THEODOROS commited on
Commit
de53cac
·
1 Parent(s): a00a960

Delete model/trainer_state.json

Browse files
Files changed (1) hide show
  1. model/trainer_state.json +0 -1196
model/trainer_state.json DELETED
@@ -1,1196 +0,0 @@
1
- {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 2.0,
5
- "global_step": 97022,
6
- "is_hyper_param_search": false,
7
- "is_local_process_zero": true,
8
- "is_world_process_zero": true,
9
- "log_history": [
10
- {
11
- "epoch": 0.01,
12
- "learning_rate": 1.25e-06,
13
- "loss": 0.3381,
14
- "step": 500
15
- },
16
- {
17
- "epoch": 0.02,
18
- "learning_rate": 2.5e-06,
19
- "loss": 0.1717,
20
- "step": 1000
21
- },
22
- {
23
- "epoch": 0.03,
24
- "learning_rate": 3.7500000000000005e-06,
25
- "loss": 0.1167,
26
- "step": 1500
27
- },
28
- {
29
- "epoch": 0.04,
30
- "learning_rate": 5e-06,
31
- "loss": 0.0962,
32
- "step": 2000
33
- },
34
- {
35
- "epoch": 0.05,
36
- "learning_rate": 4.9896073662987675e-06,
37
- "loss": 0.0848,
38
- "step": 2500
39
- },
40
- {
41
- "epoch": 0.06,
42
- "learning_rate": 4.979214732597535e-06,
43
- "loss": 0.0796,
44
- "step": 3000
45
- },
46
- {
47
- "epoch": 0.07,
48
- "learning_rate": 4.968822098896302e-06,
49
- "loss": 0.0758,
50
- "step": 3500
51
- },
52
- {
53
- "epoch": 0.08,
54
- "learning_rate": 4.95842946519507e-06,
55
- "loss": 0.0732,
56
- "step": 4000
57
- },
58
- {
59
- "epoch": 0.09,
60
- "learning_rate": 4.948036831493837e-06,
61
- "loss": 0.071,
62
- "step": 4500
63
- },
64
- {
65
- "epoch": 0.1,
66
- "learning_rate": 4.937644197792605e-06,
67
- "loss": 0.0697,
68
- "step": 5000
69
- },
70
- {
71
- "epoch": 0.11,
72
- "learning_rate": 4.927251564091372e-06,
73
- "loss": 0.0681,
74
- "step": 5500
75
- },
76
- {
77
- "epoch": 0.12,
78
- "learning_rate": 4.91685893039014e-06,
79
- "loss": 0.0678,
80
- "step": 6000
81
- },
82
- {
83
- "epoch": 0.13,
84
- "learning_rate": 4.906466296688907e-06,
85
- "loss": 0.0656,
86
- "step": 6500
87
- },
88
- {
89
- "epoch": 0.14,
90
- "learning_rate": 4.896073662987675e-06,
91
- "loss": 0.0651,
92
- "step": 7000
93
- },
94
- {
95
- "epoch": 0.15,
96
- "learning_rate": 4.885681029286442e-06,
97
- "loss": 0.0644,
98
- "step": 7500
99
- },
100
- {
101
- "epoch": 0.16,
102
- "learning_rate": 4.87528839558521e-06,
103
- "loss": 0.0638,
104
- "step": 8000
105
- },
106
- {
107
- "epoch": 0.18,
108
- "learning_rate": 4.864895761883977e-06,
109
- "loss": 0.0627,
110
- "step": 8500
111
- },
112
- {
113
- "epoch": 0.19,
114
- "learning_rate": 4.854503128182744e-06,
115
- "loss": 0.0622,
116
- "step": 9000
117
- },
118
- {
119
- "epoch": 0.2,
120
- "learning_rate": 4.844110494481512e-06,
121
- "loss": 0.0615,
122
- "step": 9500
123
- },
124
- {
125
- "epoch": 0.21,
126
- "learning_rate": 4.833717860780279e-06,
127
- "loss": 0.061,
128
- "step": 10000
129
- },
130
- {
131
- "epoch": 0.22,
132
- "learning_rate": 4.823325227079046e-06,
133
- "loss": 0.0607,
134
- "step": 10500
135
- },
136
- {
137
- "epoch": 0.23,
138
- "learning_rate": 4.812932593377814e-06,
139
- "loss": 0.0603,
140
- "step": 11000
141
- },
142
- {
143
- "epoch": 0.24,
144
- "learning_rate": 4.802539959676581e-06,
145
- "loss": 0.0597,
146
- "step": 11500
147
- },
148
- {
