craa commited on
Commit
5521f5a
·
verified ·
1 Parent(s): 6f471e4

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:e2649d43c7fc7792c518322fba4f4b6fc0d05b878feabff86a5bbac8189c7014
3
  size 503128704
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef5a6a18ad797c2307d1789e9cf2286d9c6eb15f254cd235cf3eb566c9e52553
3
  size 503128704
checkpoint-10000/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6220bf647422ca97741ee52e9023675439a3233ecc1737dfcc428d081dbfc038
3
  size 1006351290
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec09cbe899d5c53f98e5cc027566161df8e10359c528fb944a69fcbc46148c18
3
  size 1006351290
checkpoint-10000/rng_state.pth CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7ea6aba18815b6bff32277260718425e74ea8e5c5878dbd8a44da097ca7838de
3
  size 14244
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:605a1de232910c1d5eec54037787d6ff72227dda3499910f6c21b52ad36787f1
3
  size 14244
checkpoint-10000/scheduler.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:90c311f0069114271a5dba5d72dcffc82e7a33081dd5f150aa8c9b2278c1c497
3
  size 1064
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:11371a619cce2ab697b4adbd009e5c127db38cd6dc9bce6b5c33de02dedd6b1d
3
  size 1064
checkpoint-10000/trainer_state.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "best_metric": 3.7489354610443115,
3
- "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_8397/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": 1.3847299814224243,
14
- "learning_rate": 0.0003,
15
- "loss": 8.4655,
16
  "step": 50
17
  },
18
  {
19
- "epoch": 0.010763104079216447,
20
- "grad_norm": 1.0632425546646118,
21
- "learning_rate": 0.0006,
22
- "loss": 6.8249,
23
  "step": 100
24
  },
25
  {
26
- "epoch": 0.01614465611882467,
27
- "grad_norm": 2.088347911834717,
28
- "learning_rate": 0.0005996767589699385,
29
- "loss": 6.4221,
30
  "step": 150
31
  },
32
  {
33
- "epoch": 0.021526208158432893,
34
- "grad_norm": 1.0710954666137695,
35
- "learning_rate": 0.0005993535179398771,
36
- "loss": 6.1909,
37
  "step": 200
38
  },
39
  {
40
- "epoch": 0.026907760198041114,
41
- "grad_norm": 1.5236896276474,
42
- "learning_rate": 0.0005990302769098158,
43
- "loss": 6.0442,
44
  "step": 250
45
  },
46
  {
47
- "epoch": 0.03228931223764934,
48
- "grad_norm": 1.8247500658035278,
49
- "learning_rate": 0.0005987070358797543,
50
- "loss": 5.9513,
51
  "step": 300
52
  },
53
  {
54
- "epoch": 0.03767086427725756,
55
- "grad_norm": 1.2985302209854126,
56
- "learning_rate": 0.0005983837948496929,
57
- "loss": 5.8643,
58
  "step": 350
59
  },
60
  {
61
- "epoch": 0.04305241631686579,
62
- "grad_norm": 1.256401538848877,
63
- "learning_rate": 0.0005980605538196314,
64
- "loss": 5.8018,
65
  "step": 400
66
  },
67
  {
68
- "epoch": 0.048433968356474004,
69
- "grad_norm": 0.8569295406341553,
70
- "learning_rate": 0.0005977373127895701,
71
- "loss": 5.7053,
72
  "step": 450
73
  },
74
  {
75
- "epoch": 0.05381552039608223,
76
- "grad_norm": 0.860461413860321,
77
- "learning_rate": 0.0005974140717595086,
78
- "loss": 5.6413,
79
  "step": 500
80
  },
81
  {
82
- "epoch": 0.05919707243569045,
83
- "grad_norm": 1.6800360679626465,
84
- "learning_rate": 0.0005970908307294472,
85
- "loss": 5.5789,
86
  "step": 550
87
  },
88
  {
89
- "epoch": 0.06457862447529868,
90
- "grad_norm": 1.408766508102417,
91
- "learning_rate": 0.0005967675896993858,
92
- "loss": 5.4957,
93
  "step": 600
94
  },
95
  {
96
- "epoch": 0.0699601765149069,
97
- "grad_norm": 1.263355016708374,
98
- "learning_rate": 0.0005964443486693243,
99
- "loss": 5.421,
100
  "step": 650
101
  },
102
  {
103
- "epoch": 0.07534172855451512,
104
- "grad_norm": 1.617833137512207,
105
- "learning_rate": 0.000596121107639263,
106
- "loss": 5.334,
107
  "step": 700
108
  },
109
  {
110
- "epoch": 0.08072328059412334,
111
- "grad_norm": 1.2163816690444946,
112
- "learning_rate": 0.0005957978666092015,
113
- "loss": 5.2925,
114
  "step": 750
115
  },
116
  {
117
- "epoch": 0.08610483263373157,
118
- "grad_norm": 1.4054591655731201,
119
- "learning_rate": 0.0005954746255791401,
120
- "loss": 5.222,
121
  "step": 800
122
  },
123
  {
124
- "epoch": 0.09148638467333979,
125
- "grad_norm": 1.0674018859863281,
126
- "learning_rate": 0.0005951513845490787,
127
- "loss": 5.2071,
128
  "step": 850
129
  },
130
  {
131
- "epoch": 0.09686793671294801,
132
- "grad_norm": 1.4989951848983765,
133
- "learning_rate": 0.0005948281435190174,
134
- "loss": 5.1526,
135
  "step": 900
136
  },
137
  {
138
- "epoch": 0.10224948875255624,
139
- "grad_norm": 1.028799057006836,
140
- "learning_rate": 0.0005945049024889559,
141
- "loss": 5.1098,
142
  "step": 950
143
  },
144
  {
145
- "epoch": 0.10763104079216446,
146
- "grad_norm": 0.955634355545044,
147
- "learning_rate": 0.0005941816614588944,
148
- "loss": 5.0633,
149
  "step": 1000
150
  },
151
  {
152
- "epoch": 0.10763104079216446,
153
- "eval_accuracy": 0.22915491630956988,
154
- "eval_loss": 4.996466636657715,
155
- "eval_runtime": 202.5064,
156
- "eval_samples_per_second": 88.94,
157
- "eval_steps_per_second": 5.56,
158
  "step": 1000
159
  },
160
  {
161
- "epoch": 0.11301259283177269,
162
- "grad_norm": 1.2216582298278809,
163
- "learning_rate": 0.000593858420428833,
164
- "loss": 5.0219,
165
  "step": 1050
166
  },
167
  {
168
- "epoch": 0.1183941448713809,
169
- "grad_norm": 1.2919893264770508,
170
- "learning_rate": 0.0005935351793987716,
171
- "loss": 4.9962,
172
  "step": 1100
173
  },
174
  {
175
- "epoch": 0.12377569691098914,
176
- "grad_norm": 1.842529058456421,
177
- "learning_rate": 0.0005932119383687103,
178
- "loss": 4.9736,
179
  "step": 1150
180
  },
181
  {
182
- "epoch": 0.12915724895059735,
183
- "grad_norm": 0.9416630268096924,
184
- "learning_rate": 0.0005928886973386488,
185
- "loss": 4.9687,
186
  "step": 1200
187
  },
188
  {
189
- "epoch": 0.13453880099020557,
190
- "grad_norm": 1.340468406677246,
191
- "learning_rate": 0.0005925654563085874,
192
- "loss": 4.9251,
193
  "step": 1250
194
  },
195
  {
196
- "epoch": 0.1399203530298138,
197
- "grad_norm": 1.4057867527008057,
198
- "learning_rate": 0.000592242215278526,
199
- "loss": 4.8782,
200
  "step": 1300
201
  },
202
  {
203
- "epoch": 0.14530190506942203,
204
- "grad_norm": 1.3462789058685303,
205
- "learning_rate": 0.0005919189742484645,
206
- "loss": 4.8431,
207
  "step": 1350
208
  },
209
  {
210
- "epoch": 0.15068345710903025,
211
- "grad_norm": 0.9291279911994934,
212
- "learning_rate": 0.0005915957332184032,
213
- "loss": 4.8316,
214
  "step": 1400
215
  },
216
  {
217
- "epoch": 0.15606500914863847,
218
- "grad_norm": 0.7789126634597778,
219
- "learning_rate": 0.0005912724921883417,
220
- "loss": 4.8028,
221
  "step": 1450
222
  },
223
  {
224
- "epoch": 0.16144656118824668,
225
- "grad_norm": 1.2111430168151855,
226
- "learning_rate": 0.0005909492511582803,
227
- "loss": 4.7837,
228
  "step": 1500
229
  },
230
  {
231
- "epoch": 0.1668281132278549,
232
- "grad_norm": 1.1703121662139893,
233
- "learning_rate": 0.0005906260101282189,
234
- "loss": 4.7456,
235
  "step": 1550
236
  },
237
  {
238
- "epoch": 0.17220966526746315,
239
- "grad_norm": 0.9755436182022095,
240
- "learning_rate": 0.0005903027690981575,
241
- "loss": 4.7303,
242
  "step": 1600
243
  },
244
  {
245
- "epoch": 0.17759121730707136,
246
- "grad_norm": 0.9121830463409424,
247
- "learning_rate": 0.000589979528068096,
248
- "loss": 4.7437,
249
  "step": 1650
250
  },
251
  {
252
- "epoch": 0.18297276934667958,
253
- "grad_norm": 1.2337779998779297,
254
- "learning_rate": 0.0005896562870380347,
255
- "loss": 4.6991,
256
  "step": 1700
257
  },
258
  {
259
- "epoch": 0.1883543213862878,
260
- "grad_norm": 0.7923302054405212,
261
- "learning_rate": 0.0005893330460079732,
262
- "loss": 4.6819,
263
  "step": 1750
264
  },
265
  {
266
- "epoch": 0.19373587342589602,
267
- "grad_norm": 0.9414947032928467,
268
- "learning_rate": 0.0005890098049779118,
269
- "loss": 4.6401,
270
  "step": 1800
271
  },
272
  {
273
- "epoch": 0.19911742546550426,
274
- "grad_norm": 0.8575401306152344,
275
- "learning_rate": 0.0005886865639478504,
276
- "loss": 4.6503,
277
  "step": 1850
278
  },
279
  {
280
- "epoch": 0.20449897750511248,
281
- "grad_norm": 0.874113917350769,
282
- "learning_rate": 0.0005883633229177889,
283
- "loss": 4.6262,
284
  "step": 1900
285
  },
286
  {
287
- "epoch": 0.2098805295447207,
288
- "grad_norm": 0.7579442262649536,
289
- "learning_rate": 0.0005880400818877276,
290
- "loss": 4.5946,
291
  "step": 1950
292
  },
293
  {
294
- "epoch": 0.2152620815843289,
295
- "grad_norm": 0.8062272071838379,
296
- "learning_rate": 0.0005877168408576662,
297
- "loss": 4.579,
298
  "step": 2000
299
  },
300
  {
301
- "epoch": 0.2152620815843289,
302
- "eval_accuracy": 0.2712283567521957,
303
- "eval_loss": 4.50425910949707,
304
- "eval_runtime": 211.1194,
305
- "eval_samples_per_second": 85.312,
306
- "eval_steps_per_second": 5.333,
307
  "step": 2000
308
  },
309
  {
310
- "epoch": 0.22064363362393713,
311
- "grad_norm": 1.1572167873382568,
312
- "learning_rate": 0.0005873935998276048,
313
- "loss": 4.5564,
314
  "step": 2050
315
  },
316
  {
317
- "epoch": 0.22602518566354537,
318
- "grad_norm": 1.1494579315185547,
319
- "learning_rate": 0.0005870703587975433,
320
- "loss": 4.5425,
321
  "step": 2100
322
  },
323
  {
324
- "epoch": 0.2314067377031536,
325
- "grad_norm": 0.9105241298675537,
326
- "learning_rate": 0.0005867471177674818,
327
- "loss": 4.5083,
328
  "step": 2150
329
  },
330
  {
331
- "epoch": 0.2367882897427618,
332
- "grad_norm": 0.9526411890983582,
333
- "learning_rate": 0.0005864238767374205,
334
- "loss": 4.5128,
335
  "step": 2200
336
  },
337
  {
338
- "epoch": 0.24216984178237003,
339
- "grad_norm": 0.7726457715034485,
340
- "learning_rate": 0.0005861006357073591,
341
- "loss": 4.4899,
342
  "step": 2250
343
  },
344
  {
345
- "epoch": 0.24755139382197827,
