craa commited on
Commit
fbecc55
·
verified ·
1 Parent(s): 32ba9bd

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:ff371d044c592a7f5e755118b3e47ae6dfeba30cf34d548e91b8b7d86cc25a69
3
  size 503128704
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1186ab15b37f3af75e8593c0df8d10b21b1f492224b93a24637c5efbbfde26f0
3
  size 503128704
checkpoint-10000/optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:acc92f6e8213312c0abd91cd32d8fbd9d593a81d955060cce1c1fd00d5a21ea7
3
  size 1006351290
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:886faa3bb3143878fa180462540bc948cea769582c57602afc7efc673397fbd0
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:d5a1fe5bee24227b7dce0d4a17d5f8cc7090fcf2e93c899e3ee53fcaed2d3402
3
  size 14244
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:95d415f5a2495efee5f0bd162340bac975693e203cab718136e9d616a832b275
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": 3.754211902618408,
3
- "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_634/checkpoint-10000",
4
- "epoch": 1.0763104079216446,
5
  "eval_steps": 1000,
6
  "global_step": 10000,
7
  "is_hyper_param_search": false,
@@ -9,1498 +9,1498 @@
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.005381552039608223,
13
- "grad_norm": 2.294102191925049,
14
  "learning_rate": 0.0003,
15
- "loss": 8.4956,
16
  "step": 50
17
  },
18
  {
19
- "epoch": 0.010763104079216447,
20
- "grad_norm": 3.396308660507202,
21
  "learning_rate": 0.0006,
22
- "loss": 6.9359,
23
  "step": 100
24
  },
25
  {
26
- "epoch": 0.01614465611882467,
27
- "grad_norm": 5.035550117492676,
28
- "learning_rate": 0.0005996767589699385,
29
- "loss": 6.4715,
30
  "step": 150
31
  },
32
  {
33
- "epoch": 0.021526208158432893,
34
- "grad_norm": 1.6986103057861328,
35
- "learning_rate": 0.0005993535179398771,
36
- "loss": 6.2706,
37
  "step": 200
38
  },
39
  {
40
- "epoch": 0.026907760198041114,
41
- "grad_norm": 1.4676276445388794,
42
- "learning_rate": 0.0005990302769098158,
43
- "loss": 6.0985,
44
  "step": 250
45
  },
46
  {
47
- "epoch": 0.03228931223764934,
48
- "grad_norm": 1.4348349571228027,
49
- "learning_rate": 0.0005987070358797543,
50
- "loss": 5.9704,
51
  "step": 300
52
  },
53
  {
54
- "epoch": 0.03767086427725756,
55
- "grad_norm": 1.31458580493927,
56
- "learning_rate": 0.0005983837948496929,
57
- "loss": 5.8747,
58
  "step": 350
59
  },
60
  {
61
- "epoch": 0.04305241631686579,
62
- "grad_norm": 1.2392261028289795,
63
- "learning_rate": 0.0005980605538196314,
64
- "loss": 5.8011,
65
  "step": 400
66
  },
67
  {
68
- "epoch": 0.048433968356474004,
69
- "grad_norm": 1.4158153533935547,
70
- "learning_rate": 0.0005977373127895701,
71
- "loss": 5.7301,
72
  "step": 450
73
  },
74
  {
75
- "epoch": 0.05381552039608223,
76
- "grad_norm": 1.0133819580078125,
77
- "learning_rate": 0.0005974140717595086,
78
- "loss": 5.6779,
79
  "step": 500
80
  },
81
  {
82
- "epoch": 0.05919707243569045,
83
- "grad_norm": 0.9000638723373413,
84
- "learning_rate": 0.0005970908307294472,
85
- "loss": 5.6116,
86
  "step": 550
87
  },
88
  {
89
- "epoch": 0.06457862447529868,
90
- "grad_norm": 0.8626642823219299,
91
- "learning_rate": 0.0005967675896993858,
92
- "loss": 5.5394,
93
  "step": 600
94
  },
95
  {
96
- "epoch": 0.0699601765149069,
97
- "grad_norm": 1.1039506196975708,
98
- "learning_rate": 0.0005964443486693243,
99
- "loss": 5.4684,
100
  "step": 650
101
  },
102
  {
103
- "epoch": 0.07534172855451512,
104
- "grad_norm": 1.2943150997161865,
105
- "learning_rate": 0.000596121107639263,
106
- "loss": 5.3917,
107
  "step": 700
108
  },
109
  {
110
- "epoch": 0.08072328059412334,
111
- "grad_norm": 1.446522831916809,
112
- "learning_rate": 0.0005957978666092015,
113
- "loss": 5.3201,
114
  "step": 750
115
  },
116
  {
117
- "epoch": 0.08610483263373157,
118
- "grad_norm": 1.090991735458374,
119
- "learning_rate": 0.0005954746255791401,
120
- "loss": 5.2782,
121
  "step": 800
122
  },
123
  {
124
- "epoch": 0.09148638467333979,
125
- "grad_norm": 1.2055628299713135,
126
- "learning_rate": 0.0005951513845490787,
127
- "loss": 5.2487,
128
  "step": 850
129
  },
130
  {
131
- "epoch": 0.09686793671294801,
132
- "grad_norm": 1.0582399368286133,
133
- "learning_rate": 0.0005948281435190174,
134
- "loss": 5.1728,
135
  "step": 900
136
  },
137
  {
138
- "epoch": 0.10224948875255624,
139
- "grad_norm": 1.1412220001220703,
140
- "learning_rate": 0.0005945049024889559,
141
- "loss": 5.1545,
142
  "step": 950
143
  },
144
  {
145
- "epoch": 0.10763104079216446,
146
- "grad_norm": 1.7075587511062622,
147
- "learning_rate": 0.0005941816614588944,
148
- "loss": 5.1117,
149
  "step": 1000
150
  },
151
  {
152
- "epoch": 0.10763104079216446,
153
- "eval_accuracy": 0.225164095740144,
154
- "eval_loss": 5.035466194152832,
155
- "eval_runtime": 180.4781,
156
- "eval_samples_per_second": 99.796,
157
- "eval_steps_per_second": 6.239,
158
  "step": 1000
159
  },
160
  {
161
- "epoch": 0.11301259283177269,
162
- "grad_norm": 0.9164252877235413,
163
- "learning_rate": 0.000593858420428833,
164
- "loss": 5.0652,
165
  "step": 1050
166
  },
167
  {
168
- "epoch": 0.1183941448713809,
169
- "grad_norm": 1.255126714706421,
170
- "learning_rate": 0.0005935351793987716,
171
- "loss": 5.0574,
172
  "step": 1100
173
  },
174
  {
175
- "epoch": 0.12377569691098914,
176
- "grad_norm": 1.0072600841522217,
177
- "learning_rate": 0.0005932119383687103,
178
- "loss": 5.002,
179
  "step": 1150
180
  },
181
  {
182
- "epoch": 0.12915724895059735,
183
- "grad_norm": 0.8915390968322754,
184
- "learning_rate": 0.0005928886973386488,
185
- "loss": 4.9898,
186
  "step": 1200
187
  },
188
  {
189
- "epoch": 0.13453880099020557,
190
- "grad_norm": 0.6740373373031616,
191
- "learning_rate": 0.0005925654563085874,
192
- "loss": 4.9383,
193
  "step": 1250
194
  },
195
  {
196
- "epoch": 0.1399203530298138,
197
- "grad_norm": 1.7642277479171753,
198
- "learning_rate": 0.000592242215278526,
199
- "loss": 4.9236,
200
  "step": 1300
201
  },
202
  {
203
- "epoch": 0.14530190506942203,
204
- "grad_norm": 0.900189995765686,
205
- "learning_rate": 0.0005919189742484645,
206
- "loss": 4.8889,
207
  "step": 1350
208
  },
209
  {
210
- "epoch": 0.15068345710903025,
211
- "grad_norm": 1.2182579040527344,
212
- "learning_rate": 0.0005915957332184032,
213
- "loss": 4.8465,
214
  "step": 1400
215
  },
216
  {
217
- "epoch": 0.15606500914863847,
218
- "grad_norm": 0.9330576062202454,
219
- "learning_rate": 0.0005912724921883417,
220
- "loss": 4.8151,
221
  "step": 1450
222
  },
223
  {
224
- "epoch": 0.16144656118824668,
225
- "grad_norm": 1.1736196279525757,
226
- "learning_rate": 0.0005909492511582803,
227
- "loss": 4.8093,
228
  "step": 1500
229
  },
230
  {
231
- "epoch": 0.1668281132278549,
232
- "grad_norm": 1.142849087715149,
233
- "learning_rate": 0.0005906260101282189,
234
- "loss": 4.7839,
235
  "step": 1550
236
  },
237
  {
238
- "epoch": 0.17220966526746315,
239
- "grad_norm": 0.9821141958236694,
240
- "learning_rate": 0.0005903027690981575,
241
- "loss": 4.7579,
242
  "step": 1600
243
  },
244
  {
245
- "epoch": 0.17759121730707136,
246
- "grad_norm": 0.786464512348175,
247
- "learning_rate": 0.000589979528068096,
248
- "loss": 4.7527,
249
  "step": 1650
250
  },
251
  {
252
- "epoch": 0.18297276934667958,
253
- "grad_norm": 1.4257930517196655,
254
- "learning_rate": 0.0005896562870380347,
255
- "loss": 4.7186,
256
  "step": 1700
257
  },
258
  {
259
- "epoch": 0.1883543213862878,
260
- "grad_norm": 0.895348072052002,
261
- "learning_rate": 0.0005893330460079732,
262
- "loss": 4.6987,
263
  "step": 1750
264
  },
265
  {
266
- "epoch": 0.19373587342589602,
267
- "grad_norm": 0.9841587543487549,
268
- "learning_rate": 0.0005890098049779118,
269
- "loss": 4.6786,
270
  "step": 1800
271
  },
272
  {
273
- "epoch": 0.19911742546550426,
274
- "grad_norm": 0.9945094585418701,
275
- "learning_rate": 0.0005886865639478504,
276
- "loss": 4.6727,
277
  "step": 1850
278
  },
279
  {
280
- "epoch": 0.20449897750511248,
281
- "grad_norm": 0.9739004969596863,
282
- "learning_rate": 0.0005883633229177889,
283
- "loss": 4.6508,
284
  "step": 1900
285
  },
286
  {
287
- "epoch": 0.2098805295447207,
288
- "grad_norm": 1.296126127243042,
289
- "learning_rate": 0.0005880400818877276,
290
- "loss": 4.6205,
291
  "step": 1950
292
  },
293
  {
294
- "epoch": 0.2152620815843289,
295
- "grad_norm": 1.191341519355774,
296
- "learning_rate": 0.0005877168408576662,
297
- "loss": 4.6104,
298
  "step": 2000
299
  },
300
  {
301
- "epoch": 0.2152620815843289,
302
- "eval_accuracy": 0.26726350422295747,
303
- "eval_loss": 4.536436557769775,
304
- "eval_runtime": 180.36,
305
- "eval_samples_per_second": 99.861,
306
- "eval_steps_per_second": 6.243,
307
  "step": 2000
308
  },
309
  {
310
- "epoch": 0.22064363362393713,
311
- "grad_norm": 0.8131170272827148,
312
- "learning_rate": 0.0005873935998276048,
313
- "loss": 4.5915,
314
  "step": 2050
315
  },
316
  {
317
- "epoch": 0.22602518566354537,
318
- "grad_norm": 0.9375983476638794,
319
- "learning_rate": 0.0005870703587975433,
320
- "loss": 4.559,
321
  "step": 2100
322
  },
323
  {
324
- "epoch": 0.2314067377031536,
325
- "grad_norm": 0.8687012791633606,
326
- "learning_rate": 0.0005867471177674818,
327
- "loss": 4.5487,
328
  "step": 2150
329
  },
330
  {
331
- "epoch": 0.2367882897427618,
332
- "grad_norm": 0.8746796250343323,
333
- "learning_rate": 0.0005864238767374205,
334
- "loss": 4.5421,
335
  "step": 2200
336
  },
337
  {
338
- "epoch": 0.24216984178237003,
339
- "grad_norm": 0.9759683012962341,
340
- "learning_rate": 0.0005861006357073591,
341
- "loss": 4.508,
342
  "step": 2250
343
  },
344
  {
345
- "epoch": 0.24755139382197827,
346
- "grad_norm": 0.8916616439819336,
347
- "learning_rate": 0.0005857773946772977,
348
- "loss": 4.502,
349
  "step": 2300
350
  },
351
  {
352
- "epoch": 0.2529329458615865,