149
- "epoch": 0.25,
150
- "learning_rate": 4.792147325975349e-06,
151
- "loss": 0.0593,
152
- "step": 12000
153
- },
154
- {
155
- "epoch": 0.26,
156
- "learning_rate": 4.781754692274117e-06,
157
- "loss": 0.0588,
158
- "step": 12500
159
- },
160
- {
161
- "epoch": 0.27,
162
- "learning_rate": 4.771362058572884e-06,
163
- "loss": 0.0587,
164
- "step": 13000
165
- },
166
- {
167
- "epoch": 0.28,
168
- "learning_rate": 4.760969424871652e-06,
169
- "loss": 0.058,
170
- "step": 13500
171
- },
172
- {
173
- "epoch": 0.29,
174
- "learning_rate": 4.750576791170419e-06,
175
- "loss": 0.0578,
176
- "step": 14000
177
- },
178
- {
179
- "epoch": 0.3,
180
- "learning_rate": 4.740184157469187e-06,
181
- "loss": 0.0574,
182
- "step": 14500
183
- },
184
- {
185
- "epoch": 0.31,
186
- "learning_rate": 4.729791523767954e-06,
187
- "loss": 0.0572,
188
- "step": 15000
189
- },
190
- {
191
- "epoch": 0.32,
192
- "learning_rate": 4.719398890066721e-06,
193
- "loss": 0.0566,
194
- "step": 15500
195
- },
196
- {
197
- "epoch": 0.33,
198
- "learning_rate": 4.709006256365489e-06,
199
- "loss": 0.0566,
200
- "step": 16000
201
- },
202
- {
203
- "epoch": 0.34,
204
- "learning_rate": 4.698613622664256e-06,
205
- "loss": 0.0562,
206
- "step": 16500
207
- },
208
- {
209
- "epoch": 0.35,
210
- "learning_rate": 4.688220988963023e-06,
211
- "loss": 0.0555,
212
- "step": 17000
213
- },
214
- {
215
- "epoch": 0.36,
216
- "learning_rate": 4.677828355261791e-06,
217
- "loss": 0.0554,
218
- "step": 17500
219
- },
220
- {
221
- "epoch": 0.37,
222
- "learning_rate": 4.667435721560558e-06,
223
- "loss": 0.0559,
224
- "step": 18000
225
- },
226
- {
227
- "epoch": 0.38,
228
- "learning_rate": 4.657043087859326e-06,
229
- "loss": 0.0553,
230
- "step": 18500
231
- },
232
- {
233
- "epoch": 0.39,
234
- "learning_rate": 4.646650454158093e-06,
235
- "loss": 0.0548,
236
- "step": 19000
237
- },
238
- {
239
- "epoch": 0.4,
240
- "learning_rate": 4.636257820456861e-06,
241
- "loss": 0.0548,
242
- "step": 19500
243
- },
244
- {
245
- "epoch": 0.41,
246
- "learning_rate": 4.6258651867556285e-06,
247
- "loss": 0.0544,
248
- "step": 20000
249
- },
250
- {
251
- "epoch": 0.42,
252
- "learning_rate": 4.6154725530543956e-06,
253
- "loss": 0.0542,
254
- "step": 20500
255
- },
256
- {
257
- "epoch": 0.43,
258
- "learning_rate": 4.6050799193531635e-06,
259
- "loss": 0.0542,
260
- "step": 21000
261
- },
262
- {
263
- "epoch": 0.44,
264
- "learning_rate": 4.5946872856519305e-06,
265
- "loss": 0.0539,
266
- "step": 21500
267
- },
268
- {
269
- "epoch": 0.45,
270
- "learning_rate": 4.5842946519506976e-06,
271
- "loss": 0.0538,
272
- "step": 22000
273
- },
274
- {
275
- "epoch": 0.46,
276
- "learning_rate": 4.5739020182494655e-06,
277
- "loss": 0.0535,
278
- "step": 22500
279
- },
280
- {
281
- "epoch": 0.47,
282
- "learning_rate": 4.5635093845482325e-06,
283
- "loss": 0.0534,
284
- "step": 23000
285
- },
286
- {
287
- "epoch": 0.48,
288
- "learning_rate": 4.5531167508469996e-06,
289
- "loss": 0.0535,
290
- "step": 23500
291
- },
292
- {
293
- "epoch": 0.49,
294
- "learning_rate": 4.5427241171457675e-06,
295
- "loss": 0.0533,
296
- "step": 24000
297
- },
298
- {
299
- "epoch": 0.51,
300