346
- "grad_norm": 1.0442702770233154,
347
- "learning_rate": 0.0005857773946772977,
348
- "loss": 4.4716,
349
  "step": 2300
350
  },
351
  {
352
- "epoch": 0.2529329458615865,
353
- "grad_norm": 1.1448233127593994,
354
- "learning_rate": 0.0005854541536472362,
355
- "loss": 4.4607,
356
  "step": 2350
357
  },
358
  {
359
- "epoch": 0.2583144979011947,
360
- "grad_norm": 0.9515467286109924,
361
- "learning_rate": 0.0005851309126171749,
362
- "loss": 4.4544,
363
  "step": 2400
364
  },
365
  {
366
- "epoch": 0.2636960499408029,
367
- "grad_norm": 0.7929104566574097,
368
- "learning_rate": 0.0005848076715871134,
369
- "loss": 4.4392,
370
  "step": 2450
371
  },
372
  {
373
- "epoch": 0.26907760198041114,
374
- "grad_norm": 1.1213116645812988,
375
- "learning_rate": 0.000584484430557052,
376
- "loss": 4.4407,
377
  "step": 2500
378
  },
379
  {
380
- "epoch": 0.27445915402001936,
381
- "grad_norm": 1.006108045578003,
382
- "learning_rate": 0.0005841611895269906,
383
- "loss": 4.3997,
384
  "step": 2550
385
  },
386
  {
387
- "epoch": 0.2798407060596276,
388
- "grad_norm": 0.7756773829460144,
389
- "learning_rate": 0.0005838379484969291,
390
- "loss": 4.3982,
391
  "step": 2600
392
  },
393
  {
394
- "epoch": 0.2852222580992358,
395
- "grad_norm": 0.7442255020141602,
396
- "learning_rate": 0.0005835147074668678,
397
- "loss": 4.3751,
398
  "step": 2650
399
  },
400
  {
401
- "epoch": 0.29060381013884407,
402
- "grad_norm": 0.7645350694656372,
403
- "learning_rate": 0.0005831914664368063,
404
- "loss": 4.3873,
405
  "step": 2700
406
  },
407
  {
408
- "epoch": 0.2959853621784523,
409
- "grad_norm": 0.7028666734695435,
410
- "learning_rate": 0.0005828682254067449,
411
- "loss": 4.3659,
412
  "step": 2750
413
  },
414
  {
415
- "epoch": 0.3013669142180605,
416
- "grad_norm": 0.8504071831703186,
417
- "learning_rate": 0.0005825449843766835,
418
- "loss": 4.3617,
419
  "step": 2800
420
  },
421
  {
422
- "epoch": 0.3067484662576687,
423
- "grad_norm": 0.880102276802063,
424
- "learning_rate": 0.0005822217433466221,
425
- "loss": 4.3115,
426
  "step": 2850
427
  },
428
  {
429
- "epoch": 0.31213001829727693,
430
- "grad_norm": 0.7290977835655212,
431
- "learning_rate": 0.0005818985023165607,
432
- "loss": 4.3215,
433
  "step": 2900
434
  },
435
  {
436
- "epoch": 0.31751157033688515,
437
- "grad_norm": 0.8019910454750061,
438
- "learning_rate": 0.0005815752612864992,
439
- "loss": 4.3149,
440
  "step": 2950
441
  },
442
  {
443
- "epoch": 0.32289312237649337,
444
- "grad_norm": 0.7739085555076599,
445
- "learning_rate": 0.0005812520202564378,
446
- "loss": 4.322,
447
  "step": 3000
448
  },
449
  {
450
- "epoch": 0.32289312237649337,
451
- "eval_accuracy": 0.2992396144952079,
452
- "eval_loss": 4.226711750030518,
453
- "eval_runtime": 199.2176,
454
- "eval_samples_per_second": 90.409,
455
- "eval_steps_per_second": 5.652,
456
  "step": 3000
457
  },
458
  {
459
- "epoch": 0.3282746744161016,
460
- "grad_norm": 1.0869840383529663,
461
- "learning_rate": 0.0005809287792263764,
462
- "loss": 4.3072,
463
  "step": 3050
464
  },
465
  {
466
- "epoch": 0.3336562264557098,
467
- "grad_norm": 0.7337630391120911,
468
- "learning_rate": 0.0005806055381963151,
469
- "loss": 4.2759,
470
  "step": 3100
471
  },
472
  {
473
- "epoch": 0.3390377784953181,
474
- "grad_norm": 0.7336916923522949,
475
- "learning_rate": 0.0005802822971662536,
476
- "loss": 4.2838,
477
  "step": 3150
478
  },
479
  {
480
- "epoch": 0.3444193305349263,
481
- "grad_norm": 0.7901497483253479,
482
- "learning_rate": 0.0005799590561361922,
483
- "loss": 4.2757,
484
  "step": 3200
485
  },
486
  {
487
- "epoch": 0.3498008825745345,
488
- "grad_norm": 0.9568474888801575,
489
- "learning_rate": 0.0005796358151061307,
490
- "loss": 4.2574,
491
  "step": 3250
492
  },
493
  {
494
- "epoch": 0.35518243461414273,
495
- "grad_norm": 0.9219122529029846,
496
- "learning_rate": 0.0005793125740760694,
497
- "loss": 4.2423,
498
  "step": 3300
499
  },
500
  {
501
- "epoch": 0.36056398665375095,
502
- "grad_norm": 0.6552030444145203,
503
- "learning_rate": 0.0005789893330460079,
504
- "loss": 4.2339,
505
  "step": 3350
506
  },
507
  {
508
- "epoch": 0.36594553869335916,
509
- "grad_norm": 0.6318998336791992,
510
- "learning_rate": 0.0005786660920159465,
511
- "loss": 4.2253,
512
  "step": 3400
513
  },
514
  {
515
- "epoch": 0.3713270907329674,
516
- "grad_norm": 0.7610264420509338,
517
- "learning_rate": 0.0005783428509858851,
518
- "loss": 4.2369,
519
  "step": 3450
520
  },
521
  {
522
- "epoch": 0.3767086427725756,
523
- "grad_norm": 0.6418978571891785,
524
- "learning_rate": 0.0005780196099558237,
525
- "loss": 4.2239,
526
  "step": 3500
527
  },
528
  {
529
- "epoch": 0.3820901948121838,
530
- "grad_norm": 0.7791884541511536,
531
- "learning_rate": 0.0005776963689257623,
532
- "loss": 4.2253,
533
  "step": 3550
534
  },
535
  {
536
- "epoch": 0.38747174685179203,
537
- "grad_norm": 0.7014104127883911,
538
- "learning_rate": 0.0005773731278957008,
539
- "loss": 4.1888,
540
  "step": 3600
541
  },
542
  {
543
- "epoch": 0.3928532988914003,
544
- "grad_norm": 0.7691745758056641,
545
- "learning_rate": 0.0005770498868656394,
546
- "loss": 4.2055,
547
  "step": 3650
548
  },
549
  {
550
- "epoch": 0.3982348509310085,
551
- "grad_norm": 0.7330165505409241,
552
- "learning_rate": 0.000576726645835578,
553
- "loss": 4.1917,
554
  "step": 3700
555
  },
556
  {
557
- "epoch": 0.40361640297061674,
558
- "grad_norm": 0.6409156918525696,
559
- "learning_rate": 0.0005764034048055167,
560
- "loss": 4.1739,
561
  "step": 3750
562
  },
563
  {
564
- "epoch": 0.40899795501022496,
565
- "grad_norm": 0.751600444316864,
566
- "learning_rate": 0.0005760801637754552,
567
- "loss": 4.1815,
568
  "step": 3800
569
  },
570
  {
571
- "epoch": 0.4143795070498332,
572
- "grad_norm": 0.6341384649276733,
573
- "learning_rate": 0.0005757569227453937,
574
- "loss": 4.1752,
575
  "step": 3850
576
  },
577
  {
578
- "epoch": 0.4197610590894414,
579
- "grad_norm": 0.8716601729393005,
580
- "learning_rate": 0.0005754336817153324,
581
- "loss": 4.1772,
582
  "step": 3900
583
  },
584
  {
585
- "epoch": 0.4251426111290496,
586
- "grad_norm": 0.7268335223197937,
587
- "learning_rate": 0.0005751104406852709,
588
- "loss": 4.1579,
589
  "step": 3950
590
  },
591
  {
592
- "epoch": 0.4305241631686578,
593
- "grad_norm": 0.7490660548210144,
594
- "learning_rate": 0.0005747871996552096,
595
- "loss": 4.1439,
596
  "step": 4000
597
  },
598
  {
599
- "epoch": 0.4305241631686578,
600
- "eval_accuracy": 0.31218462820231296,
601
- "eval_loss": 4.092499256134033,
602
- "eval_runtime": 199.4467,
603
- "eval_samples_per_second": 90.305,
604
- "eval_steps_per_second": 5.646,
605
  "step": 4000
606
  },
607
  {
608
- "epoch": 0.43590571520826604,
609
- "grad_norm": 0.9424974918365479,
610
- "learning_rate": 0.0005744639586251481,
611
- "loss": 4.1592,
612
  "step": 4050
613
  },
614
  {
615
- "epoch": 0.44128726724787426,
616
- "grad_norm": 0.771391749382019,
617
- "learning_rate": 0.0005741407175950867,
618
- "loss": 4.1455,
619
  "step": 4100
620
  },
621
  {
622
- "epoch": 0.44666881928748253,
623
- "grad_norm": 0.7916195392608643,
624
- "learning_rate": 0.0005738174765650253,
625
- "loss": 4.1329,
626
  "step": 4150
627
  },
628
  {
629
- "epoch": 0.45205037132709075,
630
- "grad_norm": 0.7015873789787292,
631
- "learning_rate": 0.0005734942355349638,
632
- "loss": 4.1338,
633
  "step": 4200
634
  },
635
  {
636
- "epoch": 0.45743192336669897,
637
- "grad_norm": 0.62185138463974,
638
- "learning_rate": 0.0005731709945049025,
639
- "loss": 4.1517,
640
  "step": 4250
641
  },
642
  {
643
- "epoch": 0.4628134754063072,
644
- "grad_norm": 0.6499165296554565,
645
- "learning_rate": 0.000572847753474841,
646
- "loss": 4.1249,
647
  "step": 4300
648
  },
649
  {
650
- "epoch": 0.4681950274459154,
651
- "grad_norm": 0.8142028450965881,
652
- "learning_rate": 0.0005725245124447796,
653
- "loss": 4.1094,
654
  "step": 4350
655
  },
656
  {
657
- "epoch": 0.4735765794855236,
658
- "grad_norm": 0.686265230178833,
659
- "learning_rate": 0.0005722012714147182,
660
- "loss": 4.1156,
661
  "step": 4400
662
  },
663
  {
664
- "epoch": 0.47895813152513184,
665
- "grad_norm": 0.6861261129379272,
666
- "learning_rate": 0.0005718780303846568,
667
- "loss": 4.0911,
668
  "step": 4450
669
  },
670
  {
671
- "epoch": 0.48433968356474005,
672
- "grad_norm": 0.717899739742279,
673
- "learning_rate": 0.0005715547893545953,
674
- "loss": 4.0984,
675
  "step": 4500
676
  },
677
  {
678
- "epoch": 0.48972123560434827,
679
- "grad_norm": 0.635183572769165,
680
- "learning_rate": 0.000571231548324534,
681
- "loss": 4.0915,
682
  "step": 4550
683
  },
684
  {
685
- "epoch": 0.49510278764395654,
686
- "grad_norm": 0.6602552533149719,
687
- "learning_rate": 0.0005709083072944725,
688
- "loss": 4.0886,
689
  "step": 4600
690
  },
691
  {
692
- "epoch": 0.5004843396835648,
693
- "grad_norm": 0.7243141531944275,
694
- "learning_rate": 0.0005705850662644111,
695
- "loss": 4.0928,
696
  "step": 4650
697
  },
698
  {
699
- "epoch": 0.505865891723173,
700
- "grad_norm": 0.6310231685638428,
701
- "learning_rate": 0.0005702618252343497,
702
- "loss": 4.0986,
703
  "step": 4700
704
  },
705
  {
706
- "epoch": 0.5112474437627812,
707
- "grad_norm": 0.5868064761161804,
708
- "learning_rate": 0.0005699385842042882,
709
- "loss": 4.0687,
710
  "step": 4750
711
  },
712
  {
713
- "epoch": 0.5166289958023894,
714
- "grad_norm": 0.6389384269714355,
715
- "learning_rate": 0.0005696153431742269,
716
- "loss": 4.0789,
717
  "step": 4800
718
  },
719
  {
720
- "epoch": 0.5220105478419976,
721
- "grad_norm": 0.7776073217391968,
722
- "learning_rate": 0.0005692921021441655,
723
- "loss": 4.0977,
724
  "step": 4850
725
  },
726
  {
727
- "epoch": 0.5273920998816058,
728
- "grad_norm": 0.6234896779060364,
729
- "learning_rate": 0.0005689688611141041,