353
- "grad_norm": 0.9090847373008728,
354
- "learning_rate": 0.0005854541536472362,
355
- "loss": 4.499,
356
  "step": 2350
357
  },
358
  {
359
- "epoch": 0.2583144979011947,
360
- "grad_norm": 0.9077420830726624,
361
- "learning_rate": 0.0005851309126171749,
362
- "loss": 4.4837,
363
  "step": 2400
364
  },
365
  {
366
- "epoch": 0.2636960499408029,
367
- "grad_norm": 0.7771352529525757,
368
- "learning_rate": 0.0005848076715871134,
369
- "loss": 4.4573,
370
  "step": 2450
371
  },
372
  {
373
- "epoch": 0.26907760198041114,
374
- "grad_norm": 0.9242041707038879,
375
- "learning_rate": 0.000584484430557052,
376
- "loss": 4.4445,
377
  "step": 2500
378
  },
379
  {
380
- "epoch": 0.27445915402001936,
381
- "grad_norm": 0.8875961899757385,
382
- "learning_rate": 0.0005841611895269906,
383
- "loss": 4.4196,
384
  "step": 2550
385
  },
386
  {
387
- "epoch": 0.2798407060596276,
388
- "grad_norm": 1.0621362924575806,
389
- "learning_rate": 0.0005838379484969291,
390
- "loss": 4.424,
391
  "step": 2600
392
  },
393
  {
394
- "epoch": 0.2852222580992358,
395
- "grad_norm": 1.0123540163040161,
396
- "learning_rate": 0.0005835147074668678,
397
- "loss": 4.4031,
398
  "step": 2650
399
  },
400
  {
401
- "epoch": 0.29060381013884407,
402
- "grad_norm": 1.0421875715255737,
403
- "learning_rate": 0.0005831914664368063,
404
- "loss": 4.3787,
405
  "step": 2700
406
  },
407
  {
408
- "epoch": 0.2959853621784523,
409
- "grad_norm": 0.643415093421936,
410
- "learning_rate": 0.0005828682254067449,
411
- "loss": 4.3686,
412
  "step": 2750
413
  },
414
  {
415
- "epoch": 0.3013669142180605,
416
- "grad_norm": 0.7626638412475586,
417
- "learning_rate": 0.0005825449843766835,
418
- "loss": 4.3589,
419
  "step": 2800
420
  },
421
  {
422
- "epoch": 0.3067484662576687,
423
- "grad_norm": 0.9982723593711853,
424
- "learning_rate": 0.0005822217433466221,
425
- "loss": 4.3466,
426
  "step": 2850
427
  },
428
  {
429
- "epoch": 0.31213001829727693,
430
- "grad_norm": 0.7696640491485596,
431
- "learning_rate": 0.0005818985023165607,
432
- "loss": 4.328,
433
  "step": 2900
434
  },
435
  {
436
- "epoch": 0.31751157033688515,
437
- "grad_norm": 0.8518490195274353,
438
- "learning_rate": 0.0005815752612864992,
439
- "loss": 4.3432,
440
  "step": 2950
441
  },
442
  {
443
- "epoch": 0.32289312237649337,
444
- "grad_norm": 0.7042719125747681,
445
- "learning_rate": 0.0005812520202564378,
446
- "loss": 4.3309,
447
  "step": 3000
448
  },
449
  {
450
- "epoch": 0.32289312237649337,
451
- "eval_accuracy": 0.2979527296919332,
452
- "eval_loss": 4.2461700439453125,
453
- "eval_runtime": 179.8358,
454
- "eval_samples_per_second": 100.152,
455
- "eval_steps_per_second": 6.261,
456
  "step": 3000
457
  },
458
  {
459
- "epoch": 0.3282746744161016,
460
- "grad_norm": 0.8338158130645752,
461
- "learning_rate": 0.0005809287792263764,
462
- "loss": 4.3163,
463
  "step": 3050
464
  },
465
  {
466
- "epoch": 0.3336562264557098,
467
- "grad_norm": 0.8053140640258789,
468
- "learning_rate": 0.0005806055381963151,
469
- "loss": 4.3087,
470
  "step": 3100
471
  },
472
  {
473
- "epoch": 0.3390377784953181,
474
- "grad_norm": 0.8919686079025269,
475
- "learning_rate": 0.0005802822971662536,
476
- "loss": 4.2929,
477
  "step": 3150
478
  },
479
  {
480
- "epoch": 0.3444193305349263,
481
- "grad_norm": 0.9998557567596436,
482
- "learning_rate": 0.0005799590561361922,
483
- "loss": 4.2964,
484
  "step": 3200
485
  },
486
  {
487
- "epoch": 0.3498008825745345,
488
- "grad_norm": 0.8141055703163147,
489
- "learning_rate": 0.0005796358151061307,
490
- "loss": 4.2749,
491
  "step": 3250
492
  },
493
  {
494
- "epoch": 0.35518243461414273,
495
- "grad_norm": 0.9093042016029358,
496
- "learning_rate": 0.0005793125740760694,
497
- "loss": 4.2734,
498
  "step": 3300
499
  },
500
  {
501
- "epoch": 0.36056398665375095,
502
- "grad_norm": 0.7931560277938843,
503
- "learning_rate": 0.0005789893330460079,
504
- "loss": 4.2897,
505
  "step": 3350
506
  },
507
  {
508
- "epoch": 0.36594553869335916,
509
- "grad_norm": 0.8265503644943237,
510
- "learning_rate": 0.0005786660920159465,
511
- "loss": 4.2547,
512
  "step": 3400
513
  },
514
  {
515
- "epoch": 0.3713270907329674,
516
- "grad_norm": 0.6019992232322693,
517
- "learning_rate": 0.0005783428509858851,
518
- "loss": 4.2259,
519
  "step": 3450
520
  },
521
  {
522
- "epoch": 0.3767086427725756,
523
- "grad_norm": 0.899000883102417,
524
- "learning_rate": 0.0005780196099558237,
525
- "loss": 4.2346,
526
  "step": 3500
527
  },
528
  {
529
- "epoch": 0.3820901948121838,
530
- "grad_norm": 0.7591268420219421,
531
- "learning_rate": 0.0005776963689257623,
532
- "loss": 4.2446,
533
  "step": 3550
534
  },
535
  {
536
- "epoch": 0.38747174685179203,
537
- "grad_norm": 0.8733161687850952,
538
- "learning_rate": 0.0005773731278957008,
539
- "loss": 4.2186,
540
  "step": 3600
541
  },
542
  {
543
- "epoch": 0.3928532988914003,
544
- "grad_norm": 0.667341947555542,
545
- "learning_rate": 0.0005770498868656394,
546
- "loss": 4.2291,
547
  "step": 3650
548
  },
549
  {
550
- "epoch": 0.3982348509310085,
551
- "grad_norm": 0.7878106832504272,
552
- "learning_rate": 0.000576726645835578,
553
- "loss": 4.2193,
554
  "step": 3700
555
  },
556
  {
557
- "epoch": 0.40361640297061674,
558
- "grad_norm": 0.7062278389930725,
559
- "learning_rate": 0.0005764034048055167,
560
- "loss": 4.2084,
561
  "step": 3750
562
  },
563
  {
564
- "epoch": 0.40899795501022496,
565
- "grad_norm": 0.7661077380180359,
566
- "learning_rate": 0.0005760801637754552,
567
- "loss": 4.1841,
568
  "step": 3800
569
  },
570
  {
571
- "epoch": 0.4143795070498332,
572
- "grad_norm": 0.7327717542648315,
573
- "learning_rate": 0.0005757569227453937,
574
- "loss": 4.1827,
575
  "step": 3850
576
  },
577
  {
578
- "epoch": 0.4197610590894414,
579
- "grad_norm": 0.7855814695358276,
580
- "learning_rate": 0.0005754336817153324,
581
- "loss": 4.1715,
582
  "step": 3900
583
  },
584
  {
585
- "epoch": 0.4251426111290496,
586
- "grad_norm": 0.6937480568885803,
587
- "learning_rate": 0.0005751104406852709,
588
- "loss": 4.1816,
589
  "step": 3950
590
  },
591
  {
592
- "epoch": 0.4305241631686578,
593
- "grad_norm": 0.6745932102203369,
594
- "learning_rate": 0.0005747871996552096,
595
- "loss": 4.1676,
596
  "step": 4000
597
  },
598
  {
599
- "epoch": 0.4305241631686578,
600
- "eval_accuracy": 0.31133887412356503,
601
- "eval_loss": 4.100213050842285,
602
- "eval_runtime": 179.8152,
603
- "eval_samples_per_second": 100.164,
604
- "eval_steps_per_second": 6.262,
605
  "step": 4000
606
  },
607
  {
608
- "epoch": 0.43590571520826604,
609
- "grad_norm": 0.6731157302856445,
610
- "learning_rate": 0.0005744639586251481,
611
- "loss": 4.1742,
612
  "step": 4050
613
  },
614
  {
615
- "epoch": 0.44128726724787426,
616
- "grad_norm": 0.7058612704277039,
617
- "learning_rate": 0.0005741407175950867,
618
- "loss": 4.1643,
619
  "step": 4100
620
  },
621
  {
622
- "epoch": 0.44666881928748253,
623
- "grad_norm": 0.8731959462165833,
624
- "learning_rate": 0.0005738174765650253,
625
- "loss": 4.149,
626
  "step": 4150
627
  },
628
  {
629
- "epoch": 0.45205037132709075,
630
- "grad_norm": 0.6776666641235352,
631
- "learning_rate": 0.0005734942355349638,
632
- "loss": 4.1449,
633
  "step": 4200
634
  },
635
  {
636
- "epoch": 0.45743192336669897,
637
- "grad_norm": 0.5808095932006836,
638
- "learning_rate": 0.0005731709945049025,
639
- "loss": 4.1439,
640
  "step": 4250
641
  },
642
  {
643
- "epoch": 0.4628134754063072,
644
- "grad_norm": 0.6728511452674866,
645
- "learning_rate": 0.000572847753474841,
646
- "loss": 4.1451,
647
  "step": 4300
648
  },
649
  {
650
- "epoch": 0.4681950274459154,
651
- "grad_norm": 0.7343346476554871,
652
- "learning_rate": 0.0005725245124447796,
653
- "loss": 4.1303,
654
  "step": 4350
655
  },
656
  {
657
- "epoch": 0.4735765794855236,
658
- "grad_norm": 0.6369326114654541,
659
- "learning_rate": 0.0005722012714147182,
660
- "loss": 4.1164,
661
  "step": 4400
662
  },
663
  {
664
- "epoch": 0.47895813152513184,
665
- "grad_norm": 0.7540531158447266,
666
- "learning_rate": 0.0005718780303846568,
667
- "loss": 4.1296,
668
  "step": 4450
669
  },
670
  {
671
- "epoch": 0.48433968356474005,
672
- "grad_norm": 0.7194720506668091,
673
- "learning_rate": 0.0005715547893545953,
674
- "loss": 4.1096,
675
  "step": 4500
676
  },
677
  {
678
- "epoch": 0.48972123560434827,
679
- "grad_norm": 0.826496958732605,
680
- "learning_rate": 0.000571231548324534,
681
- "loss": 4.1017,
682
  "step": 4550
683
  },
684
  {
685
- "epoch": 0.49510278764395654,
686
- "grad_norm": 0.6337757706642151,
687
- "learning_rate": 0.0005709083072944725,
688
- "loss": 4.0974,
689
  "step": 4600
690
  },
691
  {
692
- "epoch": 0.5004843396835648,
693
- "grad_norm": 0.8857693076133728,
694
- "learning_rate": 0.0005705850662644111,
695
- "loss": 4.1239,
696
  "step": 4650
697
  },
698
  {
699
- "epoch": 0.505865891723173,
700
- "grad_norm": 0.6664023995399475,
701
- "learning_rate": 0.0005702618252343497,
702
- "loss": 4.1111,
703
  "step": 4700
704
  },
705
  {
706
- "epoch": 0.5112474437627812,
707
- "grad_norm": 0.6844426393508911,
708
- "learning_rate": 0.0005699385842042882,
709
- "loss": 4.1124,
710
  "step": 4750
711
  },
712
  {
713
- "epoch": 0.5166289958023894,
714
- "grad_norm": 0.6479195356369019,
715
- "learning_rate": 0.0005696153431742269,
716
- "loss": 4.0813,
717
  "step": 4800
718
  },
719
  {
720
- "epoch": 0.5220105478419976,
721
- "grad_norm": 0.7041398286819458,
722
- "learning_rate": 0.0005692921021441655,
723
- "loss": 4.094,
724
  "step": 4850
725
  },
726
  {
727
- "epoch": 0.5273920998816058,
728
- "grad_norm": 0.7125672698020935,
729
- "learning_rate": 0.0005689688611141041,
730
- "loss": 4.091,
731
  "step": 4900
732
  },
733
  {
734
- "epoch": 0.5327736519212141,
735
- "grad_norm": 0.6662024259567261,
736