- "learning_rate": 4.5323314834445345e-06,
301
- "loss": 0.053,
302
- "step": 24500
303
- },
304
- {
305
- "epoch": 0.52,
306
- "learning_rate": 4.521938849743302e-06,
307
- "loss": 0.0529,
308
- "step": 25000
309
- },
310
- {
311
- "epoch": 0.53,
312
- "learning_rate": 4.5115462160420694e-06,
313
- "loss": 0.0527,
314
- "step": 25500
315
- },
316
- {
317
- "epoch": 0.54,
318
- "learning_rate": 4.501153582340837e-06,
319
- "loss": 0.0524,
320
- "step": 26000
321
- },
322
- {
323
- "epoch": 0.55,
324
- "learning_rate": 4.490760948639604e-06,
325
- "loss": 0.052,
326
- "step": 26500
327
- },
328
- {
329
- "epoch": 0.56,
330
- "learning_rate": 4.480368314938372e-06,
331
- "loss": 0.0522,
332
- "step": 27000
333
- },
334
- {
335
- "epoch": 0.57,
336
- "learning_rate": 4.469975681237139e-06,
337
- "loss": 0.0522,
338
- "step": 27500
339
- },
340
- {
341
- "epoch": 0.58,
342
- "learning_rate": 4.459583047535907e-06,
343
- "loss": 0.0519,
344
- "step": 28000
345
- },
346
- {
347
- "epoch": 0.59,
348
- "learning_rate": 4.449190413834674e-06,
349
- "loss": 0.0521,
350
- "step": 28500
351
- },
352
- {
353
- "epoch": 0.6,
354
- "learning_rate": 4.438797780133442e-06,
355
- "loss": 0.0518,
356
- "step": 29000
357
- },
358
- {
359
- "epoch": 0.61,
360
- "learning_rate": 4.428405146432209e-06,
361
- "loss": 0.0515,
362
- "step": 29500
363
- },
364
- {
365
- "epoch": 0.62,
366
- "learning_rate": 4.418012512730976e-06,
367
- "loss": 0.0514,
368
- "step": 30000
369
- },
370
- {
371
- "epoch": 0.63,
372
- "learning_rate": 4.407619879029744e-06,
373
- "loss": 0.0511,
374
- "step": 30500
375
- },
376
- {
377
- "epoch": 0.64,
378
- "learning_rate": 4.397227245328511e-06,
379
- "loss": 0.0512,
380
- "step": 31000
381
- },
382
- {
383
- "epoch": 0.65,
384
- "learning_rate": 4.386834611627279e-06,
385
- "loss": 0.0509,
386
- "step": 31500
387
- },
388
- {
389
- "epoch": 0.66,
390
- "learning_rate": 4.376441977926046e-06,
391
- "loss": 0.0513,
392
- "step": 32000
393
- },
394
- {
395
- "epoch": 0.67,
396
- "learning_rate": 4.366049344224814e-06,
397
- "loss": 0.0507,
398
- "step": 32500
399
- },
400
- {
401
- "epoch": 0.68,
402
- "learning_rate": 4.355656710523581e-06,
403
- "loss": 0.0506,
404
- "step": 33000
405
- },
406
- {
407
- "epoch": 0.69,
408
- "learning_rate": 4.345264076822349e-06,
409
- "loss": 0.0503,
410
- "step": 33500
411
- },
412
- {
413
- "epoch": 0.7,
414
- "learning_rate": 4.334871443121116e-06,
415
- "loss": 0.0501,
416
- "step": 34000
417
- },
418
- {
419
- "epoch": 0.71,
420
- "learning_rate": 4.324478809419884e-06,
421
- "loss": 0.0502,
422
- "step": 34500
423
- },
424
- {
425
- "epoch": 0.72,
426
- "learning_rate": 4.314086175718651e-06,
427
- "loss": 0.0505,
428
- "step": 35000
429
- },
430
- {
431
- "epoch": 0.73,
432
- "learning_rate": 4.303693542017419e-06,
433
- "loss": 0.0505,
434
- "step": 35500
435
- },
436
- {
437
- "epoch": 0.74,
438
- "learning_rate": 4.293300908316186e-06,
439
- "loss": 0.0501,
440
- "step": 36000
441
- },
442
- {
443
- "epoch": 0.75,
444
- "learning_rate": 4.282908274614953e-06,
445
- "loss": 0.05,
446
- "step": 36500
447
- },
448
- {
449
- "epoch": 0.76,
450
- "learning_rate": 4.272515640913721e-06,
451
- "loss": 0.05,