730
- "loss": 4.0547,
731
  "step": 4900
732
  },
733
  {
734
- "epoch": 0.5327736519212141,
735
- "grad_norm": 0.8750380277633667,
736
- "learning_rate": 0.0005686456200840426,
737
- "loss": 4.0718,
738
  "step": 4950
739
  },
740
  {
741
- "epoch": 0.5381552039608223,
742
- "grad_norm": 0.7700529098510742,
743
- "learning_rate": 0.0005683223790539811,
744
- "loss": 4.043,
745
  "step": 5000
746
  },
747
  {
748
- "epoch": 0.5381552039608223,
749
- "eval_accuracy": 0.3214325100957547,
750
- "eval_loss": 3.9926369190216064,
751
- "eval_runtime": 225.5059,
752
- "eval_samples_per_second": 79.869,
753
- "eval_steps_per_second": 4.993,
754
  "step": 5000
755
  },
756
  {
757
- "epoch": 0.5435367560004305,
758
- "grad_norm": 0.7643857002258301,
759
- "learning_rate": 0.0005679991380239198,
760
- "loss": 4.0605,
761
  "step": 5050
762
  },
763
  {
764
- "epoch": 0.5489183080400387,
765
- "grad_norm": 0.554442286491394,
766
- "learning_rate": 0.0005676758969938584,
767
- "loss": 4.0446,
768
  "step": 5100
769
  },
770
  {
771
- "epoch": 0.5542998600796469,
772
- "grad_norm": 0.7216308116912842,
773
- "learning_rate": 0.000567352655963797,
774
- "loss": 4.0313,
775
  "step": 5150
776
  },
777
  {
778
- "epoch": 0.5596814121192552,
779
- "grad_norm": 0.6057455539703369,
780
- "learning_rate": 0.0005670294149337355,
781
- "loss": 4.0337,
782
  "step": 5200
783
  },
784
  {
785
- "epoch": 0.5650629641588634,
786
- "grad_norm": 0.535860002040863,
787
- "learning_rate": 0.0005667061739036742,
788
- "loss": 4.0392,
789
  "step": 5250
790
  },
791
  {
792
- "epoch": 0.5704445161984716,
793
- "grad_norm": 0.647204577922821,
794
- "learning_rate": 0.0005663829328736127,
795
- "loss": 4.0421,
796
  "step": 5300
797
  },
798
  {
799
- "epoch": 0.5758260682380799,
800
- "grad_norm": 0.5126392841339111,
801
- "learning_rate": 0.0005660596918435512,
802
- "loss": 4.0318,
803
  "step": 5350
804
  },
805
  {
806
- "epoch": 0.5812076202776881,
807
- "grad_norm": 0.6040687561035156,
808
- "learning_rate": 0.0005657364508134899,
809
- "loss": 4.0117,
810
  "step": 5400
811
  },
812
  {
813
- "epoch": 0.5865891723172963,
814
- "grad_norm": 0.6971921324729919,
815
- "learning_rate": 0.0005654132097834284,
816
- "loss": 4.025,
817
  "step": 5450
818
  },
819
  {
820
- "epoch": 0.5919707243569046,
821
- "grad_norm": 0.5343753099441528,
822
- "learning_rate": 0.0005650899687533671,
823
- "loss": 4.015,
824
  "step": 5500
825
  },
826
  {
827
- "epoch": 0.5973522763965128,
828
- "grad_norm": 0.5600801706314087,
829
- "learning_rate": 0.0005647667277233056,
830
- "loss": 4.0087,
831
  "step": 5550
832
  },
833
  {
834
- "epoch": 0.602733828436121,
835
- "grad_norm": 0.6542683243751526,
836
- "learning_rate": 0.0005644434866932442,
837
- "loss": 3.9957,
838
  "step": 5600
839
  },
840
  {
841
- "epoch": 0.6081153804757292,
842
- "grad_norm": 0.6249606609344482,
843
- "learning_rate": 0.0005641202456631828,
844
- "loss": 4.0136,
845
  "step": 5650
846
  },
847
  {
848
- "epoch": 0.6134969325153374,
849
- "grad_norm": 0.6572969555854797,
850
- "learning_rate": 0.0005637970046331214,
851
- "loss": 4.0046,
852
  "step": 5700
853
  },
854
  {
855
- "epoch": 0.6188784845549457,
856
- "grad_norm": 0.6926669478416443,
857
- "learning_rate": 0.00056347376360306,
858
- "loss": 3.9869,
859
  "step": 5750
860
  },
861
  {
862
- "epoch": 0.6242600365945539,
863
- "grad_norm": 0.6491366624832153,
864
- "learning_rate": 0.0005631505225729985,
865
- "loss": 4.0004,
866
  "step": 5800
867
  },
868
  {
869
- "epoch": 0.6296415886341621,
870
- "grad_norm": 0.6138956546783447,
871
- "learning_rate": 0.0005628272815429371,
872
- "loss": 3.9814,
873
  "step": 5850
874
  },
875
  {
876
- "epoch": 0.6350231406737703,
877
- "grad_norm": 0.6467788219451904,
878
- "learning_rate": 0.0005625040405128757,
879
- "loss": 3.9771,
880
  "step": 5900
881
  },
882
  {
883
- "epoch": 0.6404046927133785,
884
- "grad_norm": 0.6469590067863464,
885
- "learning_rate": 0.0005621807994828143,
886
- "loss": 3.9882,
887
  "step": 5950
888
  },
889
  {
890
- "epoch": 0.6457862447529867,
891
- "grad_norm": 0.7396846413612366,
892
- "learning_rate": 0.0005618575584527529,
893
- "loss": 3.9759,
894
  "step": 6000
895
  },
896
  {
897
- "epoch": 0.6457862447529867,
898
- "eval_accuracy": 0.32833664054615025,
899
- "eval_loss": 3.919649839401245,
900
- "eval_runtime": 200.1239,
901
- "eval_samples_per_second": 89.999,
902
- "eval_steps_per_second": 5.627,
903
  "step": 6000
904
  },
905
  {
906
- "epoch": 0.651167796792595,
907
- "grad_norm": 0.7368170619010925,
908
- "learning_rate": 0.0005615343174226915,
909
  "loss": 3.9789,
910
  "step": 6050
911
  },
912
  {
913
- "epoch": 0.6565493488322032,
914
- "grad_norm": 0.593928873538971,
915
- "learning_rate": 0.00056121107639263,
916
- "loss": 3.962,
917
  "step": 6100
918
  },
919
  {
920
- "epoch": 0.6619309008718114,
921
- "grad_norm": 0.5571395754814148,
922
- "learning_rate": 0.0005608878353625687,
923
- "loss": 3.9732,
924
  "step": 6150
925
  },
926
  {
927
- "epoch": 0.6673124529114196,
928
- "grad_norm": 0.6184373497962952,
929
- "learning_rate": 0.0005605645943325072,
930
- "loss": 3.9618,
931
  "step": 6200
932
  },
933
  {
934
- "epoch": 0.6726940049510278,
935
- "grad_norm": 0.6731172800064087,
936
- "learning_rate": 0.0005602413533024458,
937
- "loss": 3.9629,
938
  "step": 6250
939
  },
940
  {
941
- "epoch": 0.6780755569906362,
942
- "grad_norm": Infinity,
943
- "learning_rate": 0.0005599245770929855,
944
- "loss": 3.9661,
945
  "step": 6300
946
  },
947
  {
948
- "epoch": 0.6834571090302444,
949
- "grad_norm": 0.6674894690513611,
950
- "learning_rate": 0.0005596013360629242,
951
- "loss": 3.9598,
952
  "step": 6350
953
  },
954
  {
955
- "epoch": 0.6888386610698526,
956
- "grad_norm": 0.6270789504051208,
957
- "learning_rate": 0.0005592780950328628,
958
- "loss": 3.9443,
959
  "step": 6400
960
  },
961
  {
962
- "epoch": 0.6942202131094608,
963
- "grad_norm": 0.6163814067840576,
964
- "learning_rate": 0.0005589548540028014,
965
- "loss": 3.9506,
966
  "step": 6450
967
  },
968
  {
969
- "epoch": 0.699601765149069,
970
- "grad_norm": 0.6060782670974731,
971
- "learning_rate": 0.0005586316129727399,
972
- "loss": 3.9645,
973
  "step": 6500
974
  },
975
  {
976
- "epoch": 0.7049833171886772,
977
- "grad_norm": 0.6246522068977356,
978
- "learning_rate": 0.0005583083719426786,
979
- "loss": 3.9614,
980
  "step": 6550
981
  },
982
  {
983
- "epoch": 0.7103648692282855,
984
- "grad_norm": 0.603921115398407,
985
- "learning_rate": 0.0005579851309126171,
986
- "loss": 3.94,
987
  "step": 6600
988
  },
989
  {
990
- "epoch": 0.7157464212678937,
991
- "grad_norm": 0.5655505061149597,
992
- "learning_rate": 0.0005576618898825558,
993
- "loss": 3.9318,
994
  "step": 6650
995
  },
996
  {
997
- "epoch": 0.7211279733075019,
998
- "grad_norm": 0.604542076587677,
999
- "learning_rate": 0.0005573386488524943,
1000
- "loss": 3.9307,
1001
  "step": 6700
1002
  },
1003
  {
1004
- "epoch": 0.7265095253471101,
1005
- "grad_norm": 0.600004255771637,
1006
- "learning_rate": 0.0005570154078224328,
1007
- "loss": 3.9442,
1008
  "step": 6750
1009
  },
1010
  {
1011
- "epoch": 0.7318910773867183,
1012
- "grad_norm": 0.6960250735282898,
1013
- "learning_rate": 0.0005566921667923715,
1014
- "loss": 3.9425,
1015
  "step": 6800
1016
  },
1017
  {
1018
- "epoch": 0.7372726294263265,
1019
- "grad_norm": 0.5619109869003296,
1020
- "learning_rate": 0.00055636892576231,
1021
- "loss": 3.9073,
1022
  "step": 6850
1023
  },
1024
  {
1025
- "epoch": 0.7426541814659348,
1026
- "grad_norm": 0.6036113500595093,
1027
- "learning_rate": 0.0005560456847322487,
1028
- "loss": 3.9305,
1029
  "step": 6900
1030
  },
1031
  {
1032
- "epoch": 0.748035733505543,
1033
- "grad_norm": 0.6234127283096313,
1034
- "learning_rate": 0.0005557224437021872,
1035
- "loss": 3.9285,
1036
  "step": 6950
1037
  },
1038
  {
1039
- "epoch": 0.7534172855451512,
1040
- "grad_norm": 0.6131523251533508,
1041
- "learning_rate": 0.0005553992026721258,
1042
- "loss": 3.922,
1043
  "step": 7000
1044
  },
1045
  {
1046
- "epoch": 0.7534172855451512,
1047
- "eval_accuracy": 0.3334209437785411,
1048
- "eval_loss": 3.861107349395752,
1049
- "eval_runtime": 205.1306,
1050
- "eval_samples_per_second": 87.803,
1051
- "eval_steps_per_second": 5.489,
1052
  "step": 7000
1053
  },
1054
  {
1055
- "epoch": 0.7587988375847594,
1056
- "grad_norm": 0.6058487296104431,
1057
- "learning_rate": 0.0005550759616420644,
1058
- "loss": 3.9083,
1059
  "step": 7050
1060
  },
1061
  {
1062
- "epoch": 0.7641803896243676,
1063
- "grad_norm": 0.624620258808136,
1064
- "learning_rate": 0.000554752720612003,
1065
- "loss": 3.9349,
1066
  "step": 7100
1067
  },
1068
  {
1069
- "epoch": 0.7695619416639758,
1070
- "grad_norm": 0.5525732040405273,
1071
- "learning_rate": 0.0005544294795819415,
1072
- "loss": 3.9261,
1073
  "step": 7150
1074
  },
1075
  {
1076
- "epoch": 0.7749434937035841,
1077
- "grad_norm": 0.5950748324394226,
1078
- "learning_rate": 0.0005541062385518801,
1079
- "loss": 3.9067,
1080
  "step": 7200
1081
  },
1082
  {
1083
- "epoch": 0.7803250457431924,
1084
- "grad_norm": 0.5052813291549683,
1085
- "learning_rate": 0.0005537829975218188,
1086
- "loss": 3.9004,
1087
  "step": 7250
1088
  },
1089
  {
1090
- "epoch": 0.7857065977828006,
1091
- "grad_norm": 0.5610913038253784,
1092
- "learning_rate": 0.0005534662213123586,
1093
- "loss": 3.8911,
1094
  "step": 7300
1095
  },
1096
  {
1097
- "epoch": 0.7910881498224088,
1098
- "grad_norm": 0.5639758110046387,
1099
- "learning_rate": 0.0005531429802822971,
1100
- "loss": 3.931,
1101
  "step": 7350
1102
  },
1103
  {
1104
- "epoch": 0.796469701862017,
1105
- "grad_norm": 0.6221727728843689,
1106
- "learning_rate": 0.0005528197392522357,
1107
- "loss": 3.9288,
1108
  "step": 7400
1109
  },
1110
  {
1111
- "epoch": 0.8018512539016253,
