661042809486389,
340
+ "learning_rate": 0.0005860831084727468,
341
+ "loss": 4.4936,
342
  "step": 2250
343
  },
344
  {
345
+ "epoch": 0.24797843665768193,
346
+ "grad_norm": 0.9031612873077393,
347
+ "learning_rate": 0.0005857593092282784,
348
+ "loss": 4.4732,
349
  "step": 2300
350
  },
351
  {
352
+ "epoch": 0.25336927223719674,
353
+ "grad_norm": 0.8872079849243164,
354
+ "learning_rate": 0.00058543550998381,
355
+ "loss": 4.468,
356
  "step": 2350
357
  },
358
  {
359
+ "epoch": 0.2587601078167116,
360
+ "grad_norm": 1.007972002029419,
361
+ "learning_rate": 0.0005851117107393415,
362
+ "loss": 4.4588,
363
  "step": 2400
364
  },
365
  {
366
+ "epoch": 0.2641509433962264,
367
+ "grad_norm": 0.9107605218887329,
368
+ "learning_rate": 0.0005847879114948731,
369
+ "loss": 4.4611,
370
  "step": 2450
371
  },
372
  {
373
+ "epoch": 0.2695417789757412,
374
+ "grad_norm": 0.8212679624557495,
375
+ "learning_rate": 0.0005844641122504048,
376
+ "loss": 4.4549,
377
  "step": 2500
378
  },
379
  {
380
+ "epoch": 0.2749326145552561,
381
+ "grad_norm": 0.8247270584106445,
382
+ "learning_rate": 0.0005841403130059363,
383
+ "loss": 4.4192,
384
  "step": 2550
385
  },
386
  {
387
+ "epoch": 0.2803234501347709,
388
+ "grad_norm": 0.8410664796829224,
389
+ "learning_rate": 0.0005838165137614679,
390
+ "loss": 4.4058,
391
  "step": 2600
392
  },
393
  {
394
+ "epoch": 0.2857142857142857,
395
+ "grad_norm": 0.8341009616851807,
396
+ "learning_rate": 0.0005834927145169994,
397
+ "loss": 4.4019,
398
  "step": 2650
399
  },
400
  {
401
+ "epoch": 0.29110512129380056,
402
+ "grad_norm": 0.7168475985527039,
403
+ "learning_rate": 0.000583168915272531,
404
+ "loss": 4.3961,
405
  "step": 2700
406
  },
407
  {
408
+ "epoch": 0.29649595687331537,
409
+ "grad_norm": 0.9485689997673035,
410
+ "learning_rate": 0.0005828451160280626,
411
+ "loss": 4.3549,
412
  "step": 2750
413
  },
414
  {
415
+ "epoch": 0.3018867924528302,
416
+ "grad_norm": 0.7494977116584778,
417
+ "learning_rate": 0.0005825213167835941,
418
+ "loss": 4.394,
419
  "step": 2800
420
  },
421
  {
422
+ "epoch": 0.30727762803234504,
423
+ "grad_norm": 0.7997616529464722,
424
+ "learning_rate": 0.0005821975175391257,
425
+ "loss": 4.3639,
426
  "step": 2850
427
  },
428
  {
429
+ "epoch": 0.31266846361185985,
430
+ "grad_norm": 0.7571331262588501,
431
+ "learning_rate": 0.0005818737182946573,
432
+ "loss": 4.3431,
433
  "step": 2900
434
  },
435
  {
436
+ "epoch": 0.31805929919137466,
437
+ "grad_norm": 0.7700881958007812,
438
+ "learning_rate": 0.0005815499190501888,
439
+ "loss": 4.3449,
440
  "step": 2950
441
  },
442
  {
443
+ "epoch": 0.32345013477088946,
444
+ "grad_norm": 0.670320451259613,
445
+ "learning_rate": 0.0005812261198057204,
446
+ "loss": 4.3425,
447
  "step": 3000
448
  },
449
  {
450
+ "epoch": 0.32345013477088946,
451
+ "eval_accuracy": 0.2970671000033574,
452
+ "eval_loss": 4.244440078735352,
453
+ "eval_runtime": 181.2734,
454
+ "eval_samples_per_second": 99.358,
455
+ "eval_steps_per_second": 6.212,
456
  "step": 3000
457
  },
458
  {
459
+ "epoch": 0.3288409703504043,
460
+ "grad_norm": 0.7345431447029114,
461
+ "learning_rate": 0.0005809023205612519,
462
+ "loss": 4.3171,
463
  "step": 3050
464
  },
465
  {
466
+ "epoch": 0.33423180592991913,
467
+ "grad_norm": 0.7953465580940247,
468
+ "learning_rate": 0.0005805785213167836,
469
+ "loss": 4.3147,
470
  "step": 3100
471
  },
472
  {
473
+ "epoch": 0.33962264150943394,
474
+ "grad_norm": 0.7929427027702332,
475
+ "learning_rate": 0.0005802547220723151,
476
+ "loss": 4.2801,
477
  "step": 3150
478
  },
479
  {
480
+ "epoch": 0.3450134770889488,
481
+ "grad_norm": 0.8242636322975159,
482
+ "learning_rate": 0.0005799309228278467,
483
+ "loss": 4.2861,
484
  "step": 3200
485
  },
486
  {
487
+ "epoch": 0.3504043126684636,
488
+ "grad_norm": 0.9226231575012207,
489