- "learning_rate": 0.0005686456200840426,
737
- "loss": 4.0695,
738
  "step": 4950
739
  },
740
  {
741
- "epoch": 0.5381552039608223,
742
- "grad_norm": 0.7144469022750854,
743
- "learning_rate": 0.0005683223790539811,
744
- "loss": 4.0578,
745
  "step": 5000
746
  },
747
  {
748
- "epoch": 0.5381552039608223,
749
- "eval_accuracy": 0.3208241625768814,
750
- "eval_loss": 3.9998252391815186,
751
- "eval_runtime": 179.9108,
752
- "eval_samples_per_second": 100.111,
753
- "eval_steps_per_second": 6.259,
754
  "step": 5000
755
  },
756
  {
757
- "epoch": 0.5435367560004305,
758
- "grad_norm": 0.7586029171943665,
759
- "learning_rate": 0.0005679991380239198,
760
- "loss": 4.0765,
761
  "step": 5050
762
  },
763
  {
764
- "epoch": 0.5489183080400387,
765
- "grad_norm": 0.6306418776512146,
766
- "learning_rate": 0.0005676758969938584,
767
- "loss": 4.0695,
768
  "step": 5100
769
  },
770
  {
771
- "epoch": 0.5542998600796469,
772
- "grad_norm": 0.659724771976471,
773
- "learning_rate": 0.000567352655963797,
774
- "loss": 4.056,
775
  "step": 5150
776
  },
777
  {
778
- "epoch": 0.5596814121192552,
779
- "grad_norm": 0.688515305519104,
780
- "learning_rate": 0.0005670294149337355,
781
- "loss": 4.0575,
782
  "step": 5200
783
  },
784
  {
785
- "epoch": 0.5650629641588634,
786
- "grad_norm": 0.6124618053436279,
787
- "learning_rate": 0.0005667061739036742,
788
- "loss": 4.0477,
789
  "step": 5250
790
  },
791
  {
792
- "epoch": 0.5704445161984716,
793
- "grad_norm": 0.7170193791389465,
794
- "learning_rate": 0.0005663829328736127,
795
- "loss": 4.0722,
796
  "step": 5300
797
  },
798
  {
799
- "epoch": 0.5758260682380799,
800
- "grad_norm": 0.6210808753967285,
801
- "learning_rate": 0.0005660596918435512,
802
- "loss": 4.052,
803
  "step": 5350
804
  },
805
  {
806
- "epoch": 0.5812076202776881,
807
- "grad_norm": 0.6430804133415222,
808
- "learning_rate": 0.0005657364508134899,
809
- "loss": 4.0487,
810
  "step": 5400
811
  },
812
  {
813
- "epoch": 0.5865891723172963,
814
- "grad_norm": 0.5880076289176941,
815
- "learning_rate": 0.0005654132097834284,
816
- "loss": 4.0476,
817
  "step": 5450
818
  },
819
  {
820
- "epoch": 0.5919707243569046,
821
- "grad_norm": 0.749442458152771,
822
- "learning_rate": 0.0005650899687533671,
823
- "loss": 4.0475,
824
  "step": 5500
825
  },
826
  {
827
- "epoch": 0.5973522763965128,
828
- "grad_norm": 0.6980018019676208,
829
- "learning_rate": 0.0005647667277233056,
830
- "loss": 4.0195,
831
  "step": 5550
832
  },
833
  {
834
- "epoch": 0.602733828436121,
835
- "grad_norm": 0.5972914695739746,
836
- "learning_rate": 0.0005644434866932442,
837
- "loss": 4.0375,
838
  "step": 5600
839
  },
840
  {
841
- "epoch": 0.6081153804757292,
842
- "grad_norm": 0.6123340725898743,
843
- "learning_rate": 0.0005641202456631828,
844
- "loss": 4.0233,
845
  "step": 5650
846
  },
847
  {
848
- "epoch": 0.6134969325153374,
849
- "grad_norm": 0.7154938578605652,
850
- "learning_rate": 0.0005637970046331214,
851
- "loss": 4.0195,
852
  "step": 5700
853
  },
854
  {
855
- "epoch": 0.6188784845549457,
856
- "grad_norm": 0.6205494403839111,
857
- "learning_rate": 0.00056347376360306,
858
- "loss": 4.0221,
859
  "step": 5750
860
  },
861
  {
862
- "epoch": 0.6242600365945539,
863
- "grad_norm": 0.8683055639266968,
864
- "learning_rate": 0.0005631505225729985,
865
- "loss": 3.9937,
866
  "step": 5800
867
  },
868
  {
869
- "epoch": 0.6296415886341621,
870
- "grad_norm": 0.65059894323349,
871
- "learning_rate": 0.0005628272815429371,
872
- "loss": 4.0127,
873
  "step": 5850
874
  },
875
  {
876
- "epoch": 0.6350231406737703,
877
- "grad_norm": 0.6759449243545532,
878
- "learning_rate": 0.0005625040405128757,
879
- "loss": 4.0054,
880
  "step": 5900
881
  },
882
  {
883
- "epoch": 0.6404046927133785,
884
- "grad_norm": 0.7678074240684509,
885
- "learning_rate": 0.0005621807994828143,
886
- "loss": 3.9864,
887
  "step": 5950
888
  },
889
  {
890
- "epoch": 0.6457862447529867,
891
- "grad_norm": 0.5478359460830688,
892
- "learning_rate": 0.0005618575584527529,
893
- "loss": 3.9918,
894
  "step": 6000
895
  },
896
  {
897
- "epoch": 0.6457862447529867,
898
- "eval_accuracy": 0.3270703997915603,
899
- "eval_loss": 3.9250869750976562,
900
- "eval_runtime": 179.6502,
901
- "eval_samples_per_second": 100.256,
902
- "eval_steps_per_second": 6.268,
903
  "step": 6000
904
  },
905
  {
906
- "epoch": 0.651167796792595,
907
- "grad_norm": 0.7417057156562805,
908
- "learning_rate": 0.0005615343174226915,
909
- "loss": 4.0032,
910
  "step": 6050
911
  },
912
  {
913
- "epoch": 0.6565493488322032,
914
- "grad_norm": 0.6435531973838806,
915
- "learning_rate": 0.00056121107639263,
916
- "loss": 4.0167,
917
  "step": 6100
918
  },
919
  {
920
- "epoch": 0.6619309008718114,
921
- "grad_norm": 0.6030955910682678,
922
- "learning_rate": 0.0005608878353625687,
923
- "loss": 3.9908,
924
  "step": 6150
925
  },
926
  {
927
- "epoch": 0.6673124529114196,
928
- "grad_norm": 0.6270812749862671,
929
- "learning_rate": 0.0005605645943325072,
930
- "loss": 3.9982,
931
  "step": 6200
932
  },
933
  {
934
- "epoch": 0.6726940049510278,
935
- "grad_norm": 0.6616153717041016,
936
- "learning_rate": 0.0005602413533024458,
937
- "loss": 3.9777,
938
  "step": 6250
939
  },
940
  {
941
- "epoch": 0.6780755569906362,
942
- "grad_norm": 0.599023699760437,
943
- "learning_rate": 0.0005599245770929855,
944
- "loss": 3.9858,
945
  "step": 6300
946
  },
947
  {
948
- "epoch": 0.6834571090302444,
949
- "grad_norm": 0.6816189289093018,
950
- "learning_rate": 0.0005596013360629242,
951
- "loss": 3.9869,
952
  "step": 6350
953
  },
954
  {
955
- "epoch": 0.6888386610698526,
956
- "grad_norm": 0.5487352013587952,
957
- "learning_rate": 0.0005592780950328628,
958
- "loss": 3.953,
959
  "step": 6400
960
  },
961
  {
962
- "epoch": 0.6942202131094608,
963
- "grad_norm": 0.6630198359489441,
964
- "learning_rate": 0.0005589548540028014,
965
- "loss": 3.9784,
966
  "step": 6450
967
  },
968
  {
969
- "epoch": 0.699601765149069,
970
- "grad_norm": 0.5939404368400574,
971
- "learning_rate": 0.0005586316129727399,
972
- "loss": 3.9889,
973
  "step": 6500
974
  },
975
  {
976
- "epoch": 0.7049833171886772,
977
- "grad_norm": 0.5617170929908752,
978
- "learning_rate": 0.0005583083719426786,
979
- "loss": 3.9585,
980
  "step": 6550
981
  },
982
  {
983
- "epoch": 0.7103648692282855,
984
- "grad_norm": 0.5913287997245789,
985
- "learning_rate": 0.0005579851309126171,
986
- "loss": 3.9621,
987
  "step": 6600
988
  },
989
  {
990
- "epoch": 0.7157464212678937,
991
- "grad_norm": 0.600002110004425,
992
- "learning_rate": 0.0005576618898825558,
993
- "loss": 3.9674,
994
  "step": 6650
995
  },
996
  {
997
- "epoch": 0.7211279733075019,
998
- "grad_norm": 0.67916339635849,
999
- "learning_rate": 0.0005573386488524943,
1000
- "loss": 3.9554,
1001
  "step": 6700
1002
  },
1003
  {
1004
- "epoch": 0.7265095253471101,
1005
- "grad_norm": 0.6280518770217896,
1006
- "learning_rate": 0.0005570154078224328,
1007
- "loss": 3.9578,
1008
  "step": 6750
1009
  },
1010
  {
1011
- "epoch": 0.7318910773867183,
1012
- "grad_norm": 0.6209385395050049,
1013
- "learning_rate": 0.0005566921667923715,
1014
- "loss": 3.9503,
1015
  "step": 6800
1016
  },
1017
  {
1018
- "epoch": 0.7372726294263265,
1019
- "grad_norm": 0.6955409646034241,
1020
- "learning_rate": 0.00055636892576231,
1021
- "loss": 3.9534,
1022
  "step": 6850
1023
  },
1024
  {
1025
- "epoch": 0.7426541814659348,
1026
- "grad_norm": 0.5965715646743774,
1027
- "learning_rate": 0.0005560456847322487,
1028
- "loss": 3.9575,
1029
  "step": 6900
1030
  },
1031
  {
1032
- "epoch": 0.748035733505543,
1033
- "grad_norm": 0.7255541682243347,
1034
- "learning_rate": 0.0005557224437021872,
1035
- "loss": 3.9444,
1036
  "step": 6950
1037
  },
1038
  {
1039
- "epoch": 0.7534172855451512,
1040
- "grad_norm": 0.6498315334320068,
1041
- "learning_rate": 0.0005553992026721258,
1042
- "loss": 3.9315,
1043
  "step": 7000
1044
  },
1045
  {
1046
- "epoch": 0.7534172855451512,
1047
- "eval_accuracy": 0.33304120193562947,
1048
- "eval_loss": 3.866844892501831,
1049
- "eval_runtime": 180.0856,
1050
- "eval_samples_per_second": 100.014,
1051
- "eval_steps_per_second": 6.253,
1052
  "step": 7000
1053
  },
1054
  {
1055
- "epoch": 0.7587988375847594,
1056
- "grad_norm": 0.6578865051269531,
1057
- "learning_rate": 0.0005550759616420644,
1058
- "loss": 3.9476,
1059
  "step": 7050
1060
  },
1061
  {
1062
- "epoch": 0.7641803896243676,
1063
- "grad_norm": 0.6418218016624451,
1064
- "learning_rate": 0.000554752720612003,
1065
- "loss": 3.9236,
1066
  "step": 7100
1067
  },
1068
  {
1069
- "epoch": 0.7695619416639758,
1070
- "grad_norm": 0.5991949439048767,
1071
- "learning_rate": 0.0005544294795819415,
1072
- "loss": 3.9395,
1073
  "step": 7150
1074
  },
1075
  {
1076
- "epoch": 0.7749434937035841,
1077
- "grad_norm": 0.707750141620636,
1078
- "learning_rate": 0.0005541062385518801,
1079
- "loss": 3.9288,
1080
  "step": 7200
1081
  },
1082
  {
1083
- "epoch": 0.7803250457431924,
1084
- "grad_norm": 0.5853620767593384,
1085
- "learning_rate": 0.0005537829975218188,
1086
- "loss": 3.9419,
1087
  "step": 7250
1088
  },
1089
  {
1090
- "epoch": 0.7857065977828006,
1091
- "grad_norm": 0.731132447719574,
1092
- "learning_rate": 0.0005534597564917573,
1093
- "loss": 3.9326,
1094
  "step": 7300
1095
  },
1096
  {
1097
- "epoch": 0.7910881498224088,
1098
- "grad_norm": 0.5787903666496277,
1099
- "learning_rate": 0.0005531365154616959,
1100
- "loss": 3.9306,
1101
  "step": 7350
1102
  },
1103
  {
1104
- "epoch": 0.796469701862017,
1105
- "grad_norm": 0.591372549533844,
1106
- "learning_rate": 0.0005528132744316344,
1107
- "loss": 3.9172,
1108
  "step": 7400
1109
  },
1110
  {
1111
- "epoch": 0.8018512539016253,
1112
- "grad_norm": 0.6081792116165161,
1113
- "learning_rate": 0.0005524900334015731,
1114
