craa commited on
Commit
ae51ce1
·
verified ·
1 Parent(s): fa9c41d

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 sha256:2405eae117066751ff23c186f811e19203a3b71f1c05da8d9f288835ee20489b
3
  size 503128704
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a52019639e5b010548251e53e17ad42557c6d18ad0ebd0e2e37ee81922dfb7ca
3
  size 503128704
checkpoint-10000/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6f7528136a4570bc56a2b2e27b7ac42b527007ef2c8a1e277f837fd77abab827
3
  size 1006351290
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:11409cb1233ff5c63e4d6c0a241ba3f14c2b33cfe0ae5cbf79c57785bbe0e3c6
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:ce39e4abd0143f1ded75f13fa94c7409bfd3ac498f2698ca791e4e4f9ab28414
3
  size 14244
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:53a47b99e43a46630effbfb45e7223c8a493d5b48b00aebac2470d03eafb7364
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:baaddc0dc177f093cb7927eae851400a36fde6f45129bcaf2df443bdde5269f4
3
  size 1064
checkpoint-10000/trainer_state.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "best_metric": 3.7558341026306152,
3
- "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_1208/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": 8.489361763000488,
14
  "learning_rate": 0.0003,
15
- "loss": 8.5006,
16
  "step": 50
17
  },
18
  {
19
- "epoch": 0.010763104079216447,
20
- "grad_norm": 1.9615029096603394,
21
  "learning_rate": 0.0006,
22
- "loss": 6.8831,
23
  "step": 100
24
  },
25
  {
26
- "epoch": 0.01614465611882467,
27
- "grad_norm": 1.9948406219482422,
28
- "learning_rate": 0.0005996767589699385,
29
- "loss": 6.4033,
30
  "step": 150
31
  },
32
  {
33
- "epoch": 0.021526208158432893,
34
- "grad_norm": 2.2949752807617188,
35
- "learning_rate": 0.0005993535179398771,
36
- "loss": 6.2058,
37
  "step": 200
38
  },
39
  {
40
- "epoch": 0.026907760198041114,
41
- "grad_norm": 1.0898317098617554,
42
- "learning_rate": 0.0005990302769098158,
43
- "loss": 6.0385,
44
  "step": 250
45
  },
46
  {
47
- "epoch": 0.03228931223764934,
48
- "grad_norm": 1.9062098264694214,
49
- "learning_rate": 0.0005987070358797543,
50
- "loss": 5.972,
51
  "step": 300
52
  },
53
  {
54
- "epoch": 0.03767086427725756,
55
- "grad_norm": 1.7062020301818848,
56
- "learning_rate": 0.0005983837948496929,
57
- "loss": 5.8878,
58
  "step": 350
59
  },
60
  {
61
- "epoch": 0.04305241631686579,
62
- "grad_norm": 1.0963656902313232,
63
- "learning_rate": 0.0005980605538196314,
64
- "loss": 5.7736,
65
  "step": 400
66
  },
67
  {
68
- "epoch": 0.048433968356474004,
69
- "grad_norm": 1.9562582969665527,
70
- "learning_rate": 0.0005977373127895701,
71
- "loss": 5.7226,
72
  "step": 450
73
  },
74
  {
75
- "epoch": 0.05381552039608223,
76
- "grad_norm": 1.2749587297439575,
77
- "learning_rate": 0.0005974140717595086,
78
- "loss": 5.6243,
79
  "step": 500
80
  },
81
  {
82
- "epoch": 0.05919707243569045,
83
- "grad_norm": 1.6608922481536865,
84
- "learning_rate": 0.0005970908307294472,
85
- "loss": 5.5516,
86
  "step": 550
87
  },
88
  {
89
- "epoch": 0.06457862447529868,
90
- "grad_norm": 1.5849041938781738,
91
- "learning_rate": 0.0005967675896993858,
92
- "loss": 5.4851,
93
  "step": 600
94
  },
95
  {
96
- "epoch": 0.0699601765149069,
97
- "grad_norm": 1.1615055799484253,
98
- "learning_rate": 0.0005964443486693243,
99
- "loss": 5.4325,
100
  "step": 650
101
  },
102
  {
103
- "epoch": 0.07534172855451512,
104
- "grad_norm": 1.5066120624542236,
105
- "learning_rate": 0.000596121107639263,
106
- "loss": 5.3674,
107
  "step": 700
108
  },
109
  {
110
- "epoch": 0.08072328059412334,
111
- "grad_norm": 1.1698895692825317,
112
- "learning_rate": 0.0005957978666092015,
113
- "loss": 5.2951,
114
  "step": 750
115
  },
116
  {
117
- "epoch": 0.08610483263373157,
118
- "grad_norm": 1.969469428062439,
119
- "learning_rate": 0.0005954746255791401,
120
- "loss": 5.256,
121
  "step": 800
122
  },
123
  {
124
- "epoch": 0.09148638467333979,
125
- "grad_norm": 1.214189052581787,
126
- "learning_rate": 0.0005951513845490787,
127
- "loss": 5.1884,
128
  "step": 850
129
  },
130
  {
131
- "epoch": 0.09686793671294801,
132
- "grad_norm": 1.052661657333374,
133
- "learning_rate": 0.0005948281435190174,
134
- "loss": 5.1527,
135
  "step": 900
136
  },
137
  {
138
- "epoch": 0.10224948875255624,
139
- "grad_norm": 1.2582467794418335,
140
- "learning_rate": 0.0005945049024889559,
141
- "loss": 5.1205,
142
  "step": 950
143
  },
144
  {
145
- "epoch": 0.10763104079216446,
146
- "grad_norm": 0.8772917985916138,
147
- "learning_rate": 0.0005941816614588944,
148
- "loss": 5.0934,
149
  "step": 1000
150
  },
151
  {
152
- "epoch": 0.10763104079216446,
153
- "eval_accuracy": 0.22830242575177748,
154
- "eval_loss": 5.011170387268066,
155
- "eval_runtime": 184.392,
156
- "eval_samples_per_second": 97.678,
157
- "eval_steps_per_second": 6.107,
158
  "step": 1000
159
  },
160
  {
161
- "epoch": 0.11301259283177269,
162
- "grad_norm": 1.2470630407333374,
163
- "learning_rate": 0.000593858420428833,
164
- "loss": 5.0303,
165
  "step": 1050
166
  },
167
  {
168
- "epoch": 0.1183941448713809,
169
- "grad_norm": 1.3724963665008545,
170
- "learning_rate": 0.0005935351793987716,
171
- "loss": 5.0012,
172
  "step": 1100
173
  },
174
  {
175
- "epoch": 0.12377569691098914,
176
- "grad_norm": 1.0659003257751465,
177
- "learning_rate": 0.0005932119383687103,
178
- "loss": 5.0045,
179
  "step": 1150
180
  },
181
  {
182
- "epoch": 0.12915724895059735,
183
- "grad_norm": 1.0786244869232178,
184
- "learning_rate": 0.0005928886973386488,
185
- "loss": 4.9516,
186
  "step": 1200
187
  },
188
  {
189
- "epoch": 0.13453880099020557,
190
- "grad_norm": 0.7411894798278809,
191
- "learning_rate": 0.0005925654563085874,
192
- "loss": 4.9101,
193
  "step": 1250
194
  },
195
  {
196
- "epoch": 0.1399203530298138,
197
- "grad_norm": 0.8150563836097717,
198
- "learning_rate": 0.000592242215278526,
199
- "loss": 4.8947,
200
  "step": 1300
201
  },
202
  {
203
- "epoch": 0.14530190506942203,
204
- "grad_norm": 0.8947958946228027,
205
- "learning_rate": 0.0005919189742484645,
206
- "loss": 4.8786,
207
  "step": 1350
208
  },
209
  {
210
- "epoch": 0.15068345710903025,
211
- "grad_norm": 1.3871676921844482,
212
- "learning_rate": 0.0005915957332184032,
213
- "loss": 4.8246,
214
  "step": 1400
215
  },
216
  {
217
- "epoch": 0.15606500914863847,
218
- "grad_norm": 0.9359856247901917,
219
- "learning_rate": 0.0005912724921883417,
220
- "loss": 4.8174,
221
  "step": 1450
222
  },
223
  {
224
- "epoch": 0.16144656118824668,
225
- "grad_norm": 1.100040078163147,
226
- "learning_rate": 0.0005909492511582803,
227
- "loss": 4.7943,
228
  "step": 1500
229
  },
230
  {
231
- "epoch": 0.1668281132278549,
232
- "grad_norm": 0.9592791795730591,
233
- "learning_rate": 0.0005906260101282189,
234
- "loss": 4.7722,
235
  "step": 1550
236
  },
237
  {
238
- "epoch": 0.17220966526746315,
239
- "grad_norm": 0.9244337677955627,
240
- "learning_rate": 0.0005903027690981575,
241
- "loss": 4.744,
242
  "step": 1600
243
  },
244
  {
245
- "epoch": 0.17759121730707136,
246
- "grad_norm": 1.0609703063964844,
247
- "learning_rate": 0.000589979528068096,
248
- "loss": 4.704,
249
  "step": 1650
250
  },
251
  {
252
- "epoch": 0.18297276934667958,
253
- "grad_norm": 1.3286961317062378,
254
- "learning_rate": 0.0005896562870380347,
255
- "loss": 4.6926,
256
  "step": 1700
257
  },
258
  {
259
- "epoch": 0.1883543213862878,
260
- "grad_norm": 0.9569286108016968,
261
- "learning_rate": 0.0005893330460079732,
262
- "loss": 4.6732,
263
  "step": 1750
264
  },
265
  {
266
- "epoch": 0.19373587342589602,
267
- "grad_norm": 0.9205374121665955,
268
- "learning_rate": 0.0005890098049779118,
269
- "loss": 4.6424,
270
  "step": 1800
271
  },
272
  {
273
- "epoch": 0.19911742546550426,
274
- "grad_norm": 1.1031231880187988,
275
- "learning_rate": 0.0005886865639478504,
276
- "loss": 4.6293,
277
  "step": 1850
278
  },
279
  {
280
- "epoch": 0.20449897750511248,
281
- "grad_norm": 1.0414140224456787,
282
- "learning_rate": 0.0005883633229177889,
283
- "loss": 4.6116,
284
  "step": 1900
285
  },
286
  {
287
- "epoch": 0.2098805295447207,
288
- "grad_norm": 0.8618186712265015,
289
- "learning_rate": 0.0005880400818877276,
290
- "loss": 4.5951,
291
  "step": 1950
292
  },
293
  {
294
- "epoch": 0.2152620815843289,
295
- "grad_norm": 0.8570399880409241,
296
- "learning_rate": 0.0005877168408576662,
297
- "loss": 4.5741,
298
  "step": 2000
299
  },
300
  {
301
- "epoch": 0.2152620815843289,
302
- "eval_accuracy": 0.27234704688513356,
303
- "eval_loss": 4.498940467834473,
304
- "eval_runtime": 184.347,
305
- "eval_samples_per_second": 97.702,
306
- "eval_steps_per_second": 6.108,
307
  "step": 2000
308
  },
309
  {
310
- "epoch": 0.22064363362393713,
311
- "grad_norm": 0.8821980357170105,
312
- "learning_rate": 0.0005873935998276048,
313
- "loss": 4.5753,
314
  "step": 2050
315
  },
316
  {
317
- "epoch": 0.22602518566354537,
318
- "grad_norm": 0.8663649559020996,
319
- "learning_rate": 0.0005870703587975433,
320
- "loss": 4.5247,
321
  "step": 2100
322
  },
323
  {
324
- "epoch": 0.2314067377031536,
325
- "grad_norm": 0.9565311670303345,
326
- "learning_rate": 0.0005867471177674818,
327
- "loss": 4.5213,
328
  "step": 2150
329
  },
330
  {
331
- "epoch": 0.2367882897427618,
332
- "grad_norm": 0.880970299243927,
333
- "learning_rate": 0.0005864238767374205,
334
- "loss": 4.5089,
335
  "step": 2200
336
  },
337
  {
338
- "epoch": 0.24216984178237003,
339
- "grad_norm": 0.9244751930236816,
340
- "learning_rate": 0.0005861006357073591,
341
- "loss": 4.4794,
342
  "step": 2250
343
  },
344
  {
345
- "epoch": 0.24755139382197827,
346
- "grad_norm": 0.824410617351532,
347
- "learning_rate": 0.0005857773946772977,
348
- "loss": 4.4594,
349
  "step": 2300
350
  },
351
  {
352
- "epoch": 0.2529329458615865,
353