452
- "step": 37000
453
- },
454
- {
455
- "epoch": 0.77,
456
- "learning_rate": 4.262123007212488e-06,
457
- "loss": 0.0498,
458
- "step": 37500
459
- },
460
- {
461
- "epoch": 0.78,
462
- "learning_rate": 4.251730373511256e-06,
463
- "loss": 0.0498,
464
- "step": 38000
465
- },
466
- {
467
- "epoch": 0.79,
468
- "learning_rate": 4.241337739810023e-06,
469
- "loss": 0.0496,
470
- "step": 38500
471
- },
472
- {
473
- "epoch": 0.8,
474
- "learning_rate": 4.23094510610879e-06,
475
- "loss": 0.0496,
476
- "step": 39000
477
- },
478
- {
479
- "epoch": 0.81,
480
- "learning_rate": 4.220552472407558e-06,
481
- "loss": 0.0493,
482
- "step": 39500
483
- },
484
- {
485
- "epoch": 0.82,
486
- "learning_rate": 4.210159838706326e-06,
487
- "loss": 0.0495,
488
- "step": 40000
489
- },
490
- {
491
- "epoch": 0.83,
492
- "learning_rate": 4.199767205005093e-06,
493
- "loss": 0.049,
494
- "step": 40500
495
- },
496
- {
497
- "epoch": 0.85,
498
- "learning_rate": 4.189374571303861e-06,
499
- "loss": 0.0493,
500
- "step": 41000
501
- },
502
- {
503
- "epoch": 0.86,
504
- "learning_rate": 4.178981937602628e-06,
505
- "loss": 0.0491,
506
- "step": 41500
507
- },
508
- {
509
- "epoch": 0.87,
510
- "learning_rate": 4.1685893039013956e-06,
511
- "loss": 0.0492,
512
- "step": 42000
513
- },
514
- {
515
- "epoch": 0.88,
516
- "learning_rate": 4.158196670200163e-06,
517
- "loss": 0.0492,
518
- "step": 42500
519
- },
520
- {
521
- "epoch": 0.89,
522
- "learning_rate": 4.14780403649893e-06,
523
- "loss": 0.0487,
524
- "step": 43000
525
- },
526
- {
527
- "epoch": 0.9,
528
- "learning_rate": 4.1374114027976976e-06,
529
- "loss": 0.049,
530
- "step": 43500
531
- },
532
- {
533
- "epoch": 0.91,
534
- "learning_rate": 4.127018769096465e-06,
535
- "loss": 0.0487,
536
- "step": 44000
537
- },
538
- {
539
- "epoch": 0.92,
540
- "learning_rate": 4.116626135395232e-06,
541
- "loss": 0.0489,
542
- "step": 44500
543
- },
544
- {
545
- "epoch": 0.93,
546
- "learning_rate": 4.1062335016939995e-06,
547
- "loss": 0.0485,
548
- "step": 45000
549
- },
550
- {
551
- "epoch": 0.94,
552
- "learning_rate": 4.095840867992767e-06,
553
- "loss": 0.0486,
554
- "step": 45500
555
- },
556
- {
557
- "epoch": 0.95,
558
- "learning_rate": 4.0854482342915345e-06,
559
- "loss": 0.0485,
560
- "step": 46000
561
- },
562
- {
563
- "epoch": 0.96,
564
- "learning_rate": 4.0750556005903015e-06,
565
- "loss": 0.0487,
566
- "step": 46500
567
- },
568
- {
569
- "epoch": 0.97,
570
- "learning_rate": 4.0646629668890694e-06,
571
- "loss": 0.0482,
572
- "step": 47000
573
- },
574
- {
575
- "epoch": 0.98,
576
- "learning_rate": 4.054270333187837e-06,
577
- "loss": 0.0482,
578
- "step": 47500
579
- },
580
- {
581
- "epoch": 0.99,
582
- "learning_rate": 4.043877699486604e-06,
583
- "loss": 0.0481,
584
- "step": 48000
585
- },
586
- {
587
- "epoch": 1.0,
588
- "learning_rate": 4.033485065785372e-06,
589
- "loss": 0.0481,
590
- "step": 48500
591
- },
592
- {
593
- "epoch": 1.0,
594
- "eval_loss": 0.04874278977513313,
595
- "eval_runtime": 1479.0093,
596
- "eval_samples_per_second": 13.382,
597
- "eval_steps_per_second": 0.836,
598
- "step": 48511
599
- },
600
- {
601
- "epoch": 1.01,
602