- "loss": 3.9169,
1115
  "step": 7450
1116
  },
1117
  {
1118
- "epoch": 0.8072328059412335,
1119
- "grad_norm": 0.6385130882263184,
1120
- "learning_rate": 0.0005521667923715117,
1121
- "loss": 3.9087,
1122
  "step": 7500
1123
  },
1124
  {
1125
- "epoch": 0.8126143579808417,
1126
- "grad_norm": 0.5836992859840393,
1127
- "learning_rate": 0.0005518435513414502,
1128
- "loss": 3.9171,
1129
  "step": 7550
1130
  },
1131
  {
1132
- "epoch": 0.8179959100204499,
1133
- "grad_norm": 0.6080932021141052,
1134
- "learning_rate": 0.0005515203103113888,
1135
- "loss": 3.8927,
1136
  "step": 7600
1137
  },
1138
  {
1139
- "epoch": 0.8233774620600581,
1140
- "grad_norm": 0.654813289642334,
1141
- "learning_rate": 0.0005511970692813274,
1142
- "loss": 3.8979,
1143
  "step": 7650
1144
  },
1145
  {
1146
- "epoch": 0.8287590140996663,
1147
- "grad_norm": 0.5635251402854919,
1148
- "learning_rate": 0.000550873828251266,
1149
- "loss": 3.892,
1150
  "step": 7700
1151
  },
1152
  {
1153
- "epoch": 0.8341405661392746,
1154
- "grad_norm": 0.6040819883346558,
1155
- "learning_rate": 0.0005505505872212045,
1156
- "loss": 3.9078,
1157
  "step": 7750
1158
  },
1159
  {
1160
- "epoch": 0.8395221181788828,
1161
- "grad_norm": 0.6047422289848328,
1162
- "learning_rate": 0.0005502273461911432,
1163
- "loss": 3.914,
1164
  "step": 7800
1165
  },
1166
  {
1167
- "epoch": 0.844903670218491,
1168
- "grad_norm": 0.6018202304840088,
1169
- "learning_rate": 0.0005499041051610817,
1170
- "loss": 3.874,
1171
  "step": 7850
1172
  },
1173
  {
1174
- "epoch": 0.8502852222580992,
1175
- "grad_norm": 0.6617713570594788,
1176
- "learning_rate": 0.0005495808641310204,
1177
- "loss": 3.8877,
1178
  "step": 7900
1179
  },
1180
  {
1181
- "epoch": 0.8556667742977074,
1182
- "grad_norm": 0.7833295464515686,
1183
- "learning_rate": 0.0005492576231009589,
1184
- "loss": 3.9032,
1185
  "step": 7950
1186
  },
1187
  {
1188
- "epoch": 0.8610483263373157,
1189
- "grad_norm": 0.537348210811615,
1190
- "learning_rate": 0.0005489343820708974,
1191
- "loss": 3.9053,
1192
  "step": 8000
1193
  },
1194
  {
1195
- "epoch": 0.8610483263373157,
1196
- "eval_accuracy": 0.3369586818057806,
1197
- "eval_loss": 3.823279857635498,
1198
- "eval_runtime": 179.8613,
1199
- "eval_samples_per_second": 100.138,
1200
- "eval_steps_per_second": 6.26,
1201
  "step": 8000
1202
  },
1203
  {
1204
- "epoch": 0.8664298783769239,
1205
- "grad_norm": 0.620795488357544,
1206
- "learning_rate": 0.0005486111410408361,
1207
- "loss": 3.8719,
1208
  "step": 8050
1209
  },
1210
  {
1211
- "epoch": 0.8718114304165321,
1212
- "grad_norm": 0.6803261041641235,
1213
- "learning_rate": 0.0005482879000107746,
1214
- "loss": 3.8846,
1215
  "step": 8100
1216
  },
1217
  {
1218
- "epoch": 0.8771929824561403,
1219
- "grad_norm": 0.6699694395065308,
1220
- "learning_rate": 0.0005479646589807133,
1221
- "loss": 3.8805,
1222
  "step": 8150
1223
  },
1224
  {
1225
- "epoch": 0.8825745344957485,
1226
- "grad_norm": 0.6116239428520203,
1227
- "learning_rate": 0.0005476414179506518,
1228
- "loss": 3.8947,
1229
  "step": 8200
1230
  },
1231
  {
1232
- "epoch": 0.8879560865353568,
1233
- "grad_norm": 0.6413629651069641,
1234
- "learning_rate": 0.0005473181769205904,
1235
- "loss": 3.8797,
1236
  "step": 8250
1237
  },
1238
  {
1239
- "epoch": 0.8933376385749651,
1240
- "grad_norm": 0.553399920463562,
1241
- "learning_rate": 0.000546994935890529,
1242
- "loss": 3.882,
1243
  "step": 8300
1244
  },
1245
  {
1246
- "epoch": 0.8987191906145733,
1247
- "grad_norm": 0.6522381901741028,
1248
- "learning_rate": 0.0005466716948604677,
1249
- "loss": 3.8845,
1250
  "step": 8350
1251
  },
1252
  {
1253
- "epoch": 0.9041007426541815,
1254
- "grad_norm": 0.7224483489990234,
1255
- "learning_rate": 0.0005463484538304062,
1256
- "loss": 3.847,
1257
  "step": 8400
1258
  },
1259
  {
1260
- "epoch": 0.9094822946937897,
1261
- "grad_norm": 0.5507568717002869,
1262
- "learning_rate": 0.0005460252128003447,
1263
- "loss": 3.8674,
1264
  "step": 8450
1265
  },
1266
  {
1267
- "epoch": 0.9148638467333979,
1268
- "grad_norm": 0.5474512577056885,
1269
- "learning_rate": 0.0005457019717702833,
1270
- "loss": 3.8637,
1271
  "step": 8500
1272
  },
1273
  {
1274
- "epoch": 0.9202453987730062,
1275
- "grad_norm": 0.6044310927391052,
1276
- "learning_rate": 0.0005453787307402219,
1277
- "loss": 3.8567,
1278
  "step": 8550
1279
  },
1280
  {
1281
- "epoch": 0.9256269508126144,
1282
- "grad_norm": 0.558305025100708,
1283
- "learning_rate": 0.0005450554897101605,
1284
- "loss": 3.8654,
1285
  "step": 8600
1286
  },
1287
  {
1288
- "epoch": 0.9310085028522226,
1289
- "grad_norm": 0.5505136251449585,
1290
- "learning_rate": 0.0005447322486800991,
1291
- "loss": 3.8563,
1292
  "step": 8650
1293
  },
1294
  {
1295
- "epoch": 0.9363900548918308,
1296
- "grad_norm": 0.5557839274406433,
1297
- "learning_rate": 0.0005444090076500377,
1298
- "loss": 3.8643,
1299
  "step": 8700
1300
  },
1301
  {
1302
- "epoch": 0.941771606931439,
1303
- "grad_norm": 0.670755922794342,
1304
- "learning_rate": 0.0005440857666199763,
1305
- "loss": 3.8776,
1306
  "step": 8750
1307
  },
1308
  {
1309
- "epoch": 0.9471531589710472,
1310
- "grad_norm": 0.6244636178016663,
1311
- "learning_rate": 0.0005437625255899148,
1312
- "loss": 3.852,
1313
  "step": 8800
1314
  },
1315
  {
1316
- "epoch": 0.9525347110106555,
1317
- "grad_norm": 0.5825973749160767,
1318
- "learning_rate": 0.0005434392845598534,
1319
- "loss": 3.8492,
1320
  "step": 8850
1321
  },
1322
  {
1323
- "epoch": 0.9579162630502637,
1324
- "grad_norm": 0.6198579668998718,
1325
- "learning_rate": 0.000543116043529792,
1326
- "loss": 3.8572,
1327
  "step": 8900
1328
  },
1329
  {
1330
- "epoch": 0.9632978150898719,
1331
- "grad_norm": 0.6518577337265015,
1332
- "learning_rate": 0.0005427928024997306,
1333
- "loss": 3.8591,
1334
  "step": 8950
1335
  },
1336
  {
1337
- "epoch": 0.9686793671294801,
1338
- "grad_norm": 0.6103642582893372,
1339
- "learning_rate": 0.0005424695614696692,
1340
- "loss": 3.8444,
1341
  "step": 9000
1342
  },
1343
  {
1344
- "epoch": 0.9686793671294801,
1345
- "eval_accuracy": 0.3405125004604166,
1346
- "eval_loss": 3.7843289375305176,
1347
- "eval_runtime": 179.4726,
1348
- "eval_samples_per_second": 100.355,
1349
- "eval_steps_per_second": 6.274,
1350
  "step": 9000
1351
  },
1352
  {
1353
- "epoch": 0.9740609191690883,
1354
- "grad_norm": 0.5499141216278076,
1355
- "learning_rate": 0.0005421463204396078,
1356
- "loss": 3.863,
1357
  "step": 9050
1358
  },
1359
  {
1360
- "epoch": 0.9794424712086965,
1361
- "grad_norm": 0.4996525049209595,
1362
- "learning_rate": 0.0005418230794095463,
1363
- "loss": 3.8328,
1364
  "step": 9100
1365
  },
1366
  {
1367
- "epoch": 0.9848240232483048,
1368
- "grad_norm": 0.5601321458816528,
1369
- "learning_rate": 0.000541499838379485,
1370
- "loss": 3.8549,
1371
  "step": 9150
1372
  },
1373
  {
1374
- "epoch": 0.9902055752879131,
1375
- "grad_norm": 0.5210615992546082,
1376
- "learning_rate": 0.0005411765973494235,
1377
- "loss": 3.8508,
1378
  "step": 9200
1379
  },
1380
  {
1381
- "epoch": 0.9955871273275213,
1382
- "grad_norm": 0.601481556892395,
1383
- "learning_rate": 0.0005408533563193621,
1384
- "loss": 3.8402,
1385
  "step": 9250
1386
  },
1387
  {
1388
- "epoch": 1.0009686793671295,
1389
- "grad_norm": 0.5878329873085022,
1390
- "learning_rate": 0.0005405301152893007,
1391
- "loss": 3.8311,
1392
  "step": 9300
1393
  },
1394
  {
1395
- "epoch": 1.0063502314067376,
1396
- "grad_norm": 0.5495496392250061,
1397
- "learning_rate": 0.0005402068742592392,
1398
- "loss": 3.783,
1399
  "step": 9350
1400
  },
1401
  {
1402
- "epoch": 1.011731783446346,
1403
- "grad_norm": 0.5719645619392395,
1404
- "learning_rate": 0.0005398836332291779,
1405
- "loss": 3.7853,
1406
  "step": 9400
1407
  },
1408
  {
1409
- "epoch": 1.017113335485954,
1410
- "grad_norm": 0.6130096912384033,
1411
- "learning_rate": 0.0005395603921991164,
1412
- "loss": 3.7837,
1413
  "step": 9450
1414
  },
1415
  {
1416
- "epoch": 1.0224948875255624,
1417
- "grad_norm": 0.5828426480293274,
1418
- "learning_rate": 0.0005392371511690551,
1419
- "loss": 3.7777,
1420
  "step": 9500
1421
  },
1422
  {
1423
- "epoch": 1.0278764395651705,
1424
- "grad_norm": 0.5985187888145447,
1425
- "learning_rate": 0.0005389139101389936,
1426
- "loss": 3.7825,
1427
  "step": 9550
1428
  },
1429
  {
1430
- "epoch": 1.0332579916047788,
1431
- "grad_norm": 0.6139611005783081,
1432
- "learning_rate": 0.0005385906691089321,
1433
- "loss": 3.7628,
1434
  "step": 9600
1435
  },
1436
  {
1437
- "epoch": 1.0386395436443872,
1438
- "grad_norm": 0.5611709952354431,
1439
- "learning_rate": 0.0005382674280788708,
1440
- "loss": 3.7829,
1441
  "step": 9650
1442
  },
1443
  {
1444
- "epoch": 1.0440210956839953,
1445
- "grad_norm": 0.5671082735061646,
1446
- "learning_rate": 0.0005379506518694106,
1447
- "loss": 3.7744,
1448
  "step": 9700
1449
  },
1450
  {
1451
- "epoch": 1.0494026477236036,
1452
- "grad_norm": 0.7483752965927124,
1453
- "learning_rate": 0.0005376274108393491,
1454
- "loss": 3.7686,
1455
  "step": 9750
1456
  },
1457
  {
1458
- "epoch": 1.0547841997632117,
1459
- "grad_norm": 0.5836924910545349,
1460
- "learning_rate": 0.0005373041698092877,
1461
- "loss": 3.7736,
1462
  "step": 9800
1463
  },
1464
  {
1465
- "epoch": 1.06016575180282,
1466
- "grad_norm": 0.6311327815055847,
1467
- "learning_rate": 0.0005369809287792263,
1468
- "loss": 3.7538,
1469
  "step": 9850
1470
  },
1471
  {
1472
- "epoch": 1.0655473038424281,
1473
- "grad_norm": 0.6404955983161926,
1474
- "learning_rate": 0.000536657687749165,
1475
- "loss": 3.7843,
1476
  "step": 9900
1477
  },
1478
  {
1479
- "epoch": 1.0709288558820365,
1480
- "grad_norm": 0.5680845975875854,
1481
- "learning_rate": 0.0005363344467191035,
1482
- "loss": 3.7835,
1483
  "step": 9950
1484
  },
1485
  {
1486