1112
- "grad_norm": 0.5807225108146667,
1113
- "learning_rate": 0.0005524964982221743,
1114
- "loss": 3.9206,
1115
  "step": 7450
1116
  },
1117
  {
1118
- "epoch": 0.8072328059412335,
1119
- "grad_norm": 0.6214258074760437,
1120
- "learning_rate": 0.0005521732571921129,
1121
- "loss": 3.9216,
1122
  "step": 7500
1123
  },
1124
  {
1125
- "epoch": 0.8126143579808417,
1126
- "grad_norm": 0.6088549494743347,
1127
- "learning_rate": 0.0005518500161620514,
1128
- "loss": 3.9102,
1129
  "step": 7550
1130
  },
1131
  {
1132
- "epoch": 0.8179959100204499,
1133
- "grad_norm": 0.615315318107605,
1134
- "learning_rate": 0.00055152677513199,
1135
- "loss": 3.8932,
1136
  "step": 7600
1137
  },
1138
  {
1139
- "epoch": 0.8233774620600581,
1140
- "grad_norm": 0.5932325124740601,
1141
- "learning_rate": 0.0005512035341019286,
1142
- "loss": 3.9037,
1143
  "step": 7650
1144
  },
1145
  {
1146
- "epoch": 0.8287590140996663,
1147
- "grad_norm": 0.632127583026886,
1148
- "learning_rate": 0.0005508802930718672,
1149
- "loss": 3.9074,
1150
  "step": 7700
1151
  },
1152
  {
1153
- "epoch": 0.8341405661392746,
1154
- "grad_norm": 0.5906286239624023,
1155
- "learning_rate": 0.0005505570520418058,
1156
- "loss": 3.8969,
1157
  "step": 7750
1158
  },
1159
  {
1160
- "epoch": 0.8395221181788828,
1161
- "grad_norm": 0.7419958114624023,
1162
- "learning_rate": 0.0005502338110117443,
1163
- "loss": 3.911,
1164
  "step": 7800
1165
  },
1166
  {
1167
- "epoch": 0.844903670218491,
1168
- "grad_norm": 0.5959650874137878,
1169
- "learning_rate": 0.000549910569981683,
1170
- "loss": 3.885,
1171
  "step": 7850
1172
  },
1173
  {
1174
- "epoch": 0.8502852222580992,
1175
- "grad_norm": 0.5370450615882874,
1176
- "learning_rate": 0.0005495873289516215,
1177
- "loss": 3.8883,
1178
  "step": 7900
1179
  },
1180
  {
1181
- "epoch": 0.8556667742977074,
1182
- "grad_norm": 0.6142526865005493,
1183
- "learning_rate": 0.0005492640879215602,
1184
- "loss": 3.8891,
1185
  "step": 7950
1186
  },
1187
  {
1188
- "epoch": 0.8610483263373157,
1189
- "grad_norm": 0.5250054001808167,
1190
- "learning_rate": 0.0005489408468914987,
1191
- "loss": 3.8731,
1192
  "step": 8000
1193
  },
1194
  {
1195
- "epoch": 0.8610483263373157,
1196
- "eval_accuracy": 0.3380041398923315,
1197
- "eval_loss": 3.814573049545288,
1198
- "eval_runtime": 218.7029,
1199
- "eval_samples_per_second": 82.354,
1200
- "eval_steps_per_second": 5.149,
1201
  "step": 8000
1202
  },
1203
  {
1204
- "epoch": 0.8664298783769239,
1205
- "grad_norm": 0.5592741370201111,
1206
- "learning_rate": 0.0005486176058614372,
1207
- "loss": 3.8661,
1208
  "step": 8050
1209
  },
1210
  {
1211
- "epoch": 0.8718114304165321,
1212
- "grad_norm": 0.558358371257782,
1213
- "learning_rate": 0.0005482943648313759,
1214
- "loss": 3.8665,
1215
  "step": 8100
1216
  },
1217
  {
1218
- "epoch": 0.8771929824561403,
1219
- "grad_norm": 0.6065065860748291,
1220
- "learning_rate": 0.0005479711238013145,
1221
- "loss": 3.8815,
1222
  "step": 8150
1223
  },
1224
  {
1225
- "epoch": 0.8825745344957485,
1226
- "grad_norm": 0.566906750202179,
1227
- "learning_rate": 0.0005476478827712531,
1228
- "loss": 3.8788,
1229
  "step": 8200
1230
  },
1231
  {
1232
- "epoch": 0.8879560865353568,
1233
- "grad_norm": 0.5542231202125549,
1234
- "learning_rate": 0.0005473246417411916,
1235
- "loss": 3.8757,
1236
  "step": 8250
1237
  },
1238
  {
1239
- "epoch": 0.8933376385749651,
1240
- "grad_norm": 0.5650736093521118,
1241
- "learning_rate": 0.0005470014007111302,
1242
- "loss": 3.867,
1243
  "step": 8300
1244
  },
1245
  {
1246
- "epoch": 0.8987191906145733,
1247
- "grad_norm": 0.6359168887138367,
1248
- "learning_rate": 0.0005466781596810688,
1249
- "loss": 3.8704,
1250
  "step": 8350
1251
  },
1252
  {
1253
- "epoch": 0.9041007426541815,
1254
- "grad_norm": 0.5609973669052124,
1255
- "learning_rate": 0.0005463549186510073,
1256
- "loss": 3.8704,
1257
  "step": 8400
1258
  },
1259
  {
1260
- "epoch": 0.9094822946937897,
1261
- "grad_norm": 0.5966047048568726,
1262
- "learning_rate": 0.000546031677620946,
1263
- "loss": 3.8649,
1264
  "step": 8450
1265
  },
1266
  {
1267
- "epoch": 0.9148638467333979,
1268
- "grad_norm": 0.555415689945221,
1269
- "learning_rate": 0.0005457084365908845,
1270
- "loss": 3.8775,
1271
  "step": 8500
1272
  },
1273
  {
1274
- "epoch": 0.9202453987730062,
1275
- "grad_norm": 0.5719990134239197,
1276
- "learning_rate": 0.0005453851955608232,
1277
- "loss": 3.8601,
1278
  "step": 8550
1279
  },
1280
  {
1281
- "epoch": 0.9256269508126144,
1282
- "grad_norm": 0.5464230179786682,
1283
- "learning_rate": 0.0005450619545307617,
1284
- "loss": 3.8599,
1285
  "step": 8600
1286
  },
1287
  {
1288
- "epoch": 0.9310085028522226,
1289
- "grad_norm": 0.6069023013114929,
1290
- "learning_rate": 0.0005447387135007003,
1291
- "loss": 3.8584,
1292
  "step": 8650
1293
  },
1294
  {
1295
- "epoch": 0.9363900548918308,
1296
- "grad_norm": 0.556610643863678,
1297
- "learning_rate": 0.0005444154724706389,
1298
- "loss": 3.8636,
1299
  "step": 8700
1300
  },
1301
  {
1302
- "epoch": 0.941771606931439,
1303
- "grad_norm": 0.5988168120384216,
1304
- "learning_rate": 0.0005440922314405775,
1305
- "loss": 3.8366,
1306
  "step": 8750
1307
  },
1308
  {
1309
- "epoch": 0.9471531589710472,
1310
- "grad_norm": 0.5634347200393677,
1311
- "learning_rate": 0.0005437689904105161,
1312
- "loss": 3.8604,
1313
  "step": 8800
1314
  },
1315
  {
1316
- "epoch": 0.9525347110106555,
1317
- "grad_norm": 0.5349180102348328,
1318
- "learning_rate": 0.0005434457493804546,
1319
- "loss": 3.8395,
1320
  "step": 8850
1321
  },
1322
  {
1323
- "epoch": 0.9579162630502637,
1324
- "grad_norm": 0.5537461042404175,
1325
- "learning_rate": 0.0005431225083503932,
1326
- "loss": 3.8644,
1327
  "step": 8900
1328
  },
1329
  {
1330
- "epoch": 0.9632978150898719,
1331
- "grad_norm": 0.5711461901664734,
1332
- "learning_rate": 0.0005427992673203318,
1333
- "loss": 3.8352,
1334
  "step": 8950
1335
  },
1336
  {
1337
- "epoch": 0.9686793671294801,
1338
- "grad_norm": 0.6128414273262024,
1339
- "learning_rate": 0.0005424760262902704,
1340
- "loss": 3.8405,
1341
  "step": 9000
1342
  },
1343
  {
1344
- "epoch": 0.9686793671294801,
1345
- "eval_accuracy": 0.3417073562677125,
1346
- "eval_loss": 3.779106378555298,
1347
- "eval_runtime": 210.7209,
1348
- "eval_samples_per_second": 85.473,
1349
- "eval_steps_per_second": 5.344,
1350
  "step": 9000
1351
  },
1352
  {
1353
- "epoch": 0.9740609191690883,
1354
- "grad_norm": 0.6156434416770935,
1355
- "learning_rate": 0.000542152785260209,
1356
- "loss": 3.8333,
1357
  "step": 9050
1358
  },
1359
  {
1360
- "epoch": 0.9794424712086965,
1361
- "grad_norm": 0.5652556419372559,
1362
- "learning_rate": 0.0005418295442301476,
1363
- "loss": 3.8327,
1364
  "step": 9100
1365
  },
1366
  {
1367
- "epoch": 0.9848240232483048,
1368
- "grad_norm": 0.6443074941635132,
1369
- "learning_rate": 0.0005415063032000861,
1370
- "loss": 3.842,
1371
  "step": 9150
1372
  },
1373
  {
1374
- "epoch": 0.9902055752879131,
1375
- "grad_norm": 0.632297158241272,
1376
- "learning_rate": 0.0005411830621700248,
1377
- "loss": 3.8312,
1378
  "step": 9200
1379
  },
1380
  {
1381
- "epoch": 0.9955871273275213,
1382
- "grad_norm": 0.5863285064697266,
1383
- "learning_rate": 0.0005408598211399633,
1384
- "loss": 3.8298,
1385
  "step": 9250
1386
  },
1387
  {
1388
- "epoch": 1.0009686793671295,
1389
- "grad_norm": 0.5393086671829224,
1390
- "learning_rate": 0.0005405365801099019,
1391
- "loss": 3.8069,
1392
  "step": 9300
1393
  },
1394
  {
1395
- "epoch": 1.0063502314067376,
1396
- "grad_norm": 0.5285312533378601,
1397
- "learning_rate": 0.0005402133390798405,
1398
- "loss": 3.7619,
1399
  "step": 9350
1400
  },
1401
  {
1402
- "epoch": 1.011731783446346,
1403
- "grad_norm": 0.5921708941459656,
1404
- "learning_rate": 0.000539890098049779,
1405
- "loss": 3.7701,
1406
  "step": 9400
1407
  },
1408
  {
1409
- "epoch": 1.017113335485954,
1410
- "grad_norm": 0.5384525060653687,
1411
- "learning_rate": 0.0005395668570197177,
1412
- "loss": 3.7657,
1413
  "step": 9450
1414
  },
1415
  {
1416
- "epoch": 1.0224948875255624,
1417
- "grad_norm": 0.5624929666519165,
1418
- "learning_rate": 0.0005392436159896562,
1419
- "loss": 3.7754,
1420
  "step": 9500
1421
  },
1422
  {
1423
- "epoch": 1.0278764395651705,
1424
- "grad_norm": 0.5882208943367004,
1425
- "learning_rate": 0.0005389203749595948,
1426
- "loss": 3.7848,
1427
  "step": 9550
1428
  },
1429
  {
1430
- "epoch": 1.0332579916047788,
1431
- "grad_norm": 0.5588924884796143,
1432
- "learning_rate": 0.0005385971339295334,
1433
- "loss": 3.7603,
1434
  "step": 9600
1435
  },
1436
  {
1437
- "epoch": 1.0386395436443872,
1438
- "grad_norm": 0.5837873220443726,
1439
- "learning_rate": 0.000538273892899472,
1440
- "loss": 3.7785,
1441
  "step": 9650
1442
  },
1443
  {
1444
- "epoch": 1.0440210956839953,
1445
- "grad_norm": 0.5880899429321289,
1446
- "learning_rate": 0.0005379506518694106,
1447
- "loss": 3.7506,
1448
  "step": 9700
1449
  },
1450
  {
1451
- "epoch": 1.0494026477236036,
1452
- "grad_norm": 0.5579642653465271,
1453
- "learning_rate": 0.0005376274108393491,
1454
- "loss": 3.7714,
1455
  "step": 9750
1456
  },
1457
  {
1458
- "epoch": 1.0547841997632117,
1459
- "grad_norm": 0.576766312122345,
1460
- "learning_rate": 0.0005373041698092877,
1461
- "loss": 3.7653,
1462
  "step": 9800
1463
  },
1464
  {
1465
- "epoch": 1.06016575180282,
1466
- "grad_norm": 0.5471925735473633,
1467
- "learning_rate": 0.0005369809287792263,
1468
- "loss": 3.7686,
1469
  "step": 9850
1470
  },
1471
  {
1472
- "epoch": 1.0655473038424281,
1473
- "grad_norm": 0.59092116355896,
1474
- "learning_rate": 0.000536657687749165,
1475
- "loss": 3.7667,
1476
  "step": 9900
1477
  },
1478
  {
1479
- "epoch": 1.0709288558820365,
1480
- "grad_norm": 0.638380765914917,
1481
- "learning_rate": 0.0005363344467191035,
1482
