MindscapeRAG commited on
Commit
6ad4911
·
verified ·
1 Parent(s): f1846de

Delete trainer_state.json

Browse files
Files changed (1) hide show
  1. trainer_state.json +0 -923
trainer_state.json DELETED
@@ -1,923 +0,0 @@
1
- {
2
- "best_global_step": null,
3
- "best_metric": null,
4
- "best_model_checkpoint": null,
5
- "epoch": 3.0,
6
- "eval_steps": 500,
7
- "global_step": 1272,
8
- "is_hyper_param_search": false,
9
- "is_local_process_zero": true,
10
- "is_world_process_zero": true,
11
- "log_history": [
12
- {
13
- "epoch": 0.02358490566037736,
14
- "grad_norm": 111.64971904040075,
15
- "learning_rate": 4.745611564124804e-05,
16
- "loss": 53.4188,
17
- "step": 10
18
- },
19
- {
20
- "epoch": 0.04716981132075472,
21
- "grad_norm": 36.50653706838016,
22
- "learning_rate": 6.174182992696232e-05,
23
- "loss": 19.9875,
24
- "step": 20
25
- },
26
- {
27
- "epoch": 0.07075471698113207,
28
- "grad_norm": 22.585478402078195,
29
- "learning_rate": 7.009843708012171e-05,
30
- "loss": 13.9594,
31
- "step": 30
32
- },
33
- {
34
- "epoch": 0.09433962264150944,
35
- "grad_norm": 26.901765162938684,
36
- "learning_rate": 7.602754421267661e-05,
37
- "loss": 12.9,
38
- "step": 40
39
- },
40
- {
41
- "epoch": 0.1179245283018868,
42
- "grad_norm": 15.203549791437116,
43
- "learning_rate": 8.062651699678178e-05,
44
- "loss": 11.9906,
45
- "step": 50
46
- },
47
- {
48
- "epoch": 0.14150943396226415,
49
- "grad_norm": 19.669683180170903,
50
- "learning_rate": 8.438415136583599e-05,
51
- "loss": 12.1125,
52
- "step": 60
53
- },
54
- {
55
- "epoch": 0.1650943396226415,
56
- "grad_norm": 20.617026952721357,
57
- "learning_rate": 8.756118595635669e-05,
58
- "loss": 11.7344,
59
- "step": 70
60
- },
61
- {
62
- "epoch": 0.18867924528301888,
63
- "grad_norm": 14.56388559298774,
64
- "learning_rate": 9.031325849839089e-05,
65
- "loss": 11.1172,
66
- "step": 80
67
- },
68
- {
69
- "epoch": 0.21226415094339623,
70
- "grad_norm": 17.076659337879114,
71
- "learning_rate": 9.274075851899536e-05,
72
- "loss": 11.9187,
73
- "step": 90
74
- },
75
- {
76
- "epoch": 0.2358490566037736,
77
- "grad_norm": 32.07124156790194,
78
- "learning_rate": 9.491223128249609e-05,
79
- "loss": 11.6969,
80
- "step": 100
81
- },
82
- {
83
- "epoch": 0.25943396226415094,
84
- "grad_norm": 17.8530210729174,
85
- "learning_rate": 9.687656733606659e-05,
86
- "loss": 11.8938,
87
- "step": 110
88
- },
89
- {
90
- "epoch": 0.2830188679245283,
91
- "grad_norm": 27.003004374393104,
92
- "learning_rate": 9.866986565155028e-05,
93
- "loss": 12.5906,
94
- "step": 120
95
- },
96
- {
97
- "epoch": 0.30660377358490565,
98
- "grad_norm": 19.002162024515545,
99
- "learning_rate": 9.991258741258742e-05,
100
- "loss": 11.8063,
101
- "step": 130
102
- },
103
- {
104
- "epoch": 0.330188679245283,
105
- "grad_norm": 14.220172545505502,
106
- "learning_rate": 9.903846153846155e-05,
107
- "loss": 10.5687,
108
- "step": 140
109
- },
110
- {
111