- "epoch": 1.0763104079216446,
1487
- "grad_norm": 0.5650578141212463,
1488
- "learning_rate": 0.000536011205689042,
1489
- "loss": 3.7939,
1490
  "step": 10000
1491
  },
1492
  {
1493
- "epoch": 1.0763104079216446,
1494
- "eval_accuracy": 0.344210175538519,
1495
- "eval_loss": 3.754211902618408,
1496
- "eval_runtime": 179.651,
1497
- "eval_samples_per_second": 100.256,
1498
- "eval_steps_per_second": 6.268,
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": 3.7440903186798096,
3
+ "best_model_checkpoint": "/scratch/cl5625/exceptions/models/100M_low_2000_634/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": 1.3841798305511475,
14
  "learning_rate": 0.0003,
15
+ "loss": 8.6258,
16
  "step": 50
17
  },
18
  {
19
+ "epoch": 0.01078167115902965,
20
+ "grad_norm": 2.7142674922943115,
21
  "learning_rate": 0.0006,
22
+ "loss": 6.9518,
23
  "step": 100
24
  },
25
  {
26
+ "epoch": 0.016172506738544475,
27
+ "grad_norm": 1.7501500844955444,
28
+ "learning_rate": 0.0005996762007555315,
29
+ "loss": 6.4766,
30
  "step": 150
31
  },
32
  {
33
+ "epoch": 0.0215633423180593,
34
+ "grad_norm": 2.089492082595825,
35
+ "learning_rate": 0.000599352401511063,
36
+ "loss": 6.2242,
37
  "step": 200
38
  },
39
  {
40
+ "epoch": 0.026954177897574125,
41
+ "grad_norm": 1.0795507431030273,
42
+ "learning_rate": 0.0005990286022665946,
43
+ "loss": 6.0814,
44
  "step": 250
45
  },
46
  {
47
+ "epoch": 0.03234501347708895,
48
+ "grad_norm": 2.468144655227661,
49
+ "learning_rate": 0.0005987048030221263,
50
+ "loss": 5.9861,
51
  "step": 300
52
  },
53
  {
54
+ "epoch": 0.03773584905660377,
55
+ "grad_norm": 1.4130438566207886,
56
+ "learning_rate": 0.0005983810037776578,
57
+ "loss": 5.8758,
58
  "step": 350
59
  },
60
  {
61
+ "epoch": 0.0431266846361186,
62
+ "grad_norm": 1.5332931280136108,
63
+ "learning_rate": 0.0005980572045331894,
64
+ "loss": 5.7848,
65
  "step": 400
66
  },
67
  {
68
+ "epoch": 0.04851752021563342,
69
+ "grad_norm": 1.451593279838562,
70
+ "learning_rate": 0.0005977334052887209,
71
+ "loss": 5.713,
72
  "step": 450
73
  },
74
  {
75
+ "epoch": 0.05390835579514825,
76
+ "grad_norm": 1.3739091157913208,
77
+ "learning_rate": 0.0005974096060442526,
78
+ "loss": 5.6508,
79
  "step": 500
80
  },
81
  {
82
+ "epoch": 0.05929919137466307,
83
+ "grad_norm": 1.2970751523971558,
84
+ "learning_rate": 0.0005970858067997841,
85
+ "loss": 5.5723,
86
  "step": 550
87
  },
88
  {
89
+ "epoch": 0.0646900269541779,
90
+ "grad_norm": 1.2038410902023315,
91
+ "learning_rate": 0.0005967620075553157,
92
+ "loss": 5.5093,
93
  "step": 600
94
  },
95
  {
96
+ "epoch": 0.07008086253369272,
97
+ "grad_norm": 1.2364015579223633,
98
+ "learning_rate": 0.0005964382083108472,
99
+ "loss": 5.4167,
100
  "step": 650
101
  },
102
  {
103
+ "epoch": 0.07547169811320754,
104
+ "grad_norm": 1.1776434183120728,
105
+ "learning_rate": 0.0005961144090663788,
106
+ "loss": 5.3687,
107
  "step": 700
108
  },
109
  {
110
+ "epoch": 0.08086253369272237,
111
+ "grad_norm": 1.2404123544692993,
112
+ "learning_rate": 0.0005957906098219104,
113
+ "loss": 5.2965,
114
  "step": 750
115
  },
116
  {
117
+ "epoch": 0.0862533692722372,
118
+ "grad_norm": 0.8565037250518799,
119
+ "learning_rate": 0.0005954668105774419,
120
+ "loss": 5.2536,
121
  "step": 800
122
  },
123
  {
124
+ "epoch": 0.09164420485175202,
125
+ "grad_norm": 0.8625679612159729,
126
+ "learning_rate": 0.0005951430113329735,
127
+ "loss": 5.1917,
128
  "step": 850
129
  },
130
  {
131
+ "epoch": 0.09703504043126684,
132
+ "grad_norm": 1.4676177501678467,
133
+ "learning_rate": 0.0005948192120885051,
134
+ "loss": 5.1705,
135
  "step": 900
136
  },
137
  {
138
+ "epoch": 0.10242587601078167,
139
+ "grad_norm": 1.3620243072509766,
140
+ "learning_rate": 0.0005944954128440366,
141
+ "loss": 5.129,
142
  "step": 950
143
  },
144
  {
145
+ "epoch": 0.1078167115902965,
146
+ "grad_norm": 1.802571177482605,
147
+ "learning_rate": 0.0005941716135995682,
148
+ "loss": 5.0781,
149
  "step": 1000
150
  },
151
  {
152
+ "epoch": 0.1078167115902965,
153
+ "eval_accuracy": 0.22607515020446844,
154
+ "eval_loss": 5.030185699462891,
155
+ "eval_runtime": 184.9577,
156
+ "eval_samples_per_second": 97.379,
157
+ "eval_steps_per_second": 6.088,
158
  "step": 1000
159
  },
160
  {
161
+ "epoch": 0.11320754716981132,
162
+ "grad_norm": 1.0640990734100342,
163
+ "learning_rate": 0.0005938478143550997,
164
+ "loss": 5.0384,
165
  "step": 1050
166
  },
167
  {
168
+ "epoch": 0.11859838274932614,
169
+ "grad_norm": 0.9512579441070557,
170
+ "learning_rate": 0.0005935240151106314,
171
+ "loss": 5.01,
172
  "step": 1100
173
  },
174
  {
175
+ "epoch": 0.12398921832884097,
176
+ "grad_norm": 1.4336494207382202,
177
+ "learning_rate": 0.0005932002158661629,
178
+ "loss": 4.986,
179
  "step": 1150
180
  },
181
  {
182
+ "epoch": 0.1293800539083558,
183
+ "grad_norm": 0.986595094203949,
184
+ "learning_rate": 0.0005928764166216945,
185
+ "loss": 4.9237,
186
  "step": 1200
187
  },
188
  {
189
+ "epoch": 0.1347708894878706,
190
+ "grad_norm": 0.9924365878105164,
191
+ "learning_rate": 0.000592552617377226,
192
+ "loss": 4.8993,
193
  "step": 1250
194
  },
195
  {
196
+ "epoch": 0.14016172506738545,
197
+ "grad_norm": 1.405576229095459,
198
+ "learning_rate": 0.0005922288181327577,
199
+ "loss": 4.8778,
200
  "step": 1300
201
  },
202
  {
203
+ "epoch": 0.14555256064690028,
204
+ "grad_norm": 0.7532362341880798,
205
+ "learning_rate": 0.0005919050188882893,
206
+ "loss": 4.8412,
207
  "step": 1350
208
  },
209
  {
210
+ "epoch": 0.1509433962264151,
211
+ "grad_norm": 0.9290658235549927,
212
+ "learning_rate": 0.0005915812196438207,
213
+ "loss": 4.8361,
214
  "step": 1400
215
  },
216
  {
217
+ "epoch": 0.15633423180592992,
218
+ "grad_norm": 1.0212785005569458,
219
+ "learning_rate": 0.0005912574203993524,
220
+ "loss": 4.8431,
221
  "step": 1450
222
  },
223
  {
224
+ "epoch": 0.16172506738544473,
225
+ "grad_norm": 1.3435090780258179,
226
+ "learning_rate": 0.0005909336211548839,
227
+ "loss": 4.801,
228
  "step": 1500
229
  },
230
  {
231
+ "epoch": 0.16711590296495957,
232
+ "grad_norm": 1.0386897325515747,
233
+ "learning_rate": 0.0005906098219104155,
234
+ "loss": 4.7558,
235
  "step": 1550
236
  },
237
  {
238
+ "epoch": 0.1725067385444744,
239
+ "grad_norm": 0.8441452383995056,
240
+ "learning_rate": 0.000590286022665947,
241
+ "loss": 4.749,
242
  "step": 1600
243
  },
244
  {
245
+ "epoch": 0.1778975741239892,
246
+ "grad_norm": 0.9853036999702454,
247
+ "learning_rate": 0.0005899622234214787,
248
+ "loss": 4.7201,
249
  "step": 1650
250
  },
251
  {
252
+ "epoch": 0.18328840970350405,
253
+ "grad_norm": 0.9956970810890198,
254
+ "learning_rate": 0.0005896384241770102,
255
+ "loss": 4.6933,
256
  "step": 1700
257
  },
258
  {
259
+ "epoch": 0.18867924528301888,
260
+ "grad_norm": 0.8019234538078308,
261
+ "learning_rate": 0.0005893146249325418,
262
+ "loss": 4.7021,
263
  "step": 1750
264
  },
265
  {
266
+ "epoch": 0.1940700808625337,
267
+ "grad_norm": 0.8837598562240601,
268
+ "learning_rate": 0.0005889908256880733,
269
+ "loss": 4.6704,
270
  "step": 1800
271
  },
272
  {
273
+ "epoch": 0.19946091644204852,
274
+ "grad_norm": 1.046995997428894,
275
+ "learning_rate": 0.0005886670264436049,
276
+ "loss": 4.6374,
277
  "step": 1850
278
  },
279
  {
280
+ "epoch": 0.20485175202156333,
281
+ "grad_norm": 1.0166348218917847,
282
+ "learning_rate": 0.0005883432271991365,
283
+ "loss": 4.6279,
284
  "step": 1900
285
  },
286
  {
287
+ "epoch": 0.21024258760107817,
288
+ "grad_norm": 0.8063094615936279,
289
+ "learning_rate": 0.0005880194279546681,
290
+ "loss": 4.6256,
291
  "step": 1950
292
  },
293
  {
294
+ "epoch": 0.215633423180593,
295
+ "grad_norm": 0.9398548603057861,
296
+ "learning_rate": 0.0005876956287101996,
297
+ "loss": 4.5763,
298
  "step": 2000
299
  },
300
  {
301
+ "epoch": 0.215633423180593,
302
+ "eval_accuracy": 0.2714252357849155,
303
+ "eval_loss": 4.502427577972412,
304
+ "eval_runtime": 183.5351,
305
+ "eval_samples_per_second": 98.134,
306
+ "eval_steps_per_second": 6.135,
307
  "step": 2000
308
  },
309
  {
310
+ "epoch": 0.2210242587601078,
311
+ "grad_norm": 0.980785608291626,
312
+ "learning_rate": 0.0005873718294657312,
313
+ "loss": 4.5575,
314
  "step": 2050
315
  },
316
  {
317
+ "epoch": 0.22641509433962265,
318
+ "grad_norm": 0.8288772106170654,
319
+ "learning_rate": 0.0005870480302212628,
320
+ "loss": 4.5462,
321
  "step": 2100
322
  },
323
  {
324
+ "epoch": 0.23180592991913745,
325
+ "grad_norm": 0.9006871581077576,
326
+ "learning_rate": 0.0005867242309767943,
327
+ "loss": 4.5289,
328
  "step": 2150
329
  },
330
  {
331
+ "epoch": 0.2371967654986523,
332
+ "grad_norm": 0.7568821310997009,
333
+ "learning_rate": 0.0005864004317323259,
334
+ "loss": 4.4877,
335
  "step": 2200
336
  },
337
  {
338
+ "epoch": 0.24258760107816713,
339
+ "grad_norm": 0.8699207305908203,
340
+ "learning_rate": 0.0005860766324878575,
341
+ "loss": 4.5022,
342
  "step": 2250
343
  },
344
  {
345
+ "epoch": 0.24797843665768193,
346
+ "grad_norm": 0.9681375026702881,
347
+ "learning_rate": 0.000585752833243389,
348
+ "loss": 4.4781,
349
  "step": 2300
350
  },
351
  {
352
+ "epoch": 0.25336927223719674,
353
+ "grad_norm": 1.0356428623199463,
354
+ "learning_rate": 0.0005854290339989206,
355
+ "loss": 4.4547,
356
  "step": 2350
357
  },
358
  {
359
+ "epoch": 0.2587601078167116,
360
+ "grad_norm": 1.0217968225479126,
361
+ "learning_rate": 0.0005851052347544521,
362
+ "loss": 4.4496,
363
  "step": 2400
364
  },
365
  {
366
+ "epoch": 0.2641509433962264,
367
+ "grad_norm": 0.9390143752098083,
368
+ "learning_rate": 0.0005847814355099838,
369
+ "loss": 4.4324,
370
  "step": 2450
371
  },
372
  {
373
+ "epoch": 0.2695417789757412,
374
+ "grad_norm": 0.8502791523933411,
375
+ "learning_rate": 0.0005844576362655154,
376
+ "loss": 4.4238,
377