- "loss": 3.7722,
1483
  "step": 9950
1484
  },
1485
  {
1486
- "epoch": 1.0763104079216446,
1487
- "grad_norm": 0.5594536066055298,
1488
- "learning_rate": 0.000536011205689042,
1489
- "loss": 3.7655,
1490
  "step": 10000
1491
  },
1492
  {
1493
- "epoch": 1.0763104079216446,
1494
- "eval_accuracy": 0.3451049320696713,
1495
- "eval_loss": 3.7489354610443115,
1496
- "eval_runtime": 202.8767,
1497
- "eval_samples_per_second": 88.778,
1498
- "eval_steps_per_second": 5.55,
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": 3.7572708129882812,
3
+ "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_high_2000_8397/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": 1.6020084619522095,
14
+ "learning_rate": 0.00028199999999999997,
15
+ "loss": 8.5757,
16
  "step": 50
17
  },
18
  {
19
+ "epoch": 0.01078167115902965,
20
+ "grad_norm": 1.8397823572158813,
21
+ "learning_rate": 0.0005819999999999999,
22
+ "loss": 6.94,
23
  "step": 100
24
  },
25
  {
26
+ "epoch": 0.016172506738544475,
27
+ "grad_norm": 2.0749616622924805,
28
+ "learning_rate": 0.0005996956287101997,
29
+ "loss": 6.498,
30
  "step": 150
31
  },
32
  {
33
+ "epoch": 0.0215633423180593,
34
+ "grad_norm": 2.3127713203430176,
35
+ "learning_rate": 0.0005993718294657311,
36
+ "loss": 6.2576,
37
  "step": 200
38
  },
39
  {
40
+ "epoch": 0.026954177897574125,
41
+ "grad_norm": 0.8723729848861694,
42
+ "learning_rate": 0.0005990480302212627,
43
+ "loss": 6.1064,
44
  "step": 250
45
  },
46
  {
47
+ "epoch": 0.03234501347708895,
48
+ "grad_norm": 1.36876380443573,
49
+ "learning_rate": 0.0005987242309767944,
50
+ "loss": 6.0069,
51
  "step": 300
52
  },
53
  {
54
+ "epoch": 0.03773584905660377,
55
+ "grad_norm": 1.1116160154342651,
56
+ "learning_rate": 0.0005984004317323259,
57
+ "loss": 5.8938,
58
  "step": 350
59
  },
60
  {
61
+ "epoch": 0.0431266846361186,
62
+ "grad_norm": 1.7246553897857666,
63
+ "learning_rate": 0.0005980766324878575,
64
+ "loss": 5.8482,
65
  "step": 400
66
  },
67
  {
68
+ "epoch": 0.04851752021563342,
69
+ "grad_norm": 1.5949625968933105,
70
+ "learning_rate": 0.000597752833243389,
71
+ "loss": 5.7717,
72
  "step": 450
73
  },
74
  {
75
+ "epoch": 0.05390835579514825,
76
+ "grad_norm": 1.2913446426391602,
77
+ "learning_rate": 0.0005974290339989207,
78
+ "loss": 5.6782,
79
  "step": 500
80
  },
81
  {
82
+ "epoch": 0.05929919137466307,
83
+ "grad_norm": 1.3474897146224976,
84
+ "learning_rate": 0.0005971052347544522,
85
+ "loss": 5.6001,
86
  "step": 550
87
  },
88
  {
89
+ "epoch": 0.0646900269541779,
90
+ "grad_norm": 1.230842113494873,
91
+ "learning_rate": 0.0005967814355099838,
92
+ "loss": 5.5233,
93
  "step": 600
94
  },
95
  {
96
+ "epoch": 0.07008086253369272,
97
+ "grad_norm": 1.3157235383987427,
98
+ "learning_rate": 0.0005964576362655153,
99
+ "loss": 5.4574,
100
  "step": 650
101
  },
102
  {
103
+ "epoch": 0.07547169811320754,
104
+ "grad_norm": 1.1696827411651611,
105
+ "learning_rate": 0.0005961338370210469,
106
+ "loss": 5.4063,
107
  "step": 700
108
  },
109
  {
110
+ "epoch": 0.08086253369272237,
111
+ "grad_norm": 1.3056632280349731,
112
+ "learning_rate": 0.0005958100377765785,
113
+ "loss": 5.3331,
114
  "step": 750
115
  },
116
  {
117
+ "epoch": 0.0862533692722372,
118
+ "grad_norm": 1.6241542100906372,
119
+ "learning_rate": 0.00059548623853211,
120
+ "loss": 5.2849,
121
  "step": 800
122
  },
123
  {
124
+ "epoch": 0.09164420485175202,
125
+ "grad_norm": 1.0650733709335327,
126
+ "learning_rate": 0.0005951624392876416,
127
+ "loss": 5.2456,
128
  "step": 850
129
  },
130
  {
131
+ "epoch": 0.09703504043126684,
132
+ "grad_norm": 1.5372745990753174,
133
+ "learning_rate": 0.0005948386400431732,
134
+ "loss": 5.191,
135
  "step": 900
136
  },
137
  {
138
+ "epoch": 0.10242587601078167,
139
+ "grad_norm": 1.4312268495559692,
140
+ "learning_rate": 0.0005945148407987047,
141
+ "loss": 5.1659,
142
  "step": 950
143
  },
144
  {
145
+ "epoch": 0.1078167115902965,
146
+ "grad_norm": 0.9435037970542908,
147
+ "learning_rate": 0.0005941910415542363,
148
+ "loss": 5.1266,
149
  "step": 1000
150
  },
151
  {
152
+ "epoch": 0.1078167115902965,
153
+ "eval_accuracy": 0.22553308094716198,
154
+ "eval_loss": 5.041633605957031,
155
+ "eval_runtime": 152.6311,
156
+ "eval_samples_per_second": 118.004,
157
+ "eval_steps_per_second": 7.377,
158
  "step": 1000
159
  },
160
  {
161
+ "epoch": 0.11320754716981132,
162
+ "grad_norm": 1.088536024093628,
163
+ "learning_rate": 0.0005938672423097679,
164
+ "loss": 5.0725,
165
  "step": 1050
166
  },
167
  {
168
+ "epoch": 0.11859838274932614,
169
+ "grad_norm": 0.9872896075248718,
170
+ "learning_rate": 0.0005935434430652995,
171
+ "loss": 5.0443,
172
  "step": 1100
173
  },
174
  {
175
+ "epoch": 0.12398921832884097,
176
+ "grad_norm": 0.9727984070777893,
177
+ "learning_rate": 0.000593219643820831,
178
+ "loss": 5.0164,
179
  "step": 1150
180
  },
181
  {
182
+ "epoch": 0.1293800539083558,
183
+ "grad_norm": 1.0641847848892212,
184
+ "learning_rate": 0.0005928958445763626,
185
+ "loss": 4.978,
186
  "step": 1200
187
  },
188
  {
189
+ "epoch": 0.1347708894878706,
190
+ "grad_norm": 0.8890565633773804,
191
+ "learning_rate": 0.0005925720453318941,
192
+ "loss": 4.9292,
193
  "step": 1250
194
  },
195
  {
196
+ "epoch": 0.14016172506738545,
197
+ "grad_norm": 1.0303993225097656,
198
+ "learning_rate": 0.0005922482460874258,
199
+ "loss": 4.9,
200
  "step": 1300
201
  },
202
  {
203
+ "epoch": 0.14555256064690028,
204
+ "grad_norm": 1.0149179697036743,
205
+ "learning_rate": 0.0005919244468429573,
206
+ "loss": 4.8846,
207
  "step": 1350
208
  },
209
  {
210
+ "epoch": 0.1509433962264151,
211
+ "grad_norm": 1.0036956071853638,
212
+ "learning_rate": 0.0005916006475984889,
213
+ "loss": 4.8728,
214
  "step": 1400
215
  },
216
  {
217
+ "epoch": 0.15633423180592992,
218
+ "grad_norm": 0.8696756958961487,
219
+ "learning_rate": 0.0005912768483540205,
220
+ "loss": 4.8206,
221
  "step": 1450
222
  },
223
  {
224
+ "epoch": 0.16172506738544473,
225
+ "grad_norm": 1.1622956991195679,
226
+ "learning_rate": 0.0005909530491095521,
227
+ "loss": 4.8359,
228
  "step": 1500
229
  },
230
  {
231
+ "epoch": 0.16711590296495957,
232
+ "grad_norm": 1.1141048669815063,
233
+ "learning_rate": 0.0005906292498650836,
234
+ "loss": 4.7878,
235
  "step": 1550
236
  },
237
  {
238
+ "epoch": 0.1725067385444744,
239
+ "grad_norm": 1.274352788925171,
240
+ "learning_rate": 0.0005903054506206151,
241
+ "loss": 4.7862,
242
  "step": 1600
243
  },
244
  {
245
+ "epoch": 0.1778975741239892,
246
+ "grad_norm": 0.8880785703659058,
247
+ "learning_rate": 0.0005899816513761468,
248
+ "loss": 4.7445,
249
  "step": 1650
250
  },
251
  {
252
+ "epoch": 0.18328840970350405,
253
+ "grad_norm": 0.9446322321891785,
254
+ "learning_rate": 0.0005896578521316783,
255
+ "loss": 4.7346,
256
  "step": 1700
257
  },
258
  {
259
+ "epoch": 0.18867924528301888,
260
+ "grad_norm": 0.9463135600090027,
261
+ "learning_rate": 0.0005893340528872099,
262
+ "loss": 4.6952,
263
  "step": 1750
264
  },
265
  {
266
+ "epoch": 0.1940700808625337,
267
+ "grad_norm": 0.8582187294960022,
268
+ "learning_rate": 0.0005890102536427414,
269
+ "loss": 4.7029,
270
  "step": 1800
271
  },
272
  {
273
+ "epoch": 0.19946091644204852,
274
+ "grad_norm": 0.9058123826980591,
275
+ "learning_rate": 0.0005886864543982731,
276
+ "loss": 4.6735,
277
  "step": 1850
278
  },
279
  {
280
+ "epoch": 0.20485175202156333,
281
+ "grad_norm": 0.9203746318817139,
282
+ "learning_rate": 0.0005883626551538046,
283
+ "loss": 4.6596,
284
  "step": 1900
285
  },
286
  {
287
+ "epoch": 0.21024258760107817,
288
+ "grad_norm": 1.0241466760635376,
289
+ "learning_rate": 0.0005880388559093362,
290
+ "loss": 4.6282,
291
  "step": 1950
292
  },
293
  {
294
+ "epoch": 0.215633423180593,
295
+ "grad_norm": 1.1103599071502686,
296
+ "learning_rate": 0.0005877150566648677,
297
+ "loss": 4.6194,
298
  "step": 2000
299
  },
300
  {
301
+ "epoch": 0.215633423180593,
302
+ "eval_accuracy": 0.2670949835939572,
303
+ "eval_loss": 4.5360541343688965,
304
+ "eval_runtime": 152.732,
305
+ "eval_samples_per_second": 117.926,
306
+ "eval_steps_per_second": 7.372,
307
  "step": 2000
308
  },
309
  {
310
+ "epoch": 0.2210242587601078,
311
+ "grad_norm": 0.8629128336906433,
312
+ "learning_rate": 0.0005873912574203993,
313
+ "loss": 4.5954,
314
  "step": 2050
315
  },
316
  {
317
+ "epoch": 0.22641509433962265,
318
+ "grad_norm": 0.7517758011817932,
319
+ "learning_rate": 0.0005870674581759309,
320
+ "loss": 4.5774,
321
  "step": 2100
322
  },
323
  {
324
+ "epoch": 0.23180592991913745,
325
+ "grad_norm": 1.1525517702102661,
326
+ "learning_rate": 0.0005867436589314624,
327
+ "loss": 4.5567,
328
  "step": 2150
329
  },
330
  {
331
+ "epoch": 0.2371967654986523,
332
+ "grad_norm": 1.0665569305419922,
333
+ "learning_rate": 0.000586419859686994,
334
+ "loss": 4.5313,
335
  "step": 2200
336
  },
337
  {
338
+ "epoch": 0.24258760107816713,
339
+ "grad_norm": 0.8165591359138489,
340
+ "learning_rate": 0.0005860960604425256,
341
+ "loss": 4.5156,
342
  "step": 2250
343
  },
344
  {
345
+ "epoch": 0.24797843665768193,
346
+ "grad_norm": 1.0752588510513306,
347
+ "learning_rate": 0.0005857722611980571,
348
+ "loss": 4.526,
349
  "step": 2300
350
  },
351
  {
352
+ "epoch": 0.25336927223719674,
353
+ "grad_norm": 0.9010254740715027,
354
+ "learning_rate": 0.0005854484619535887,
355
+ "loss": 4.4768,
356
  "step": 2350
357
  },
358
  {
359
+ "epoch": 0.2587601078167116,
360
+ "grad_norm": 0.813308596611023,
361
+ "learning_rate": 0.0005851246627091202,
362
+ "loss": 4.486,
363
  "step": 2400
364
  },
365
  {
366
+ "epoch": 0.2641509433962264,
367
+ "grad_norm": 0.997901976108551,
368
+ "learning_rate": 0.0005848008634646519,
369
+ "loss": 4.4602,
370
  "step": 2450
371
  },
372
  {
373
+ "epoch": 0.2695417789757412,
374
+ "grad_norm": 1.0612961053848267,
375