+ "learning_rate": 0.0005796071235833782,
490
+ "loss": 4.2607,
491
  "step": 3250
492
  },
493
  {
494
+ "epoch": 0.3557951482479784,
495
+ "grad_norm": 0.8058574199676514,
496
+ "learning_rate": 0.0005792833243389099,
497
+ "loss": 4.2833,
498
  "step": 3300
499
  },
500
  {
501
+ "epoch": 0.3611859838274933,
502
+ "grad_norm": 0.5928198099136353,
503
+ "learning_rate": 0.0005789595250944414,
504
+ "loss": 4.262,
505
  "step": 3350
506
  },
507
  {
508
+ "epoch": 0.3665768194070081,
509
+ "grad_norm": 0.7675554752349854,
510
+ "learning_rate": 0.000578635725849973,
511
+ "loss": 4.2542,
512
  "step": 3400
513
  },
514
  {
515
+ "epoch": 0.3719676549865229,
516
+ "grad_norm": 0.611778974533081,
517
+ "learning_rate": 0.0005783119266055045,
518
+ "loss": 4.2414,
519
  "step": 3450
520
  },
521
  {
522
+ "epoch": 0.37735849056603776,
523
+ "grad_norm": 0.6161898970603943,
524
+ "learning_rate": 0.0005779881273610361,
525
+ "loss": 4.2189,
526
  "step": 3500
527
  },
528
  {
529
+ "epoch": 0.38274932614555257,
530
+ "grad_norm": 0.739580512046814,
531
+ "learning_rate": 0.0005776643281165676,
532
+ "loss": 4.2229,
533
  "step": 3550
534
  },
535
  {
536
+ "epoch": 0.3881401617250674,
537
+ "grad_norm": 0.8803327679634094,
538
+ "learning_rate": 0.0005773405288720992,
539
+ "loss": 4.2268,
540
  "step": 3600
541
  },
542
  {
543
+ "epoch": 0.3935309973045822,
544
+ "grad_norm": 1.1461224555969238,
545
+ "learning_rate": 0.0005770167296276308,
546
+ "loss": 4.2344,
547
  "step": 3650
548
  },
549
  {
550
+ "epoch": 0.39892183288409705,
551
+ "grad_norm": 0.9922895431518555,
552
+ "learning_rate": 0.0005766929303831624,
553
+ "loss": 4.2043,
554
  "step": 3700
555
  },
556
  {
557
+ "epoch": 0.40431266846361186,
558
+ "grad_norm": 0.7285529375076294,
559
+ "learning_rate": 0.000576369131138694,
560
+ "loss": 4.2012,
561
  "step": 3750
562
  },
563
  {
564
+ "epoch": 0.40970350404312667,
565
+ "grad_norm": 0.809013307094574,
566
+ "learning_rate": 0.0005760453318942255,
567
+ "loss": 4.2178,
568
  "step": 3800
569
  },
570
  {
571
+ "epoch": 0.41509433962264153,
572
+ "grad_norm": 0.8063985109329224,
573
+ "learning_rate": 0.000575721532649757,
574
+ "loss": 4.2018,
575
  "step": 3850
576
  },
577
  {
578
+ "epoch": 0.42048517520215634,
579
+ "grad_norm": 0.847189724445343,
580
+ "learning_rate": 0.0005753977334052887,
581
+ "loss": 4.1669,
582
  "step": 3900
583
  },
584
  {
585
+ "epoch": 0.42587601078167114,
586
+ "grad_norm": 0.6911249756813049,
587
+ "learning_rate": 0.0005750739341608203,
588
+ "loss": 4.1661,
589
  "step": 3950
590
  },
591
  {
592
+ "epoch": 0.431266846361186,
593
+ "grad_norm": 0.7047509551048279,
594
+ "learning_rate": 0.0005747501349163518,
595
+ "loss": 4.1711,
596
  "step": 4000
597
  },
598
  {
599
+ "epoch": 0.431266846361186,
600
+ "eval_accuracy": 0.31155715777518433,
601
+ "eval_loss": 4.099443435668945,
602
+ "eval_runtime": 181.2897,
603
+ "eval_samples_per_second": 99.349,
604
+ "eval_steps_per_second": 6.211,
605
  "step": 4000
606
  },
607
  {
608
+ "epoch": 0.4366576819407008,
609
+ "grad_norm": 0.6963608264923096,
610
+ "learning_rate": 0.0005744263356718834,
611
+ "loss": 4.1712,
612
  "step": 4050
613
  },
614
  {
615
+ "epoch": 0.4420485175202156,
616
+ "grad_norm": 0.5528955459594727,
617
+ "learning_rate": 0.000574102536427415,
618
+ "loss": 4.1704,
619
  "step": 4100
620
  },
621
  {
622
+ "epoch": 0.4474393530997305,
623
+ "grad_norm": 0.6239518523216248,
624
+ "learning_rate": 0.0005737787371829465,
625
+ "loss": 4.1571,
626
  "step": 4150
627
  },
628
  {
629
+ "epoch": 0.4528301886792453,
630
+ "grad_norm": 0.6668544411659241,
631
+ "learning_rate": 0.0005734549379384781,
632
+ "loss": 4.1498,
633
  "step": 4200
634
  },
635
  {
636
+ "epoch": 0.4582210242587601,
637
+ "grad_norm": 0.6054083704948425,
638
+ "learning_rate": 0.0005731311386940097,
639