  "step": 2500
378
  },
379
  {
380
+ "epoch": 0.2749326145552561,
381
+ "grad_norm": 0.9871562719345093,
382
+ "learning_rate": 0.0005841338370210469,
383
+ "loss": 4.3975,
384
  "step": 2550
385
  },
386
  {
387
+ "epoch": 0.2803234501347709,
388
+ "grad_norm": 0.9953110814094543,
389
+ "learning_rate": 0.0005838100377765785,
390
+ "loss": 4.3725,
391
  "step": 2600
392
  },
393
  {
394
+ "epoch": 0.2857142857142857,
395
+ "grad_norm": 0.8370280861854553,
396
+ "learning_rate": 0.0005834862385321101,
397
+ "loss": 4.3714,
398
  "step": 2650
399
  },
400
  {
401
+ "epoch": 0.29110512129380056,
402
+ "grad_norm": 0.8906695246696472,
403
+ "learning_rate": 0.0005831624392876417,
404
+ "loss": 4.3636,
405
  "step": 2700
406
  },
407
  {
408
+ "epoch": 0.29649595687331537,
409
+ "grad_norm": 0.8926903009414673,
410
+ "learning_rate": 0.0005828386400431731,
411
+ "loss": 4.3712,
412
  "step": 2750
413
  },
414
  {
415
+ "epoch": 0.3018867924528302,
416
+ "grad_norm": 0.6380210518836975,
417
+ "learning_rate": 0.0005825148407987048,
418
+ "loss": 4.3652,
419
  "step": 2800
420
  },
421
  {
422
+ "epoch": 0.30727762803234504,
423
+ "grad_norm": 0.9012649655342102,
424
+ "learning_rate": 0.0005821910415542363,
425
+ "loss": 4.3264,
426
  "step": 2850
427
  },
428
  {
429
+ "epoch": 0.31266846361185985,
430
+ "grad_norm": 0.8436964750289917,
431
+ "learning_rate": 0.0005818672423097679,
432
+ "loss": 4.3127,
433
  "step": 2900
434
  },
435
  {
436
+ "epoch": 0.31805929919137466,
437
+ "grad_norm": 0.8248408436775208,
438
+ "learning_rate": 0.0005815434430652994,
439
+ "loss": 4.3424,
440
  "step": 2950
441
  },
442
  {
443
+ "epoch": 0.32345013477088946,
444
+ "grad_norm": 0.7463982105255127,
445
+ "learning_rate": 0.0005812196438208311,
446
+ "loss": 4.3107,
447
  "step": 3000
448
  },
449
  {
450
+ "epoch": 0.32345013477088946,
451
+ "eval_accuracy": 0.2991547565898248,
452
+ "eval_loss": 4.231179237365723,
453
+ "eval_runtime": 183.5046,
454
+ "eval_samples_per_second": 98.15,
455
+ "eval_steps_per_second": 6.136,
456
  "step": 3000
457
  },
458
  {
459
+ "epoch": 0.3288409703504043,
460
+ "grad_norm": 0.7356308102607727,
461
+ "learning_rate": 0.0005808958445763626,
462
+ "loss": 4.3028,
463
  "step": 3050
464
  },
465
  {
466
+ "epoch": 0.33423180592991913,
467
+ "grad_norm": 0.7332613468170166,
468
+ "learning_rate": 0.0005805720453318942,
469
+ "loss": 4.281,
470
  "step": 3100
471
  },
472
  {
473
+ "epoch": 0.33962264150943394,
474
+ "grad_norm": 0.9393066763877869,
475
+ "learning_rate": 0.0005802482460874257,
476
+ "loss": 4.2707,
477
  "step": 3150
478
  },
479
  {
480
+ "epoch": 0.3450134770889488,
481
+ "grad_norm": 0.7228361964225769,
482
+ "learning_rate": 0.0005799244468429573,
483
+ "loss": 4.2769,
484
  "step": 3200
485
  },
486
  {
487
+ "epoch": 0.3504043126684636,
488
+ "grad_norm": 0.7841880917549133,
489
+ "learning_rate": 0.0005796006475984889,
490
+ "loss": 4.2499,
491
  "step": 3250
492
  },
493
  {
494
+ "epoch": 0.3557951482479784,
495
+ "grad_norm": 0.7926186919212341,
496
+ "learning_rate": 0.0005792768483540205,
497
+ "loss": 4.2548,
498
  "step": 3300
499
  },
500
  {
501
+ "epoch": 0.3611859838274933,
502
+ "grad_norm": 0.8690990805625916,
503
+ "learning_rate": 0.000578953049109552,
504
+ "loss": 4.254,
505
  "step": 3350
506
  },
507
  {
508
+ "epoch": 0.3665768194070081,
509
+ "grad_norm": 0.7845708727836609,
510
+ "learning_rate": 0.0005786292498650836,
511
+ "loss": 4.2325,
512
  "step": 3400
513
  },
514
  {
515
+ "epoch": 0.3719676549865229,
516
+ "grad_norm": 0.8615543246269226,
517
+ "learning_rate": 0.0005783054506206152,
518
+ "loss": 4.2247,
519
  "step": 3450
520
  },
521
  {
522
+ "epoch": 0.37735849056603776,
523
+ "grad_norm": 0.7429792284965515,
524
+ "learning_rate": 0.0005779816513761467,
525
+ "loss": 4.216,
526
  "step": 3500
527
  },
528
  {
529
+ "epoch": 0.38274932614555257,
530
+ "grad_norm": 0.7726993560791016,
531
+ "learning_rate": 0.0005776578521316782,
532
+ "loss": 4.1962,
533
  "step": 3550
534
  },
535
  {
536
+ "epoch": 0.3881401617250674,
537
+ "grad_norm": 0.7264295220375061,
538
+ "learning_rate": 0.0005773340528872099,
539
+ "loss": 4.2044,
540
  "step": 3600
541
  },
542
  {
543
+ "epoch": 0.3935309973045822,
544
+ "grad_norm": 0.7918074131011963,
545
+ "learning_rate": 0.0005770102536427414,
546
+ "loss": 4.219,
547
  "step": 3650
548
  },
549
  {
550
+ "epoch": 0.39892183288409705,
551
+ "grad_norm": 0.7423751950263977,
552
+ "learning_rate": 0.000576686454398273,
553
+ "loss": 4.2045,
554
  "step": 3700
555
  },
556
  {
557
+ "epoch": 0.40431266846361186,
558
+ "grad_norm": 0.6395682096481323,
559
+ "learning_rate": 0.0005763626551538045,
560
+ "loss": 4.1897,
561
  "step": 3750
562
  },
563
  {
564
+ "epoch": 0.40970350404312667,
565
+ "grad_norm": 0.6632921695709229,
566
+ "learning_rate": 0.0005760388559093362,
567
+ "loss": 4.182,
568
  "step": 3800
569
  },
570
  {
571
+ "epoch": 0.41509433962264153,
572
+ "grad_norm": 0.6251309514045715,
573
+ "learning_rate": 0.0005757150566648678,
574
+ "loss": 4.1772,
575
  "step": 3850
576
  },
577
  {
578
+ "epoch": 0.42048517520215634,
579
+ "grad_norm": 0.8533205986022949,
580
+ "learning_rate": 0.0005753912574203993,
581
+ "loss": 4.1699,
582
  "step": 3900
583
  },
584
  {
585
+ "epoch": 0.42587601078167114,
586
+ "grad_norm": 0.707554817199707,
587
+ "learning_rate": 0.0005750674581759309,
588
+ "loss": 4.1668,
589
  "step": 3950
590
  },
591
  {
592
+ "epoch": 0.431266846361186,
593
+ "grad_norm": 0.7947620153427124,
594
+ "learning_rate": 0.0005747436589314624,
595
+ "loss": 4.1524,
596
  "step": 4000
597
  },
598
  {
599
+ "epoch": 0.431266846361186,
600
+ "eval_accuracy": 0.3127696153503061,
601
+ "eval_loss": 4.084148406982422,
602
+ "eval_runtime": 183.63,
603
+ "eval_samples_per_second": 98.083,
604
+ "eval_steps_per_second": 6.132,
605
  "step": 4000
606
  },
607
  {
608
+ "epoch": 0.4366576819407008,
609
+ "grad_norm": 0.5997743606567383,
610
+ "learning_rate": 0.0005744198596869941,
611
+ "loss": 4.1636,
612
  "step": 4050
613
  },
614
  {
615
+ "epoch": 0.4420485175202156,
616
+ "grad_norm": 0.7036544680595398,
617
+ "learning_rate": 0.0005740960604425255,
618
+ "loss": 4.156,
619
  "step": 4100
620
  },
621
  {
622
+ "epoch": 0.4474393530997305,
623
+ "grad_norm": 0.7143358588218689,
624
+ "learning_rate": 0.0005737722611980572,
625
+ "loss": 4.1384,
626
  "step": 4150
627
  },
628
  {
629
+ "epoch": 0.4528301886792453,
630
+ "grad_norm": 0.6587009429931641,
631
+ "learning_rate": 0.0005734484619535887,
632
+ "loss": 4.113,
633
  "step": 4200
634
  },
635
  {
636
+ "epoch": 0.4582210242587601,
637
+ "grad_norm": 0.704142153263092,
638
+ "learning_rate": 0.0005731246627091203,
639
+ "loss": 4.1487,
640
  "step": 4250
641
  },
642
  {
643
+ "epoch": 0.4636118598382749,
644
+ "grad_norm": 0.5820618867874146,
645
+ "learning_rate": 0.0005728008634646518,
646
+ "loss": 4.1219,
647
  "step": 4300
648
  },
649
  {
650
+ "epoch": 0.46900269541778977,
651
+ "grad_norm": 0.6929563879966736,
652
+ "learning_rate": 0.0005724770642201835,
653
+ "loss": 4.1263,
654
  "step": 4350
655
  },
656
  {
657
+ "epoch": 0.4743935309973046,
658
+ "grad_norm": 0.7341722846031189,
659
+ "learning_rate": 0.000572153264975715,
660
+ "loss": 4.1214,
661
  "step": 4400
662
  },
663
  {
664
+ "epoch": 0.4797843665768194,
665
+ "grad_norm": 0.5978211164474487,
666
+ "learning_rate": 0.0005718294657312466,
667
+ "loss": 4.0946,
668
  "step": 4450
669
  },
670
  {
671
+ "epoch": 0.48517520215633425,
672
+ "grad_norm": 0.6326689124107361,
673
+ "learning_rate": 0.0005715056664867781,
674
+ "loss": 4.0929,
675
  "step": 4500
676
  },
677
  {
678
+ "epoch": 0.49056603773584906,
679
+ "grad_norm": 0.6336254477500916,
680
+ "learning_rate": 0.0005711818672423097,
681
+ "loss": 4.0867,
682
  "step": 4550
683
  },
684
  {
685
+ "epoch": 0.49595687331536387,
686
+ "grad_norm": 0.6374770998954773,
687
+ "learning_rate": 0.0005708580679978413,
688
+ "loss": 4.0936,
689
  "step": 4600
690
  },
691
  {
692
+ "epoch": 0.5013477088948787,
693
+ "grad_norm": 0.7158097624778748,
694
+ "learning_rate": 0.0005705342687533729,
695
+ "loss": 4.0868,
696
  "step": 4650
697
  },
698
  {
699
+ "epoch": 0.5067385444743935,
700
+ "grad_norm": 0.6559752225875854,
701
+ "learning_rate": 0.0005702104695089044,
702
+ "loss": 4.0694,
703
  "step": 4700
704
  },
705
  {
706
+ "epoch": 0.5121293800539084,
707
+ "grad_norm": 0.6569370627403259,
708
+ "learning_rate": 0.000569886670264436,
709
+ "loss": 4.0812,
710
  "step": 4750
711
  },
712
  {
713
+ "epoch": 0.5175202156334232,
714
+ "grad_norm": 0.6718046069145203,
715
+ "learning_rate": 0.0005695628710199675,
716
+ "loss": 4.0682,
717
  "step": 4800
718
  },
719
  {
720
+ "epoch": 0.522911051212938,
721
+ "grad_norm": 0.6529590487480164,
722
+ "learning_rate": 0.0005692390717754991,
723
+ "loss": 4.0635,
724
  "step": 4850
725
  },
726
  {
727
+ "epoch": 0.5283018867924528,
728
+ "grad_norm": 0.6065393686294556,
729
+ "learning_rate": 0.0005689152725310306,
730
+ "loss": 4.0566,
731
  "step": 4900
732
  },
733
  {
734
+ "epoch": 0.5336927223719676,
735
+ "grad_norm": 0.711345911026001,
736
+ "learning_rate": 0.0005685914732865623,
737
+ "loss": 4.0593,
738
  "step": 4950
739
  },
740
  {
741
+ "epoch": 0.5390835579514824,
742
+ "grad_norm": 0.6914499998092651,
743
+ "learning_rate": 0.0005682676740420939,
744
+ "loss": 4.0724,
745
  "step": 5000
746
  },
747
  {
748
+ "epoch": 0.5390835579514824,
749
+ "eval_accuracy": 0.32181616344262765,
750
+ "eval_loss": 3.98882794380188,
751
+ "eval_runtime": 183.5537,
752
+ "eval_samples_per_second": 98.124,
753
+ "eval_steps_per_second": 6.134,
754
  "step": 5000
755
  },
756
  {
757
+ "epoch": 0.5444743935309974,
758
+ "grad_norm": 0.6853660345077515,
759
+ "learning_rate": 0.0005679438747976254,