- "grad_norm": 0.7963064908981323,
354
- "learning_rate": 0.0005854541536472362,
355
- "loss": 4.4382,
356
  "step": 2350
357
  },
358
  {
359
- "epoch": 0.2583144979011947,
360
- "grad_norm": 0.9127968549728394,
361
- "learning_rate": 0.0005851309126171749,
362
- "loss": 4.4464,
363
  "step": 2400
364
  },
365
  {
366
- "epoch": 0.2636960499408029,
367
- "grad_norm": 0.7743198275566101,
368
- "learning_rate": 0.0005848076715871134,
369
- "loss": 4.4167,
370
  "step": 2450
371
  },
372
  {
373
- "epoch": 0.26907760198041114,
374
- "grad_norm": 1.0491642951965332,
375
- "learning_rate": 0.000584484430557052,
376
- "loss": 4.4227,
377
  "step": 2500
378
  },
379
  {
380
- "epoch": 0.27445915402001936,
381
- "grad_norm": 0.9228188395500183,
382
- "learning_rate": 0.0005841611895269906,
383
- "loss": 4.4175,
384
  "step": 2550
385
  },
386
  {
387
- "epoch": 0.2798407060596276,
388
- "grad_norm": 0.837192177772522,
389
- "learning_rate": 0.0005838379484969291,
390
- "loss": 4.3804,
391
  "step": 2600
392
  },
393
  {
394
- "epoch": 0.2852222580992358,
395
- "grad_norm": 0.7594670653343201,
396
- "learning_rate": 0.0005835147074668678,
397
- "loss": 4.3766,
398
  "step": 2650
399
  },
400
  {
401
- "epoch": 0.29060381013884407,
402
- "grad_norm": 0.7400719523429871,
403
- "learning_rate": 0.0005831914664368063,
404
- "loss": 4.3444,
405
  "step": 2700
406
  },
407
  {
408
- "epoch": 0.2959853621784523,
409
- "grad_norm": 0.8986117243766785,
410
- "learning_rate": 0.0005828682254067449,
411
- "loss": 4.353,
412
  "step": 2750
413
  },
414
  {
415
- "epoch": 0.3013669142180605,
416
- "grad_norm": 0.8467971682548523,
417
- "learning_rate": 0.0005825449843766835,
418
- "loss": 4.3674,
419
  "step": 2800
420
  },
421
  {
422
- "epoch": 0.3067484662576687,
423
- "grad_norm": 0.6583292484283447,
424
- "learning_rate": 0.0005822217433466221,
425
- "loss": 4.356,
426
  "step": 2850
427
  },
428
  {
429
- "epoch": 0.31213001829727693,
430
- "grad_norm": 0.9095297455787659,
431
- "learning_rate": 0.0005818985023165607,
432
- "loss": 4.3273,
433
  "step": 2900
434
  },
435
  {
436
- "epoch": 0.31751157033688515,
437
- "grad_norm": 0.7011778354644775,
438
- "learning_rate": 0.0005815752612864992,
439
- "loss": 4.3155,
440
  "step": 2950
441
  },
442
  {
443
- "epoch": 0.32289312237649337,
444
- "grad_norm": 0.8573430776596069,
445
- "learning_rate": 0.0005812520202564378,
446
- "loss": 4.2905,
447
  "step": 3000
448
  },
449
  {
450
- "epoch": 0.32289312237649337,
451
- "eval_accuracy": 0.2991464989703509,
452
- "eval_loss": 4.2332539558410645,
453
- "eval_runtime": 184.5016,
454
- "eval_samples_per_second": 97.62,
455
- "eval_steps_per_second": 6.103,
456
  "step": 3000
457
  },
458
  {
459
- "epoch": 0.3282746744161016,
460
- "grad_norm": 0.6798554062843323,
461
- "learning_rate": 0.0005809287792263764,
462
- "loss": 4.3078,
463
  "step": 3050
464
  },
465
  {
466
- "epoch": 0.3336562264557098,
467
- "grad_norm": 1.0313102006912231,
468
- "learning_rate": 0.0005806055381963151,
469
- "loss": 4.2915,
470
  "step": 3100
471
  },
472
  {
473
- "epoch": 0.3390377784953181,
474
- "grad_norm": 0.7952976822853088,
475
- "learning_rate": 0.0005802822971662536,
476
- "loss": 4.267,
477
  "step": 3150
478
  },
479
  {
480
- "epoch": 0.3444193305349263,
481
- "grad_norm": 0.6104669570922852,
482
- "learning_rate": 0.0005799590561361922,
483
- "loss": 4.2788,
484
  "step": 3200
485
  },
486
  {
487
- "epoch": 0.3498008825745345,
488
- "grad_norm": 0.7965825796127319,
489
- "learning_rate": 0.0005796358151061307,
490
- "loss": 4.2722,
491
  "step": 3250
492
  },
493
  {
494
- "epoch": 0.35518243461414273,
495
- "grad_norm": 0.6662214398384094,
496
- "learning_rate": 0.0005793125740760694,
497
- "loss": 4.263,
498
  "step": 3300
499
  },
500
  {
501
- "epoch": 0.36056398665375095,
502
- "grad_norm": 0.7256644368171692,
503
- "learning_rate": 0.0005789893330460079,
504
- "loss": 4.2343,
505
  "step": 3350
506
  },
507
  {
508
- "epoch": 0.36594553869335916,
509
- "grad_norm": 0.6321602463722229,
510
- "learning_rate": 0.0005786660920159465,
511
- "loss": 4.2483,
512
  "step": 3400
513
  },
514
  {
515
- "epoch": 0.3713270907329674,
516
- "grad_norm": 0.6723715662956238,
517
- "learning_rate": 0.0005783428509858851,
518
- "loss": 4.2271,
519
  "step": 3450
520
  },
521
  {
522
- "epoch": 0.3767086427725756,
523
- "grad_norm": 0.757610023021698,
524
- "learning_rate": 0.0005780196099558237,
525
- "loss": 4.2201,
526
  "step": 3500
527
  },
528
  {
529
- "epoch": 0.3820901948121838,
530
- "grad_norm": 0.6822935938835144,
531
- "learning_rate": 0.0005776963689257623,
532
- "loss": 4.2124,
533
  "step": 3550
534
  },
535
  {
536
- "epoch": 0.38747174685179203,
537
- "grad_norm": 0.6186670660972595,
538
- "learning_rate": 0.0005773731278957008,
539
- "loss": 4.2126,
540
  "step": 3600
541
  },
542
  {
543
- "epoch": 0.3928532988914003,
544
- "grad_norm": 0.7107723355293274,
545
- "learning_rate": 0.0005770498868656394,
546
- "loss": 4.1955,
547
  "step": 3650
548
  },
549
  {
550
- "epoch": 0.3982348509310085,
551
- "grad_norm": 0.6134611964225769,
552
- "learning_rate": 0.000576726645835578,
553
- "loss": 4.187,
554
  "step": 3700
555
  },
556
  {
557
- "epoch": 0.40361640297061674,
558
- "grad_norm": 0.7130006551742554,
559
- "learning_rate": 0.0005764034048055167,
560
- "loss": 4.2103,
561
  "step": 3750
562
  },
563
  {
564
- "epoch": 0.40899795501022496,
565
- "grad_norm": 0.7081485986709595,
566
- "learning_rate": 0.0005760801637754552,
567
- "loss": 4.1975,
568
  "step": 3800
569
  },
570
  {
571
- "epoch": 0.4143795070498332,
572
- "grad_norm": 0.8055673241615295,
573
- "learning_rate": 0.0005757569227453937,
574
- "loss": 4.1592,
575
  "step": 3850
576
  },
577
  {
578
- "epoch": 0.4197610590894414,
579
- "grad_norm": 0.9049285054206848,
580
- "learning_rate": 0.0005754336817153324,
581
- "loss": 4.155,
582
  "step": 3900
583
  },
584
  {
585
- "epoch": 0.4251426111290496,
586
- "grad_norm": 0.7098596096038818,
587
- "learning_rate": 0.0005751104406852709,
588
- "loss": 4.1541,
589
  "step": 3950
590
  },
591
  {
592
- "epoch": 0.4305241631686578,
593
- "grad_norm": 0.6744562387466431,
594
- "learning_rate": 0.0005747871996552096,
595
- "loss": 4.1453,
596
  "step": 4000
597
  },
598
  {
599
- "epoch": 0.4305241631686578,
600
- "eval_accuracy": 0.3129022805263276,
601
- "eval_loss": 4.090099811553955,
602
- "eval_runtime": 184.4208,
603
- "eval_samples_per_second": 97.663,
604
- "eval_steps_per_second": 6.106,
605
  "step": 4000
606
  },
607
  {
608
- "epoch": 0.43590571520826604,
609
- "grad_norm": 0.6521379947662354,
610
- "learning_rate": 0.0005744639586251481,
611
- "loss": 4.1527,
612
  "step": 4050
613
  },
614
  {
615
- "epoch": 0.44128726724787426,
616
- "grad_norm": 0.6583467721939087,
617
- "learning_rate": 0.0005741407175950867,
618
- "loss": 4.1572,
619
  "step": 4100
620
  },
621
  {
622
- "epoch": 0.44666881928748253,
623
- "grad_norm": 0.6938270330429077,
624
- "learning_rate": 0.0005738174765650253,
625
- "loss": 4.1557,
626
  "step": 4150
627
  },
628
  {
629
- "epoch": 0.45205037132709075,
630
- "grad_norm": 0.6435059309005737,
631
- "learning_rate": 0.0005734942355349638,
632
- "loss": 4.1395,
633
  "step": 4200
634
  },
635
  {
636
- "epoch": 0.45743192336669897,
637
- "grad_norm": 0.6761099696159363,
638
- "learning_rate": 0.0005731709945049025,
639
- "loss": 4.1328,
640
  "step": 4250
641
  },
642
  {
643
- "epoch": 0.4628134754063072,
644
- "grad_norm": 0.6372486352920532,
645
- "learning_rate": 0.000572847753474841,
646
- "loss": 4.138,
647
  "step": 4300
648
  },
649
  {
650
- "epoch": 0.4681950274459154,
651
- "grad_norm": 0.7407318949699402,
652
- "learning_rate": 0.0005725245124447796,
653
- "loss": 4.1446,
654
  "step": 4350
655
  },
656
  {
657
- "epoch": 0.4735765794855236,
658
- "grad_norm": 0.6796604990959167,
659
- "learning_rate": 0.0005722012714147182,
660
- "loss": 4.1164,
661
  "step": 4400
662
  },
663
  {
664
- "epoch": 0.47895813152513184,
665
- "grad_norm": 0.6598204970359802,
666
- "learning_rate": 0.0005718780303846568,
667
- "loss": 4.1197,
668
  "step": 4450
669
  },
670
  {
671
- "epoch": 0.48433968356474005,
672
- "grad_norm": 0.6131014823913574,
673
- "learning_rate": 0.0005715547893545953,
674
- "loss": 4.106,
675
  "step": 4500
676
  },
677
  {
678
- "epoch": 0.48972123560434827,
679
- "grad_norm": 0.6856024265289307,
680
- "learning_rate": 0.000571231548324534,
681
- "loss": 4.0877,
682
  "step": 4550
683
  },
684
  {
685
- "epoch": 0.49510278764395654,
686
- "grad_norm": 0.6875824928283691,
687
- "learning_rate": 0.0005709083072944725,
688
- "loss": 4.0935,
689
  "step": 4600
690
  },
691
  {
692
- "epoch": 0.5004843396835648,
693
- "grad_norm": 0.5760625004768372,
694
- "learning_rate": 0.0005705850662644111,
695
- "loss": 4.1158,
696
  "step": 4650
697
  },
698
  {
699
- "epoch": 0.505865891723173,
700
- "grad_norm": 0.8604832291603088,
701
- "learning_rate": 0.0005702618252343497,
702
- "loss": 4.0911,
703
  "step": 4700
704
  },
705
  {
706
- "epoch": 0.5112474437627812,
707
- "grad_norm": 0.6754705905914307,
708
- "learning_rate": 0.0005699385842042882,
709
- "loss": 4.1016,
710
  "step": 4750
711
  },
712
  {
713
- "epoch": 0.5166289958023894,
714
- "grad_norm": 0.5712053775787354,
715
- "learning_rate": 0.0005696153431742269,
716
- "loss": 4.0716,
717
  "step": 4800
718
  },
719
  {
720
- "epoch": 0.5220105478419976,
721
- "grad_norm": 0.6515219807624817,
722
- "learning_rate": 0.0005692921021441655,
723
- "loss": 4.0726,
724
  "step": 4850
725
  },
726
  {
727
- "epoch": 0.5273920998816058,
728
- "grad_norm": 0.6000773906707764,
729
- "learning_rate": 0.0005689688611141041,
730
- "loss": 4.0808,
731
  "step": 4900
732
  },
733
  {
734
- "epoch": 0.5327736519212141,
735
- "grad_norm": 0.6588471531867981,
736
- "learning_rate": 0.0005686456200840426,