- "epoch": 0.35377358490566035,
112
- "grad_norm": 14.994374088898928,
113
- "learning_rate": 9.816433566433567e-05,
114
- "loss": 11.4469,
115
- "step": 150
116
- },
117
- {
118
- "epoch": 0.37735849056603776,
119
- "grad_norm": 28.610439819975692,
120
- "learning_rate": 9.729020979020979e-05,
121
- "loss": 11.6031,
122
- "step": 160
123
- },
124
- {
125
- "epoch": 0.4009433962264151,
126
- "grad_norm": 13.755122715578342,
127
- "learning_rate": 9.641608391608393e-05,
128
- "loss": 11.7562,
129
- "step": 170
130
- },
131
- {
132
- "epoch": 0.42452830188679247,
133
- "grad_norm": 13.235152354159759,
134
- "learning_rate": 9.554195804195805e-05,
135
- "loss": 11.5156,
136
- "step": 180
137
- },
138
- {
139
- "epoch": 0.4481132075471698,
140
- "grad_norm": 12.55770261202009,
141
- "learning_rate": 9.466783216783217e-05,
142
- "loss": 11.2953,
143
- "step": 190
144
- },
145
- {
146
- "epoch": 0.4716981132075472,
147
- "grad_norm": 13.582038438276099,
148
- "learning_rate": 9.37937062937063e-05,
149
- "loss": 11.1062,
150
- "step": 200
151
- },
152
- {
153
- "epoch": 0.49528301886792453,
154
- "grad_norm": 11.770352793407033,
155
- "learning_rate": 9.291958041958042e-05,
156
- "loss": 10.7781,
157
- "step": 210
158
- },
159
- {
160
- "epoch": 0.5188679245283019,
161
- "grad_norm": 14.820093896670828,
162
- "learning_rate": 9.204545454545454e-05,
163
- "loss": 11.4578,
164
- "step": 220
165
- },
166
- {
167
- "epoch": 0.5424528301886793,
168
- "grad_norm": 16.03397508198143,
169
- "learning_rate": 9.117132867132868e-05,
170
- "loss": 11.3906,
171
- "step": 230
172
- },
173
- {
174
- "epoch": 0.5660377358490566,
175
- "grad_norm": 12.665474151765666,
176
- "learning_rate": 9.029720279720281e-05,
177
- "loss": 11.2812,
178
- "step": 240
179
- },
180
- {
181
- "epoch": 0.589622641509434,
182
- "grad_norm": 12.380296703659218,
183
- "learning_rate": 8.942307692307693e-05,
184
- "loss": 10.925,
185
- "step": 250
186
- },
187
- {
188
- "epoch": 0.6132075471698113,
189
- "grad_norm": 11.609348406547522,
190
- "learning_rate": 8.854895104895105e-05,
191
- "loss": 10.9625,
192
- "step": 260
193
- },
194
- {
195
- "epoch": 0.6367924528301887,
196
- "grad_norm": 41.38762916621647,
197
- "learning_rate": 8.767482517482519e-05,
198
- "loss": 11.3031,
199
- "step": 270
200
- },
201
- {
202
- "epoch": 0.660377358490566,
203
- "grad_norm": 11.71078545247518,
204
- "learning_rate": 8.68006993006993e-05,
205
- "loss": 10.5437,
206
- "step": 280
207
- },
208
- {
209
- "epoch": 0.6839622641509434,
210
- "grad_norm": 14.740681537649765,
211
- "learning_rate": 8.592657342657343e-05,
212
- "loss": 11.0656,
213
- "step": 290
214
- },
215
- {
216
- "epoch": 0.7075471698113207,
217
- "grad_norm": 13.407050612496178,
218
- "learning_rate": 8.505244755244756e-05,
219
- "loss": 10.275,
220
- "step": 300
221
- },
222
- {
223
- "epoch": 0.7311320754716981,
224
- "grad_norm": 17.285671485791312,
225
- "learning_rate": 8.417832167832168e-05,