+ "loss": 4.1415,
640
  "step": 4250
641
  },
642
  {
643
+ "epoch": 0.4636118598382749,
644
+ "grad_norm": 0.6274952292442322,
645
+ "learning_rate": 0.0005728073394495412,
646
+ "loss": 4.1351,
647
  "step": 4300
648
  },
649
  {
650
+ "epoch": 0.46900269541778977,
651
+ "grad_norm": 0.6964167952537537,
652
+ "learning_rate": 0.0005724835402050728,
653
+ "loss": 4.1408,
654
  "step": 4350
655
  },
656
  {
657
+ "epoch": 0.4743935309973046,
658
+ "grad_norm": 0.7939971685409546,
659
+ "learning_rate": 0.0005721597409606043,
660
+ "loss": 4.1304,
661
  "step": 4400
662
  },
663
  {
664
+ "epoch": 0.4797843665768194,
665
+ "grad_norm": 0.7325452566146851,
666
+ "learning_rate": 0.000571835941716136,
667
+ "loss": 4.1334,
668
  "step": 4450
669
  },
670
  {
671
+ "epoch": 0.48517520215633425,
672
+ "grad_norm": 0.6969997882843018,
673
+ "learning_rate": 0.0005715121424716675,
674
+ "loss": 4.1142,
675
  "step": 4500
676
  },
677
  {
678
+ "epoch": 0.49056603773584906,
679
+ "grad_norm": 0.6884226202964783,
680
+ "learning_rate": 0.0005711883432271991,
681
+ "loss": 4.1028,
682
  "step": 4550
683
  },
684
  {
685
+ "epoch": 0.49595687331536387,
686
+ "grad_norm": 0.6321718096733093,
687
+ "learning_rate": 0.0005708645439827306,
688
+ "loss": 4.1078,
689
  "step": 4600
690
  },
691
  {
692
+ "epoch": 0.5013477088948787,
693
+ "grad_norm": 0.701016366481781,
694
+ "learning_rate": 0.0005705407447382623,
695
+ "loss": 4.0883,
696
  "step": 4650
697
  },
698
  {
699
+ "epoch": 0.5067385444743935,
700
+ "grad_norm": 0.7334260940551758,
701
+ "learning_rate": 0.0005702169454937938,
702
+ "loss": 4.0958,
703
  "step": 4700
704
  },
705
  {
706
+ "epoch": 0.5121293800539084,
707
+ "grad_norm": 0.5910300612449646,
708
+ "learning_rate": 0.0005698931462493253,
709
+ "loss": 4.075,
710
  "step": 4750
711
  },
712
  {
713
+ "epoch": 0.5175202156334232,
714
+ "grad_norm": 0.6013472080230713,
715
+ "learning_rate": 0.0005695693470048569,
716
+ "loss": 4.0922,
717
  "step": 4800
718
  },
719
  {
720
+ "epoch": 0.522911051212938,
721
+ "grad_norm": 0.6121587753295898,
722
+ "learning_rate": 0.0005692455477603885,
723
+ "loss": 4.0841,
724
  "step": 4850
725
  },
726
  {
727
+ "epoch": 0.5283018867924528,
728
+ "grad_norm": 0.613993227481842,
729
+ "learning_rate": 0.00056892174851592,
730
+ "loss": 4.0931,
731
  "step": 4900
732
  },
733
  {
734
+ "epoch": 0.5336927223719676,
735
+ "grad_norm": 0.6811482310295105,
736
+ "learning_rate": 0.0005685979492714516,
737
+ "loss": 4.0664,
738
  "step": 4950
739
  },
740
  {
741
+ "epoch": 0.5390835579514824,
742
+ "grad_norm": 0.6156988739967346,
743
+ "learning_rate": 0.0005682741500269833,
744
+ "loss": 4.0584,
745
  "step": 5000
746
  },
747
  {
748
+ "epoch": 0.5390835579514824,
749
+ "eval_accuracy": 0.3207276788125022,
750
+ "eval_loss": 3.9989216327667236,
751
+ "eval_runtime": 181.0844,
752
+ "eval_samples_per_second": 99.462,
753
+ "eval_steps_per_second": 6.218,
754
  "step": 5000
755
  },
756
  {
757
+ "epoch": 0.5444743935309974,
758
+ "grad_norm": 0.5171638131141663,
759
+ "learning_rate": 0.0005679503507825148,
760
+ "loss": 4.0485,
761
  "step": 5050
762
  },
763
  {
764
+ "epoch": 0.5498652291105122,
765
+ "grad_norm": 0.8664311170578003,
766
+ "learning_rate": 0.0005676265515380464,
767
+ "loss": 4.0568,
768
  "step": 5100
769
  },
770
  {
771
+ "epoch": 0.555256064690027,
772
+ "grad_norm": 0.5895658731460571,
773
+ "learning_rate": 0.0005673027522935779,
774
+ "loss": 4.0612,
775
  "step": 5150
776
  },
777
  {
778
+ "epoch": 0.5606469002695418,
779
+ "grad_norm": 0.7551826238632202,
780
+ "learning_rate": 0.0005669789530491095,
781
+ "loss": 4.0396,
782
  "step": 5200
783
  },
784
  {
785
+ "epoch": 0.5660377358490566,
786
+ "grad_norm": 0.6080760359764099,
787
+ "learning_rate": 0.0005666551538046411,
788
+ "loss": 4.0427,
789
  "step": 5250
790
  },