760
+ "loss": 4.0452,
761
  "step": 5050
762
  },
763
  {
764
+ "epoch": 0.5498652291105122,
765
+ "grad_norm": 0.6142929792404175,
766
+ "learning_rate": 0.000567620075553157,
767
+ "loss": 4.0442,
768
  "step": 5100
769
  },
770
  {
771
+ "epoch": 0.555256064690027,
772
+ "grad_norm": 0.6437095999717712,
773
+ "learning_rate": 0.0005672962763086886,
774
+ "loss": 4.0475,
775
  "step": 5150
776
  },
777
  {
778
+ "epoch": 0.5606469002695418,
779
+ "grad_norm": 0.7355647683143616,
780
+ "learning_rate": 0.0005669724770642202,
781
+ "loss": 4.0336,
782
  "step": 5200
783
  },
784
  {
785
+ "epoch": 0.5660377358490566,
786
+ "grad_norm": 0.6523656249046326,
787
+ "learning_rate": 0.0005666486778197517,
788
+ "loss": 4.0297,
789
  "step": 5250
790
  },
791
  {
792
+ "epoch": 0.5714285714285714,
793
+ "grad_norm": 0.6674039959907532,
794
+ "learning_rate": 0.0005663248785752833,
795
+ "loss": 4.0411,
796
  "step": 5300
797
  },
798
  {
799
+ "epoch": 0.5768194070080862,
800
+ "grad_norm": 0.599259078502655,
801
+ "learning_rate": 0.0005660010793308148,
802
+ "loss": 4.0249,
803
  "step": 5350
804
  },
805
  {
806
+ "epoch": 0.5822102425876011,
807
+ "grad_norm": 0.6059934496879578,
808
+ "learning_rate": 0.0005656772800863465,
809
+ "loss": 4.0416,
810
  "step": 5400
811
  },
812
  {
813
+ "epoch": 0.5876010781671159,
814
+ "grad_norm": 0.5647556185722351,
815
+ "learning_rate": 0.0005653534808418779,
816
+ "loss": 4.0184,
817
  "step": 5450
818
  },
819
  {
820
+ "epoch": 0.5929919137466307,
821
+ "grad_norm": 0.6619639992713928,
822
+ "learning_rate": 0.0005650296815974096,
823
+ "loss": 4.0231,
824
  "step": 5500
825
  },
826
  {
827
+ "epoch": 0.5983827493261455,
828
+ "grad_norm": 0.6255214810371399,
829
+ "learning_rate": 0.0005647058823529411,
830
+ "loss": 4.0157,
831
  "step": 5550
832
  },
833
  {
834
+ "epoch": 0.6037735849056604,
835
+ "grad_norm": 0.5285488963127136,
836
+ "learning_rate": 0.0005643820831084727,
837
+ "loss": 4.005,
838
  "step": 5600
839
  },
840
  {
841
+ "epoch": 0.6091644204851752,
842
+ "grad_norm": 0.8316670656204224,
843
+ "learning_rate": 0.0005640582838640042,
844
+ "loss": 4.0181,
845
  "step": 5650
846
  },
847
  {
848
+ "epoch": 0.6145552560646901,
849
+ "grad_norm": 0.7264182567596436,
850
+ "learning_rate": 0.0005637344846195358,
851
+ "loss": 4.0123,
852
  "step": 5700
853
  },
854
  {
855
+ "epoch": 0.6199460916442049,
856
+ "grad_norm": 0.6089359521865845,
857
+ "learning_rate": 0.0005634106853750674,
858
+ "loss": 4.0165,
859
  "step": 5750
860
  },
861
  {
862
+ "epoch": 0.6253369272237197,
863
+ "grad_norm": 0.769210696220398,
864
+ "learning_rate": 0.000563086886130599,
865
+ "loss": 3.9894,
866
  "step": 5800
867
  },
868
  {
869
+ "epoch": 0.6307277628032345,
870
+ "grad_norm": 0.7284867763519287,
871
+ "learning_rate": 0.0005627630868861305,
872
+ "loss": 3.9899,
873
  "step": 5850
874
  },
875
  {
876
+ "epoch": 0.6361185983827493,
877
+ "grad_norm": 0.604616105556488,
878
+ "learning_rate": 0.0005624392876416621,
879
+ "loss": 3.9758,
880
  "step": 5900
881
  },
882
  {
883
+ "epoch": 0.6415094339622641,
884
+ "grad_norm": 0.685613751411438,
885
+ "learning_rate": 0.0005621154883971937,
886
+ "loss": 3.9787,
887
  "step": 5950
888
  },
889
  {
890
+ "epoch": 0.6469002695417789,
891
+ "grad_norm": 0.685235321521759,
892
+ "learning_rate": 0.0005617916891527253,
893
+ "loss": 3.9609,
894
  "step": 6000
895
  },
896
  {
897
+ "epoch": 0.6469002695417789,
898
+ "eval_accuracy": 0.32830632639045004,
899
+ "eval_loss": 3.917071580886841,
900
+ "eval_runtime": 183.57,
901
+ "eval_samples_per_second": 98.115,
902
+ "eval_steps_per_second": 6.134,
903
  "step": 6000
904
  },
905
  {
906
+ "epoch": 0.6522911051212938,
907
+ "grad_norm": 0.6348211169242859,
908
+ "learning_rate": 0.0005614678899082568,
909
+ "loss": 3.9973,
910
  "step": 6050
911
  },
912
  {
913
+ "epoch": 0.6576819407008087,
914
+ "grad_norm": 0.5842711925506592,
915
+ "learning_rate": 0.0005611440906637884,
916
+ "loss": 3.9818,
917
  "step": 6100
918
  },
919
  {
920
+ "epoch": 0.6630727762803235,
921
+ "grad_norm": 0.6440193057060242,
922
+ "learning_rate": 0.00056082029141932,
923
+ "loss": 3.9752,
924
  "step": 6150
925
  },
926
  {
927
+ "epoch": 0.6684636118598383,
928
+ "grad_norm": 0.6132529377937317,
929
+ "learning_rate": 0.0005604964921748515,
930
+ "loss": 3.9854,
931
  "step": 6200
932
  },
933
  {
934
+ "epoch": 0.6738544474393531,
935
+ "grad_norm": 0.6459117531776428,
936
+ "learning_rate": 0.000560172692930383,
937
+ "loss": 3.9462,
938
  "step": 6250
939
  },
940
  {
941
+ "epoch": 0.6792452830188679,
942
+ "grad_norm": 0.6339119672775269,
943
+ "learning_rate": 0.0005598488936859147,
944
+ "loss": 3.9565,
945
  "step": 6300
946
  },
947
  {
948
+ "epoch": 0.6846361185983828,
949
+ "grad_norm": 0.6577672958374023,
950
+ "learning_rate": 0.0005595250944414463,
951
+ "loss": 3.9497,
952
  "step": 6350
953
  },
954
  {
955
+ "epoch": 0.6900269541778976,
956
+ "grad_norm": 0.6435871720314026,
957
+ "learning_rate": 0.0005592012951969778,
958
+ "loss": 3.9424,
959
  "step": 6400
960
  },
961
  {
962
+ "epoch": 0.6954177897574124,
963
+ "grad_norm": 0.7063788771629333,
964
+ "learning_rate": 0.0005588774959525094,
965
+ "loss": 3.9694,
966
  "step": 6450
967
  },
968
  {
969
+ "epoch": 0.7008086253369272,
970
+ "grad_norm": 0.5613532662391663,
971
+ "learning_rate": 0.000558553696708041,
972
+ "loss": 3.9543,
973
  "step": 6500
974
  },
975
  {
976
+ "epoch": 0.706199460916442,
977
+ "grad_norm": 0.559969425201416,
978
+ "learning_rate": 0.0005582298974635726,
979
+ "loss": 3.9512,
980
  "step": 6550
981
  },
982
  {
983
+ "epoch": 0.7115902964959568,
984
+ "grad_norm": 0.58429354429245,
985
+ "learning_rate": 0.0005579060982191041,
986
+ "loss": 3.9474,
987
  "step": 6600
988
  },
989
  {
990
+ "epoch": 0.7169811320754716,
991
+ "grad_norm": Infinity,
992
+ "learning_rate": 0.0005575887749595251,
993
+ "loss": 3.9414,
994
  "step": 6650
995
  },
996
  {
997
+ "epoch": 0.7223719676549866,
998
+ "grad_norm": 0.589440107345581,
999
+ "learning_rate": 0.0005572649757150566,
1000
+ "loss": 3.9341,
1001
  "step": 6700
1002
  },
1003
  {
1004
+ "epoch": 0.7277628032345014,
1005
+ "grad_norm": 0.7343171834945679,
1006
+ "learning_rate": 0.0005569411764705882,
1007
+ "loss": 3.9383,
1008
  "step": 6750
1009
  },
1010
  {
1011
+ "epoch": 0.7331536388140162,
1012
+ "grad_norm": 0.6029019355773926,
1013
+ "learning_rate": 0.0005566173772261198,
1014
+ "loss": 3.9469,
1015
  "step": 6800
1016
  },
1017
  {
1018
+ "epoch": 0.738544474393531,
1019
+ "grad_norm": 0.5020064115524292,
1020
+ "learning_rate": 0.0005562935779816513,
1021
+ "loss": 3.931,
1022
  "step": 6850
1023
  },
1024
  {
1025
+ "epoch": 0.7439353099730458,
1026
+ "grad_norm": 0.6130984425544739,
1027
+ "learning_rate": 0.0005559697787371828,
1028
+ "loss": 3.9264,
1029
  "step": 6900
1030
  },
1031
  {
1032
+ "epoch": 0.7493261455525606,
1033
+ "grad_norm": 0.6600217819213867,
1034
+ "learning_rate": 0.0005556459794927145,
1035
+ "loss": 3.9137,
1036
  "step": 6950
1037
  },
1038
  {
1039
+ "epoch": 0.7547169811320755,
1040
+ "grad_norm": 0.5343778133392334,
1041
+ "learning_rate": 0.000555322180248246,
1042
+ "loss": 3.9274,
1043
  "step": 7000
1044
  },
1045
  {
1046
+ "epoch": 0.7547169811320755,
1047
+ "eval_accuracy": 0.33360293736563035,
1048
+ "eval_loss": 3.858804941177368,
1049
+ "eval_runtime": 183.535,
1050
+ "eval_samples_per_second": 98.134,
1051
+ "eval_steps_per_second": 6.135,
1052
  "step": 7000
1053
  },
1054
  {
1055
+ "epoch": 0.7601078167115903,
1056
+ "grad_norm": 0.6115443110466003,
1057
+ "learning_rate": 0.0005549983810037776,
1058
+ "loss": 3.931,
1059
  "step": 7050
1060
  },
1061
  {
1062
+ "epoch": 0.7654986522911051,
1063
+ "grad_norm": 0.5609748959541321,
1064
+ "learning_rate": 0.0005546745817593091,
1065
+ "loss": 3.9311,
1066
  "step": 7100
1067
  },
1068
  {
1069
+ "epoch": 0.77088948787062,
1070
+ "grad_norm": 0.6355463266372681,
1071
+ "learning_rate": 0.0005543507825148408,
1072
+ "loss": 3.9063,
1073
  "step": 7150
1074
  },
1075
  {
1076
+ "epoch": 0.7762803234501348,
1077
+ "grad_norm": 0.6557473540306091,
1078
+ "learning_rate": 0.0005540269832703723,
1079
+ "loss": 3.9159,
1080
  "step": 7200
1081
  },
1082
  {
1083
+ "epoch": 0.7816711590296496,
1084
+ "grad_norm": 0.6337311863899231,
1085
+ "learning_rate": 0.0005537031840259039,
1086
+ "loss": 3.9184,
1087
  "step": 7250
1088
  },
1089
  {
1090
+ "epoch": 0.7870619946091644,
1091
+ "grad_norm": 0.5891764760017395,
1092
+ "learning_rate": 0.0005533793847814354,
1093
+ "loss": 3.9032,
1094
  "step": 7300
1095
  },
1096
  {
1097
+ "epoch": 0.7924528301886793,
1098
+ "grad_norm": 0.6272363066673279,
1099
+ "learning_rate": 0.000553055585536967,
1100
+ "loss": 3.9148,
1101
  "step": 7350
1102
  },
1103
  {
1104
+ "epoch": 0.7978436657681941,
1105
+ "grad_norm": 0.6207209229469299,
1106
+ "learning_rate": 0.0005527317862924987,
1107
+ "loss": 3.8796,
1108
  "step": 7400
1109
  },
1110
  {
1111
+ "epoch": 0.8032345013477089,
1112
+ "grad_norm": 0.5580083131790161,
1113
+ "learning_rate": 0.0005524079870480301,
1114
+ "loss": 3.8865,
1115
  "step": 7450
1116
  },
1117
  {
1118
+ "epoch": 0.8086253369272237,
1119
+ "grad_norm": 0.7376517057418823,
1120
+ "learning_rate": 0.0005520841878035618,
1121
+ "loss": 3.9018,
1122
  "step": 7500
1123
  },
1124
  {
1125
+ "epoch": 0.8140161725067385,
1126
+ "grad_norm": 0.6539864540100098,
1127
+ "learning_rate": 0.0005517603885590933,
1128
+ "loss": 3.8971,
1129
  "step": 7550
1130
  },
1131
  {
1132
+ "epoch": 0.8194070080862533,
1133
+ "grad_norm": 0.640784502029419,
1134
+ "learning_rate": 0.0005514365893146249,
1135
+ "loss": 3.9161,
1136
  "step": 7600
1137
  },
1138
  {
1139
+ "epoch": 0.8247978436657682,
1140
+ "grad_norm": 0.5456549525260925,