226
- "loss": 10.9094,
227
- "step": 310
228
- },
229
- {
230
- "epoch": 0.7547169811320755,
231
- "grad_norm": 12.234499943045662,
232
- "learning_rate": 8.33041958041958e-05,
233
- "loss": 10.3812,
234
- "step": 320
235
- },
236
- {
237
- "epoch": 0.7783018867924528,
238
- "grad_norm": 13.844957482436712,
239
- "learning_rate": 8.243006993006994e-05,
240
- "loss": 10.6016,
241
- "step": 330
242
- },
243
- {
244
- "epoch": 0.8018867924528302,
245
- "grad_norm": 14.359721900450984,
246
- "learning_rate": 8.155594405594406e-05,
247
- "loss": 10.7719,
248
- "step": 340
249
- },
250
- {
251
- "epoch": 0.8254716981132075,
252
- "grad_norm": 13.685695871660895,
253
- "learning_rate": 8.068181818181818e-05,
254
- "loss": 10.9281,
255
- "step": 350
256
- },
257
- {
258
- "epoch": 0.8490566037735849,
259
- "grad_norm": 11.817457808585134,
260
- "learning_rate": 7.980769230769231e-05,
261
- "loss": 10.5906,
262
- "step": 360
263
- },
264
- {
265
- "epoch": 0.8726415094339622,
266
- "grad_norm": 14.157887783746466,
267
- "learning_rate": 7.893356643356644e-05,
268
- "loss": 10.4375,
269
- "step": 370
270
- },
271
- {
272
- "epoch": 0.8962264150943396,
273
- "grad_norm": 16.666070129232146,
274
- "learning_rate": 7.805944055944057e-05,
275
- "loss": 10.2781,
276
- "step": 380
277
- },
278
- {
279
- "epoch": 0.9198113207547169,
280
- "grad_norm": 13.667548285529008,
281
- "learning_rate": 7.718531468531469e-05,
282
- "loss": 10.2438,
283
- "step": 390
284
- },
285
- {
286
- "epoch": 0.9433962264150944,
287
- "grad_norm": 12.066751708653879,
288
- "learning_rate": 7.631118881118882e-05,
289
- "loss": 10.0094,
290
- "step": 400
291
- },
292
- {
293
- "epoch": 0.9669811320754716,
294
- "grad_norm": 12.248491316508403,
295
- "learning_rate": 7.543706293706294e-05,
296
- "loss": 10.2328,
297
- "step": 410
298
- },
299
- {
300
- "epoch": 0.9905660377358491,
301
- "grad_norm": 18.024158226856255,
302
- "learning_rate": 7.456293706293706e-05,
303
- "loss": 9.9656,
304
- "step": 420
305
- },
306
- {
307
- "epoch": 1.0141509433962264,
308
- "grad_norm": 11.870175278410656,
309
- "learning_rate": 7.36888111888112e-05,
310
- "loss": 9.6953,
311
- "step": 430
312
- },
313
- {
314
- "epoch": 1.0377358490566038,
315
- "grad_norm": 11.950018950483969,
316
- "learning_rate": 7.281468531468531e-05,
317
- "loss": 10.1656,
318
- "step": 440
319
- },
320
- {
321
- "epoch": 1.0613207547169812,
322
- "grad_norm": 16.767199122744902,
323
- "learning_rate": 7.194055944055944e-05,
324
- "loss": 10.1594,
325
- "step": 450
326
- },
327
- {
328
- "epoch": 1.0849056603773586,
329
- "grad_norm": 15.290808887183745,
330
- "learning_rate": 7.106643356643357e-05,
331
- "loss": 9.6875,
332
- "step": 460
333
- },
334
- {
335
- "epoch": 1.1084905660377358,
336
- "grad_norm": 10.022244420640183,
337
- "learning_rate": 7.019230769230769e-05,
338
- "loss": 10.3281,
339
- "step": 470
340
- },
341
- {
342