- "learning_rate": 4.023092432084139e-06,
603
- "loss": 0.0479,
604
- "step": 49000
605
- },
606
- {
607
- "epoch": 1.02,
608
- "learning_rate": 4.012699798382906e-06,
609
- "loss": 0.0478,
610
- "step": 49500
611
- },
612
- {
613
- "epoch": 1.03,
614
- "learning_rate": 4.002307164681674e-06,
615
- "loss": 0.048,
616
- "step": 50000
617
- },
618
- {
619
- "epoch": 1.04,
620
- "learning_rate": 3.991914530980441e-06,
621
- "loss": 0.0477,
622
- "step": 50500
623
- },
624
- {
625
- "epoch": 1.05,
626
- "learning_rate": 3.981521897279208e-06,
627
- "loss": 0.0478,
628
- "step": 51000
629
- },
630
- {
631
- "epoch": 1.06,
632
- "learning_rate": 3.971129263577976e-06,
633
- "loss": 0.0476,
634
- "step": 51500
635
- },
636
- {
637
- "epoch": 1.07,
638
- "learning_rate": 3.960736629876743e-06,
639
- "loss": 0.0475,
640
- "step": 52000
641
- },
642
- {
643
- "epoch": 1.08,
644
- "learning_rate": 3.950343996175511e-06,
645
- "loss": 0.0473,
646
- "step": 52500
647
- },
648
- {
649
- "epoch": 1.09,
650
- "learning_rate": 3.939951362474278e-06,
651
- "loss": 0.0477,
652
- "step": 53000
653
- },
654
- {
655
- "epoch": 1.1,
656
- "learning_rate": 3.929558728773046e-06,
657
- "loss": 0.0472,
658
- "step": 53500
659
- },
660
- {
661
- "epoch": 1.11,
662
- "learning_rate": 3.919166095071813e-06,
663
- "loss": 0.0471,
664
- "step": 54000
665
- },
666
- {
667
- "epoch": 1.12,
668
- "learning_rate": 3.908773461370581e-06,
669
- "loss": 0.0472,
670
- "step": 54500
671
- },
672
- {
673
- "epoch": 1.13,
674
- "learning_rate": 3.898380827669349e-06,
675
- "loss": 0.0472,
676
- "step": 55000
677
- },
678
- {
679
- "epoch": 1.14,
680
- "learning_rate": 3.887988193968116e-06,
681
- "loss": 0.0471,
682
- "step": 55500
683
- },
684
- {
685
- "epoch": 1.15,
686
- "learning_rate": 3.877595560266883e-06,
687
- "loss": 0.0472,
688
- "step": 56000
689
- },
690
- {
691
- "epoch": 1.16,
692
- "learning_rate": 3.867202926565651e-06,
693
- "loss": 0.0471,
694
- "step": 56500
695
- },
696
- {
697
- "epoch": 1.17,
698
- "learning_rate": 3.856810292864418e-06,
699
- "loss": 0.047,
700
- "step": 57000
701
- },
702
- {
703
- "epoch": 1.19,
704
- "learning_rate": 3.846417659163185e-06,
705
- "loss": 0.0472,
706
- "step": 57500
707
- },
708
- {
709
- "epoch": 1.2,
710
- "learning_rate": 3.836025025461953e-06,
711
- "loss": 0.0471,
712
- "step": 58000
713
- },
714
- {
715
- "epoch": 1.21,
716
- "learning_rate": 3.82563239176072e-06,
717
- "loss": 0.0468,
718
- "step": 58500
719
- },
720
- {
721
- "epoch": 1.22,
722
- "learning_rate": 3.815239758059488e-06,
723
- "loss": 0.047,
724
- "step": 59000
725
- },
726
- {
727
- "epoch": 1.23,
728
- "learning_rate": 3.8048471243582554e-06,
729
- "loss": 0.0467,
730
- "step": 59500
731
- },
732
- {
733
- "epoch": 1.24,
734
- "learning_rate": 3.7944544906570224e-06,
735
- "loss": 0.0465,
736
- "step": 60000
737
- },
738
- {
739
- "epoch": 1.25,
740
- "learning_rate": 3.7840618569557903e-06,
741
- "loss": 0.0467,
742
- "step": 60500
743
- },
744
- {
745
- "epoch": 1.26,
746
- "learning_rate": 3.7736692232545574e-06,
747
- "loss": 0.0465,
748
- "step": 61000
749
- },
750
- {
751
- "epoch": 1.27,
752
- "learning_rate": 3.763276589553325e-06,
753
- "loss": 0.0467,