+ "learning_rate": 0.0005844770642201834,
376
+ "loss": 4.4441,
377
  "step": 2500
378
  },
379
  {
380
+ "epoch": 0.2749326145552561,
381
+ "grad_norm": 0.7074645757675171,
382
+ "learning_rate": 0.000584153264975715,
383
+ "loss": 4.4268,
384
  "step": 2550
385
  },
386
  {
387
+ "epoch": 0.2803234501347709,
388
+ "grad_norm": 0.7893711924552917,
389
+ "learning_rate": 0.0005838294657312465,
390
+ "loss": 4.3983,
391
  "step": 2600
392
  },
393
  {
394
+ "epoch": 0.2857142857142857,
395
+ "grad_norm": 0.8888121843338013,
396
+ "learning_rate": 0.0005835056664867782,
397
+ "loss": 4.4127,
398
  "step": 2650
399
  },
400
  {
401
+ "epoch": 0.29110512129380056,
402
+ "grad_norm": 0.8462523221969604,
403
+ "learning_rate": 0.0005831818672423098,
404
+ "loss": 4.4129,
405
  "step": 2700
406
  },
407
  {
408
+ "epoch": 0.29649595687331537,
409
+ "grad_norm": 0.8949580192565918,
410
+ "learning_rate": 0.0005828580679978413,
411
+ "loss": 4.3751,
412
  "step": 2750
413
  },
414
  {
415
+ "epoch": 0.3018867924528302,
416
+ "grad_norm": 0.7590197324752808,
417
+ "learning_rate": 0.0005825342687533729,
418
+ "loss": 4.3503,
419
  "step": 2800
420
  },
421
  {
422
+ "epoch": 0.30727762803234504,
423
+ "grad_norm": 0.6843862533569336,
424
+ "learning_rate": 0.0005822104695089044,
425
+ "loss": 4.3515,
426
  "step": 2850
427
  },
428
  {
429
+ "epoch": 0.31266846361185985,
430
+ "grad_norm": 0.8599426746368408,
431
+ "learning_rate": 0.000581886670264436,
432
+ "loss": 4.342,
433
  "step": 2900
434
  },
435
  {
436
+ "epoch": 0.31805929919137466,
437
+ "grad_norm": 0.893551766872406,
438
+ "learning_rate": 0.0005815628710199675,
439
+ "loss": 4.3359,
440
  "step": 2950
441
  },
442
  {
443
+ "epoch": 0.32345013477088946,
444
+ "grad_norm": 0.8023635149002075,
445
+ "learning_rate": 0.0005812390717754992,
446
+ "loss": 4.3296,
447
  "step": 3000
448
  },
449
  {
450
+ "epoch": 0.32345013477088946,
451
+ "eval_accuracy": 0.29539221573769714,
452
+ "eval_loss": 4.2696380615234375,
453
+ "eval_runtime": 152.7589,
454
+ "eval_samples_per_second": 117.905,
455
+ "eval_steps_per_second": 7.371,
456
  "step": 3000
457
  },
458
  {
459
+ "epoch": 0.3288409703504043,
460
+ "grad_norm": 0.6987361907958984,
461
+ "learning_rate": 0.0005809152725310307,
462
+ "loss": 4.3208,
463
  "step": 3050
464
  },
465
  {
466
+ "epoch": 0.33423180592991913,
467
+ "grad_norm": 0.7668080925941467,
468
+ "learning_rate": 0.0005805914732865623,
469
+ "loss": 4.3246,
470
  "step": 3100
471
  },
472
  {
473
+ "epoch": 0.33962264150943394,
474
+ "grad_norm": 0.7270567417144775,
475
+ "learning_rate": 0.0005802676740420938,
476
+ "loss": 4.2823,
477
  "step": 3150
478
  },
479
  {
480
+ "epoch": 0.3450134770889488,
481
+ "grad_norm": 0.6801213622093201,
482
+ "learning_rate": 0.0005799438747976255,
483
+ "loss": 4.2811,
484
  "step": 3200
485
  },
486
  {
487
+ "epoch": 0.3504043126684636,
488
+ "grad_norm": 0.7790431976318359,
489
+ "learning_rate": 0.000579620075553157,
490
+ "loss": 4.28,
491
  "step": 3250
492
  },
493
  {
494
+ "epoch": 0.3557951482479784,
495
+ "grad_norm": 0.8592586517333984,
496
+ "learning_rate": 0.0005792962763086886,
497
+ "loss": 4.2634,
498
  "step": 3300
499
  },
500
  {
501
+ "epoch": 0.3611859838274933,
502
+ "grad_norm": 0.7063053846359253,
503
+ "learning_rate": 0.0005789724770642201,
504
+ "loss": 4.2784,
505
  "step": 3350
506
  },
507
  {
508
+ "epoch": 0.3665768194070081,
509
+ "grad_norm": 0.7619624733924866,
510
+ "learning_rate": 0.0005786486778197517,
511
+ "loss": 4.2507,
512
  "step": 3400
513
  },
514
  {
515
+ "epoch": 0.3719676549865229,
516
+ "grad_norm": 0.7230337858200073,
517
+ "learning_rate": 0.0005783248785752833,
518
+ "loss": 4.2259,
519
  "step": 3450
520
  },
521
  {
522
+ "epoch": 0.37735849056603776,
523
+ "grad_norm": 0.7475656270980835,
524
+ "learning_rate": 0.0005780010793308148,
525
+ "loss": 4.2353,
526
  "step": 3500
527
  },
528
  {
529
+ "epoch": 0.38274932614555257,
530
+ "grad_norm": 0.7857275009155273,
531
+ "learning_rate": 0.0005776772800863464,
532
+ "loss": 4.2317,
533
  "step": 3550
534
  },
535
  {
536
+ "epoch": 0.3881401617250674,
537
+ "grad_norm": 0.6113690733909607,
538
+ "learning_rate": 0.000577353480841878,
539
+ "loss": 4.2344,
540
  "step": 3600
541
  },
542
  {
543
+ "epoch": 0.3935309973045822,
544
+ "grad_norm": 0.8256933093070984,
545
+ "learning_rate": 0.0005770296815974095,
546
+ "loss": 4.2054,
547
  "step": 3650
548
  },
549
  {
550
+ "epoch": 0.39892183288409705,
551
+ "grad_norm": 0.8064551949501038,
552
+ "learning_rate": 0.0005767058823529411,
553
+ "loss": 4.2078,
554
  "step": 3700
555
  },
556
  {
557
+ "epoch": 0.40431266846361186,
558
+ "grad_norm": 0.6000850200653076,
559
+ "learning_rate": 0.0005763820831084726,
560
+ "loss": 4.2117,
561
  "step": 3750
562
  },
563
  {
564
+ "epoch": 0.40970350404312667,
565
+ "grad_norm": 0.6357919573783875,
566
+ "learning_rate": 0.0005760582838640043,
567
+ "loss": 4.205,
568
  "step": 3800
569
  },
570
  {
571
+ "epoch": 0.41509433962264153,
572
+ "grad_norm": 0.6840759515762329,
573
+ "learning_rate": 0.0005757344846195359,
574
+ "loss": 4.1844,
575
  "step": 3850
576
  },
577
  {
578
+ "epoch": 0.42048517520215634,
579
+ "grad_norm": 0.7748379111289978,
580
+ "learning_rate": 0.0005754106853750674,
581
+ "loss": 4.1815,
582
  "step": 3900
583
  },
584
  {
585
+ "epoch": 0.42587601078167114,
586
+ "grad_norm": 0.6725760698318481,
587
+ "learning_rate": 0.000575086886130599,
588
+ "loss": 4.1742,
589
  "step": 3950
590
  },
591
  {
592
+ "epoch": 0.431266846361186,
593
+ "grad_norm": 0.8388188481330872,
594
+ "learning_rate": 0.0005747630868861306,
595
+ "loss": 4.1751,
596
  "step": 4000
597
  },
598
  {
599
+ "epoch": 0.431266846361186,
600
+ "eval_accuracy": 0.31143709633414934,
601
+ "eval_loss": 4.099897384643555,
602
+ "eval_runtime": 152.799,
603
+ "eval_samples_per_second": 117.874,
604
+ "eval_steps_per_second": 7.369,
605
  "step": 4000
606
  },
607
  {
608
+ "epoch": 0.4366576819407008,
609
+ "grad_norm": 0.5372797250747681,
610
+ "learning_rate": 0.0005744392876416622,
611
+ "loss": 4.1637,
612
  "step": 4050
613
  },
614
  {
615
+ "epoch": 0.4420485175202156,
616
+ "grad_norm": 0.7316073775291443,
617
+ "learning_rate": 0.0005741154883971936,
618
+ "loss": 4.1648,
619
  "step": 4100
620
  },
621
  {
622
+ "epoch": 0.4474393530997305,
623
+ "grad_norm": 0.6784662008285522,
624
+ "learning_rate": 0.0005737916891527253,
625
+ "loss": 4.1653,
626
  "step": 4150
627
  },
628
  {
629
+ "epoch": 0.4528301886792453,
630
+ "grad_norm": 0.6223421692848206,
631
+ "learning_rate": 0.0005734678899082568,
632
+ "loss": 4.1481,
633
  "step": 4200
634
  },
635
  {
636
+ "epoch": 0.4582210242587601,
637
+ "grad_norm": 0.6360011100769043,
638
+ "learning_rate": 0.0005731440906637884,
639
+ "loss": 4.1347,
640
  "step": 4250
641
  },
642
  {
643
+ "epoch": 0.4636118598382749,
644
+ "grad_norm": 0.9023606181144714,
645
+ "learning_rate": 0.0005728202914193199,
646
+ "loss": 4.1318,
647
  "step": 4300
648
  },
649
  {
650
+ "epoch": 0.46900269541778977,
651
+ "grad_norm": 0.5858056545257568,
652
+ "learning_rate": 0.0005724964921748516,
653
+ "loss": 4.1316,
654
  "step": 4350
655
  },
656
  {
657
+ "epoch": 0.4743935309973046,
658
+ "grad_norm": 0.7229328751564026,
659
+ "learning_rate": 0.0005721726929303831,
660
+ "loss": 4.1268,
661
  "step": 4400
662
  },
663
  {
664
+ "epoch": 0.4797843665768194,
665
+ "grad_norm": 0.8891165256500244,
666
+ "learning_rate": 0.0005718488936859147,
667
+ "loss": 4.1138,
668
  "step": 4450
669
  },
670
  {
671
+ "epoch": 0.48517520215633425,
672
+ "grad_norm": 0.6901257038116455,
673
+ "learning_rate": 0.0005715250944414462,
674
+ "loss": 4.1243,
675
  "step": 4500
676
  },
677
  {
678
+ "epoch": 0.49056603773584906,
679
+ "grad_norm": 0.6422321200370789,
680
+ "learning_rate": 0.0005712012951969778,
681
+ "loss": 4.1152,
682
  "step": 4550
683
  },
684
  {
685
+ "epoch": 0.49595687331536387,
686
+ "grad_norm": 0.7133573293685913,
687
+ "learning_rate": 0.0005708774959525094,
688
+ "loss": 4.0984,
689
  "step": 4600
690
  },
691
  {
692
+ "epoch": 0.5013477088948787,
693
+ "grad_norm": 0.625822126865387,
694
+ "learning_rate": 0.000570553696708041,
695
+ "loss": 4.0931,
696
  "step": 4650
697
  },
698
  {
699
+ "epoch": 0.5067385444743935,
700
+ "grad_norm": 0.6901353001594543,
701
+ "learning_rate": 0.0005702298974635725,
702
+ "loss": 4.1064,
703
  "step": 4700
704
  },
705
  {
706
+ "epoch": 0.5121293800539084,
707
+ "grad_norm": 0.6352577805519104,
708
+ "learning_rate": 0.0005699060982191041,
709
+ "loss": 4.0958,
710
  "step": 4750
711
  },
712
  {
713
+ "epoch": 0.5175202156334232,
714
+ "grad_norm": 0.7718018293380737,
715
+ "learning_rate": 0.0005695822989746357,
716
+ "loss": 4.0929,
717
  "step": 4800
718
  },
719
  {
720
+ "epoch": 0.522911051212938,
721
+ "grad_norm": 0.6640039682388306,
722
+ "learning_rate": 0.0005692584997301672,
723
+ "loss": 4.0793,
724
  "step": 4850
725
  },
726
  {
727
+ "epoch": 0.5283018867924528,
728
+ "grad_norm": 0.5986649990081787,
729
+ "learning_rate": 0.0005689347004856988,
730
+ "loss": 4.0624,
731
  "step": 4900
732
  },
733
  {
734
+ "epoch": 0.5336927223719676,
735
+ "grad_norm": 0.6509894132614136,
736
+ "learning_rate": 0.0005686109012412304,
737
+ "loss": 4.08,
738
  "step": 4950
739
  },
740
  {
741
+ "epoch": 0.5390835579514824,
742
+ "grad_norm": 0.5881455540657043,
743
+ "learning_rate": 0.000568287101996762,
744
+ "loss": 4.0618,
745
  "step": 5000
746
  },
747
  {
748
+ "epoch": 0.5390835579514824,
749
+ "eval_accuracy": 0.3202667732623931,
750
+ "eval_loss": 4.000813007354736,
751
+ "eval_runtime": 152.9398,
752
+ "eval_samples_per_second": 117.765,
753
+ "eval_steps_per_second": 7.362,
754
  "step": 5000
755
  },
756
  {
757
+ "epoch": 0.5444743935309974,
758
+ "grad_norm": 0.6732229590415955,