- "epoch": 1.1320754716981132,
343
- "grad_norm": 12.951298055638265,
344
- "learning_rate": 6.931818181818182e-05,
345
- "loss": 9.6766,
346
- "step": 480
347
- },
348
- {
349
- "epoch": 1.1556603773584906,
350
- "grad_norm": 13.127907045424376,
351
- "learning_rate": 6.844405594405596e-05,
352
- "loss": 10.1594,
353
- "step": 490
354
- },
355
- {
356
- "epoch": 1.179245283018868,
357
- "grad_norm": 11.892546949840662,
358
- "learning_rate": 6.756993006993008e-05,
359
- "loss": 9.9844,
360
- "step": 500
361
- },
362
- {
363
- "epoch": 1.2028301886792452,
364
- "grad_norm": 12.563312310051254,
365
- "learning_rate": 6.66958041958042e-05,
366
- "loss": 9.8906,
367
- "step": 510
368
- },
369
- {
370
- "epoch": 1.2264150943396226,
371
- "grad_norm": 12.048069831318715,
372
- "learning_rate": 6.582167832167833e-05,
373
- "loss": 9.9938,
374
- "step": 520
375
- },
376
- {
377
- "epoch": 1.25,
378
- "grad_norm": 21.579035414990887,
379
- "learning_rate": 6.494755244755245e-05,
380
- "loss": 9.4234,
381
- "step": 530
382
- },
383
- {
384
- "epoch": 1.2735849056603774,
385
- "grad_norm": 12.065467497517545,
386
- "learning_rate": 6.407342657342657e-05,
387
- "loss": 10.0563,
388
- "step": 540
389
- },
390
- {
391
- "epoch": 1.2971698113207548,
392
- "grad_norm": 11.549132237202269,
393
- "learning_rate": 6.31993006993007e-05,
394
- "loss": 9.8984,
395
- "step": 550
396
- },
397
- {
398
- "epoch": 1.320754716981132,
399
- "grad_norm": 11.938528591591838,
400
- "learning_rate": 6.232517482517483e-05,
401
- "loss": 9.9969,
402
- "step": 560
403
- },
404
- {
405
- "epoch": 1.3443396226415094,
406
- "grad_norm": 14.929825222343986,
407
- "learning_rate": 6.145104895104895e-05,
408
- "loss": 9.7594,
409
- "step": 570
410
- },
411
- {
412
- "epoch": 1.3679245283018868,
413
- "grad_norm": 12.842065570056926,
414
- "learning_rate": 6.0576923076923076e-05,
415
- "loss": 9.5531,
416
- "step": 580
417
- },
418
- {
419
- "epoch": 1.3915094339622642,
420
- "grad_norm": 15.540618048359113,
421
- "learning_rate": 5.97027972027972e-05,
422
- "loss": 10.1094,
423
- "step": 590
424
- },
425
- {
426
- "epoch": 1.4150943396226414,
427
- "grad_norm": 17.595607626825647,
428
- "learning_rate": 5.882867132867134e-05,
429
- "loss": 9.4891,
430
- "step": 600
431
- },
432
- {
433
- "epoch": 1.4386792452830188,
434
- "grad_norm": 10.493212987412385,
435
- "learning_rate": 5.7954545454545464e-05,
436
- "loss": 9.6406,
437
- "step": 610
438
- },
439
- {
440
- "epoch": 1.4622641509433962,
441
- "grad_norm": 24.84577511132654,
442
- "learning_rate": 5.7080419580419585e-05,
443
- "loss": 9.7812,
444
- "step": 620
445
- },
446
- {
447
- "epoch": 1.4858490566037736,
448
- "grad_norm": 13.697240248862046,
449
- "learning_rate": 5.620629370629371e-05,
450
- "loss": 9.8063,
451
- "step": 630
452
- },
453
- {
454
- "epoch": 1.509433962264151,
455
- "grad_norm": 11.714963877097121,
456
- "learning_rate": 5.533216783216784e-05,
457