791
  {
792
+ "epoch": 0.5714285714285714,
793
+ "grad_norm": 0.6198018193244934,
794
+ "learning_rate": 0.0005663313545601727,
795
+ "loss": 4.0237,
796
  "step": 5300
797
  },
798
  {
799
+ "epoch": 0.5768194070080862,
800
+ "grad_norm": 0.6986076235771179,
801
+ "learning_rate": 0.0005660075553157042,
802
+ "loss": 4.037,
803
  "step": 5350
804
  },
805
  {
806
+ "epoch": 0.5822102425876011,
807
+ "grad_norm": 0.6296870708465576,
808
+ "learning_rate": 0.0005656837560712358,
809
+ "loss": 4.0294,
810
  "step": 5400
811
  },
812
  {
813
+ "epoch": 0.5876010781671159,
814
+ "grad_norm": 0.7363589406013489,
815
+ "learning_rate": 0.0005653599568267674,
816
+ "loss": 4.0366,
817
  "step": 5450
818
  },
819
  {
820
+ "epoch": 0.5929919137466307,
821
+ "grad_norm": 0.5726944804191589,
822
+ "learning_rate": 0.0005650361575822989,
823
+ "loss": 4.038,
824
  "step": 5500
825
  },
826
  {
827
+ "epoch": 0.5983827493261455,
828
+ "grad_norm": 0.6348768472671509,
829
+ "learning_rate": 0.0005647123583378305,
830
+ "loss": 4.0435,
831
  "step": 5550
832
  },
833
  {
834
+ "epoch": 0.6037735849056604,
835
+ "grad_norm": 0.6801629662513733,
836
+ "learning_rate": 0.0005643885590933621,
837
+ "loss": 4.0221,
838
  "step": 5600
839
  },
840
  {
841
+ "epoch": 0.6091644204851752,
842
+ "grad_norm": 0.5451263785362244,
843
+ "learning_rate": 0.0005640647598488936,
844
+ "loss": 4.0095,
845
  "step": 5650
846
  },
847
  {
848
+ "epoch": 0.6145552560646901,
849
+ "grad_norm": 0.6517798900604248,
850
+ "learning_rate": 0.0005637409606044252,
851
+ "loss": 4.0171,
852
  "step": 5700
853
  },
854
  {
855
+ "epoch": 0.6199460916442049,
856
+ "grad_norm": 0.5931031107902527,
857
+ "learning_rate": 0.0005634171613599567,
858
+ "loss": 4.0271,
859
  "step": 5750
860
  },
861
  {
862
+ "epoch": 0.6253369272237197,
863
+ "grad_norm": 0.6558272838592529,
864
+ "learning_rate": 0.0005630933621154884,
865
+ "loss": 4.0187,
866
  "step": 5800
867
  },
868
  {
869
+ "epoch": 0.6307277628032345,
870
+ "grad_norm": 0.6062823534011841,
871
+ "learning_rate": 0.0005627695628710199,
872
+ "loss": 4.0266,
873
  "step": 5850
874
  },
875
  {
876
+ "epoch": 0.6361185983827493,
877
+ "grad_norm": 0.6019988656044006,
878
+ "learning_rate": 0.0005624457636265515,
879
+ "loss": 3.996,
880
  "step": 5900
881
  },
882
  {
883
+ "epoch": 0.6415094339622641,
884
+ "grad_norm": 0.6201125383377075,
885
+ "learning_rate": 0.000562121964382083,
886
+ "loss": 3.9898,
887
  "step": 5950
888
  },
889
  {
890
+ "epoch": 0.6469002695417789,
891
+ "grad_norm": 0.6050779819488525,
892
+ "learning_rate": 0.0005617981651376146,
893
+ "loss": 4.0004,
894
  "step": 6000
895
  },
896
  {
897
+ "epoch": 0.6469002695417789,
898
+ "eval_accuracy": 0.32684505370223305,
899
+ "eval_loss": 3.9261674880981445,
900
+ "eval_runtime": 185.9413,
901
+ "eval_samples_per_second": 96.864,
902
+ "eval_steps_per_second": 6.056,
903
  "step": 6000
904
  },
905
  {
906
+ "epoch": 0.6522911051212938,
907
+ "grad_norm": 0.6865056753158569,
908
+ "learning_rate": 0.0005614743658931463,
909
+ "loss": 3.9731,
910
  "step": 6050
911
  },
912
  {
913
+ "epoch": 0.6576819407008087,
914
+ "grad_norm": 0.7325416803359985,
915
+ "learning_rate": 0.0005611505666486777,
916
+ "loss": 4.0069,
917
  "step": 6100
918
  },
919
  {
920
+ "epoch": 0.6630727762803235,
921
+ "grad_norm": 0.5644854307174683,
922
+ "learning_rate": 0.0005608267674042094,
923
+ "loss": 3.9862,
924
  "step": 6150
925
  },
926
  {
927
+ "epoch": 0.6684636118598383,
928
+ "grad_norm": 0.6596018075942993,
929
+ "learning_rate": 0.0005605029681597409,
930
+ "loss": 3.9661,
931
  "step": 6200
932
  },
933
  {
934
+ "epoch": 0.6738544474393531,
935
+ "grad_norm": 0.6351631879806519,
936
+ "learning_rate": 0.0005601791689152725,
937
+ "loss": 3.9823,
938
  "step": 6250
939
  },
940
  {
941
+ "epoch": 0.6792452830188679,