754
- "step": 61500
755
- },
756
- {
757
- "epoch": 1.28,
758
- "learning_rate": 3.7528839558520923e-06,
759
- "loss": 0.0468,
760
- "step": 62000
761
- },
762
- {
763
- "epoch": 1.29,
764
- "learning_rate": 3.7424913221508598e-06,
765
- "loss": 0.0469,
766
- "step": 62500
767
- },
768
- {
769
- "epoch": 1.3,
770
- "learning_rate": 3.7320986884496272e-06,
771
- "loss": 0.0465,
772
- "step": 63000
773
- },
774
- {
775
- "epoch": 1.31,
776
- "learning_rate": 3.7217060547483947e-06,
777
- "loss": 0.0462,
778
- "step": 63500
779
- },
780
- {
781
- "epoch": 1.32,
782
- "learning_rate": 3.7113134210471618e-06,
783
- "loss": 0.0467,
784
- "step": 64000
785
- },
786
- {
787
- "epoch": 1.33,
788
- "learning_rate": 3.7009207873459297e-06,
789
- "loss": 0.0465,
790
- "step": 64500
791
- },
792
- {
793
- "epoch": 1.34,
794
- "learning_rate": 3.6905281536446967e-06,
795
- "loss": 0.0464,
796
- "step": 65000
797
- },
798
- {
799
- "epoch": 1.35,
800
- "learning_rate": 3.6801355199434646e-06,
801
- "loss": 0.0465,
802
- "step": 65500
803
- },
804
- {
805
- "epoch": 1.36,
806
- "learning_rate": 3.6697428862422316e-06,
807
- "loss": 0.0465,
808
- "step": 66000
809
- },
810
- {
811
- "epoch": 1.37,
812
- "learning_rate": 3.659350252540999e-06,
813
- "loss": 0.0463,
814
- "step": 66500
815
- },
816
- {
817
- "epoch": 1.38,
818
- "learning_rate": 3.648957618839767e-06,
819
- "loss": 0.0461,
820
- "step": 67000
821
- },
822
- {
823
- "epoch": 1.39,
824
- "learning_rate": 3.638564985138534e-06,
825
- "loss": 0.0461,
826
- "step": 67500
827
- },
828
- {
829
- "epoch": 1.4,
830
- "learning_rate": 3.628172351437301e-06,
831
- "loss": 0.0462,
832
- "step": 68000
833
- },
834
- {
835
- "epoch": 1.41,
836
- "learning_rate": 3.617779717736069e-06,
837
- "loss": 0.0459,
838
- "step": 68500
839
- },
840
- {
841
- "epoch": 1.42,
842
- "learning_rate": 3.6073870840348365e-06,
843
- "loss": 0.0462,
844
- "step": 69000
845
- },
846
- {
847
- "epoch": 1.43,
848
- "learning_rate": 3.596994450333604e-06,
849
- "loss": 0.0459,
850
- "step": 69500
851
- },
852
- {
853
- "epoch": 1.44,
854
- "learning_rate": 3.5866018166323714e-06,
855
- "loss": 0.0459,
856
- "step": 70000
857
- },
858
- {
859
- "epoch": 1.45,
860
- "learning_rate": 3.5762091829311385e-06,
861
- "loss": 0.0457,
862
- "step": 70500
863
- },
864
- {
865
- "epoch": 1.46,
866
- "learning_rate": 3.5658165492299064e-06,
867
- "loss": 0.0457,
868
- "step": 71000
869
- },
870
- {
871
- "epoch": 1.47,
872
- "learning_rate": 3.5554239155286734e-06,
873
- "loss": 0.0456,
874
- "step": 71500
875
- },
876
- {
877
- "epoch": 1.48,
878
- "learning_rate": 3.5450312818274413e-06,
879
- "loss": 0.0457,
880
- "step": 72000
881
- },
882
- {
883
- "epoch": 1.49,
884
- "learning_rate": 3.5346386481262083e-06,
885
- "loss": 0.0461,
886
- "step": 72500
887
- },
888
- {
889
- "epoch": 1.5,
890
- "learning_rate": 3.524246014424976e-06,
891
- "loss": 0.046,
892
- "step": 73000
893
- },
894
- {
895
- "epoch": 1.52,
896
- "learning_rate": 3.5138533807237433e-06,
897
- "loss": 0.0456,
898
- "step": 73500
899
- },
900
- {
901
- "epoch": 1.53,
902
- "learning_rate": 3.5034607470225108e-06,
903
- "loss": 0.0456,
904
- "step": 74000
905