- "loss": 9.85,
458
- "step": 640
459
- },
460
- {
461
- "epoch": 1.5330188679245285,
462
- "grad_norm": 15.925762627169787,
463
- "learning_rate": 5.445804195804196e-05,
464
- "loss": 9.5437,
465
- "step": 650
466
- },
467
- {
468
- "epoch": 1.5566037735849056,
469
- "grad_norm": 10.21229978672529,
470
- "learning_rate": 5.358391608391609e-05,
471
- "loss": 9.7359,
472
- "step": 660
473
- },
474
- {
475
- "epoch": 1.580188679245283,
476
- "grad_norm": 12.299504271971967,
477
- "learning_rate": 5.2709790209790214e-05,
478
- "loss": 10.0203,
479
- "step": 670
480
- },
481
- {
482
- "epoch": 1.6037735849056602,
483
- "grad_norm": 10.756878837227157,
484
- "learning_rate": 5.1835664335664335e-05,
485
- "loss": 8.9719,
486
- "step": 680
487
- },
488
- {
489
- "epoch": 1.6273584905660377,
490
- "grad_norm": 47.89124644047696,
491
- "learning_rate": 5.096153846153846e-05,
492
- "loss": 9.8281,
493
- "step": 690
494
- },
495
- {
496
- "epoch": 1.650943396226415,
497
- "grad_norm": 13.321980781383072,
498
- "learning_rate": 5.008741258741258e-05,
499
- "loss": 9.5188,
500
- "step": 700
501
- },
502
- {
503
- "epoch": 1.6745283018867925,
504
- "grad_norm": 12.375904788348723,
505
- "learning_rate": 4.9213286713286716e-05,
506
- "loss": 10.3156,
507
- "step": 710
508
- },
509
- {
510
- "epoch": 1.6981132075471699,
511
- "grad_norm": 13.04878290460268,
512
- "learning_rate": 4.8339160839160843e-05,
513
- "loss": 9.3875,
514
- "step": 720
515
- },
516
- {
517
- "epoch": 1.7216981132075473,
518
- "grad_norm": 13.835045116571802,
519
- "learning_rate": 4.7465034965034964e-05,
520
- "loss": 10.0781,
521
- "step": 730
522
- },
523
- {
524
- "epoch": 1.7452830188679245,
525
- "grad_norm": 13.127688060476284,
526
- "learning_rate": 4.659090909090909e-05,
527
- "loss": 9.5703,
528
- "step": 740
529
- },
530
- {
531
- "epoch": 1.7688679245283019,
532
- "grad_norm": 11.61972720292625,
533
- "learning_rate": 4.571678321678322e-05,
534
- "loss": 9.3,
535
- "step": 750
536
- },
537
- {
538
- "epoch": 1.7924528301886793,
539
- "grad_norm": 11.880032085014191,
540
- "learning_rate": 4.4842657342657346e-05,
541
- "loss": 9.2141,
542
- "step": 760
543
- },
544
- {
545
- "epoch": 1.8160377358490565,
546
- "grad_norm": 12.679739149762932,
547
- "learning_rate": 4.396853146853147e-05,
548
- "loss": 10.0031,
549
- "step": 770
550
- },
551
- {
552
- "epoch": 1.8396226415094339,
553
- "grad_norm": 12.587125834021709,
554
- "learning_rate": 4.309440559440559e-05,
555
- "loss": 10.0437,
556
- "step": 780
557
- },
558
- {
559
- "epoch": 1.8632075471698113,
560
- "grad_norm": 9.743520340164396,
561
- "learning_rate": 4.222027972027972e-05,
562
- "loss": 9.6719,
563
- "step": 790
564
- },
565
- {
566
- "epoch": 1.8867924528301887,
567
- "grad_norm": 13.971700558144239,
568
- "learning_rate": 4.134615384615385e-05,
569
- "loss": 9.3938,
570
- "step": 800
571
- },
572
- {
573
- "epoch": 1.9103773584905661,
574