942
+ "grad_norm": 0.7349021434783936,
943
+ "learning_rate": 0.000559855369670804,
944
+ "loss": 4.0002,
945
  "step": 6300
946
  },
947
  {
948
+ "epoch": 0.6846361185983828,
949
+ "grad_norm": 0.6522649526596069,
950
+ "learning_rate": 0.0005595315704263357,
951
+ "loss": 3.9623,
952
  "step": 6350
953
  },
954
  {
955
+ "epoch": 0.6900269541778976,
956
+ "grad_norm": 0.5307708978652954,
957
+ "learning_rate": 0.0005592077711818672,
958
+ "loss": 3.9669,
959
  "step": 6400
960
  },
961
  {
962
+ "epoch": 0.6954177897574124,
963
+ "grad_norm": 0.5935479998588562,
964
+ "learning_rate": 0.0005588839719373988,
965
+ "loss": 3.9545,
966
  "step": 6450
967
  },
968
  {
969
+ "epoch": 0.7008086253369272,
970
+ "grad_norm": 0.6537756323814392,
971
+ "learning_rate": 0.0005585601726929303,
972
+ "loss": 3.9651,
973
  "step": 6500
974
  },
975
  {
976
+ "epoch": 0.706199460916442,
977
+ "grad_norm": 0.5890859961509705,
978
+ "learning_rate": 0.0005582363734484619,
979
+ "loss": 3.9888,
980
  "step": 6550
981
  },
982
  {
983
+ "epoch": 0.7115902964959568,
984
+ "grad_norm": 0.6212690472602844,
985
+ "learning_rate": 0.0005579125742039935,
986
+ "loss": 3.9448,
987
  "step": 6600
988
  },
989
  {
990
+ "epoch": 0.7169811320754716,
991
+ "grad_norm": 0.6124947667121887,
992
+ "learning_rate": 0.0005575887749595251,
993
+ "loss": 3.9736,
994
  "step": 6650
995
  },
996
  {
997
+ "epoch": 0.7223719676549866,
998
+ "grad_norm": 0.5879746675491333,
999
+ "learning_rate": 0.0005572649757150566,
1000
+ "loss": 3.9494,
1001
  "step": 6700
1002
  },
1003
  {
1004
+ "epoch": 0.7277628032345014,
1005
+ "grad_norm": 0.5893728733062744,
1006
+ "learning_rate": 0.0005569411764705882,
1007
+ "loss": 3.9548,
1008
  "step": 6750
1009
  },
1010
  {
1011
+ "epoch": 0.7331536388140162,
1012
+ "grad_norm": 0.626844584941864,
1013
+ "learning_rate": 0.0005566173772261198,
1014
+ "loss": 3.9442,
1015
  "step": 6800
1016
  },
1017
  {
1018
+ "epoch": 0.738544474393531,
1019
+ "grad_norm": 0.6960020065307617,
1020
+ "learning_rate": 0.0005562935779816513,
1021
+ "loss": 3.9258,
1022
  "step": 6850
1023
  },
1024
  {
1025
+ "epoch": 0.7439353099730458,
1026
+ "grad_norm": 0.6511778235435486,
1027
+ "learning_rate": 0.0005559697787371828,
1028
+ "loss": 3.9527,
1029
  "step": 6900
1030
  },
1031
  {
1032
+ "epoch": 0.7493261455525606,
1033
+ "grad_norm": 0.6318475604057312,
1034
+ "learning_rate": 0.0005556459794927145,
1035
+ "loss": 3.9468,
1036
  "step": 6950
1037
  },
1038
  {
1039
+ "epoch": 0.7547169811320755,
1040
+ "grad_norm": 0.7971281409263611,
1041
+ "learning_rate": 0.000555322180248246,
1042
+ "loss": 3.9359,
1043
  "step": 7000
1044
  },
1045
  {
1046
+ "epoch": 0.7547169811320755,
1047
+ "eval_accuracy": 0.3327894531945633,
1048
+ "eval_loss": 3.869997501373291,
1049
+ "eval_runtime": 181.2103,
1050
+ "eval_samples_per_second": 99.393,
1051
+ "eval_steps_per_second": 6.214,
1052
  "step": 7000
1053
  },
1054
  {
1055
+ "epoch": 0.7601078167115903,
1056
+ "grad_norm": 0.6572127342224121,
1057
+ "learning_rate": 0.0005549983810037776,
1058
+ "loss": 3.9277,
1059
  "step": 7050
1060
  },
1061
  {
1062
+ "epoch": 0.7654986522911051,
1063
+ "grad_norm": 0.5682153105735779,
1064
+ "learning_rate": 0.0005546745817593091,
1065
+ "loss": 3.9367,
1066
  "step": 7100
1067
  },
1068
  {
1069
+ "epoch": 0.77088948787062,
1070
+ "grad_norm": 0.5704119801521301,
1071
+ "learning_rate": 0.0005543507825148408,
1072
+ "loss": 3.9349,
1073
  "step": 7150
1074
  },
1075
  {
1076
+ "epoch": 0.7762803234501348,
1077
+ "grad_norm": 0.6218463182449341,
1078
+ "learning_rate": 0.0005540269832703723,
1079
+ "loss": 3.9385,
1080
  "step": 7200
1081
  },
1082
  {
1083
+ "epoch": 0.7816711590296496,
1084
+ "grad_norm": 0.5849258303642273,
1085
+ "learning_rate": 0.0005537031840259039,
1086
+ "loss": 3.8932,
1087
  "step": 7250
1088