737
- "loss": 4.0791,
738
  "step": 4950
739
  },
740
  {
741
- "epoch": 0.5381552039608223,
742
- "grad_norm": 0.6547403931617737,
743
- "learning_rate": 0.0005683223790539811,
744
- "loss": 4.0703,
745
  "step": 5000
746
  },
747
  {
748
- "epoch": 0.5381552039608223,
749
- "eval_accuracy": 0.3209675843887965,
750
- "eval_loss": 3.9936656951904297,
751
- "eval_runtime": 184.5678,
752
- "eval_samples_per_second": 97.585,
753
- "eval_steps_per_second": 6.101,
754
  "step": 5000
755
  },
756
  {
757
- "epoch": 0.5435367560004305,
758
- "grad_norm": 0.649213969707489,
759
- "learning_rate": 0.0005679991380239198,
760
- "loss": 4.0619,
761
  "step": 5050
762
  },
763
  {
764
- "epoch": 0.5489183080400387,
765
- "grad_norm": 0.6468812823295593,
766
- "learning_rate": 0.0005676758969938584,
767
- "loss": 4.0439,
768
  "step": 5100
769
  },
770
  {
771
- "epoch": 0.5542998600796469,
772
- "grad_norm": 0.6261686682701111,
773
- "learning_rate": 0.000567352655963797,
774
- "loss": 4.0405,
775
  "step": 5150
776
  },
777
  {
778
- "epoch": 0.5596814121192552,
779
- "grad_norm": 0.7181739211082458,
780
- "learning_rate": 0.0005670294149337355,
781
- "loss": 4.0494,
782
  "step": 5200
783
  },
784
  {
785
- "epoch": 0.5650629641588634,
786
- "grad_norm": 0.5509181022644043,
787
- "learning_rate": 0.0005667061739036742,
788
- "loss": 4.0364,
789
  "step": 5250
790
  },
791
  {
792
- "epoch": 0.5704445161984716,
793
- "grad_norm": 0.7940995693206787,
794
- "learning_rate": 0.0005663829328736127,
795
- "loss": 4.0348,
796
  "step": 5300
797
  },
798
  {
799
- "epoch": 0.5758260682380799,
800
- "grad_norm": 0.6015467047691345,
801
- "learning_rate": 0.0005660596918435512,
802
- "loss": 4.0591,
803
  "step": 5350
804
  },
805
  {
806
- "epoch": 0.5812076202776881,
807
- "grad_norm": 0.6166031360626221,
808
- "learning_rate": 0.0005657364508134899,
809
- "loss": 4.0306,
810
  "step": 5400
811
  },
812
  {
813
- "epoch": 0.5865891723172963,
814
- "grad_norm": 0.7661638259887695,
815
- "learning_rate": 0.0005654132097834284,
816
- "loss": 4.0307,
817
  "step": 5450
818
  },
819
  {
820
- "epoch": 0.5919707243569046,
821
- "grad_norm": 0.6555174589157104,
822
- "learning_rate": 0.0005650899687533671,
823
- "loss": 4.024,
824
  "step": 5500
825
  },
826
  {
827
- "epoch": 0.5973522763965128,
828
- "grad_norm": 0.595856785774231,
829
- "learning_rate": 0.0005647667277233056,
830
- "loss": 4.0182,
831
  "step": 5550
832
  },
833
  {
834
- "epoch": 0.602733828436121,
835
- "grad_norm": 0.6452604532241821,
836
- "learning_rate": 0.0005644434866932442,
837
  "loss": 4.0241,
838
  "step": 5600
839
  },
840
  {
841
- "epoch": 0.6081153804757292,
842
- "grad_norm": 0.6375003457069397,
843
- "learning_rate": 0.0005641202456631828,
844
- "loss": 3.9957,
845
  "step": 5650
846
  },
847
  {
848
- "epoch": 0.6134969325153374,
849
- "grad_norm": 0.6434177160263062,
850
- "learning_rate": 0.0005637970046331214,
851
- "loss": 3.9934,
852
  "step": 5700
853
  },
854
  {
855
- "epoch": 0.6188784845549457,
856
- "grad_norm": 0.6278543472290039,
857
- "learning_rate": 0.00056347376360306,
858
- "loss": 4.0028,
859
  "step": 5750
860
  },
861
  {
862
- "epoch": 0.6242600365945539,
863
- "grad_norm": 0.6161077618598938,
864
- "learning_rate": 0.0005631505225729985,
865
- "loss": 4.0068,
866
  "step": 5800
867
  },
868
  {
869
- "epoch": 0.6296415886341621,
870
- "grad_norm": 0.5981731414794922,
871
- "learning_rate": 0.0005628272815429371,
872
- "loss": 4.0076,
873
  "step": 5850
874
  },
875
  {
876
- "epoch": 0.6350231406737703,
877
- "grad_norm": 0.5853173136711121,
878
- "learning_rate": 0.0005625040405128757,
879
- "loss": 4.0131,
880
  "step": 5900
881
  },
882
  {
883
- "epoch": 0.6404046927133785,
884
- "grad_norm": 0.7181535959243774,
885
- "learning_rate": 0.0005621807994828143,
886
- "loss": 3.9811,
887
  "step": 5950
888
  },
889
  {
890
- "epoch": 0.6457862447529867,
891
- "grad_norm": 0.7647160887718201,
892
- "learning_rate": 0.0005618575584527529,
893
- "loss": 3.9912,
894
  "step": 6000
895
  },
896
  {
897
- "epoch": 0.6457862447529867,
898
- "eval_accuracy": 0.32738592777777353,
899
- "eval_loss": 3.926551342010498,
900
- "eval_runtime": 184.2978,
901
- "eval_samples_per_second": 97.728,
902
- "eval_steps_per_second": 6.11,
903
  "step": 6000
904
  },
905
  {
906
- "epoch": 0.651167796792595,
907
- "grad_norm": 0.6108874082565308,
908
- "learning_rate": 0.0005615343174226915,
909
- "loss": 3.9942,
910
  "step": 6050
911
  },
912
  {
913
- "epoch": 0.6565493488322032,
914
- "grad_norm": 0.539562463760376,
915
- "learning_rate": 0.00056121107639263,
916
- "loss": 3.9947,
917
  "step": 6100
918
  },
919
  {
920
- "epoch": 0.6619309008718114,
921
- "grad_norm": 0.6353080868721008,
922
- "learning_rate": 0.0005608878353625687,
923
- "loss": 3.9713,
924
  "step": 6150
925
  },
926
  {
927
- "epoch": 0.6673124529114196,
928
- "grad_norm": 0.6244759559631348,
929
- "learning_rate": 0.0005605645943325072,
930
- "loss": 3.9771,
931
  "step": 6200
932
  },
933
  {
934
- "epoch": 0.6726940049510278,
935
- "grad_norm": 0.5772109627723694,
936
- "learning_rate": 0.0005602413533024458,
937
- "loss": 3.9849,
938
  "step": 6250
939
  },
940
  {
941
- "epoch": 0.6780755569906362,
942
- "grad_norm": 0.5281596183776855,
943
- "learning_rate": 0.0005599181122723844,
944
- "loss": 3.9597,
945
  "step": 6300
946
  },
947
  {
948
- "epoch": 0.6834571090302444,
949
- "grad_norm": 0.5710061192512512,
950
- "learning_rate": 0.000559594871242323,
951
- "loss": 3.9619,
952
  "step": 6350
953
  },
954
  {
955
- "epoch": 0.6888386610698526,
956
- "grad_norm": 0.605676531791687,
957
- "learning_rate": 0.0005592716302122616,
958
- "loss": 3.9618,
959
  "step": 6400
960
  },
961
  {
962
- "epoch": 0.6942202131094608,
963
- "grad_norm": 0.6223154067993164,
964
- "learning_rate": 0.0005589483891822001,
965
- "loss": 3.9637,
966
  "step": 6450
967
  },
968
  {
969
- "epoch": 0.699601765149069,
970
- "grad_norm": 0.776096761226654,
971
- "learning_rate": 0.0005586251481521387,
972
- "loss": 3.9487,
973
  "step": 6500
974
  },
975
  {
976
- "epoch": 0.7049833171886772,
977
- "grad_norm": 0.5951947569847107,
978
- "learning_rate": 0.0005583019071220773,
979
- "loss": 3.9713,
980
  "step": 6550
981
  },
982
  {
983
- "epoch": 0.7103648692282855,
984
- "grad_norm": 0.6330660581588745,
985
- "learning_rate": 0.000557978666092016,
986
- "loss": 3.9564,
987
  "step": 6600
988
  },
989
  {
990
- "epoch": 0.7157464212678937,
991
- "grad_norm": 0.6448531746864319,
992
- "learning_rate": 0.0005576554250619545,
993
- "loss": 3.9346,
994
  "step": 6650
995
  },
996
  {
997
- "epoch": 0.7211279733075019,
998
- "grad_norm": 0.6924978494644165,
999
- "learning_rate": 0.000557332184031893,
1000
- "loss": 3.9377,
1001
  "step": 6700
1002
  },
1003
  {
1004
- "epoch": 0.7265095253471101,
1005
- "grad_norm": 0.7174770832061768,
1006
- "learning_rate": 0.0005570089430018317,
1007
- "loss": 3.9641,
1008
  "step": 6750
1009
  },
1010
  {
1011
- "epoch": 0.7318910773867183,
1012
- "grad_norm": 0.5341421365737915,
1013
- "learning_rate": 0.0005566857019717702,
1014
- "loss": 3.9428,
1015
  "step": 6800
1016
  },
1017
  {
1018
- "epoch": 0.7372726294263265,
1019
- "grad_norm": 0.6674601435661316,
1020
- "learning_rate": 0.0005563624609417089,
1021
- "loss": 3.9446,
1022
  "step": 6850
1023
  },
1024
  {
1025
- "epoch": 0.7426541814659348,
1026
- "grad_norm": 0.6375956535339355,
1027
- "learning_rate": 0.0005560392199116474,
1028
- "loss": 3.9471,
1029
  "step": 6900
1030
  },
1031
  {
1032
- "epoch": 0.748035733505543,
1033
- "grad_norm": 0.6670394539833069,
1034
- "learning_rate": 0.000555715978881586,
1035
- "loss": 3.9303,
1036
  "step": 6950
1037
  },
1038
  {
1039
- "epoch": 0.7534172855451512,
1040
- "grad_norm": 0.7189449667930603,
1041
- "learning_rate": 0.0005553927378515246,
1042
- "loss": 3.9195,
1043
  "step": 7000
1044
  },
1045
  {
1046
- "epoch": 0.7534172855451512,
1047
- "eval_accuracy": 0.3326128922518648,
1048
- "eval_loss": 3.872013568878174,
1049
- "eval_runtime": 184.7751,
1050
- "eval_samples_per_second": 97.475,
1051
- "eval_steps_per_second": 6.094,
1052
  "step": 7000
1053
  },
1054
  {
1055
- "epoch": 0.7587988375847594,
1056
- "grad_norm": 0.6056647896766663,
1057
- "learning_rate": 0.0005550694968214631,
1058
- "loss": 3.9423,
1059
  "step": 7050
1060
  },
1061
  {
1062
- "epoch": 0.7641803896243676,
1063
- "grad_norm": 0.6418593525886536,
1064
- "learning_rate": 0.0005547462557914018,
1065
- "loss": 3.9169,
1066
  "step": 7100
1067
  },
1068
  {
1069
- "epoch": 0.7695619416639758,
1070
- "grad_norm": 0.5867264866828918,
1071
- "learning_rate": 0.0005544230147613403,
1072
- "loss": 3.9172,
1073
  "step": 7150
1074
  },
1075
  {
1076
- "epoch": 0.7749434937035841,
1077
- "grad_norm": 0.5679484009742737,
1078
- "learning_rate": 0.0005540997737312789,
1079
- "loss": 3.9287,
1080
  "step": 7200
1081
  },
1082
  {
1083
- "epoch": 0.7803250457431924,
1084
- "grad_norm": 0.6049643754959106,
1085
- "learning_rate": 0.0005537829975218188,
1086
- "loss": 3.9073,
1087
  "step": 7250
1088
  },
1089
  {
1090
- "epoch": 0.7857065977828006,
1091
- "grad_norm": 0.5454378724098206,
1092
- "learning_rate": 0.0005534597564917573,
1093
- "loss": 3.931,
1094
  "step": 7300
1095
  },
1096
  {
1097
- "epoch": 0.7910881498224088,
1098
- "grad_norm": 0.6361756324768066,
1099
- "learning_rate": 0.0005531365154616959,
1100
- "loss": 3.9185,
1101
  "step": 7350
1102
  },
1103
  {
1104
- "epoch": 0.796469701862017,
1105
- "grad_norm": 0.5948432087898254,
1106
- "learning_rate": 0.0005528132744316344,
1107
- "loss": 3.9223,
1108
  "step": 7400
1109
  },
1110
  {
1111
- "epoch": 0.8018512539016253,
1112
- "grad_norm": 0.6710109114646912,
1113
- "learning_rate": 0.0005524900334015731,
1114
- "loss": 3.9103,
1115
  "step": 7450
1116
  },
1117
  {
1118