- "grad_norm": 11.607356019588352,
575
- "learning_rate": 4.0472027972027975e-05,
576
- "loss": 9.3859,
577
- "step": 810
578
- },
579
- {
580
- "epoch": 1.9339622641509435,
581
- "grad_norm": 12.35296887661871,
582
- "learning_rate": 3.95979020979021e-05,
583
- "loss": 9.3297,
584
- "step": 820
585
- },
586
- {
587
- "epoch": 1.9575471698113207,
588
- "grad_norm": 10.899779163585386,
589
- "learning_rate": 3.872377622377622e-05,
590
- "loss": 9.7781,
591
- "step": 830
592
- },
593
- {
594
- "epoch": 1.9811320754716981,
595
- "grad_norm": 11.971338017749327,
596
- "learning_rate": 3.784965034965035e-05,
597
- "loss": 10.1312,
598
- "step": 840
599
- },
600
- {
601
- "epoch": 2.0047169811320753,
602
- "grad_norm": 10.878937392831533,
603
- "learning_rate": 3.697552447552448e-05,
604
- "loss": 9.0563,
605
- "step": 850
606
- },
607
- {
608
- "epoch": 2.0283018867924527,
609
- "grad_norm": 15.806159570601048,
610
- "learning_rate": 3.61013986013986e-05,
611
- "loss": 8.5625,
612
- "step": 860
613
- },
614
- {
615
- "epoch": 2.05188679245283,
616
- "grad_norm": 13.812473930051553,
617
- "learning_rate": 3.522727272727273e-05,
618
- "loss": 9.2078,
619
- "step": 870
620
- },
621
- {
622
- "epoch": 2.0754716981132075,
623
- "grad_norm": 9.891081116732627,
624
- "learning_rate": 3.435314685314686e-05,
625
- "loss": 8.7234,
626
- "step": 880
627
- },
628
- {
629
- "epoch": 2.099056603773585,
630
- "grad_norm": 10.673891805568328,
631
- "learning_rate": 3.347902097902098e-05,
632
- "loss": 9.0563,
633
- "step": 890
634
- },
635
- {
636
- "epoch": 2.1226415094339623,
637
- "grad_norm": 10.205081198781729,
638
- "learning_rate": 3.2604895104895106e-05,
639
- "loss": 8.9688,
640
- "step": 900
641
- },
642
- {
643
- "epoch": 2.1462264150943398,
644
- "grad_norm": 11.566662730230892,
645
- "learning_rate": 3.1730769230769234e-05,
646
- "loss": 9.0531,
647
- "step": 910
648
- },
649
- {
650
- "epoch": 2.169811320754717,
651
- "grad_norm": 11.119994413592556,
652
- "learning_rate": 3.0856643356643354e-05,
653
- "loss": 9.0516,
654
- "step": 920
655
- },
656
- {
657
- "epoch": 2.1933962264150946,
658
- "grad_norm": 10.765841854970871,
659
- "learning_rate": 2.9982517482517485e-05,
660
- "loss": 8.7312,
661
- "step": 930
662
- },
663
- {
664
- "epoch": 2.2169811320754715,
665
- "grad_norm": 14.277150964403473,
666
- "learning_rate": 2.9108391608391612e-05,
667
- "loss": 8.9609,
668
- "step": 940
669
- },
670
- {
671
- "epoch": 2.240566037735849,
672
- "grad_norm": 13.46535727014516,
673
- "learning_rate": 2.8234265734265736e-05,
674
- "loss": 9.0531,
675
- "step": 950
676
- },
677
- {
678
- "epoch": 2.2641509433962264,
679
- "grad_norm": 13.360257816479617,
680
- "learning_rate": 2.736013986013986e-05,
681
- "loss": 9.4016,
682
- "step": 960
683
- },
684
- {
685
- "epoch": 2.2877358490566038,
686
- "grad_norm": 10.644001178785137,
687
- "learning_rate": 2.6486013986013987e-05,
688
- "loss": 9.4281,
689