  },
1089
  {
1090
+ "epoch": 0.7870619946091644,
1091
+ "grad_norm": 0.5835541486740112,
1092
+ "learning_rate": 0.0005533793847814354,
1093
+ "loss": 3.9344,
1094
  "step": 7300
1095
  },
1096
  {
1097
+ "epoch": 0.7924528301886793,
1098
+ "grad_norm": 0.5649335980415344,
1099
+ "learning_rate": 0.000553055585536967,
1100
+ "loss": 3.9192,
1101
  "step": 7350
1102
  },
1103
  {
1104
+ "epoch": 0.7978436657681941,
1105
+ "grad_norm": 0.6361489295959473,
1106
+ "learning_rate": 0.0005527317862924987,
1107
+ "loss": 3.9193,
1108
  "step": 7400
1109
  },
1110
  {
1111
+ "epoch": 0.8032345013477089,
1112
+ "grad_norm": 0.5219013094902039,
1113
+ "learning_rate": 0.0005524079870480301,
1114
+ "loss": 3.9079,
1115
  "step": 7450
1116
  },
1117
  {
1118
+ "epoch": 0.8086253369272237,
1119
+ "grad_norm": 0.6409909129142761,
1120
+ "learning_rate": 0.0005520841878035618,
1121
+ "loss": 3.8963,
1122
  "step": 7500
1123
  },
1124
  {
1125
+ "epoch": 0.8140161725067385,
1126
+ "grad_norm": 0.6128087043762207,
1127
+ "learning_rate": 0.0005517603885590933,
1128
+ "loss": 3.9236,
1129
  "step": 7550
1130
  },
1131
  {
1132
+ "epoch": 0.8194070080862533,
1133
+ "grad_norm": 0.6087100505828857,
1134
+ "learning_rate": 0.0005514365893146249,
1135
+ "loss": 3.9237,
1136
  "step": 7600
1137
  },
1138
  {
1139
+ "epoch": 0.8247978436657682,
1140
+ "grad_norm": 0.6042186617851257,
1141
+ "learning_rate": 0.0005511127900701564,
1142
+ "loss": 3.8995,
1143
  "step": 7650
1144
  },
1145
  {
1146
+ "epoch": 0.8301886792452831,
1147
+ "grad_norm": 0.566831111907959,
1148
+ "learning_rate": 0.000550788990825688,
1149
+ "loss": 3.905,
1150
  "step": 7700
1151
  },
1152
  {
1153
+ "epoch": 0.8355795148247979,
1154
+ "grad_norm": 0.544748842716217,
1155
+ "learning_rate": 0.0005504651915812196,
1156
+ "loss": 3.8854,
1157
  "step": 7750
1158
  },
1159
  {
1160
+ "epoch": 0.8409703504043127,
1161
+ "grad_norm": 0.5742127299308777,
1162
+ "learning_rate": 0.0005501413923367512,
1163
+ "loss": 3.9136,
1164
  "step": 7800
1165
  },
1166
  {
1167
+ "epoch": 0.8463611859838275,
1168
+ "grad_norm": 0.6579509973526001,
1169
+ "learning_rate": 0.0005498175930922827,
1170
+ "loss": 3.8913,
1171
  "step": 7850
1172
  },
1173
  {
1174
+ "epoch": 0.8517520215633423,
1175
+ "grad_norm": 0.5239347219467163,
1176
+ "learning_rate": 0.0005494937938478143,
1177
+ "loss": 3.8935,
1178
  "step": 7900
1179
  },
1180
  {
1181
+ "epoch": 0.8571428571428571,
1182
+ "grad_norm": 0.6472986340522766,
1183
+ "learning_rate": 0.0005491699946033459,
1184
+ "loss": 3.8914,
1185
  "step": 7950
1186
  },
1187
  {
1188
+ "epoch": 0.862533692722372,
1189
+ "grad_norm": 0.5899927616119385,
1190
+ "learning_rate": 0.0005488461953588775,
1191
+ "loss": 3.8808,
1192
  "step": 8000
1193
  },
1194
  {
1195
+ "epoch": 0.862533692722372,
1196
+ "eval_accuracy": 0.33687349794173405,
1197
+ "eval_loss": 3.8257834911346436,
1198
+ "eval_runtime": 181.2863,
1199
+ "eval_samples_per_second": 99.351,
1200
+ "eval_steps_per_second": 6.211,
1201
  "step": 8000
1202
  },
1203
  {
1204
+ "epoch": 0.8679245283018868,
1205
+ "grad_norm": 0.5935076475143433,
1206
+ "learning_rate": 0.000548522396114409,
1207
+ "loss": 3.8878,
1208
  "step": 8050
1209
  },
1210
  {
1211
+ "epoch": 0.8733153638814016,
1212
+ "grad_norm": 0.6126903891563416,
1213
+ "learning_rate": 0.0005481985968699406,
1214
+ "loss": 3.873,
1215
  "step": 8100
1216
  },
1217
  {
1218
+ "epoch": 0.8787061994609164,
1219
+ "grad_norm": 0.5333702564239502,
1220
+ "learning_rate": 0.0005478747976254721,
1221
+ "loss": 3.8833,
1222
  "step": 8150
1223
  },
1224
  {
1225
+ "epoch": 0.8840970350404312,
1226
+ "grad_norm": 0.542793869972229,
1227
+ "learning_rate": 0.0005475509983810037,
1228
+ "loss": 3.874,
1229
  "step": 8200
1230
  },
1231
  {
1232
+ "epoch": 0.889487870619946,
1233
+ "grad_norm": 0.5891146659851074,