- "epoch": 0.8072328059412335,
1119
- "grad_norm": 0.6390253305435181,
1120
- "learning_rate": 0.0005521667923715117,
1121
- "loss": 3.8857,
1122
  "step": 7500
1123
  },
1124
  {
1125
- "epoch": 0.8126143579808417,
1126
- "grad_norm": 0.6726111173629761,
1127
- "learning_rate": 0.0005518435513414502,
1128
- "loss": 3.8906,
1129
  "step": 7550
1130
  },
1131
  {
1132
- "epoch": 0.8179959100204499,
1133
- "grad_norm": 0.5953141450881958,
1134
- "learning_rate": 0.0005515203103113888,
1135
- "loss": 3.911,
1136
  "step": 7600
1137
  },
1138
  {
1139
- "epoch": 0.8233774620600581,
1140
- "grad_norm": 0.6084983348846436,
1141
- "learning_rate": 0.0005511970692813274,
1142
- "loss": 3.8937,
1143
  "step": 7650
1144
  },
1145
  {
1146
- "epoch": 0.8287590140996663,
1147
- "grad_norm": 0.6089417338371277,
1148
- "learning_rate": 0.000550873828251266,
1149
- "loss": 3.907,
1150
  "step": 7700
1151
  },
1152
  {
1153
- "epoch": 0.8341405661392746,
1154
- "grad_norm": 0.5650714635848999,
1155
- "learning_rate": 0.0005505505872212045,
1156
- "loss": 3.9024,
1157
  "step": 7750
1158
  },
1159
  {
1160
- "epoch": 0.8395221181788828,
1161
- "grad_norm": 0.556892991065979,
1162
- "learning_rate": 0.0005502273461911432,
1163
- "loss": 3.8743,
1164
  "step": 7800
1165
  },
1166
  {
1167
- "epoch": 0.844903670218491,
1168
- "grad_norm": 0.5664134621620178,
1169
- "learning_rate": 0.0005499041051610817,
1170
- "loss": 3.913,
1171
  "step": 7850
1172
  },
1173
  {
1174
- "epoch": 0.8502852222580992,
1175
- "grad_norm": 0.6096441745758057,
1176
- "learning_rate": 0.0005495808641310204,
1177
- "loss": 3.879,
1178
  "step": 7900
1179
  },
1180
  {
1181
- "epoch": 0.8556667742977074,
1182
- "grad_norm": 0.5594208240509033,
1183
- "learning_rate": 0.0005492576231009589,
1184
- "loss": 3.8784,
1185
  "step": 7950
1186
  },
1187
  {
1188
- "epoch": 0.8610483263373157,
1189
- "grad_norm": 0.538360595703125,
1190
- "learning_rate": 0.0005489343820708974,
1191
- "loss": 3.8902,
1192
  "step": 8000
1193
  },
1194
  {
1195
- "epoch": 0.8610483263373157,
1196
- "eval_accuracy": 0.3369882353912661,
1197
- "eval_loss": 3.822113037109375,
1198
- "eval_runtime": 184.4529,
1199
- "eval_samples_per_second": 97.646,
1200
- "eval_steps_per_second": 6.105,
1201
  "step": 8000
1202
  },
1203
  {
1204
- "epoch": 0.8664298783769239,
1205
- "grad_norm": 0.6012279987335205,
1206
- "learning_rate": 0.0005486111410408361,
1207
- "loss": 3.8879,
1208
  "step": 8050
1209
  },
1210
  {
1211
- "epoch": 0.8718114304165321,
1212
- "grad_norm": 0.5948436260223389,
1213
- "learning_rate": 0.0005482879000107746,
1214
- "loss": 3.8825,
1215
  "step": 8100
1216
  },
1217
  {
1218
- "epoch": 0.8771929824561403,
1219
- "grad_norm": 0.6777280569076538,
1220
- "learning_rate": 0.0005479646589807133,
1221
- "loss": 3.8714,
1222
  "step": 8150
1223
  },
1224
  {
1225
- "epoch": 0.8825745344957485,
1226
- "grad_norm": 0.5959411859512329,
1227
- "learning_rate": 0.0005476414179506518,
1228
- "loss": 3.8722,
1229
  "step": 8200
1230
  },
1231
  {
1232
- "epoch": 0.8879560865353568,
1233
- "grad_norm": 0.5966766476631165,
1234
- "learning_rate": 0.0005473181769205904,
1235
- "loss": 3.877,
1236
  "step": 8250
1237
  },
1238
  {
1239
- "epoch": 0.8933376385749651,
1240
- "grad_norm": 0.5824183225631714,
1241
- "learning_rate": 0.000546994935890529,
1242
- "loss": 3.8932,
1243
  "step": 8300
1244
  },
1245
  {
1246
- "epoch": 0.8987191906145733,
1247
- "grad_norm": 0.565321147441864,
1248
- "learning_rate": 0.0005466716948604677,
1249
- "loss": 3.8744,
1250
  "step": 8350
1251
  },
1252
  {
1253
- "epoch": 0.9041007426541815,
1254
- "grad_norm": 0.6127881407737732,
1255
- "learning_rate": 0.0005463484538304062,
1256
- "loss": 3.8627,
1257
  "step": 8400
1258
  },
1259
  {
1260
- "epoch": 0.9094822946937897,
1261
- "grad_norm": 0.5455643534660339,
1262
- "learning_rate": 0.0005460252128003447,
1263
- "loss": 3.8711,
1264
  "step": 8450
1265
  },
1266
  {
1267
- "epoch": 0.9148638467333979,
1268
- "grad_norm": 0.5501160621643066,
1269
- "learning_rate": 0.0005457019717702833,
1270
- "loss": 3.8719,
1271
  "step": 8500
1272
  },
1273
  {
1274
- "epoch": 0.9202453987730062,
1275
- "grad_norm": 0.5473130941390991,
1276
- "learning_rate": 0.0005453787307402219,
1277
- "loss": 3.8563,
1278
  "step": 8550
1279
  },
1280
  {
1281
- "epoch": 0.9256269508126144,
1282
- "grad_norm": 0.5105751156806946,
1283
- "learning_rate": 0.0005450554897101605,
1284
  "loss": 3.8604,
1285
  "step": 8600
1286
  },
1287
  {
1288
- "epoch": 0.9310085028522226,
1289
- "grad_norm": 0.6234155893325806,
1290
- "learning_rate": 0.0005447322486800991,
1291
- "loss": 3.8682,
1292
  "step": 8650
1293
  },
1294
  {
1295
- "epoch": 0.9363900548918308,
1296
- "grad_norm": 0.6187501549720764,
1297
- "learning_rate": 0.0005444090076500377,
1298
- "loss": 3.8721,
1299
  "step": 8700
1300
  },
1301
  {
1302
- "epoch": 0.941771606931439,
1303
- "grad_norm": 0.563278317451477,
1304
- "learning_rate": 0.0005440857666199763,
1305
- "loss": 3.8535,
1306
  "step": 8750
1307
  },
1308
  {
1309
- "epoch": 0.9471531589710472,
1310
- "grad_norm": 0.5937185883522034,
1311
- "learning_rate": 0.0005437625255899148,
1312
- "loss": 3.868,
1313
  "step": 8800
1314
  },
1315
  {
1316
- "epoch": 0.9525347110106555,
1317
- "grad_norm": 0.592932403087616,
1318
- "learning_rate": 0.0005434392845598534,
1319
- "loss": 3.8573,
1320
  "step": 8850
1321
  },
1322
  {
1323
- "epoch": 0.9579162630502637,
1324
- "grad_norm": 0.6201167106628418,
1325
- "learning_rate": 0.000543116043529792,
1326
- "loss": 3.8481,
1327
  "step": 8900
1328
  },
1329
  {
1330
- "epoch": 0.9632978150898719,
1331
- "grad_norm": 0.554725706577301,
1332
- "learning_rate": 0.0005427928024997306,
1333
- "loss": 3.8541,
1334
  "step": 8950
1335
  },
1336
  {
1337
- "epoch": 0.9686793671294801,
1338
- "grad_norm": 0.5520325899124146,
1339
- "learning_rate": 0.0005424695614696692,
1340
- "loss": 3.85,
1341
  "step": 9000
1342
  },
1343
  {
1344
- "epoch": 0.9686793671294801,
1345
- "eval_accuracy": 0.3408027123237691,
1346
- "eval_loss": 3.7828752994537354,
1347
- "eval_runtime": 184.162,
1348
- "eval_samples_per_second": 97.8,
1349
- "eval_steps_per_second": 6.114,
1350
  "step": 9000
1351
  },
1352
  {
1353
- "epoch": 0.9740609191690883,
1354
- "grad_norm": 0.5292057394981384,
1355
- "learning_rate": 0.0005421463204396078,
1356
- "loss": 3.8381,
1357
  "step": 9050
1358
  },
1359
  {
1360
- "epoch": 0.9794424712086965,
1361
- "grad_norm": 0.5707278847694397,
1362
- "learning_rate": 0.0005418230794095463,
1363
- "loss": 3.8459,
1364
  "step": 9100
1365
  },
1366
  {
1367
- "epoch": 0.9848240232483048,
1368
- "grad_norm": 0.5048800110816956,
1369
- "learning_rate": 0.000541499838379485,
1370
- "loss": 3.8483,
1371
  "step": 9150
1372
  },
1373
  {
1374
- "epoch": 0.9902055752879131,
1375
- "grad_norm": 0.6456687450408936,
1376
- "learning_rate": 0.0005411765973494235,
1377
- "loss": 3.8333,
1378
  "step": 9200
1379
  },
1380
  {
1381
- "epoch": 0.9955871273275213,
1382
- "grad_norm": 0.6426123976707458,
1383
- "learning_rate": 0.0005408533563193621,
1384
- "loss": 3.8645,
1385
  "step": 9250
1386
  },
1387
  {
1388
- "epoch": 1.0009686793671295,
1389
- "grad_norm": 0.6812326908111572,
1390
- "learning_rate": 0.0005405301152893007,
1391
- "loss": 3.8295,
1392
  "step": 9300
1393
  },
1394
  {
1395
- "epoch": 1.0063502314067376,
1396
- "grad_norm": 0.5232545733451843,
1397
- "learning_rate": 0.0005402068742592392,
1398
- "loss": 3.7647,
1399
  "step": 9350
1400
  },
1401
  {
1402
- "epoch": 1.011731783446346,
1403
- "grad_norm": 0.5876726508140564,
1404
- "learning_rate": 0.0005398836332291779,
1405
- "loss": 3.7793,
1406
  "step": 9400
1407
  },
1408
  {
1409
- "epoch": 1.017113335485954,
1410
- "grad_norm": 0.5287936925888062,
1411
- "learning_rate": 0.0005395603921991164,
1412
- "loss": 3.7686,
1413
  "step": 9450
1414
  },
1415
  {
1416
- "epoch": 1.0224948875255624,
1417
- "grad_norm": 0.5465691089630127,
1418
- "learning_rate": 0.0005392436159896562,
1419
- "loss": 3.7749,
1420
  "step": 9500
1421
  },
1422
  {
1423
- "epoch": 1.0278764395651705,
1424
- "grad_norm": 0.5962051749229431,
1425
- "learning_rate": 0.0005389203749595948,
1426
- "loss": 3.7751,
1427
  "step": 9550
1428
  },
1429
  {
1430
- "epoch": 1.0332579916047788,
1431
- "grad_norm": 0.543983519077301,
1432
- "learning_rate": 0.0005385971339295334,
1433
- "loss": 3.7671,
1434
  "step": 9600
1435
  },
1436
  {
1437
- "epoch": 1.0386395436443872,
1438
- "grad_norm": 0.5294556021690369,
1439
- "learning_rate": 0.000538273892899472,
1440
- "loss": 3.7737,
1441
  "step": 9650
1442
  },
1443
  {
1444
- "epoch": 1.0440210956839953,
1445
- "grad_norm": 0.6221946477890015,
1446
- "learning_rate": 0.0005379506518694106,
1447
- "loss": 3.7698,
1448
  "step": 9700
1449
  },
1450
  {
1451
- "epoch": 1.0494026477236036,
1452
- "grad_norm": 0.5739998817443848,
1453
- "learning_rate": 0.0005376274108393491,
1454
  "loss": 3.7579,
1455
  "step": 9750
1456
  },
1457
  {
1458
- "epoch": 1.0547841997632117,
1459
- "grad_norm": 0.6448442339897156,
1460
- "learning_rate": 0.0005373041698092877,
1461
- "loss": 3.779,
1462
  "step": 9800
1463
  },
1464
  {
1465
- "epoch": 1.06016575180282,
1466
- "grad_norm": 0.5370635986328125,
1467
- "learning_rate": 0.0005369809287792263,
1468
- "loss": 3.7778,
1469
  "step": 9850
1470
  },
1471
  {
1472
- "epoch": 1.0655473038424281,
1473
- "grad_norm": 0.5538362860679626,
1474
- "learning_rate": 0.000536657687749165,
1475
- "loss": 3.7705,
1476
  "step": 9900
1477
  },
1478
  {
1479
- "epoch": 1.0709288558820365,
1480
- "grad_norm": 0.5770606994628906,
1481
- "learning_rate": 0.0005363344467191035,
1482
- "loss": 3.7701,
1483
  "step": 9950
1484
  },
1485
  {
1486
- "epoch": 1.0763104079216446,
1487
- "grad_norm": 0.5529940128326416,
1488