1141
+ "learning_rate": 0.0005511127900701564,
1142
+ "loss": 3.8959,
1143
  "step": 7650
1144
  },
1145
  {
1146
+ "epoch": 0.8301886792452831,
1147
+ "grad_norm": 0.5484740734100342,
1148
+ "learning_rate": 0.000550788990825688,
1149
+ "loss": 3.871,
1150
  "step": 7700
1151
  },
1152
  {
1153
+ "epoch": 0.8355795148247979,
1154
+ "grad_norm": 0.605842649936676,
1155
+ "learning_rate": 0.0005504651915812196,
1156
+ "loss": 3.8834,
1157
  "step": 7750
1158
  },
1159
  {
1160
+ "epoch": 0.8409703504043127,
1161
+ "grad_norm": 0.5898163914680481,
1162
+ "learning_rate": 0.0005501413923367512,
1163
+ "loss": 3.8977,
1164
  "step": 7800
1165
  },
1166
  {
1167
+ "epoch": 0.8463611859838275,
1168
+ "grad_norm": 0.7656612992286682,
1169
+ "learning_rate": 0.0005498175930922827,
1170
+ "loss": 3.8864,
1171
  "step": 7850
1172
  },
1173
  {
1174
+ "epoch": 0.8517520215633423,
1175
+ "grad_norm": 0.8152493834495544,
1176
+ "learning_rate": 0.0005494937938478143,
1177
+ "loss": 3.8847,
1178
  "step": 7900
1179
  },
1180
  {
1181
+ "epoch": 0.8571428571428571,
1182
+ "grad_norm": 0.5602722764015198,
1183
+ "learning_rate": 0.0005491699946033459,
1184
+ "loss": 3.8864,
1185
  "step": 7950
1186
  },
1187
  {
1188
+ "epoch": 0.862533692722372,
1189
+ "grad_norm": 0.5723937153816223,
1190
+ "learning_rate": 0.0005488461953588775,
1191
+ "loss": 3.8618,
1192
  "step": 8000
1193
  },
1194
  {
1195
+ "epoch": 0.862533692722372,
1196
+ "eval_accuracy": 0.33786419497282644,
1197
+ "eval_loss": 3.8131983280181885,
1198
+ "eval_runtime": 183.2819,
1199
+ "eval_samples_per_second": 98.269,
1200
+ "eval_steps_per_second": 6.144,
1201
  "step": 8000
1202
  },
1203
  {
1204
+ "epoch": 0.8679245283018868,
1205
+ "grad_norm": 0.5539727210998535,
1206
+ "learning_rate": 0.000548522396114409,
1207
+ "loss": 3.879,
1208
  "step": 8050
1209
  },
1210
  {
1211
+ "epoch": 0.8733153638814016,
1212
+ "grad_norm": 0.7839546203613281,
1213
+ "learning_rate": 0.0005481985968699406,
1214
+ "loss": 3.8589,
1215
  "step": 8100
1216
  },
1217
  {
1218
+ "epoch": 0.8787061994609164,
1219
+ "grad_norm": 0.5483460426330566,
1220
+ "learning_rate": 0.0005478747976254721,
1221
+ "loss": 3.8715,
1222
  "step": 8150
1223
  },
1224
  {
1225
+ "epoch": 0.8840970350404312,
1226
+ "grad_norm": 0.5753620266914368,
1227
+ "learning_rate": 0.0005475509983810037,
1228
+ "loss": 3.8617,
1229
  "step": 8200
1230
  },
1231
  {
1232
+ "epoch": 0.889487870619946,
1233
+ "grad_norm": 0.5939356088638306,
1234
+ "learning_rate": 0.0005472271991365352,
1235
+ "loss": 3.8748,
1236
  "step": 8250
1237
  },
1238
  {
1239
+ "epoch": 0.894878706199461,
1240
+ "grad_norm": 0.5494759678840637,
1241
+ "learning_rate": 0.0005469033998920669,
1242
+ "loss": 3.8485,
1243
  "step": 8300
1244
  },
1245
  {
1246
+ "epoch": 0.9002695417789758,
1247
+ "grad_norm": 0.5566021800041199,
1248
+ "learning_rate": 0.0005465796006475984,
1249
+ "loss": 3.8652,
1250
  "step": 8350
1251
  },
1252
  {
1253
+ "epoch": 0.9056603773584906,
1254
+ "grad_norm": 0.5274083018302917,
1255
+ "learning_rate": 0.00054625580140313,
1256
+ "loss": 3.866,
1257
  "step": 8400
1258
  },
1259
  {
1260
+ "epoch": 0.9110512129380054,
1261
+ "grad_norm": 0.5682411193847656,
1262
+ "learning_rate": 0.0005459320021586615,
1263
+ "loss": 3.8363,
1264
  "step": 8450
1265
  },
1266
  {
1267
+ "epoch": 0.9164420485175202,
1268
+ "grad_norm": 0.5468075275421143,
1269
+ "learning_rate": 0.0005456082029141932,
1270
+ "loss": 3.8486,
1271
  "step": 8500
1272
  },
1273
  {
1274
+ "epoch": 0.921832884097035,
1275
+ "grad_norm": 0.5886979699134827,
1276
+ "learning_rate": 0.0005452844036697248,
1277
+ "loss": 3.8496,
1278
  "step": 8550
1279
  },
1280
  {
1281
+ "epoch": 0.9272237196765498,
1282
+ "grad_norm": 0.566277027130127,
1283
+ "learning_rate": 0.0005449606044252563,
1284
+ "loss": 3.8782,
1285
  "step": 8600
1286
  },
1287
  {
1288
+ "epoch": 0.9326145552560647,
1289
+ "grad_norm": 0.6201555728912354,
1290
+ "learning_rate": 0.0005446368051807879,
1291
+ "loss": 3.8507,
1292
  "step": 8650
1293
  },
1294
  {
1295
+ "epoch": 0.9380053908355795,
1296
+ "grad_norm": 0.5942217111587524,
1297
+ "learning_rate": 0.0005443130059363194,
1298
+ "loss": 3.8613,
1299
  "step": 8700
1300
  },
1301
  {
1302
+ "epoch": 0.9433962264150944,
1303
+ "grad_norm": 0.5227022767066956,
1304
+ "learning_rate": 0.0005439892066918511,
1305
+ "loss": 3.8428,
1306
  "step": 8750
1307
  },
1308
  {
1309
+ "epoch": 0.9487870619946092,
1310
+ "grad_norm": 0.6055588126182556,
1311
+ "learning_rate": 0.0005436654074473825,
1312
+ "loss": 3.8415,
1313
  "step": 8800
1314
  },
1315
  {
1316
+ "epoch": 0.954177897574124,
1317
+ "grad_norm": 0.5056185722351074,
1318
+ "learning_rate": 0.0005433416082029142,
1319
+ "loss": 3.8431,
1320
  "step": 8850
1321
  },
1322
  {
1323
+ "epoch": 0.9595687331536388,
1324
+ "grad_norm": 0.6033360958099365,
1325
+ "learning_rate": 0.0005430178089584457,
1326
+ "loss": 3.8452,
1327
  "step": 8900
1328
  },
1329
  {
1330
+ "epoch": 0.9649595687331537,
1331
+ "grad_norm": 0.6905035972595215,
1332
+ "learning_rate": 0.0005426940097139773,
1333
+ "loss": 3.8455,
1334
  "step": 8950
1335
  },
1336
  {
1337
+ "epoch": 0.9703504043126685,
1338
+ "grad_norm": 0.6233043670654297,
1339
+ "learning_rate": 0.0005423766864543982,
1340
+ "loss": 3.8625,
1341
  "step": 9000
1342
  },
1343
  {
1344
+ "epoch": 0.9703504043126685,
1345
+ "eval_accuracy": 0.34123297775951444,
1346
+ "eval_loss": 3.775101900100708,
1347
+ "eval_runtime": 183.702,
1348
+ "eval_samples_per_second": 98.045,
1349
+ "eval_steps_per_second": 6.129,
1350
  "step": 9000
1351
  },
1352
  {
1353
+ "epoch": 0.9757412398921833,
1354
+ "grad_norm": 0.6275179982185364,
1355
+ "learning_rate": 0.0005420528872099298,
1356
+ "loss": 3.839,
1357
  "step": 9050
1358
  },
1359
  {
1360
+ "epoch": 0.9811320754716981,
1361
+ "grad_norm": 0.5922442674636841,
1362
+ "learning_rate": 0.0005417290879654613,
1363
+ "loss": 3.8288,
1364
  "step": 9100
1365
  },
1366
  {
1367
+ "epoch": 0.9865229110512129,
1368
+ "grad_norm": 0.5687377452850342,
1369
+ "learning_rate": 0.000541405288720993,
1370
+ "loss": 3.8416,
1371
  "step": 9150
1372
  },
1373
  {
1374
+ "epoch": 0.9919137466307277,
1375
+ "grad_norm": 0.5579700469970703,
1376
+ "learning_rate": 0.0005410814894765245,
1377
+ "loss": 3.8436,
1378
  "step": 9200
1379
  },
1380
  {
1381
+ "epoch": 0.9973045822102425,
1382
+ "grad_norm": 0.5312648415565491,
1383
+ "learning_rate": 0.0005407576902320561,
1384
+ "loss": 3.8355,
1385
  "step": 9250
1386
  },
1387
  {
1388
+ "epoch": 1.0026954177897573,
1389
+ "grad_norm": 0.5382205247879028,
1390
+ "learning_rate": 0.0005404338909875876,
1391
+ "loss": 3.8133,
1392
  "step": 9300
1393
  },
1394
  {
1395
+ "epoch": 1.0080862533692723,
1396
+ "grad_norm": 0.6140266060829163,
1397
+ "learning_rate": 0.0005401100917431192,
1398
+ "loss": 3.7783,
1399
  "step": 9350
1400
  },
1401
  {
1402
+ "epoch": 1.013477088948787,
1403
+ "grad_norm": 0.5851580500602722,
1404
+ "learning_rate": 0.0005397862924986508,
1405
+ "loss": 3.7885,
1406
  "step": 9400
1407
  },
1408
  {
1409
+ "epoch": 1.0188679245283019,
1410
+ "grad_norm": 0.5980308651924133,
1411
+ "learning_rate": 0.0005394624932541824,
1412
+ "loss": 3.7783,
1413
  "step": 9450
1414
  },
1415
  {
1416
+ "epoch": 1.0242587601078168,
1417
+ "grad_norm": 0.5626499652862549,
1418
+ "learning_rate": 0.0005391386940097139,
1419
+ "loss": 3.7758,
1420
  "step": 9500
1421
  },
1422
  {
1423
+ "epoch": 1.0296495956873315,
1424
+ "grad_norm": 0.5909863114356995,
1425
+ "learning_rate": 0.0005388148947652455,
1426
+ "loss": 3.7869,
1427
  "step": 9550
1428
  },
1429
  {
1430
+ "epoch": 1.0350404312668464,
1431
+ "grad_norm": 0.5526650547981262,
1432
+ "learning_rate": 0.000538491095520777,
1433
+ "loss": 3.7655,
1434
  "step": 9600
1435
  },
1436
  {
1437
+ "epoch": 1.0404312668463611,
1438
+ "grad_norm": 0.5704436898231506,
1439
+ "learning_rate": 0.0005381672962763086,
1440
+ "loss": 3.7576,
1441
  "step": 9650
1442
  },
1443
  {
1444
+ "epoch": 1.045822102425876,
1445
+ "grad_norm": 0.5485429167747498,
1446
+ "learning_rate": 0.0005378434970318403,
1447
+ "loss": 3.7695,
1448
  "step": 9700
1449
  },
1450
  {
1451
+ "epoch": 1.0512129380053907,
1452
+ "grad_norm": 0.5683411955833435,
1453
+ "learning_rate": 0.0005375196977873718,
1454
+ "loss": 3.7622,
1455
  "step": 9750
1456
  },
1457
  {
1458
+ "epoch": 1.0566037735849056,
1459
+ "grad_norm": 0.5670942068099976,
1460
+ "learning_rate": 0.0005371958985429034,
1461
+ "loss": 3.7672,
1462
  "step": 9800
1463
  },
1464
  {
1465
+ "epoch": 1.0619946091644206,
1466
+ "grad_norm": 0.6690695881843567,
1467
+ "learning_rate": 0.0005368720992984349,
1468
+ "loss": 3.7521,
1469
  "step": 9850
1470
  },
1471
  {
1472
+ "epoch": 1.0673854447439353,
1473
+ "grad_norm": 0.6253776550292969,
1474
+ "learning_rate": 0.0005365483000539665,
1475
+ "loss": 3.762,
1476
  "step": 9900
1477
  },
1478
  {
1479
+ "epoch": 1.0727762803234502,
1480
+ "grad_norm": 0.5629079937934875,
1481
+ "learning_rate": 0.0005362245008094981,
1482
+ "loss": 3.7594,
1483
  "step": 9950
1484
  },
1485
  {
1486
+ "epoch": 1.0781671159029649,
1487
+ "grad_norm": 0.5760037302970886,
1488
+ "learning_rate": 0.0005359007015650297,
1489
+ "loss": 3.7656,
1490
  "step": 10000
1491
  },
1492
  {
1493
+ "epoch": 1.0781671159029649,
1494
+ "eval_accuracy": 0.34546098758303934,
1495
+ "eval_loss": 3.7440903186798096,
1496
+ "eval_runtime": 183.3277,
1497
+ "eval_samples_per_second": 98.245,
1498
+ "eval_steps_per_second": 6.142,
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 sha256:8d044a307c17651cff2657f304d62416d1cb870c073e18fb33096f4e4f260c01
3
  size 5304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6f70a211759452569c4a583d57df8ecf1cd8848934be54c26401d6bef277c4f7
3
  size 5304