1234
+ "learning_rate": 0.0005472271991365352,
1235
+ "loss": 3.8604,
1236
  "step": 8250
1237
  },
1238
  {
1239
+ "epoch": 0.894878706199461,
1240
+ "grad_norm": 0.6112461090087891,
1241
+ "learning_rate": 0.0005469033998920669,
1242
+ "loss": 3.8849,
1243
  "step": 8300
1244
  },
1245
  {
1246
+ "epoch": 0.9002695417789758,
1247
+ "grad_norm": 0.5972184538841248,
1248
+ "learning_rate": 0.0005465796006475984,
1249
+ "loss": 3.8678,
1250
  "step": 8350
1251
  },
1252
  {
1253
+ "epoch": 0.9056603773584906,
1254
+ "grad_norm": 0.685249388217926,
1255
+ "learning_rate": 0.00054625580140313,
1256
+ "loss": 3.8825,
1257
  "step": 8400
1258
  },
1259
  {
1260
+ "epoch": 0.9110512129380054,
1261
+ "grad_norm": 0.6276729702949524,
1262
+ "learning_rate": 0.0005459320021586615,
1263
+ "loss": 3.8911,
1264
  "step": 8450
1265
  },
1266
  {
1267
+ "epoch": 0.9164420485175202,
1268
+ "grad_norm": 0.5620297193527222,
1269
+ "learning_rate": 0.0005456082029141932,
1270
+ "loss": 3.8825,
1271
  "step": 8500
1272
  },
1273
  {
1274
+ "epoch": 0.921832884097035,
1275
+ "grad_norm": 0.5803284049034119,
1276
+ "learning_rate": 0.0005452844036697248,
1277
+ "loss": 3.8637,
1278
  "step": 8550
1279
  },
1280
  {
1281
+ "epoch": 0.9272237196765498,
1282
+ "grad_norm": 0.5740498900413513,
1283
+ "learning_rate": 0.0005449606044252563,
1284
+ "loss": 3.8559,
1285
  "step": 8600
1286
  },
1287
  {
1288
+ "epoch": 0.9326145552560647,
1289
+ "grad_norm": 0.5694971680641174,
1290
+ "learning_rate": 0.0005446368051807879,
1291
+ "loss": 3.8485,
1292
  "step": 8650
1293
  },
1294
  {
1295
+ "epoch": 0.9380053908355795,
1296
+ "grad_norm": 0.555069625377655,
1297
+ "learning_rate": 0.0005443130059363194,
1298
+ "loss": 3.857,
1299
  "step": 8700
1300
  },
1301
  {
1302
+ "epoch": 0.9433962264150944,
1303
+ "grad_norm": 0.5356409549713135,
1304
+ "learning_rate": 0.0005439892066918511,
1305
+ "loss": 3.8539,
1306
  "step": 8750
1307
  },
1308
  {
1309
+ "epoch": 0.9487870619946092,
1310
+ "grad_norm": 0.5425722002983093,
1311
+ "learning_rate": 0.0005436654074473825,
1312
+ "loss": 3.8465,
1313
  "step": 8800
1314
  },
1315
  {
1316
+ "epoch": 0.954177897574124,
1317
+ "grad_norm": 0.5436581373214722,
1318
+ "learning_rate": 0.0005433416082029142,
1319
+ "loss": 3.8369,
1320
  "step": 8850
1321
  },
1322
  {
1323
+ "epoch": 0.9595687331536388,
1324
+ "grad_norm": 0.5613462328910828,
1325
+ "learning_rate": 0.0005430178089584457,
1326
+ "loss": 3.8548,
1327
  "step": 8900
1328
  },
1329
  {
1330
+ "epoch": 0.9649595687331537,
1331
+ "grad_norm": 0.5960737466812134,
1332
+ "learning_rate": 0.0005426940097139773,
1333
+ "loss": 3.8608,
1334
  "step": 8950
1335
  },
1336
  {
1337
+ "epoch": 0.9703504043126685,
1338
+ "grad_norm": 0.5757009387016296,
1339
+ "learning_rate": 0.0005423702104695088,
1340
+ "loss": 3.8477,
1341
  "step": 9000
1342
  },
1343
  {
1344
+ "epoch": 0.9703504043126685,
1345
+ "eval_accuracy": 0.340650489627941,
1346
+ "eval_loss": 3.7848594188690186,
1347
+ "eval_runtime": 181.5098,
1348
+ "eval_samples_per_second": 99.229,
1349
+ "eval_steps_per_second": 6.204,
1350
  "step": 9000
1351
  },
1352
  {
1353
+ "epoch": 0.9757412398921833,
1354
+ "grad_norm": 0.5870667695999146,
1355
+ "learning_rate": 0.0005420464112250404,
1356
+ "loss": 3.8409,
1357
  "step": 9050
1358
  },
1359
  {
1360
+ "epoch": 0.9811320754716981,
1361
+ "grad_norm": 0.6312296986579895,
1362
+ "learning_rate": 0.000541722611980572,
1363
+ "loss": 3.8558,
1364
  "step": 9100
1365
  },
1366
  {
1367
+ "epoch": 0.9865229110512129,
1368
+ "grad_norm": 0.5529806613922119,
1369
+ "learning_rate": 0.0005413988127361036,
1370
+ "loss": 3.8301,
1371
  "step": 9150
1372
  },
1373
  {
1374
+ "epoch": 0.9919137466307277,
1375
+ "grad_norm": 0.5621755123138428,
1376
+ "learning_rate": 0.0005410750134916351,
1377
+ "loss": 3.8279,
1378