- "learning_rate": 0.000536011205689042,
1489
- "loss": 3.768,
1490
  "step": 10000
1491
  },
1492
  {
1493
- "epoch": 1.0763104079216446,
1494
- "eval_accuracy": 0.3442798220396081,
1495
- "eval_loss": 3.7558341026306152,
1496
- "eval_runtime": 184.4409,
1497
- "eval_samples_per_second": 97.652,
1498
- "eval_steps_per_second": 6.105,
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": 8.3609792151552e+16,
1520
  "train_batch_size": 32,
1521
  "trial_name": null,
1522
  "trial_params": null
 
1
  {
2
+ "best_metric": 3.7509524822235107,
3
+ "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_low_10_1208/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": 1.6905114650726318,
14
  "learning_rate": 0.0003,
15
+ "loss": 8.7932,
16
  "step": 50
17
  },
18
  {
19
+ "epoch": 0.01078167115902965,
20
+ "grad_norm": 1.417388677597046,
21
  "learning_rate": 0.0006,
22
+ "loss": 6.993,
23
  "step": 100
24
  },
25
  {
26
+ "epoch": 0.016172506738544475,
27
+ "grad_norm": 1.0006530284881592,
28
+ "learning_rate": 0.0005996762007555315,
29
+ "loss": 6.5351,
30
  "step": 150
31
  },
32
  {
33
+ "epoch": 0.0215633423180593,
34
+ "grad_norm": 1.0720206499099731,
35
+ "learning_rate": 0.000599352401511063,
36
+ "loss": 6.263,
37
  "step": 200
38
  },
39
  {
40
+ "epoch": 0.026954177897574125,
41
+ "grad_norm": 1.171477198600769,
42
+ "learning_rate": 0.0005990286022665946,
43
+ "loss": 6.0868,
44
  "step": 250
45
  },
46
  {
47
+ "epoch": 0.03234501347708895,
48
+ "grad_norm": 1.6523106098175049,
49
+ "learning_rate": 0.0005987048030221263,
50
+ "loss": 5.9841,
51
  "step": 300
52
  },
53
  {
54
+ "epoch": 0.03773584905660377,
55
+ "grad_norm": 1.6248761415481567,
56
+ "learning_rate": 0.0005983810037776578,
57
+ "loss": 5.905,
58
  "step": 350
59
  },
60
  {
61
+ "epoch": 0.0431266846361186,
62
+ "grad_norm": 1.0370368957519531,
63
+ "learning_rate": 0.0005980572045331894,
64
+ "loss": 5.8319,
65
  "step": 400
66
  },
67
  {
68
+ "epoch": 0.04851752021563342,
69
+ "grad_norm": 1.0454210042953491,
70
+ "learning_rate": 0.0005977334052887209,
71
+ "loss": 5.7378,
72
  "step": 450
73
  },
74
  {
75
+ "epoch": 0.05390835579514825,
76
+ "grad_norm": 1.295013666152954,
77
+ "learning_rate": 0.0005974096060442526,
78
+ "loss": 5.6894,
79
  "step": 500
80
  },
81
  {
82
+ "epoch": 0.05929919137466307,
83
+ "grad_norm": 1.084869146347046,
84
+ "learning_rate": 0.0005970858067997841,
85
+ "loss": 5.5938,
86
  "step": 550
87
  },
88
  {
89
+ "epoch": 0.0646900269541779,
90
+ "grad_norm": 0.8982395529747009,
91
+ "learning_rate": 0.0005967620075553157,
92
+ "loss": 5.5525,
93
  "step": 600
94
  },
95
  {
96
+ "epoch": 0.07008086253369272,
97
+ "grad_norm": 1.0134685039520264,
98
+ "learning_rate": 0.0005964382083108472,
99
+ "loss": 5.4505,
100
  "step": 650
101
  },
102
  {
103
+ "epoch": 0.07547169811320754,
104
+ "grad_norm": 1.2105776071548462,
105
+ "learning_rate": 0.0005961144090663788,
106
+ "loss": 5.422,
107
  "step": 700
108
  },
109
  {
110
+ "epoch": 0.08086253369272237,
111
+ "grad_norm": 1.2871284484863281,
112
+ "learning_rate": 0.0005957906098219104,
113
+ "loss": 5.3284,
114
  "step": 750
115
  },
116
  {
117
+ "epoch": 0.0862533692722372,
118
+ "grad_norm": 0.9173638224601746,
119
+ "learning_rate": 0.0005954668105774419,
120
+ "loss": 5.2824,
121
  "step": 800
122
  },
123
  {
124
+ "epoch": 0.09164420485175202,
125
+ "grad_norm": 1.3351950645446777,
126
+ "learning_rate": 0.0005951430113329735,
127
+ "loss": 5.2262,
128
  "step": 850
129
  },
130
  {
131
+ "epoch": 0.09703504043126684,
132
+ "grad_norm": 1.00580894947052,
133
+ "learning_rate": 0.0005948192120885051,
134
+ "loss": 5.1935,
135
  "step": 900
136
  },
137
  {
138
+ "epoch": 0.10242587601078167,
139
+ "grad_norm": 1.3590151071548462,
140
+ "learning_rate": 0.0005944954128440366,
141
+ "loss": 5.1385,
142
  "step": 950
143
  },
144
  {
145
+ "epoch": 0.1078167115902965,
146
+ "grad_norm": 1.0777369737625122,
147
+ "learning_rate": 0.0005941716135995682,
148
+ "loss": 5.1241,
149
  "step": 1000
150
  },
151
  {
152
+ "epoch": 0.1078167115902965,
153
+ "eval_accuracy": 0.2255982726798507,
154
+ "eval_loss": 5.041033744812012,
155
+ "eval_runtime": 183.3106,
156
+ "eval_samples_per_second": 98.254,
157
+ "eval_steps_per_second": 6.143,
158
  "step": 1000
159
  },
160
  {
161
+ "epoch": 0.11320754716981132,
162
+ "grad_norm": 1.0632835626602173,
163
+ "learning_rate": 0.0005938478143550997,
164
+ "loss": 5.0886,
165
  "step": 1050
166
  },
167
  {
168
+ "epoch": 0.11859838274932614,
169
+ "grad_norm": 0.91739422082901,
170
+ "learning_rate": 0.0005935240151106314,
171
+ "loss": 5.0135,
172
  "step": 1100
173
  },
174
  {
175
+ "epoch": 0.12398921832884097,
176
+ "grad_norm": 1.300990343093872,
177
+ "learning_rate": 0.0005932002158661629,
178
+ "loss": 5.0122,
179
  "step": 1150
180
  },
181
  {
182
+ "epoch": 0.1293800539083558,
183
+ "grad_norm": 0.8993210196495056,
184
+ "learning_rate": 0.0005928764166216945,
185
+ "loss": 4.9734,
186
  "step": 1200
187
  },
188
  {
189
+ "epoch": 0.1347708894878706,
190
+ "grad_norm": 0.8592536449432373,
191
+ "learning_rate": 0.000592552617377226,
192
+ "loss": 4.9176,
193
  "step": 1250
194
  },
195
  {
196
+ "epoch": 0.14016172506738545,
197
+ "grad_norm": 1.1306790113449097,
198
+ "learning_rate": 0.0005922288181327577,
199
+ "loss": 4.92,
200
  "step": 1300
201
  },
202
  {
203
+ "epoch": 0.14555256064690028,
204
+ "grad_norm": 0.8705931901931763,
205
+ "learning_rate": 0.0005919050188882893,
206
+ "loss": 4.8889,
207
  "step": 1350
208
  },
209
  {
210
+ "epoch": 0.1509433962264151,
211
+ "grad_norm": 1.0139693021774292,
212
+ "learning_rate": 0.0005915812196438207,
213
+ "loss": 4.8493,
214
  "step": 1400
215
  },
216
  {
217
+ "epoch": 0.15633423180592992,
218
+ "grad_norm": 0.9368259310722351,
219
+ "learning_rate": 0.0005912574203993524,
220
+ "loss": 4.8326,
221
  "step": 1450
222
  },
223
  {
224
+ "epoch": 0.16172506738544473,
225
+ "grad_norm": 0.8887364268302917,
226
+ "learning_rate": 0.0005909336211548839,
227
+ "loss": 4.8207,
228
  "step": 1500
229
  },
230
  {
231
+ "epoch": 0.16711590296495957,
232
+ "grad_norm": 0.9200729131698608,
233
+ "learning_rate": 0.0005906098219104155,
234
+ "loss": 4.7862,
235
  "step": 1550
236
  },
237
  {
238
+ "epoch": 0.1725067385444744,
239
+ "grad_norm": 1.1816471815109253,
240
+ "learning_rate": 0.000590286022665947,
241
+ "loss": 4.7976,
242
  "step": 1600
243
  },
244
  {
245
+ "epoch": 0.1778975741239892,
246
+ "grad_norm": 1.226843237876892,
247
+ "learning_rate": 0.0005899622234214787,
248
+ "loss": 4.7459,
249
  "step": 1650
250
  },
251
  {
252
+ "epoch": 0.18328840970350405,
253
+ "grad_norm": 0.8250679969787598,
254
+ "learning_rate": 0.0005896384241770102,
255
+ "loss": 4.699,
256
  "step": 1700
257
  },
258
  {
259
+ "epoch": 0.18867924528301888,
260
+ "grad_norm": 0.8124901056289673,
261
+ "learning_rate": 0.0005893146249325418,
262
+ "loss": 4.6808,
263
  "step": 1750
264
  },
265
  {
266
+ "epoch": 0.1940700808625337,
267
+ "grad_norm": 0.7820393443107605,
268
+ "learning_rate": 0.0005889908256880733,
269
+ "loss": 4.6828,
270
  "step": 1800
271
  },
272
  {
273
+ "epoch": 0.19946091644204852,
274
+ "grad_norm": 0.763322114944458,
275
+ "learning_rate": 0.0005886670264436049,
276
+ "loss": 4.691,
277
  "step": 1850
278
  },
279
  {
280
+ "epoch": 0.20485175202156333,
281
+ "grad_norm": 0.9577104449272156,
282
+ "learning_rate": 0.0005883432271991365,
283
+ "loss": 4.6278,
284
  "step": 1900
285
  },
286
  {
287
+ "epoch": 0.21024258760107817,
288
+ "grad_norm": 0.6926529407501221,
289
+ "learning_rate": 0.0005880194279546681,
290
+ "loss": 4.6146,
291
  "step": 1950
292
  },
293
  {
294
+ "epoch": 0.215633423180593,
295
+ "grad_norm": 1.0170329809188843,
296
+ "learning_rate": 0.0005876956287101996,
297
+ "loss": 4.5903,
298
  "step": 2000
299
  },
300
  {
301
+ "epoch": 0.215633423180593,
302
+ "eval_accuracy": 0.2692834700603165,
303
+ "eval_loss": 4.519266605377197,
304
+ "eval_runtime": 183.2891,
305
+ "eval_samples_per_second": 98.266,
306
+ "eval_steps_per_second": 6.143,
307
  "step": 2000
308
  },
309
  {
310
+ "epoch": 0.2210242587601078,
311
+ "grad_norm": 1.0118404626846313,
312
+ "learning_rate": 0.0005873718294657312,
313
+ "loss": 4.5812,
314
  "step": 2050
315
  },
316
  {
317
+ "epoch": 0.22641509433962265,
318
+ "grad_norm": 0.9022784233093262,
319
+ "learning_rate": 0.0005870480302212628,
320
+ "loss": 4.5477,
321
  "step": 2100
322
  },
323
  {
324
+ "epoch": 0.23180592991913745,
325
+ "grad_norm": 0.9819560647010803,
326
+ "learning_rate": 0.0005867242309767943,
327
+ "loss": 4.549,
328
  "step": 2150
329
  },
330
  {
331
+ "epoch": 0.2371967654986523,
332
+ "grad_norm": 0.8590049147605896,
333
+ "learning_rate": 0.0005864004317323259,
334
+ "loss": 4.5197,
335
  "step": 2200
336
  },
337
  {
338
+ "epoch": 0.24258760107816713,
339
+ "grad_norm": 0.889808177947998,
340
+ "learning_rate": 0.0005860766324878575,
341
+ "loss": 4.5046,
342
  "step": 2250
343
  },
344
  {
345
+ "epoch": 0.24797843665768193,
346
+ "grad_norm": 0.7039113640785217,
347
+ "learning_rate": 0.000585752833243389,
348
+ "loss": 4.4798,
349
  "step": 2300
350
  },
351
  {
352
+ "epoch": 0.25336927223719674,
353
+ "grad_norm": 0.8285587430000305,
354
+ "learning_rate": 0.0005854290339989206,
355
+ "loss": 4.4681,
356
  "step": 2350
357
  },
358
  {
359
+ "epoch": 0.2587601078167116,
360
+ "grad_norm": 0.8999916315078735,
361
+ "learning_rate": 0.0005851052347544521,
362
+ "loss": 4.461,
363
  "step": 2400
364
  },
365
  {
366
+ "epoch": 0.2641509433962264,
367
+ "grad_norm": 0.9571085572242737,
368
+ "learning_rate": 0.0005847814355099838,
369
+ "loss": 4.4459,
370
  "step": 2450
371
  },
372
  {
373
+ "epoch": 0.2695417789757412,
374
+ "grad_norm": 0.7957876920700073,
375
+ "learning_rate": 0.0005844576362655154,
376