- },
906
- {
907
- "epoch": 1.54,
908
- "learning_rate": 3.493068113321278e-06,
909
- "loss": 0.0458,
910
- "step": 74500
911
- },
912
- {
913
- "epoch": 1.55,
914
- "learning_rate": 3.4826754796200457e-06,
915
- "loss": 0.0455,
916
- "step": 75000
917
- },
918
- {
919
- "epoch": 1.56,
920
- "learning_rate": 3.4722828459188128e-06,
921
- "loss": 0.0458,
922
- "step": 75500
923
- },
924
- {
925
- "epoch": 1.57,
926
- "learning_rate": 3.4618902122175806e-06,
927
- "loss": 0.0455,
928
- "step": 76000
929
- },
930
- {
931
- "epoch": 1.58,
932
- "learning_rate": 3.451497578516348e-06,
933
- "loss": 0.0459,
934
- "step": 76500
935
- },
936
- {
937
- "epoch": 1.59,
938
- "learning_rate": 3.441104944815115e-06,
939
- "loss": 0.0455,
940
- "step": 77000
941
- },
942
- {
943
- "epoch": 1.6,
944
- "learning_rate": 3.430712311113883e-06,
945
- "loss": 0.0457,
946
- "step": 77500
947
- },
948
- {
949
- "epoch": 1.61,
950
- "learning_rate": 3.42031967741265e-06,
951
- "loss": 0.0456,
952
- "step": 78000
953
- },
954
- {
955
- "epoch": 1.62,
956
- "learning_rate": 3.409927043711417e-06,
957
- "loss": 0.0453,
958
- "step": 78500
959
- },
960
- {
961
- "epoch": 1.63,
962
- "learning_rate": 3.399534410010185e-06,
963
- "loss": 0.0454,
964
- "step": 79000
965
- },
966
- {
967
- "epoch": 1.64,
968
- "learning_rate": 3.3891417763089525e-06,
969
- "loss": 0.0453,
970
- "step": 79500
971
- },
972
- {
973
- "epoch": 1.65,
974
- "learning_rate": 3.37874914260772e-06,
975
- "loss": 0.0455,
976
- "step": 80000
977
- },
978
- {
979
- "epoch": 1.66,
980
- "learning_rate": 3.3683565089064875e-06,
981
- "loss": 0.0453,
982
- "step": 80500
983
- },
984
- {
985
- "epoch": 1.67,
986
- "learning_rate": 3.3579638752052545e-06,
987
- "loss": 0.0455,
988
- "step": 81000
989
- },
990
- {
991
- "epoch": 1.68,
992
- "learning_rate": 3.3475712415040224e-06,
993
- "loss": 0.0451,
994
- "step": 81500
995
- },
996
- {
997
- "epoch": 1.69,
998
- "learning_rate": 3.3371786078027895e-06,
999
- "loss": 0.0448,
1000
- "step": 82000
1001
- },
1002
- {
1003
- "epoch": 1.7,
1004
- "learning_rate": 3.3267859741015574e-06,
1005
- "loss": 0.0453,
1006
- "step": 82500
1007
- },
1008
- {
1009
- "epoch": 1.71,
1010
- "learning_rate": 3.3163933404003244e-06,
1011
- "loss": 0.045,
1012
- "step": 83000
1013
- },
1014
- {
1015
- "epoch": 1.72,
1016
- "learning_rate": 3.306000706699092e-06,
1017
- "loss": 0.045,
1018
- "step": 83500
1019
- },
1020
- {
1021
- "epoch": 1.73,
1022
- "learning_rate": 3.2956080729978598e-06,
1023
- "loss": 0.0453,
1024
- "step": 84000
1025
- },
1026
- {
1027
- "epoch": 1.74,
1028
- "learning_rate": 3.285215439296627e-06,
1029
- "loss": 0.045,
1030
- "step": 84500
1031
- },
1032
- {
1033
- "epoch": 1.75,
1034
- "learning_rate": 3.274822805595394e-06,
1035
- "loss": 0.0454,
1036
- "step": 85000
1037
- },
1038
- {
1039
- "epoch": 1.76,
1040
- "learning_rate": 3.2644301718941618e-06,
1041
- "loss": 0.045,
1042
- "step": 85500
1043
- },
1044
- {
1045
- "epoch": 1.77,
1046
- "learning_rate": 3.254037538192929e-06,
1047
- "loss": 0.0449,
1048
- "step": 86000
1049
- },
1050
- {
1051
- "epoch": 1.78,
1052
- "learning_rate": 3.2436449044916967e-06,
1053
- "loss": 0.045,
1054