759
+ "learning_rate": 0.0005679633027522935,
760
+ "loss": 4.068,
761
  "step": 5050
762
  },
763
  {
764
+ "epoch": 0.5498652291105122,
765
+ "grad_norm": 0.6501848697662354,
766
+ "learning_rate": 0.000567639503507825,
767
+ "loss": 4.0611,
768
  "step": 5100
769
  },
770
  {
771
+ "epoch": 0.555256064690027,
772
+ "grad_norm": 0.6546076536178589,
773
+ "learning_rate": 0.0005673157042633567,
774
+ "loss": 4.0601,
775
  "step": 5150
776
  },
777
  {
778
+ "epoch": 0.5606469002695418,
779
+ "grad_norm": 0.6216645836830139,
780
+ "learning_rate": 0.0005669919050188883,
781
+ "loss": 4.0788,
782
  "step": 5200
783
  },
784
  {
785
+ "epoch": 0.5660377358490566,
786
+ "grad_norm": 0.6688370704650879,
787
+ "learning_rate": 0.0005666681057744198,
788
+ "loss": 4.0529,
789
  "step": 5250
790
  },
791
  {
792
+ "epoch": 0.5714285714285714,
793
+ "grad_norm": 0.7299574613571167,
794
+ "learning_rate": 0.0005663443065299514,
795
+ "loss": 4.0448,
796
  "step": 5300
797
  },
798
  {
799
+ "epoch": 0.5768194070080862,
800
+ "grad_norm": 0.6612991094589233,
801
+ "learning_rate": 0.000566020507285483,
802
+ "loss": 4.0386,
803
  "step": 5350
804
  },
805
  {
806
+ "epoch": 0.5822102425876011,
807
+ "grad_norm": 0.5907670855522156,
808
+ "learning_rate": 0.0005656967080410146,
809
+ "loss": 4.0478,
810
  "step": 5400
811
  },
812
  {
813
+ "epoch": 0.5876010781671159,
814
+ "grad_norm": 0.539881706237793,
815
+ "learning_rate": 0.000565372908796546,
816
+ "loss": 4.0481,
817
  "step": 5450
818
  },
819
  {
820
+ "epoch": 0.5929919137466307,
821
+ "grad_norm": 0.6923815608024597,
822
+ "learning_rate": 0.0005650491095520777,
823
+ "loss": 4.028,
824
  "step": 5500
825
  },
826
  {
827
+ "epoch": 0.5983827493261455,
828
+ "grad_norm": 0.6604423522949219,
829
+ "learning_rate": 0.0005647253103076092,
830
+ "loss": 4.0191,
831
  "step": 5550
832
  },
833
  {
834
+ "epoch": 0.6037735849056604,
835
+ "grad_norm": 0.6762202382087708,
836
+ "learning_rate": 0.0005644015110631408,
837
+ "loss": 4.038,
838
  "step": 5600
839
  },
840
  {
841
+ "epoch": 0.6091644204851752,
842
+ "grad_norm": 0.6797250509262085,
843
+ "learning_rate": 0.0005640777118186723,
844
+ "loss": 4.015,
845
  "step": 5650
846
  },
847
  {
848
+ "epoch": 0.6145552560646901,
849
+ "grad_norm": 0.5718669891357422,
850
+ "learning_rate": 0.000563753912574204,
851
+ "loss": 4.0142,
852
  "step": 5700
853
  },
854
  {
855
+ "epoch": 0.6199460916442049,
856
+ "grad_norm": 0.6036492586135864,
857
+ "learning_rate": 0.0005634301133297355,
858
+ "loss": 3.9904,
859
  "step": 5750
860
  },
861
  {
862
+ "epoch": 0.6253369272237197,
863
+ "grad_norm": 0.6295326352119446,
864
+ "learning_rate": 0.0005631063140852671,
865
+ "loss": 3.9973,
866
  "step": 5800
867
  },
868
  {
869
+ "epoch": 0.6307277628032345,
870
+ "grad_norm": 0.6866003274917603,
871
+ "learning_rate": 0.0005627825148407986,
872
+ "loss": 4.0086,
873
  "step": 5850
874
  },
875
  {
876
+ "epoch": 0.6361185983827493,
877
+ "grad_norm": 0.6472943425178528,
878
+ "learning_rate": 0.0005624587155963302,
879
+ "loss": 4.0048,
880
  "step": 5900
881
  },
882
  {
883
+ "epoch": 0.6415094339622641,
884
+ "grad_norm": 0.6805757284164429,
885
+ "learning_rate": 0.0005621349163518618,
886
+ "loss": 3.9913,
887
  "step": 5950
888
  },
889
  {
890
+ "epoch": 0.6469002695417789,
891
+ "grad_norm": 0.6941430568695068,
892
+ "learning_rate": 0.0005618111171073934,
893
+ "loss": 4.0077,
894
  "step": 6000
895
  },
896
  {
897
+ "epoch": 0.6469002695417789,
898
+ "eval_accuracy": 0.3263130891634934,
899
+ "eval_loss": 3.9324727058410645,
900
+ "eval_runtime": 152.9066,
901
+ "eval_samples_per_second": 117.791,
902
+ "eval_steps_per_second": 7.364,
903
  "step": 6000
904
  },
905
  {
906
+ "epoch": 0.6522911051212938,
907
+ "grad_norm": 0.6350868344306946,
908
+ "learning_rate": 0.0005614873178629249,
909
  "loss": 3.9789,
910
  "step": 6050
911
  },
912
  {
913
+ "epoch": 0.6576819407008087,
914
+ "grad_norm": 0.6860512495040894,
915
+ "learning_rate": 0.0005611635186184565,
916
+ "loss": 4.0102,
917
  "step": 6100
918
  },
919
  {
920
+ "epoch": 0.6630727762803235,
921
+ "grad_norm": 0.771823525428772,
922
+ "learning_rate": 0.0005608397193739882,
923
+ "loss": 3.9907,
924
  "step": 6150
925
  },
926
  {
927
+ "epoch": 0.6684636118598383,
928
+ "grad_norm": 0.6326267123222351,
929
+ "learning_rate": 0.0005605159201295196,
930
+ "loss": 3.9633,
931
  "step": 6200
932
  },
933
  {
934
+ "epoch": 0.6738544474393531,
935
+ "grad_norm": 0.5878480672836304,
936
+ "learning_rate": 0.0005601921208850511,
937
+ "loss": 3.9821,
938
  "step": 6250
939
  },
940
  {
941
+ "epoch": 0.6792452830188679,
942
+ "grad_norm": 0.6786486506462097,
943
+ "learning_rate": 0.0005598683216405828,
944
+ "loss": 3.9856,
945
  "step": 6300
946
  },
947
  {
948
+ "epoch": 0.6846361185983828,
949
+ "grad_norm": 0.5655982494354248,
950
+ "learning_rate": 0.0005595445223961144,
951
+ "loss": 3.9732,
952
  "step": 6350
953
  },
954
  {
955
+ "epoch": 0.6900269541778976,
956
+ "grad_norm": 0.6204545497894287,
957
+ "learning_rate": 0.0005592207231516459,
958
+ "loss": 3.9792,
959
  "step": 6400
960
  },
961
  {
962
+ "epoch": 0.6954177897574124,
963
+ "grad_norm": 0.718676745891571,
964
+ "learning_rate": 0.0005588969239071775,
965
+ "loss": 3.9757,
966
  "step": 6450
967
  },
968
  {
969
+ "epoch": 0.7008086253369272,
970
+ "grad_norm": 0.5987905263900757,
971
+ "learning_rate": 0.0005585731246627091,
972
+ "loss": 3.9713,
973
  "step": 6500
974
  },
975
  {
976
+ "epoch": 0.706199460916442,
977
+ "grad_norm": 0.6304261684417725,
978
+ "learning_rate": 0.0005582493254182407,
979
+ "loss": 3.9699,
980
  "step": 6550
981
  },
982
  {
983
+ "epoch": 0.7115902964959568,
984
+ "grad_norm": 0.5621398091316223,
985
+ "learning_rate": 0.0005579255261737722,
986
+ "loss": 3.9707,
987
  "step": 6600
988
  },
989
  {
990
+ "epoch": 0.7169811320754716,
991
+ "grad_norm": 0.637505829334259,
992
+ "learning_rate": 0.0005576017269293038,
993
+ "loss": 3.9813,
994
  "step": 6650
995
  },
996
  {
997
+ "epoch": 0.7223719676549866,
998
+ "grad_norm": 0.6425164341926575,
999
+ "learning_rate": 0.0005572779276848353,
1000
+ "loss": 3.9482,
1001
  "step": 6700
1002
  },
1003
  {
1004
+ "epoch": 0.7277628032345014,
1005
+ "grad_norm": 0.624649703502655,
1006
+ "learning_rate": 0.000556954128440367,
1007
+ "loss": 3.9741,
1008
  "step": 6750
1009
  },
1010
  {
1011
+ "epoch": 0.7331536388140162,
1012
+ "grad_norm": 0.6176960468292236,
1013
+ "learning_rate": 0.0005566303291958984,
1014
+ "loss": 3.9436,
1015
  "step": 6800
1016
  },
1017
  {
1018
+ "epoch": 0.738544474393531,
1019
+ "grad_norm": 0.6119649410247803,
1020
+ "learning_rate": 0.0005563065299514301,
1021
+ "loss": 3.9442,
1022
  "step": 6850
1023
  },
1024
  {
1025
+ "epoch": 0.7439353099730458,
1026
+ "grad_norm": 0.5820748805999756,
1027
+ "learning_rate": 0.0005559827307069616,
1028
+ "loss": 3.9463,
1029
  "step": 6900
1030
  },
1031
  {
1032
+ "epoch": 0.7493261455525606,
1033
+ "grad_norm": 0.6512102484703064,
1034
+ "learning_rate": 0.0005556589314624932,
1035
+ "loss": 3.9498,
1036
  "step": 6950
1037
  },
1038
  {
1039
+ "epoch": 0.7547169811320755,
1040
+ "grad_norm": 0.548623263835907,
1041
+ "learning_rate": 0.0005553351322180247,
1042
+ "loss": 3.9306,
1043
  "step": 7000
1044
  },
1045
  {
1046
+ "epoch": 0.7547169811320755,
1047
+ "eval_accuracy": 0.3323316985782009,
1048
+ "eval_loss": 3.8741683959960938,
1049
+ "eval_runtime": 152.843,
1050
+ "eval_samples_per_second": 117.84,
1051
+ "eval_steps_per_second": 7.367,
1052
  "step": 7000
1053
  },
1054
  {
1055
+ "epoch": 0.7601078167115903,
1056
+ "grad_norm": 0.6566927433013916,
1057
+ "learning_rate": 0.0005550113329735564,
1058
+ "loss": 3.9435,
1059
  "step": 7050
1060
  },
1061
  {
1062
+ "epoch": 0.7654986522911051,
1063
+ "grad_norm": 0.8524190187454224,
1064
+ "learning_rate": 0.0005546875337290879,
1065
+ "loss": 3.9177,
1066
  "step": 7100
1067
  },
1068
  {
1069
+ "epoch": 0.77088948787062,
1070
+ "grad_norm": 0.6614950299263,
1071
+ "learning_rate": 0.0005543637344846195,
1072
+ "loss": 3.9448,
1073
  "step": 7150
1074
  },
1075
  {
1076
+ "epoch": 0.7762803234501348,
1077
+ "grad_norm": 0.5254852175712585,
1078
+ "learning_rate": 0.000554039935240151,
1079
+ "loss": 3.9164,
1080
  "step": 7200
1081
  },
1082
  {
1083
+ "epoch": 0.7816711590296496,
1084
+ "grad_norm": 0.6216305494308472,
1085
+ "learning_rate": 0.0005537161359956826,
1086
+ "loss": 3.909,
1087
  "step": 7250
1088
  },
1089
  {
1090
+ "epoch": 0.7870619946091644,
1091
+ "grad_norm": 0.6032219529151917,
1092
+ "learning_rate": 0.0005533923367512143,
1093
+ "loss": 3.9232,
1094
  "step": 7300
1095
  },
1096
  {
1097
+ "epoch": 0.7924528301886793,
1098
+ "grad_norm": 0.6744025349617004,
1099
+ "learning_rate": 0.0005530685375067458,
1100
+ "loss": 3.9224,
1101
  "step": 7350
1102
  },
1103
  {
1104
+ "epoch": 0.7978436657681941,
1105
+ "grad_norm": 0.568401038646698,
1106
+ "learning_rate": 0.0005527447382622774,
1107
+ "loss": 3.918,
1108
  "step": 7400
1109
  },
1110
  {
1111
+ "epoch": 0.8032345013477089,
1112
+ "grad_norm": 0.5871582627296448,
1113
+ "learning_rate": 0.0005524209390178089,
1114
+ "loss": 3.8955,
1115
  "step": 7450
1116
  },
1117
  {
1118
+ "epoch": 0.8086253369272237,
1119
+ "grad_norm": 0.641424834728241,
1120
+ "learning_rate": 0.0005520971397733406,
1121
+ "loss": 3.9128,
1122
  "step": 7500
1123
  },
1124
  {
1125
+ "epoch": 0.8140161725067385,
1126
+ "grad_norm": 0.60320645570755,
1127
+ "learning_rate": 0.000551773340528872,
1128
+ "loss": 3.908,
1129
  "step": 7550
1130
  },
1131
  {
1132
+ "epoch": 0.8194070080862533,
1133
+ "grad_norm": 0.6258028149604797,
1134
+ "learning_rate": 0.0005514495412844036,
1135
+ "loss": 3.9018,
1136
  "step": 7600
1137
  },
1138
  {
1139
+ "epoch": 0.8247978436657682,
1140