  "step": 9200
1379
  },
1380
  {
1381
+ "epoch": 0.9973045822102425,
1382
+ "grad_norm": 0.564895510673523,
1383
+ "learning_rate": 0.0005407512142471667,
1384
+ "loss": 3.8412,
1385
  "step": 9250
1386
  },
1387
  {
1388
+ "epoch": 1.0026954177897573,
1389
+ "grad_norm": 0.5690886974334717,
1390
+ "learning_rate": 0.0005404274150026983,
1391
+ "loss": 3.8157,
1392
  "step": 9300
1393
  },
1394
  {
1395
+ "epoch": 1.0080862533692723,
1396
+ "grad_norm": 0.544973611831665,
1397
+ "learning_rate": 0.0005401100917431192,
1398
+ "loss": 3.7794,
1399
  "step": 9350
1400
  },
1401
  {
1402
+ "epoch": 1.013477088948787,
1403
+ "grad_norm": 0.61162269115448,
1404
+ "learning_rate": 0.0005397862924986508,
1405
+ "loss": 3.774,
1406
  "step": 9400
1407
  },
1408
  {
1409
+ "epoch": 1.0188679245283019,
1410
+ "grad_norm": 0.6406456828117371,
1411
+ "learning_rate": 0.0005394624932541824,
1412
+ "loss": 3.773,
1413
  "step": 9450
1414
  },
1415
  {
1416
+ "epoch": 1.0242587601078168,
1417
+ "grad_norm": 0.6477643847465515,
1418
+ "learning_rate": 0.0005391386940097139,
1419
+ "loss": 3.7741,
1420
  "step": 9500
1421
  },
1422
  {
1423
+ "epoch": 1.0296495956873315,
1424
+ "grad_norm": 0.5913816094398499,
1425
+ "learning_rate": 0.0005388148947652455,
1426
+ "loss": 3.7717,
1427
  "step": 9550
1428
  },
1429
  {
1430
+ "epoch": 1.0350404312668464,
1431
+ "grad_norm": 0.6406410932540894,
1432
+ "learning_rate": 0.000538491095520777,
1433
+ "loss": 3.76,
1434
  "step": 9600
1435
  },
1436
  {
1437
+ "epoch": 1.0404312668463611,
1438
+ "grad_norm": 0.6126238107681274,
1439
+ "learning_rate": 0.0005381672962763086,
1440
+ "loss": 3.801,
1441
  "step": 9650
1442
  },
1443
  {
1444
+ "epoch": 1.045822102425876,
1445
+ "grad_norm": 0.5817605257034302,
1446
+ "learning_rate": 0.0005378434970318403,
1447
+ "loss": 3.7861,
1448
  "step": 9700
1449
  },
1450
  {
1451
+ "epoch": 1.0512129380053907,
1452
+ "grad_norm": 0.5051379203796387,
1453
+ "learning_rate": 0.0005375196977873718,
1454
+ "loss": 3.7861,
1455
  "step": 9750
1456
  },
1457
  {
1458
+ "epoch": 1.0566037735849056,
1459
+ "grad_norm": 0.5685185194015503,
1460
+ "learning_rate": 0.0005371958985429034,
1461
+ "loss": 3.7939,
1462
  "step": 9800
1463
  },
1464
  {
1465
+ "epoch": 1.0619946091644206,
1466
+ "grad_norm": 0.561697244644165,
1467
+ "learning_rate": 0.0005368720992984349,
1468
+ "loss": 3.788,
1469
  "step": 9850
1470
  },
1471
  {
1472
+ "epoch": 1.0673854447439353,
1473
+ "grad_norm": 0.5285218954086304,
1474
+ "learning_rate": 0.0005365483000539665,
1475
+ "loss": 3.7792,
1476
  "step": 9900
1477
  },
1478
  {
1479
+ "epoch": 1.0727762803234502,
1480
+ "grad_norm": 0.6429543495178223,
1481
+ "learning_rate": 0.0005362245008094981,
1482
+ "loss": 3.7742,
1483
  "step": 9950
1484
  },
1485
  {
1486
+ "epoch": 1.0781671159029649,
1487
+ "grad_norm": 0.6265733242034912,
1488
+ "learning_rate": 0.0005359007015650297,
1489
+ "loss": 3.7666,
1490
  "step": 10000
1491
  },
1492
  {
1493
+ "epoch": 1.0781671159029649,
1494
+ "eval_accuracy": 0.34411564752612045,
1495
+ "eval_loss": 3.753229856491089,
1496
+ "eval_runtime": 181.6252,
1497
+ "eval_samples_per_second": 99.166,
1498
+ "eval_steps_per_second": 6.2,
1499
  "step": 10000
1500
  }
1501
  ],
1502
  "logging_steps": 50,
1503
+ "max_steps": 92750,
1504
  "num_input_tokens_seen": 0,
1505
  "num_train_epochs": 10,
1506
  "save_steps": 10000,
 
1516
  "attributes": {}
1517
  }
1518
  },
1519
+ "total_flos": 8.36069179392e+16,
1520
  "train_batch_size": 32,
1521
  "trial_name": null,
1522
  "trial_params": null
checkpoint-10000/training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:afce0a72ad125f75905f2f0d22e2b9b901626403e511a920d55b73209c30e029
3
  size 5304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b9e13e50542c0805c5428fafff83c01e0b5b324f63cb6bdf0fab92b8f96e924c
3
  size 5304