+ "loss": 4.4679,
377
  "step": 2500
378
  },
379
  {
380
+ "epoch": 0.2749326145552561,
381
+ "grad_norm": 1.1164896488189697,
382
+ "learning_rate": 0.0005841338370210469,
383
+ "loss": 4.4144,
384
  "step": 2550
385
  },
386
  {
387
+ "epoch": 0.2803234501347709,
388
+ "grad_norm": 0.8796373009681702,
389
+ "learning_rate": 0.0005838100377765785,
390
+ "loss": 4.3885,
391
  "step": 2600
392
  },
393
  {
394
+ "epoch": 0.2857142857142857,
395
+ "grad_norm": 0.9395208358764648,
396
+ "learning_rate": 0.0005834862385321101,
397
+ "loss": 4.3928,
398
  "step": 2650
399
  },
400
  {
401
+ "epoch": 0.29110512129380056,
402
+ "grad_norm": 1.015073537826538,
403
+ "learning_rate": 0.0005831624392876417,
404
+ "loss": 4.382,
405
  "step": 2700
406
  },
407
  {
408
+ "epoch": 0.29649595687331537,
409
+ "grad_norm": 0.8688176870346069,
410
+ "learning_rate": 0.0005828386400431731,
411
+ "loss": 4.3686,
412
  "step": 2750
413
  },
414
  {
415
+ "epoch": 0.3018867924528302,
416
+ "grad_norm": 0.7499260902404785,
417
+ "learning_rate": 0.0005825148407987048,
418
+ "loss": 4.3649,
419
  "step": 2800
420
  },
421
  {
422
+ "epoch": 0.30727762803234504,
423
+ "grad_norm": 0.8421972393989563,
424
+ "learning_rate": 0.0005821910415542363,
425
+ "loss": 4.3666,
426
  "step": 2850
427
  },
428
  {
429
+ "epoch": 0.31266846361185985,
430
+ "grad_norm": 0.8522474765777588,
431
+ "learning_rate": 0.0005818672423097679,
432
+ "loss": 4.345,
433
  "step": 2900
434
  },
435
  {
436
+ "epoch": 0.31805929919137466,
437
+ "grad_norm": 0.8274930715560913,
438
+ "learning_rate": 0.0005815434430652994,
439
+ "loss": 4.3585,
440
  "step": 2950
441
  },
442
  {
443
+ "epoch": 0.32345013477088946,
444
+ "grad_norm": 0.7252627611160278,
445
+ "learning_rate": 0.0005812196438208311,
446
+ "loss": 4.3009,
447
  "step": 3000
448
  },
449
  {
450
+ "epoch": 0.32345013477088946,
451
+ "eval_accuracy": 0.29779442243438753,
452
+ "eval_loss": 4.241104602813721,
453
+ "eval_runtime": 183.1601,
454
+ "eval_samples_per_second": 98.335,
455
+ "eval_steps_per_second": 6.148,
456
  "step": 3000
457
  },
458
  {
459
+ "epoch": 0.3288409703504043,
460
+ "grad_norm": 0.7949138283729553,
461
+ "learning_rate": 0.0005808958445763626,
462
+ "loss": 4.3115,
463
  "step": 3050
464
  },
465
  {
466
+ "epoch": 0.33423180592991913,
467
+ "grad_norm": 0.8113821148872375,
468
+ "learning_rate": 0.0005805720453318942,
469
+ "loss": 4.2896,
470
  "step": 3100
471
  },
472
  {
473
+ "epoch": 0.33962264150943394,
474
+ "grad_norm": 0.7971798181533813,
475
+ "learning_rate": 0.0005802482460874257,
476
+ "loss": 4.2815,
477
  "step": 3150
478
  },
479
  {
480
+ "epoch": 0.3450134770889488,
481
+ "grad_norm": 0.6988677382469177,
482
+ "learning_rate": 0.0005799244468429573,
483
+ "loss": 4.304,
484
  "step": 3200
485
  },
486
  {
487
+ "epoch": 0.3504043126684636,
488
+ "grad_norm": 0.816828727722168,
489
+ "learning_rate": 0.0005796006475984889,
490
+ "loss": 4.289,
491
  "step": 3250
492
  },
493
  {
494
+ "epoch": 0.3557951482479784,
495
+ "grad_norm": 1.0474231243133545,
496
+ "learning_rate": 0.0005792768483540205,
497
+ "loss": 4.2633,
498
  "step": 3300
499
  },
500
  {
501
+ "epoch": 0.3611859838274933,
502
+ "grad_norm": 0.6509954333305359,
503
+ "learning_rate": 0.000578953049109552,
504
+ "loss": 4.2729,
505
  "step": 3350
506
  },
507
  {
508
+ "epoch": 0.3665768194070081,
509
+ "grad_norm": 0.7029348611831665,
510
+ "learning_rate": 0.0005786292498650836,
511
+ "loss": 4.2416,
512
  "step": 3400
513
  },
514
  {
515
+ "epoch": 0.3719676549865229,
516
+ "grad_norm": 0.7324013710021973,
517
+ "learning_rate": 0.0005783054506206152,
518
+ "loss": 4.2535,
519
  "step": 3450
520
  },
521
  {
522
+ "epoch": 0.37735849056603776,
523
+ "grad_norm": 0.7252773642539978,
524
+ "learning_rate": 0.0005779816513761467,
525
+ "loss": 4.233,
526
  "step": 3500
527
  },
528
  {
529
+ "epoch": 0.38274932614555257,
530
+ "grad_norm": 1.0832518339157104,
531
+ "learning_rate": 0.0005776578521316782,
532
+ "loss": 4.2134,
533
  "step": 3550
534
  },
535
  {
536
+ "epoch": 0.3881401617250674,
537
+ "grad_norm": 0.7716037034988403,
538
+ "learning_rate": 0.0005773340528872099,
539
+ "loss": 4.2353,
540
  "step": 3600
541
  },
542
  {
543
+ "epoch": 0.3935309973045822,
544
+ "grad_norm": 0.6974150538444519,
545
+ "learning_rate": 0.0005770102536427414,
546
+ "loss": 4.2306,
547
  "step": 3650
548
  },
549
  {
550
+ "epoch": 0.39892183288409705,
551
+ "grad_norm": 0.6498096585273743,
552
+ "learning_rate": 0.000576686454398273,
553
+ "loss": 4.2078,
554
  "step": 3700
555
  },
556
  {
557
+ "epoch": 0.40431266846361186,
558
+ "grad_norm": 0.7130216956138611,
559
+ "learning_rate": 0.0005763626551538045,
560
+ "loss": 4.1822,
561
  "step": 3750
562
  },
563
  {
564
+ "epoch": 0.40970350404312667,
565
+ "grad_norm": 0.7488390207290649,
566
+ "learning_rate": 0.0005760388559093362,
567
+ "loss": 4.1907,
568
  "step": 3800
569
  },
570
  {
571
+ "epoch": 0.41509433962264153,
572
+ "grad_norm": 0.6339357495307922,
573
+ "learning_rate": 0.0005757150566648678,
574
+ "loss": 4.1742,
575
  "step": 3850
576
  },
577
  {
578
+ "epoch": 0.42048517520215634,
579
+ "grad_norm": 0.7490821480751038,
580
+ "learning_rate": 0.0005753912574203993,
581
+ "loss": 4.1807,
582
  "step": 3900
583
  },
584
  {
585
+ "epoch": 0.42587601078167114,
586
+ "grad_norm": 0.780868411064148,
587
+ "learning_rate": 0.0005750674581759309,
588
+ "loss": 4.1914,
589
  "step": 3950
590
  },
591
  {
592
+ "epoch": 0.431266846361186,
593
+ "grad_norm": 0.5976775884628296,
594
+ "learning_rate": 0.0005747436589314624,
595
+ "loss": 4.1673,
596
  "step": 4000
597
  },
598
  {
599
+ "epoch": 0.431266846361186,
600
+ "eval_accuracy": 0.3114474183584917,
601
+ "eval_loss": 4.098546504974365,
602
+ "eval_runtime": 183.054,
603
+ "eval_samples_per_second": 98.392,
604
+ "eval_steps_per_second": 6.151,
605
  "step": 4000
606
  },
607
  {
608
+ "epoch": 0.4366576819407008,
609
+ "grad_norm": 0.9827622771263123,
610
+ "learning_rate": 0.0005744198596869941,
611
+ "loss": 4.1522,
612
  "step": 4050
613
  },
614
  {
615
+ "epoch": 0.4420485175202156,
616
+ "grad_norm": 0.6283232569694519,
617
+ "learning_rate": 0.0005740960604425255,
618
+ "loss": 4.1417,
619
  "step": 4100
620
  },
621
  {
622
+ "epoch": 0.4474393530997305,
623
+ "grad_norm": 0.6134821176528931,
624
+ "learning_rate": 0.0005737722611980572,
625
+ "loss": 4.1527,
626
  "step": 4150
627
  },
628
  {
629
+ "epoch": 0.4528301886792453,
630
+ "grad_norm": 0.7765461802482605,
631
+ "learning_rate": 0.0005734484619535887,
632
+ "loss": 4.1375,
633
  "step": 4200
634
  },
635
  {
636
+ "epoch": 0.4582210242587601,
637
+ "grad_norm": 0.7716192007064819,
638
+ "learning_rate": 0.0005731246627091203,
639
+ "loss": 4.1374,
640
  "step": 4250
641
  },
642
  {
643
+ "epoch": 0.4636118598382749,
644
+ "grad_norm": 0.6671397089958191,
645
+ "learning_rate": 0.0005728008634646518,
646
+ "loss": 4.1203,
647
  "step": 4300
648
  },
649
  {
650
+ "epoch": 0.46900269541778977,
651
+ "grad_norm": 0.647169828414917,
652
+ "learning_rate": 0.0005724770642201835,
653
+ "loss": 4.1192,
654
  "step": 4350
655
  },
656
  {
657
+ "epoch": 0.4743935309973046,
658
+ "grad_norm": 0.6429058313369751,
659
+ "learning_rate": 0.000572153264975715,
660
+ "loss": 4.1191,
661
  "step": 4400
662
  },
663
  {
664
+ "epoch": 0.4797843665768194,
665
+ "grad_norm": 0.7357009649276733,
666
+ "learning_rate": 0.0005718294657312466,
667
+ "loss": 4.1407,
668
  "step": 4450
669
  },
670
  {
671
+ "epoch": 0.48517520215633425,
672
+ "grad_norm": 0.6552907824516296,
673
+ "learning_rate": 0.0005715056664867781,
674
+ "loss": 4.1015,
675
  "step": 4500
676
  },
677
  {
678
+ "epoch": 0.49056603773584906,
679
+ "grad_norm": 0.6006699204444885,
680
+ "learning_rate": 0.0005711818672423097,
681
+ "loss": 4.1109,
682
  "step": 4550
683
  },
684
  {
685
+ "epoch": 0.49595687331536387,
686
+ "grad_norm": 0.7775830030441284,
687
+ "learning_rate": 0.0005708580679978413,
688
+ "loss": 4.102,
689
  "step": 4600
690
  },
691
  {
692
+ "epoch": 0.5013477088948787,
693
+ "grad_norm": 0.5961248278617859,
694
+ "learning_rate": 0.0005705342687533729,
695
+ "loss": 4.1,
696
  "step": 4650
697
  },
698
  {
699
+ "epoch": 0.5067385444743935,
700
+ "grad_norm": 0.7144160866737366,
701
+ "learning_rate": 0.0005702104695089044,
702
+ "loss": 4.1015,
703
  "step": 4700
704
  },
705
  {
706
+ "epoch": 0.5121293800539084,
707
+ "grad_norm": 0.6750192642211914,
708
+ "learning_rate": 0.000569886670264436,
709
+ "loss": 4.101,
710
  "step": 4750
711
  },
712
  {
713
+ "epoch": 0.5175202156334232,
714
+ "grad_norm": 0.6833533644676208,
715
+ "learning_rate": 0.0005695628710199675,
716
+ "loss": 4.0794,
717
  "step": 4800
718
  },
719
  {
720
+ "epoch": 0.522911051212938,
721
+ "grad_norm": 0.7269881963729858,
722
+ "learning_rate": 0.0005692390717754991,
723
+ "loss": 4.1048,
724
  "step": 4850
725
  },
726
  {
727
+ "epoch": 0.5283018867924528,
728
+ "grad_norm": 0.5909144878387451,
729
+ "learning_rate": 0.0005689152725310306,
730
+ "loss": 4.078,
731
  "step": 4900
732
  },
733
  {
734
+ "epoch": 0.5336927223719676,
735
+ "grad_norm": 0.7556242942810059,
736
+ "learning_rate": 0.0005685914732865623,
737
+ "loss": 4.0665,
738
  "step": 4950
739
  },
740
  {
741
+ "epoch": 0.5390835579514824,
742
+ "grad_norm": 0.745907187461853,
743
+ "learning_rate": 0.0005682676740420939,
744
+ "loss": 4.069,
745
  "step": 5000
746
  },
747
  {
748
+ "epoch": 0.5390835579514824,
749
+ "eval_accuracy": 0.3210300597992899,
750
+ "eval_loss": 3.9976468086242676,
751
+ "eval_runtime": 183.1773,
752
+ "eval_samples_per_second": 98.326,
753
+ "eval_steps_per_second": 6.147,
754
  "step": 5000
755
  },
756
  {
757
+ "epoch": 0.5444743935309974,
758
+ "grad_norm": 0.5867395997047424,
759
+ "learning_rate": 0.0005679438747976254,