- "step": 970
690
- },
691
- {
692
- "epoch": 2.311320754716981,
693
- "grad_norm": 13.041471877077432,
694
- "learning_rate": 2.561188811188811e-05,
695
- "loss": 8.4094,
696
- "step": 980
697
- },
698
- {
699
- "epoch": 2.3349056603773586,
700
- "grad_norm": 11.380539833099244,
701
- "learning_rate": 2.4737762237762238e-05,
702
- "loss": 8.5703,
703
- "step": 990
704
- },
705
- {
706
- "epoch": 2.358490566037736,
707
- "grad_norm": 15.24184551745,
708
- "learning_rate": 2.3863636363636365e-05,
709
- "loss": 8.7469,
710
- "step": 1000
711
- },
712
- {
713
- "epoch": 2.3820754716981134,
714
- "grad_norm": 13.296204794285742,
715
- "learning_rate": 2.298951048951049e-05,
716
- "loss": 8.6531,
717
- "step": 1010
718
- },
719
- {
720
- "epoch": 2.4056603773584904,
721
- "grad_norm": 14.553056169204998,
722
- "learning_rate": 2.2115384615384616e-05,
723
- "loss": 9.4719,
724
- "step": 1020
725
- },
726
- {
727
- "epoch": 2.4292452830188678,
728
- "grad_norm": 12.864094699757313,
729
- "learning_rate": 2.1241258741258744e-05,
730
- "loss": 8.5609,
731
- "step": 1030
732
- },
733
- {
734
- "epoch": 2.452830188679245,
735
- "grad_norm": 12.84578923442427,
736
- "learning_rate": 2.0367132867132867e-05,
737
- "loss": 9.2219,
738
- "step": 1040
739
- },
740
- {
741
- "epoch": 2.4764150943396226,
742
- "grad_norm": 13.709126969591345,
743
- "learning_rate": 1.9493006993006995e-05,
744
- "loss": 8.8297,
745
- "step": 1050
746
- },
747
- {
748
- "epoch": 2.5,
749
- "grad_norm": 12.83886757229034,
750
- "learning_rate": 1.8618881118881122e-05,
751
- "loss": 9.2188,
752
- "step": 1060
753
- },
754
- {
755
- "epoch": 2.5235849056603774,
756
- "grad_norm": 12.383754458241203,
757
- "learning_rate": 1.7744755244755246e-05,
758
- "loss": 8.3781,
759
- "step": 1070
760
- },
761
- {
762
- "epoch": 2.547169811320755,
763
- "grad_norm": 15.566078599420996,
764
- "learning_rate": 1.687062937062937e-05,
765
- "loss": 8.9109,
766
- "step": 1080
767
- },
768
- {
769
- "epoch": 2.5707547169811322,
770
- "grad_norm": 12.227977636887342,
771
- "learning_rate": 1.5996503496503497e-05,
772
- "loss": 8.5984,
773
- "step": 1090
774
- },
775
- {
776
- "epoch": 2.5943396226415096,
777
- "grad_norm": 11.392465117491689,
778
- "learning_rate": 1.5122377622377622e-05,
779
- "loss": 8.0125,
780
- "step": 1100
781
- },
782
- {
783
- "epoch": 2.617924528301887,
784
- "grad_norm": 11.211860612831199,
785
- "learning_rate": 1.4248251748251748e-05,
786
- "loss": 8.6391,
787
- "step": 1110
788
- },
789
- {
790
- "epoch": 2.641509433962264,
791
- "grad_norm": 12.379882956164083,
792
- "learning_rate": 1.3374125874125875e-05,
793
- "loss": 8.5687,
794
- "step": 1120
795
- },
796
- {
797
- "epoch": 2.6650943396226414,
798
- "grad_norm": 13.914701836881665,
799
- "learning_rate": 1.25e-05,
800
- "loss": 7.9281,
801
- "step": 1130
802
- },
803
- {
804
- "epoch": 2.688679245283019,
805
- "grad_norm": 14.688732375132988,
806