+ "grad_norm": 0.5553128123283386,
1141
+ "learning_rate": 0.0005511257420399352,
1142
+ "loss": 3.9058,
1143
  "step": 7650
1144
  },
1145
  {
1146
+ "epoch": 0.8301886792452831,
1147
+ "grad_norm": 0.6213424801826477,
1148
+ "learning_rate": 0.0005508019427954668,
1149
+ "loss": 3.9005,
1150
  "step": 7700
1151
  },
1152
  {
1153
+ "epoch": 0.8355795148247979,
1154
+ "grad_norm": 0.6496004462242126,
1155
+ "learning_rate": 0.0005504781435509983,
1156
+ "loss": 3.9025,
1157
  "step": 7750
1158
  },
1159
  {
1160
+ "epoch": 0.8409703504043127,
1161
+ "grad_norm": 0.6373844146728516,
1162
+ "learning_rate": 0.0005501543443065299,
1163
+ "loss": 3.8844,
1164
  "step": 7800
1165
  },
1166
  {
1167
+ "epoch": 0.8463611859838275,
1168
+ "grad_norm": 0.5957589745521545,
1169
+ "learning_rate": 0.0005498305450620615,
1170
+ "loss": 3.8902,
1171
  "step": 7850
1172
  },
1173
  {
1174
+ "epoch": 0.8517520215633423,
1175
+ "grad_norm": 0.6516701579093933,
1176
+ "learning_rate": 0.0005495067458175931,
1177
+ "loss": 3.8876,
1178
  "step": 7900
1179
  },
1180
  {
1181
+ "epoch": 0.8571428571428571,
1182
+ "grad_norm": 0.6374366283416748,
1183
+ "learning_rate": 0.0005491829465731246,
1184
+ "loss": 3.9173,
1185
  "step": 7950
1186
  },
1187
  {
1188
+ "epoch": 0.862533692722372,
1189
+ "grad_norm": 0.5943326354026794,
1190
+ "learning_rate": 0.0005488591473286562,
1191
+ "loss": 3.8636,
1192
  "step": 8000
1193
  },
1194
  {
1195
+ "epoch": 0.862533692722372,
1196
+ "eval_accuracy": 0.3369106572293666,
1197
+ "eval_loss": 3.824968099594116,
1198
+ "eval_runtime": 152.7242,
1199
+ "eval_samples_per_second": 117.932,
1200
+ "eval_steps_per_second": 7.373,
1201
  "step": 8000
1202
  },
1203
  {
1204
+ "epoch": 0.8679245283018868,
1205
+ "grad_norm": 0.6996535062789917,
1206
+ "learning_rate": 0.0005485353480841877,
1207
+ "loss": 3.906,
1208
  "step": 8050
1209
  },
1210
  {
1211
+ "epoch": 0.8733153638814016,
1212
+ "grad_norm": 0.6234637498855591,
1213
+ "learning_rate": 0.0005482115488397194,
1214
+ "loss": 3.8831,
1215
  "step": 8100
1216
  },
1217
  {
1218
+ "epoch": 0.8787061994609164,
1219
+ "grad_norm": 0.6020538806915283,
1220
+ "learning_rate": 0.0005478877495952508,
1221
+ "loss": 3.8864,
1222
  "step": 8150
1223
  },
1224
  {
1225
+ "epoch": 0.8840970350404312,
1226
+ "grad_norm": 0.5640770792961121,
1227
+ "learning_rate": 0.0005475639503507825,
1228
+ "loss": 3.8799,
1229
  "step": 8200
1230
  },
1231
  {
1232
+ "epoch": 0.889487870619946,
1233
+ "grad_norm": 0.6602328419685364,
1234
+ "learning_rate": 0.000547240151106314,
1235
+ "loss": 3.8788,
1236
  "step": 8250
1237
  },
1238
  {
1239
+ "epoch": 0.894878706199461,
1240
+ "grad_norm": 0.6336698532104492,
1241
+ "learning_rate": 0.0005469163518618456,
1242
+ "loss": 3.8969,
1243
  "step": 8300
1244
  },
1245
  {
1246
+ "epoch": 0.9002695417789758,
1247
+ "grad_norm": 0.562479555606842,
1248
+ "learning_rate": 0.0005465925526173771,
1249
+ "loss": 3.8686,
1250
  "step": 8350
1251
  },
1252
  {
1253
+ "epoch": 0.9056603773584906,
1254
+ "grad_norm": 0.5576388835906982,
1255
+ "learning_rate": 0.0005462687533729087,
1256
+ "loss": 3.8729,
1257
  "step": 8400
1258
  },
1259
  {
1260
+ "epoch": 0.9110512129380054,
1261
+ "grad_norm": 0.5111304521560669,
1262
+ "learning_rate": 0.0005459449541284403,
1263
+ "loss": 3.8679,
1264
  "step": 8450
1265
  },
1266
  {
1267
+ "epoch": 0.9164420485175202,
1268
+ "grad_norm": 0.6154835224151611,
1269
+ "learning_rate": 0.0005456211548839719,
1270
+ "loss": 3.8749,
1271
  "step": 8500
1272
  },
1273
  {
1274
+ "epoch": 0.921832884097035,
1275
+ "grad_norm": 0.5426989197731018,
1276
+ "learning_rate": 0.0005452973556395034,
1277
+ "loss": 3.8685,
1278
  "step": 8550
1279
  },
1280
  {
1281
+ "epoch": 0.9272237196765498,
1282
+ "grad_norm": 0.5612382888793945,
1283
+ "learning_rate": 0.000544973556395035,
1284
+ "loss": 3.8747,
1285
  "step": 8600
1286
  },
1287
  {
1288
+ "epoch": 0.9326145552560647,
1289
+ "grad_norm": 0.5568158626556396,
1290
+ "learning_rate": 0.0005446497571505667,
1291
+ "loss": 3.8547,
1292
  "step": 8650
1293
  },
1294
  {
1295
+ "epoch": 0.9380053908355795,
1296
+ "grad_norm": 0.6064712405204773,
1297
+ "learning_rate": 0.0005443259579060982,
1298
+ "loss": 3.8589,
1299
  "step": 8700
1300
  },
1301
  {
1302
+ "epoch": 0.9433962264150944,
1303
+ "grad_norm": 0.5729287266731262,
1304
+ "learning_rate": 0.0005440021586616298,
1305
+ "loss": 3.8548,
1306
  "step": 8750
1307
  },
1308
  {
1309
+ "epoch": 0.9487870619946092,
1310
+ "grad_norm": 0.5489872694015503,
1311
+ "learning_rate": 0.0005436783594171613,
1312
+ "loss": 3.8599,
1313
  "step": 8800
1314
  },
1315
  {
1316
+ "epoch": 0.954177897574124,
1317
+ "grad_norm": 0.6285766959190369,
1318
+ "learning_rate": 0.0005433545601726929,
1319
+ "loss": 3.8606,
1320
  "step": 8850
1321
  },
1322
  {
1323
+ "epoch": 0.9595687331536388,
1324
+ "grad_norm": 0.566647469997406,
1325
+ "learning_rate": 0.0005430307609282244,
1326
+ "loss": 3.8472,
1327
  "step": 8900
1328
  },
1329
  {
1330
+ "epoch": 0.9649595687331537,
1331
+ "grad_norm": 0.5826287269592285,
1332
+ "learning_rate": 0.000542706961683756,
1333
+ "loss": 3.8404,
1334
  "step": 8950
1335
  },
1336
  {
1337
+ "epoch": 0.9703504043126685,
1338
+ "grad_norm": 0.6150389313697815,
1339
+ "learning_rate": 0.0005423831624392876,
1340
+ "loss": 3.8653,
1341
  "step": 9000
1342
  },
1343
  {
1344
+ "epoch": 0.9703504043126685,
1345
+ "eval_accuracy": 0.3401243923451433,
1346
+ "eval_loss": 3.788287878036499,
1347
+ "eval_runtime": 153.2128,
1348
+ "eval_samples_per_second": 117.555,
1349
+ "eval_steps_per_second": 7.349,
1350
  "step": 9000
1351
  },
1352
  {
1353
+ "epoch": 0.9757412398921833,
1354
+ "grad_norm": 0.6907781958580017,
1355
+ "learning_rate": 0.0005420593631948192,
1356
+ "loss": 3.8789,
1357
  "step": 9050
1358
  },
1359
  {
1360
+ "epoch": 0.9811320754716981,
1361
+ "grad_norm": 0.5165227055549622,
1362
+ "learning_rate": 0.0005417355639503507,
1363
+ "loss": 3.8372,
1364
  "step": 9100
1365
  },
1366
  {
1367
+ "epoch": 0.9865229110512129,
1368
+ "grad_norm": 0.563310980796814,
1369
+ "learning_rate": 0.0005414117647058823,
1370
+ "loss": 3.848,
1371
  "step": 9150
1372
  },
1373
  {
1374
+ "epoch": 0.9919137466307277,
1375
+ "grad_norm": 0.5229910016059875,
1376
+ "learning_rate": 0.0005410879654614139,
1377
+ "loss": 3.8388,
1378
  "step": 9200
1379
  },
1380
  {
1381
+ "epoch": 0.9973045822102425,
1382
+ "grad_norm": 0.5532335638999939,
1383
+ "learning_rate": 0.0005407641662169455,
1384
+ "loss": 3.8559,
1385
  "step": 9250
1386
  },
1387
  {
1388
+ "epoch": 1.0026954177897573,
1389
+ "grad_norm": 0.5877416729927063,
1390
+ "learning_rate": 0.000540440366972477,
1391
+ "loss": 3.8053,
1392
  "step": 9300
1393
  },
1394
  {
1395
+ "epoch": 1.0080862533692723,
1396
+ "grad_norm": 0.5423709154129028,
1397
+ "learning_rate": 0.0005401165677280086,
1398
+ "loss": 3.7729,
1399
  "step": 9350
1400
  },
1401
  {
1402
+ "epoch": 1.013477088948787,
1403
+ "grad_norm": 0.6410120725631714,
1404
+ "learning_rate": 0.0005397927684835401,
1405
+ "loss": 3.7655,
1406
  "step": 9400
1407
  },
1408
  {
1409
+ "epoch": 1.0188679245283019,
1410
+ "grad_norm": 0.5875224471092224,
1411
+ "learning_rate": 0.0005394689692390718,
1412
+ "loss": 3.7643,
1413
  "step": 9450
1414
  },
1415
  {
1416
+ "epoch": 1.0242587601078168,
1417
+ "grad_norm": 0.5834726095199585,
1418
+ "learning_rate": 0.0005391451699946032,
1419
+ "loss": 3.7698,
1420
  "step": 9500
1421
  },
1422
  {
1423
+ "epoch": 1.0296495956873315,
1424
+ "grad_norm": 0.5419567227363586,
1425
+ "learning_rate": 0.0005388213707501349,
1426
+ "loss": 3.7799,
1427
  "step": 9550
1428
  },
1429
  {
1430
+ "epoch": 1.0350404312668464,
1431
+ "grad_norm": 0.6511868238449097,
1432
+ "learning_rate": 0.0005384975715056664,
1433
+ "loss": 3.7784,
1434
  "step": 9600
1435
  },
1436
  {
1437
+ "epoch": 1.0404312668463611,
1438
+ "grad_norm": 0.6302853226661682,
1439
+ "learning_rate": 0.000538173772261198,
1440
+ "loss": 3.7842,
1441
  "step": 9650
1442
  },
1443
  {
1444
+ "epoch": 1.045822102425876,
1445
+ "grad_norm": 0.531711220741272,
1446
+ "learning_rate": 0.0005378499730167295,
1447
+ "loss": 3.7661,
1448
  "step": 9700
1449
  },
1450
  {
1451
+ "epoch": 1.0512129380053907,
1452
+ "grad_norm": 0.5548348426818848,
1453
+ "learning_rate": 0.0005375261737722611,
1454
+ "loss": 3.7646,
1455
  "step": 9750
1456
  },
1457
  {
1458
+ "epoch": 1.0566037735849056,
1459
+ "grad_norm": 0.5238717794418335,
1460
+ "learning_rate": 0.0005372023745277928,
1461
+ "loss": 3.7914,
1462
  "step": 9800
1463
  },
1464
  {
1465
+ "epoch": 1.0619946091644206,
1466
+ "grad_norm": 0.5447802543640137,
1467
+ "learning_rate": 0.0005368785752833243,
1468
+ "loss": 3.7911,
1469
  "step": 9850
1470
  },
1471
  {
1472
+ "epoch": 1.0673854447439353,
1473
+ "grad_norm": 0.5512890815734863,
1474
+ "learning_rate": 0.0005365547760388559,
1475
+ "loss": 3.7828,
1476
  "step": 9900
1477
  },
1478
  {
1479
+ "epoch": 1.0727762803234502,
1480
+ "grad_norm": 0.5664125680923462,
1481
+ "learning_rate": 0.0005362309767943874,
1482
+ "loss": 3.7949,
1483
  "step": 9950
1484
  },
1485
  {
1486
+ "epoch": 1.0781671159029649,
1487
+ "grad_norm": 0.6107265949249268,
1488
+ "learning_rate": 0.0005359071775499191,
1489
+ "loss": 3.7721,
1490
  "step": 10000
1491
  },
1492
  {
1493
+ "epoch": 1.0781671159029649,
1494
+ "eval_accuracy": 0.34332498046149446,
1495
+ "eval_loss": 3.7572708129882812,
1496
+ "eval_runtime": 152.6822,
1497
+ "eval_samples_per_second": 117.964,
1498
+ "eval_steps_per_second": 7.375,
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:a8af7b84dbc09e2405215b378e28c9508606efb6c1cb6a8674665fa2271606b7
3
  size 5304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ed1ab53749d5de05256c65585525328fbc309ce2d8351fa07539a9badd09c05
3
  size 5304