760
+ "loss": 4.0661,
761
  "step": 5050
762
  },
763
  {
764
+ "epoch": 0.5498652291105122,
765
+ "grad_norm": 0.6460950374603271,
766
+ "learning_rate": 0.000567620075553157,
767
+ "loss": 4.0524,
768
  "step": 5100
769
  },
770
  {
771
+ "epoch": 0.555256064690027,
772
+ "grad_norm": 0.5987642407417297,
773
+ "learning_rate": 0.0005672962763086886,
774
+ "loss": 4.064,
775
  "step": 5150
776
  },
777
  {
778
+ "epoch": 0.5606469002695418,
779
+ "grad_norm": 0.6056843400001526,
780
+ "learning_rate": 0.0005669724770642202,
781
+ "loss": 4.0542,
782
  "step": 5200
783
  },
784
  {
785
+ "epoch": 0.5660377358490566,
786
+ "grad_norm": 0.6567562222480774,
787
+ "learning_rate": 0.0005666486778197517,
788
+ "loss": 4.0546,
789
  "step": 5250
790
  },
791
  {
792
+ "epoch": 0.5714285714285714,
793
+ "grad_norm": 0.6721335649490356,
794
+ "learning_rate": 0.0005663248785752833,
795
+ "loss": 4.0355,
796
  "step": 5300
797
  },
798
  {
799
+ "epoch": 0.5768194070080862,
800
+ "grad_norm": 0.6371747851371765,
801
+ "learning_rate": 0.0005660010793308148,
802
+ "loss": 4.0116,
803
  "step": 5350
804
  },
805
  {
806
+ "epoch": 0.5822102425876011,
807
+ "grad_norm": 0.7218015789985657,
808
+ "learning_rate": 0.0005656772800863465,
809
+ "loss": 4.0524,
810
  "step": 5400
811
  },
812
  {
813
+ "epoch": 0.5876010781671159,
814
+ "grad_norm": 0.6780861020088196,
815
+ "learning_rate": 0.0005653534808418779,
816
+ "loss": 4.0286,
817
  "step": 5450
818
  },
819
  {
820
+ "epoch": 0.5929919137466307,
821
+ "grad_norm": 0.6822832822799683,
822
+ "learning_rate": 0.0005650296815974096,
823
+ "loss": 4.0383,
824
  "step": 5500
825
  },
826
  {
827
+ "epoch": 0.5983827493261455,
828
+ "grad_norm": 0.610171377658844,
829
+ "learning_rate": 0.0005647058823529411,
830
+ "loss": 4.0229,
831
  "step": 5550
832
  },
833
  {
834
+ "epoch": 0.6037735849056604,
835
+ "grad_norm": 0.6452059149742126,
836
+ "learning_rate": 0.0005643820831084727,
837
  "loss": 4.0241,
838
  "step": 5600
839
  },
840
  {
841
+ "epoch": 0.6091644204851752,
842
+ "grad_norm": 0.6566394567489624,
843
+ "learning_rate": 0.0005640582838640042,
844
+ "loss": 3.998,
845
  "step": 5650
846
  },
847
  {
848
+ "epoch": 0.6145552560646901,
849
+ "grad_norm": 0.6727635860443115,
850
+ "learning_rate": 0.0005637344846195358,
851
+ "loss": 4.0113,
852
  "step": 5700
853
  },
854
  {
855
+ "epoch": 0.6199460916442049,
856
+ "grad_norm": 0.7007443308830261,
857
+ "learning_rate": 0.0005634106853750674,
858
+ "loss": 4.0045,
859
  "step": 5750
860
  },
861
  {
862
+ "epoch": 0.6253369272237197,
863
+ "grad_norm": 0.5535261034965515,
864
+ "learning_rate": 0.000563086886130599,
865
+ "loss": 4.0178,
866
  "step": 5800
867
  },
868
  {
869
+ "epoch": 0.6307277628032345,
870
+ "grad_norm": 0.5908109545707703,
871
+ "learning_rate": 0.0005627630868861305,
872
+ "loss": 4.0013,
873
  "step": 5850
874
  },
875
  {
876
+ "epoch": 0.6361185983827493,
877
+ "grad_norm": 0.6292212605476379,
878
+ "learning_rate": 0.0005624392876416621,
879
+ "loss": 3.9984,
880
  "step": 5900
881
  },
882
  {
883
+ "epoch": 0.6415094339622641,
884
+ "grad_norm": 0.6456776857376099,
885
+ "learning_rate": 0.0005621154883971937,
886
+ "loss": 3.9997,
887
  "step": 5950
888
  },
889
  {
890
+ "epoch": 0.6469002695417789,
891
+ "grad_norm": 0.6281388401985168,
892
+ "learning_rate": 0.0005617916891527253,
893
+ "loss": 3.9954,
894
  "step": 6000
895
  },
896
  {
897
+ "epoch": 0.6469002695417789,
898
+ "eval_accuracy": 0.32793147392749006,
899
+ "eval_loss": 3.9262521266937256,
900
+ "eval_runtime": 183.2552,
901
+ "eval_samples_per_second": 98.284,
902
+ "eval_steps_per_second": 6.144,
903
  "step": 6000
904
  },
905
  {
906
+ "epoch": 0.6522911051212938,
907
+ "grad_norm": 0.581551730632782,
908
+ "learning_rate": 0.0005614678899082568,
909
+ "loss": 3.9906,
910
  "step": 6050
911
  },
912
  {
913
+ "epoch": 0.6576819407008087,
914
+ "grad_norm": 0.6080215573310852,
915
+ "learning_rate": 0.0005611505666486777,
916
+ "loss": 4.0079,
917
  "step": 6100
918
  },
919
  {
920
+ "epoch": 0.6630727762803235,
921
+ "grad_norm": 0.6454057097434998,
922
+ "learning_rate": 0.0005608267674042094,
923
+ "loss": 3.9902,
924
  "step": 6150
925
  },
926
  {
927
+ "epoch": 0.6684636118598383,
928
+ "grad_norm": 0.61163330078125,
929
+ "learning_rate": 0.0005605029681597409,
930
+ "loss": 3.9904,
931
  "step": 6200
932
  },
933
  {
934
+ "epoch": 0.6738544474393531,
935
+ "grad_norm": 0.636034369468689,
936
+ "learning_rate": 0.0005601791689152725,
937
+ "loss": 3.9806,
938
  "step": 6250
939
  },
940
  {
941
+ "epoch": 0.6792452830188679,
942
+ "grad_norm": 0.6304333209991455,
943
+ "learning_rate": 0.000559855369670804,
944
+ "loss": 3.9816,
945
  "step": 6300
946
  },
947
  {
948
+ "epoch": 0.6846361185983828,
949
+ "grad_norm": 0.6418639421463013,
950
+ "learning_rate": 0.0005595315704263357,
951
+ "loss": 3.9651,
952
  "step": 6350
953
  },
954
  {
955
+ "epoch": 0.6900269541778976,
956
+ "grad_norm": 0.5913263559341431,
957
+ "learning_rate": 0.0005592077711818672,
958
+ "loss": 3.9812,
959
  "step": 6400
960
  },
961
  {
962
+ "epoch": 0.6954177897574124,
963
+ "grad_norm": 0.7592663168907166,
964
+ "learning_rate": 0.0005588839719373988,
965
+ "loss": 3.9551,
966
  "step": 6450
967
  },
968
  {
969
+ "epoch": 0.7008086253369272,
970
+ "grad_norm": 0.5913580656051636,
971
+ "learning_rate": 0.0005585601726929303,
972
+ "loss": 3.9616,
973
  "step": 6500
974
  },
975
  {
976
+ "epoch": 0.706199460916442,
977
+ "grad_norm": 0.6632679104804993,
978
+ "learning_rate": 0.0005582363734484619,
979
+ "loss": 3.9452,
980
  "step": 6550
981
  },
982
  {
983
+ "epoch": 0.7115902964959568,
984
+ "grad_norm": 0.5981972813606262,
985
+ "learning_rate": 0.0005579125742039935,
986
+ "loss": 3.9522,
987
  "step": 6600
988
  },
989
  {
990
+ "epoch": 0.7169811320754716,
991
+ "grad_norm": 0.6516519784927368,
992
+ "learning_rate": 0.0005575887749595251,
993
+ "loss": 3.9547,
994
  "step": 6650
995
  },
996
  {
997
+ "epoch": 0.7223719676549866,
998
+ "grad_norm": 0.6064105033874512,
999
+ "learning_rate": 0.0005572649757150566,
1000
+ "loss": 3.9576,
1001
  "step": 6700
1002
  },
1003
  {
1004
+ "epoch": 0.7277628032345014,
1005
+ "grad_norm": 0.6877552270889282,
1006
+ "learning_rate": 0.0005569411764705882,
1007
+ "loss": 3.9551,
1008
  "step": 6750
1009
  },
1010
  {
1011
+ "epoch": 0.7331536388140162,
1012
+ "grad_norm": 0.5797296166419983,
1013
+ "learning_rate": 0.0005566173772261198,
1014
+ "loss": 3.9213,
1015
  "step": 6800
1016
  },
1017
  {
1018
+ "epoch": 0.738544474393531,
1019
+ "grad_norm": 0.6150147318840027,
1020
+ "learning_rate": 0.0005562935779816513,
1021
+ "loss": 3.9364,
1022
  "step": 6850
1023
  },
1024
  {
1025
+ "epoch": 0.7439353099730458,
1026
+ "grad_norm": 0.6063942909240723,
1027
+ "learning_rate": 0.0005559697787371828,
1028
+ "loss": 3.9515,
1029
  "step": 6900
1030
  },
1031
  {
1032
+ "epoch": 0.7493261455525606,
1033
+ "grad_norm": 0.6256096959114075,
1034
+ "learning_rate": 0.0005556459794927145,
1035
+ "loss": 3.9323,
1036
  "step": 6950
1037
  },
1038
  {
1039
+ "epoch": 0.7547169811320755,
1040
+ "grad_norm": 0.615798830986023,
1041
+ "learning_rate": 0.000555322180248246,
1042
+ "loss": 3.9429,
1043
  "step": 7000
1044
  },
1045
  {
1046
+ "epoch": 0.7547169811320755,
1047
+ "eval_accuracy": 0.33324362226562787,
1048
+ "eval_loss": 3.869093894958496,
1049
+ "eval_runtime": 183.4536,
1050
+ "eval_samples_per_second": 98.177,
1051
+ "eval_steps_per_second": 6.138,
1052
  "step": 7000
1053
  },
1054
  {
1055
+ "epoch": 0.7601078167115903,
1056
+ "grad_norm": 0.7148776054382324,
1057
+ "learning_rate": 0.0005549983810037776,
1058
+ "loss": 3.9117,
1059
  "step": 7050
1060
  },
1061
  {
1062
+ "epoch": 0.7654986522911051,
1063
+ "grad_norm": 0.5761632919311523,
1064
+ "learning_rate": 0.0005546745817593091,
1065
+ "loss": 3.9146,
1066
  "step": 7100
1067
  },
1068
  {
1069
+ "epoch": 0.77088948787062,
1070
+ "grad_norm": 0.6276112794876099,
1071
+ "learning_rate": 0.0005543507825148408,
1072
+ "loss": 3.9177,
1073
  "step": 7150
1074
  },
1075
  {
1076
+ "epoch": 0.7762803234501348,
1077
+ "grad_norm": 0.5876376628875732,
1078
+ "learning_rate": 0.0005540269832703723,
1079
+ "loss": 3.899,
1080
  "step": 7200
1081
  },
1082
  {
1083
+ "epoch": 0.7816711590296496,
1084
+ "grad_norm": 0.5909850597381592,
1085
+ "learning_rate": 0.0005537031840259039,
1086
+ "loss": 3.9271,
1087
  "step": 7250
1088
  },
1089
  {
1090
+ "epoch": 0.7870619946091644,
1091
+ "grad_norm": 0.5844208002090454,
1092
+ "learning_rate": 0.0005533793847814354,
1093
+ "loss": 3.9013,
1094
  "step": 7300
1095
  },
1096
  {
1097
+ "epoch": 0.7924528301886793,
1098
+ "grad_norm": 0.6150322556495667,
1099
+ "learning_rate": 0.000553055585536967,
1100
+ "loss": 3.9195,
1101
  "step": 7350
1102
  },
1103
  {
1104
+ "epoch": 0.7978436657681941,
1105
+ "grad_norm": 0.5726950168609619,
1106
+ "learning_rate": 0.0005527317862924987,
1107
+ "loss": 3.92,
1108
  "step": 7400
1109
  },
1110
  {
1111
+ "epoch": 0.8032345013477089,
1112
+ "grad_norm": 0.5585176348686218,
1113
+ "learning_rate": 0.0005524079870480301,
1114
+ "loss": 3.9137,
1115
  "step": 7450
1116
  },
1117
  {
1118
+ "epoch": 0.8086253369272237,
1119
+ "grad_norm": 0.5901408195495605,
1120
+ "learning_rate": 0.0005520841878035618,
1121
+ "loss": 3.9295,
1122
  "step": 7500
1123
  },
1124
  {
1125
+ "epoch": 0.8140161725067385,
1126
+ "grad_norm": 0.5504985451698303,
1127
+ "learning_rate": 0.0005517603885590933,
1128
+ "loss": 3.9024,
1129
  "step": 7550
1130
  },
1131
  {
1132
+ "epoch": 0.8194070080862533,
1133
+ "grad_norm": 0.5761569142341614,
1134
+ "learning_rate": 0.0005514365893146249,
1135
+ "loss": 3.9011,
1136
  "step": 7600
1137
  },
1138
  {
1139
+ "epoch": 0.8247978436657682,
1140