- "step": 86500
1055
- },
1056
- {
1057
- "epoch": 1.79,
1058
- "learning_rate": 3.233252270790464e-06,
1059
- "loss": 0.0448,
1060
- "step": 87000
1061
- },
1062
- {
1063
- "epoch": 1.8,
1064
- "learning_rate": 3.2228596370892312e-06,
1065
- "loss": 0.0449,
1066
- "step": 87500
1067
- },
1068
- {
1069
- "epoch": 1.81,
1070
- "learning_rate": 3.212467003387999e-06,
1071
- "loss": 0.0446,
1072
- "step": 88000
1073
- },
1074
- {
1075
- "epoch": 1.82,
1076
- "learning_rate": 3.202074369686766e-06,
1077
- "loss": 0.045,
1078
- "step": 88500
1079
- },
1080
- {
1081
- "epoch": 1.83,
1082
- "learning_rate": 3.191681735985534e-06,
1083
- "loss": 0.0449,
1084
- "step": 89000
1085
- },
1086
- {
1087
- "epoch": 1.84,
1088
- "learning_rate": 3.181289102284301e-06,
1089
- "loss": 0.0446,
1090
- "step": 89500
1091
- },
1092
- {
1093
- "epoch": 1.86,
1094
- "learning_rate": 3.1708964685830686e-06,
1095
- "loss": 0.0445,
1096
- "step": 90000
1097
- },
1098
- {
1099
- "epoch": 1.87,
1100
- "learning_rate": 3.160503834881836e-06,
1101
- "loss": 0.0448,
1102
- "step": 90500
1103
- },
1104
- {
1105
- "epoch": 1.88,
1106
- "learning_rate": 3.1501112011806035e-06,
1107
- "loss": 0.0445,
1108
- "step": 91000
1109
- },
1110
- {
1111
- "epoch": 1.89,
1112
- "learning_rate": 3.1397185674793706e-06,
1113
- "loss": 0.0447,
1114
- "step": 91500
1115
- },
1116
- {
1117
- "epoch": 1.9,
1118
- "learning_rate": 3.1293259337781385e-06,
1119
- "loss": 0.0446,
1120
- "step": 92000
1121
- },
1122
- {
1123
- "epoch": 1.91,
1124
- "learning_rate": 3.1189333000769055e-06,
1125
- "loss": 0.0449,
1126
- "step": 92500
1127
- },
1128
- {
1129
- "epoch": 1.92,
1130
- "learning_rate": 3.1085406663756734e-06,
1131
- "loss": 0.0446,
1132
- "step": 93000
1133
- },
1134
- {
1135
- "epoch": 1.93,
1136
- "learning_rate": 3.0981480326744404e-06,
1137
- "loss": 0.0442,
1138
- "step": 93500
1139
- },
1140
- {
1141
- "epoch": 1.94,
1142
- "learning_rate": 3.087755398973208e-06,
1143
- "loss": 0.0445,
1144
- "step": 94000
1145
- },
1146
- {
1147
- "epoch": 1.95,
1148
- "learning_rate": 3.077362765271976e-06,
1149
- "loss": 0.0444,
1150
- "step": 94500
1151
- },
1152
- {
1153
- "epoch": 1.96,
1154
- "learning_rate": 3.066970131570743e-06,
1155
- "loss": 0.0444,
1156
- "step": 95000
1157
- },
1158
- {
1159
- "epoch": 1.97,
1160
- "learning_rate": 3.05657749786951e-06,
1161
- "loss": 0.0444,
1162
- "step": 95500
1163
- },
1164
- {
1165
- "epoch": 1.98,
1166
- "learning_rate": 3.046184864168278e-06,
1167
- "loss": 0.0444,
1168
- "step": 96000
1169
- },
1170
- {
1171
- "epoch": 1.99,
1172
- "learning_rate": 3.0357922304670453e-06,
1173
- "loss": 0.0446,
1174
- "step": 96500
1175
- },
1176
- {
1177
- "epoch": 2.0,
1178
- "learning_rate": 3.0253995967658127e-06,
1179
- "loss": 0.0443,
1180
- "step": 97000
1181
- },
1182
- {
1183
- "epoch": 2.0,
1184
- "eval_loss": 0.04501689225435257,
1185
- "eval_runtime": 1478.7113,
1186
- "eval_samples_per_second": 13.385,
1187
- "eval_steps_per_second": 0.837,
1188
- "step": 97022
1189
- }
1190
- ],
1191
- "max_steps": 242555,
1192
- "num_train_epochs": 5,
1193
- "total_flos": 1.1940942833642373e+18,
1194
- "trial_name": null,
1195
- "trial_params": null
1196
- }