- "learning_rate": 1.1625874125874126e-05,
807
- "loss": 8.3812,
808
- "step": 1140
809
- },
810
- {
811
- "epoch": 2.7122641509433962,
812
- "grad_norm": 15.688619890127729,
813
- "learning_rate": 1.0751748251748252e-05,
814
- "loss": 8.7547,
815
- "step": 1150
816
- },
817
- {
818
- "epoch": 2.7358490566037736,
819
- "grad_norm": 11.089768426746776,
820
- "learning_rate": 9.877622377622379e-06,
821
- "loss": 8.8172,
822
- "step": 1160
823
- },
824
- {
825
- "epoch": 2.759433962264151,
826
- "grad_norm": 23.92616633558165,
827
- "learning_rate": 9.003496503496504e-06,
828
- "loss": 8.0312,
829
- "step": 1170
830
- },
831
- {
832
- "epoch": 2.7830188679245285,
833
- "grad_norm": 15.521482664204598,
834
- "learning_rate": 8.12937062937063e-06,
835
- "loss": 8.9219,
836
- "step": 1180
837
- },
838
- {
839
- "epoch": 2.8066037735849054,
840
- "grad_norm": 11.858360531048184,
841
- "learning_rate": 7.2552447552447555e-06,
842
- "loss": 8.2797,
843
- "step": 1190
844
- },
845
- {
846
- "epoch": 2.830188679245283,
847
- "grad_norm": 11.546554540604026,
848
- "learning_rate": 6.381118881118882e-06,
849
- "loss": 8.8203,
850
- "step": 1200
851
- },
852
- {
853
- "epoch": 2.8537735849056602,
854
- "grad_norm": 13.017973974001212,
855
- "learning_rate": 5.5069930069930074e-06,
856
- "loss": 8.5516,
857
- "step": 1210
858
- },
859
- {
860
- "epoch": 2.8773584905660377,
861
- "grad_norm": 28.505581742223395,
862
- "learning_rate": 4.632867132867133e-06,
863
- "loss": 8.5891,
864
- "step": 1220
865
- },
866
- {
867
- "epoch": 2.900943396226415,
868
- "grad_norm": 13.282491850039023,
869
- "learning_rate": 3.7587412587412585e-06,
870
- "loss": 8.8891,
871
- "step": 1230
872
- },
873
- {
874
- "epoch": 2.9245283018867925,
875
- "grad_norm": 12.713216988295962,
876
- "learning_rate": 2.884615384615385e-06,
877
- "loss": 8.55,
878
- "step": 1240
879
- },
880
- {
881
- "epoch": 2.94811320754717,
882
- "grad_norm": 13.2259063179825,
883
- "learning_rate": 2.0104895104895104e-06,
884
- "loss": 8.4469,
885
- "step": 1250
886
- },
887
- {
888
- "epoch": 2.9716981132075473,
889
- "grad_norm": 12.767435596869587,
890
- "learning_rate": 1.1363636363636364e-06,
891
- "loss": 8.4953,
892
- "step": 1260
893
- },
894
- {
895
- "epoch": 2.9952830188679247,
896
- "grad_norm": 12.197730584858347,
897
- "learning_rate": 2.622377622377623e-07,
898
- "loss": 8.2344,
899
- "step": 1270
900
- }
901
- ],
902
- "logging_steps": 10,
903
- "max_steps": 1272,
904
- "num_input_tokens_seen": 0,
905
- "num_train_epochs": 3,
906
- "save_steps": 235,
907
- "stateful_callbacks": {
908
- "TrainerControl": {
909
- "args": {
910
- "should_epoch_stop": false,
911
- "should_evaluate": false,
912
- "should_log": false,
913
- "should_save": true,
914
- "should_training_stop": true
915
- },
916
- "attributes": {}
917
- }
918
- },
919
- "total_flos": 0.0,
920
- "train_batch_size": 4,
921
- "trial_name": null,
922
- "trial_params": null
923
- }