+ "grad_norm": 0.5243045687675476,
1141
+ "learning_rate": 0.0005511127900701564,
1142
+ "loss": 3.8925,
1143
  "step": 7650
1144
  },
1145
  {
1146
+ "epoch": 0.8301886792452831,
1147
+ "grad_norm": 0.5392599105834961,
1148
+ "learning_rate": 0.000550788990825688,
1149
+ "loss": 3.9223,
1150
  "step": 7700
1151
  },
1152
  {
1153
+ "epoch": 0.8355795148247979,
1154
+ "grad_norm": 0.6053388118743896,
1155
+ "learning_rate": 0.0005504651915812196,
1156
+ "loss": 3.9166,
1157
  "step": 7750
1158
  },
1159
  {
1160
+ "epoch": 0.8409703504043127,
1161
+ "grad_norm": 0.6524935960769653,
1162
+ "learning_rate": 0.0005501413923367512,
1163
+ "loss": 3.8809,
1164
  "step": 7800
1165
  },
1166
  {
1167
+ "epoch": 0.8463611859838275,
1168
+ "grad_norm": 0.607294499874115,
1169
+ "learning_rate": 0.0005498175930922827,
1170
+ "loss": 3.893,
1171
  "step": 7850
1172
  },
1173
  {
1174
+ "epoch": 0.8517520215633423,
1175
+ "grad_norm": 0.589516818523407,
1176
+ "learning_rate": 0.0005494937938478143,
1177
+ "loss": 3.8986,
1178
  "step": 7900
1179
  },
1180
  {
1181
+ "epoch": 0.8571428571428571,
1182
+ "grad_norm": 0.5285811424255371,
1183
+ "learning_rate": 0.0005491699946033459,
1184
+ "loss": 3.8868,
1185
  "step": 7950
1186
  },
1187
  {
1188
+ "epoch": 0.862533692722372,
1189
+ "grad_norm": 0.5825549364089966,
1190
+ "learning_rate": 0.0005488461953588775,
1191
+ "loss": 3.8867,
1192
  "step": 8000
1193
  },
1194
  {
1195
+ "epoch": 0.862533692722372,
1196
+ "eval_accuracy": 0.3376740524191511,
1197
+ "eval_loss": 3.8201420307159424,
1198
+ "eval_runtime": 183.0162,
1199
+ "eval_samples_per_second": 98.412,
1200
+ "eval_steps_per_second": 6.152,
1201
  "step": 8000
1202
  },
1203
  {
1204
+ "epoch": 0.8679245283018868,
1205
+ "grad_norm": 0.5807555913925171,
1206
+ "learning_rate": 0.000548522396114409,
1207
+ "loss": 3.8912,
1208
  "step": 8050
1209
  },
1210
  {
1211
+ "epoch": 0.8733153638814016,
1212
+ "grad_norm": 0.5636572241783142,
1213
+ "learning_rate": 0.0005481985968699406,
1214
+ "loss": 3.8969,
1215
  "step": 8100
1216
  },
1217
  {
1218
+ "epoch": 0.8787061994609164,
1219
+ "grad_norm": 0.5182288289070129,
1220
+ "learning_rate": 0.0005478812736103615,
1221
+ "loss": 3.8758,
1222
  "step": 8150
1223
  },
1224
  {
1225
+ "epoch": 0.8840970350404312,
1226
+ "grad_norm": 0.6025846004486084,
1227
+ "learning_rate": 0.0005475574743658931,
1228
+ "loss": 3.8797,
1229
  "step": 8200
1230
  },
1231
  {
1232
+ "epoch": 0.889487870619946,
1233
+ "grad_norm": 0.6195996403694153,
1234
+ "learning_rate": 0.0005472336751214246,
1235
+ "loss": 3.8745,
1236
  "step": 8250
1237
  },
1238
  {
1239
+ "epoch": 0.894878706199461,
1240
+ "grad_norm": 0.6360132694244385,
1241
+ "learning_rate": 0.0005469098758769562,
1242
+ "loss": 3.8602,
1243
  "step": 8300
1244
  },
1245
  {
1246
+ "epoch": 0.9002695417789758,
1247
+ "grad_norm": 0.5655954480171204,
1248
+ "learning_rate": 0.0005465860766324878,
1249
+ "loss": 3.8461,
1250
  "step": 8350
1251
  },
1252
  {
1253
+ "epoch": 0.9056603773584906,
1254
+ "grad_norm": 0.5116124153137207,
1255
+ "learning_rate": 0.0005462622773880194,
1256
+ "loss": 3.8923,
1257
  "step": 8400
1258
  },
1259
  {
1260
+ "epoch": 0.9110512129380054,
1261
+ "grad_norm": 0.6523976922035217,
1262
+ "learning_rate": 0.000545938478143551,
1263
+ "loss": 3.8607,
1264
  "step": 8450
1265
  },
1266
  {
1267
+ "epoch": 0.9164420485175202,
1268
+ "grad_norm": 0.5672992467880249,
1269
+ "learning_rate": 0.0005456146788990825,
1270
+ "loss": 3.8484,
1271
  "step": 8500
1272
  },
1273
  {
1274
+ "epoch": 0.921832884097035,
1275
+ "grad_norm": 0.6210187077522278,
1276
+ "learning_rate": 0.000545290879654614,
1277
+ "loss": 3.8607,
1278
  "step": 8550
1279
  },
1280
  {
1281
+ "epoch": 0.9272237196765498,
1282
+ "grad_norm": 0.6343889832496643,
1283
+ "learning_rate": 0.0005449670804101457,
1284
  "loss": 3.8604,
1285
  "step": 8600
1286
  },
1287
  {
1288
+ "epoch": 0.9326145552560647,
1289
+ "grad_norm": 0.5989923477172852,
1290
+ "learning_rate": 0.0005446432811656773,
1291
+ "loss": 3.8633,
1292
  "step": 8650
1293
  },
1294
  {
1295
+ "epoch": 0.9380053908355795,
1296
+ "grad_norm": 0.563865065574646,
1297
+ "learning_rate": 0.0005443194819212088,
1298
+ "loss": 3.8419,
1299
  "step": 8700
1300
  },
1301
  {
1302
+ "epoch": 0.9433962264150944,
1303
+ "grad_norm": 0.6404436826705933,
1304
+ "learning_rate": 0.0005439956826767404,
1305
+ "loss": 3.8581,
1306
  "step": 8750
1307
  },
1308
  {
1309
+ "epoch": 0.9487870619946092,
1310
+ "grad_norm": 0.6358834505081177,
1311
+ "learning_rate": 0.000543671883432272,
1312
+ "loss": 3.8479,
1313
  "step": 8800
1314
  },
1315
  {
1316
+ "epoch": 0.954177897574124,
1317
+ "grad_norm": 0.6391129493713379,
1318
+ "learning_rate": 0.0005433480841878035,
1319
+ "loss": 3.8407,
1320
  "step": 8850
1321
  },
1322
  {
1323
+ "epoch": 0.9595687331536388,
1324
+ "grad_norm": 0.6343260407447815,
1325
+ "learning_rate": 0.000543024284943335,
1326
+ "loss": 3.8336,
1327
  "step": 8900
1328
  },
1329
  {
1330
+ "epoch": 0.9649595687331537,
1331
+ "grad_norm": 0.6411747336387634,
1332
+ "learning_rate": 0.0005427004856988667,
1333
+ "loss": 3.8503,
1334
  "step": 8950
1335
  },
1336
  {
1337
+ "epoch": 0.9703504043126685,
1338
+ "grad_norm": 0.6860132217407227,
1339
+ "learning_rate": 0.0005423766864543982,
1340
+ "loss": 3.8499,
1341
  "step": 9000
1342
  },
1343
  {
1344
+ "epoch": 0.9703504043126685,
1345
+ "eval_accuracy": 0.3411033548643518,
1346
+ "eval_loss": 3.7823479175567627,
1347
+ "eval_runtime": 183.4606,
1348
+ "eval_samples_per_second": 98.174,
1349
+ "eval_steps_per_second": 6.138,
1350
  "step": 9000
1351
  },
1352
  {
1353
+ "epoch": 0.9757412398921833,
1354
+ "grad_norm": 0.5406085848808289,
1355
+ "learning_rate": 0.0005420528872099298,
1356
+ "loss": 3.8325,
1357
  "step": 9050
1358
  },
1359
  {
1360
+ "epoch": 0.9811320754716981,
1361
+ "grad_norm": 0.5538688898086548,
1362
+ "learning_rate": 0.0005417290879654613,
1363
+ "loss": 3.8317,
1364
  "step": 9100
1365
  },
1366
  {
1367
+ "epoch": 0.9865229110512129,
1368
+ "grad_norm": 0.5626528859138489,
1369
+ "learning_rate": 0.000541405288720993,
1370
+ "loss": 3.8459,
1371
  "step": 9150
1372
  },
1373
  {
1374
+ "epoch": 0.9919137466307277,
1375
+ "grad_norm": 0.5986374020576477,
1376
+ "learning_rate": 0.0005410814894765245,
1377
+ "loss": 3.8284,
1378
  "step": 9200
1379
  },
1380
  {
1381
+ "epoch": 0.9973045822102425,
1382
+ "grad_norm": 0.5362386703491211,
1383
+ "learning_rate": 0.0005407576902320561,
1384
+ "loss": 3.8566,
1385
  "step": 9250
1386
  },
1387
  {
1388
+ "epoch": 1.0026954177897573,
1389
+ "grad_norm": 0.6233709454536438,
1390
+ "learning_rate": 0.0005404338909875876,
1391
+ "loss": 3.7906,
1392
  "step": 9300
1393
  },
1394
  {
1395
+ "epoch": 1.0080862533692723,
1396
+ "grad_norm": 0.6044151782989502,
1397
+ "learning_rate": 0.0005401100917431192,
1398
+ "loss": 3.7677,
1399
  "step": 9350
1400
  },
1401
  {
1402
+ "epoch": 1.013477088948787,
1403
+ "grad_norm": 0.6204626560211182,
1404
+ "learning_rate": 0.0005397862924986508,
1405
+ "loss": 3.7815,
1406
  "step": 9400
1407
  },
1408
  {
1409
+ "epoch": 1.0188679245283019,
1410
+ "grad_norm": 0.5932472944259644,
1411
+ "learning_rate": 0.0005394624932541824,
1412
+ "loss": 3.7888,
1413
  "step": 9450
1414
  },
1415
  {
1416
+ "epoch": 1.0242587601078168,
1417
+ "grad_norm": 0.5732197761535645,
1418
+ "learning_rate": 0.0005391386940097139,
1419
+ "loss": 3.7574,
1420
  "step": 9500
1421
  },
1422
  {
1423
+ "epoch": 1.0296495956873315,
1424
+ "grad_norm": 0.6381683349609375,
1425
+ "learning_rate": 0.0005388148947652455,
1426
+ "loss": 3.7696,
1427
  "step": 9550
1428
  },
1429
  {
1430
+ "epoch": 1.0350404312668464,
1431
+ "grad_norm": 0.5938410758972168,
1432
+ "learning_rate": 0.000538491095520777,
1433
+ "loss": 3.796,
1434
  "step": 9600
1435
  },
1436
  {
1437
+ "epoch": 1.0404312668463611,
1438
+ "grad_norm": 0.5650463104248047,
1439
+ "learning_rate": 0.0005381672962763086,
1440
+ "loss": 3.7558,
1441
  "step": 9650
1442
  },
1443
  {
1444
+ "epoch": 1.045822102425876,
1445
+ "grad_norm": 0.5852746963500977,
1446
+ "learning_rate": 0.0005378434970318403,
1447
+ "loss": 3.7809,
1448
  "step": 9700
1449
  },
1450
  {
1451
+ "epoch": 1.0512129380053907,
1452
+ "grad_norm": 0.6100252866744995,
1453
+ "learning_rate": 0.0005375196977873718,
1454
  "loss": 3.7579,
1455
  "step": 9750
1456
  },
1457
  {
1458
+ "epoch": 1.0566037735849056,
1459
+ "grad_norm": 0.6463873386383057,
1460
+ "learning_rate": 0.0005371958985429034,
1461
+ "loss": 3.7686,
1462
  "step": 9800
1463
  },
1464
  {
1465
+ "epoch": 1.0619946091644206,
1466
+ "grad_norm": 0.5528656840324402,
1467
+ "learning_rate": 0.0005368720992984349,
1468
+ "loss": 3.7732,
1469
  "step": 9850
1470
  },
1471
  {
1472
+ "epoch": 1.0673854447439353,
1473
+ "grad_norm": 0.6037126779556274,
1474
+ "learning_rate": 0.0005365483000539665,
1475
+ "loss": 3.7688,
1476
  "step": 9900
1477
  },
1478
  {
1479
+ "epoch": 1.0727762803234502,
1480
+ "grad_norm": 0.5881588459014893,
1481
+ "learning_rate": 0.0005362245008094981,
1482
+ "loss": 3.7831,
1483
  "step": 9950
1484
  },
1485
  {
1486
+ "epoch": 1.0781671159029649,
1487
+ "grad_norm": 0.5433666110038757,
1488
+ "learning_rate": 0.0005359007015650297,
1489
+ "loss": 3.7737,
1490
  "step": 10000
1491
  },
1492
  {
1493
+ "epoch": 1.0781671159029649,
1494
+ "eval_accuracy": 0.34445040707347685,
1495
+ "eval_loss": 3.7509524822235107,
1496
+ "eval_runtime": 183.1516,
1497
+ "eval_samples_per_second": 98.339,
1498
+ "eval_steps_per_second": 6.148,
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 sha256:4a35de0cf02e0a9b32253d3eedad027187c94bd5bdaa4f4c21caf087fc92111d
3
  size 5304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ac11a1a1f76416a45394d4ceb70267762cba1a2dcdead602cb4ea8f6869e0b6
3
  size 5304