mosama commited on
Commit
3b51e33
1 Parent(s): e4a4d4a

Delete trainer_state.json

Browse files
Files changed (1) hide show
  1. trainer_state.json +0 -1096
trainer_state.json DELETED
@@ -1,1096 +0,0 @@
1
- {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 1.4354066985645932,
5
- "global_step": 1800,
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": 0.0001992025518341308,
13
- "loss": 1.2986,
14
- "step": 10
15
- },
16
- {
17
- "epoch": 0.02,
18
- "learning_rate": 0.00019840510366826158,
19
- "loss": 1.1659,
20
- "step": 20
21
- },
22
- {
23
- "epoch": 0.02,
24
- "learning_rate": 0.00019760765550239236,
25
- "loss": 1.1845,
26
- "step": 30
27
- },
28
- {
29
- "epoch": 0.03,
30
- "learning_rate": 0.00019681020733652312,
31
- "loss": 1.1455,
32
- "step": 40
33
- },
34
- {
35
- "epoch": 0.04,
36
- "learning_rate": 0.0001960127591706539,
37
- "loss": 1.173,
38
- "step": 50
39
- },
40
- {
41
- "epoch": 0.05,
42
- "learning_rate": 0.0001952153110047847,
43
- "loss": 1.164,
44
- "step": 60
45
- },
46
- {
47
- "epoch": 0.06,
48
- "learning_rate": 0.00019441786283891547,
49
- "loss": 1.1505,
50
- "step": 70
51
- },
52
- {
53
- "epoch": 0.06,
54
- "learning_rate": 0.00019362041467304626,
55
- "loss": 1.1619,
56
- "step": 80
57
- },
58
- {
59
- "epoch": 0.07,
60
- "learning_rate": 0.00019282296650717704,
61
- "loss": 1.137,
62
- "step": 90
63
- },
64
- {
65
- "epoch": 0.08,
66
- "learning_rate": 0.00019202551834130783,
67
- "loss": 1.1657,
68
- "step": 100
69
- },
70
- {
71
- "epoch": 0.09,
72
- "learning_rate": 0.0001912280701754386,
73
- "loss": 1.1437,
74
- "step": 110
75
- },
76
- {
77
- "epoch": 0.1,
78
- "learning_rate": 0.00019043062200956937,
79
- "loss": 1.1532,
80
- "step": 120
81
- },
82
- {
83
- "epoch": 0.1,
84
- "learning_rate": 0.00018963317384370016,
85
- "loss": 1.1612,
86
- "step": 130
87
- },
88
- {
89
- "epoch": 0.11,
90
- "learning_rate": 0.00018883572567783094,
91
- "loss": 1.1774,
92
- "step": 140
93
- },
94
- {
95
- "epoch": 0.12,
96
- "learning_rate": 0.00018803827751196172,
97
- "loss": 1.1317,
98
- "step": 150
99
- },
100
- {
101
- "epoch": 0.13,
102
- "learning_rate": 0.0001872408293460925,
103
- "loss": 1.1472,
104
- "step": 160
105
- },
106
- {
107
- "epoch": 0.14,
108
- "learning_rate": 0.0001864433811802233,
109
- "loss": 1.1432,
110
- "step": 170
111
- },
112
- {
113
- "epoch": 0.14,
114
- "learning_rate": 0.00018564593301435408,
115
- "loss": 1.1301,
116
- "step": 180
117
- },
118
- {
119
- "epoch": 0.15,
120
- "learning_rate": 0.00018484848484848484,
121
- "loss": 1.137,
122
- "step": 190
123
- },
124
- {
125
- "epoch": 0.16,
126
- "learning_rate": 0.00018405103668261562,
127
- "loss": 1.1175,
128
- "step": 200
129
- },
130
- {
131
- "epoch": 0.17,
132
- "learning_rate": 0.0001832535885167464,
133
- "loss": 1.1263,
134
- "step": 210
135
- },
136
- {
137
- "epoch": 0.18,
138
- "learning_rate": 0.0001824561403508772,
139
- "loss": 1.192,
140
- "step": 220
141
- },
142
- {
143
- "epoch": 0.18,
144
- "learning_rate": 0.00018165869218500797,
145
- "loss": 1.1642,
146
- "step": 230
147
- },
148
- {
149
- "epoch": 0.19,
150
- "learning_rate": 0.00018086124401913876,
151
- "loss": 1.134,
152
- "step": 240
153
- },
154
- {
155
- "epoch": 0.2,
156
- "learning_rate": 0.00018006379585326954,
157
- "loss": 1.1357,
158
- "step": 250
159
- },
160
- {
161
- "epoch": 0.21,
162
- "learning_rate": 0.00017926634768740033,
163
- "loss": 1.1093,
164
- "step": 260
165
- },
166
- {
167
- "epoch": 0.22,
168
- "learning_rate": 0.0001784688995215311,
169
- "loss": 1.122,
170
- "step": 270
171
- },
172
- {
173
- "epoch": 0.22,
174
- "learning_rate": 0.0001776714513556619,
175
- "loss": 1.1453,
176
- "step": 280
177
- },
178
- {
179
- "epoch": 0.23,
180
- "learning_rate": 0.00017687400318979268,
181
- "loss": 1.1171,
182
- "step": 290
183
- },
184
- {
185
- "epoch": 0.24,
186
- "learning_rate": 0.00017607655502392347,
187
- "loss": 1.1384,
188
- "step": 300
189
- },
190
- {
191
- "epoch": 0.25,
192
- "learning_rate": 0.00017527910685805425,
193
- "loss": 1.1345,
194
- "step": 310
195
- },
196
- {
197
- "epoch": 0.26,
198
- "learning_rate": 0.00017448165869218504,
199
- "loss": 1.1423,
200
- "step": 320
201
- },
202
- {
203
- "epoch": 0.26,
204
- "learning_rate": 0.0001736842105263158,
205
- "loss": 1.1592,
206
- "step": 330
207
- },
208
- {
209
- "epoch": 0.27,
210
- "learning_rate": 0.00017288676236044658,
211
- "loss": 1.1214,
212
- "step": 340
213
- },
214
- {
215
- "epoch": 0.28,
216
- "learning_rate": 0.00017208931419457736,
217
- "loss": 1.0846,
218
- "step": 350
219
- },
220
- {
221
- "epoch": 0.29,
222
- "learning_rate": 0.00017129186602870815,
223
- "loss": 1.0876,
224
- "step": 360
225
- },
226
- {
227
- "epoch": 0.3,
228
- "learning_rate": 0.00017049441786283893,
229
- "loss": 1.1132,
230
- "step": 370
231
- },
232
- {
233
- "epoch": 0.3,
234
- "learning_rate": 0.00016969696969696972,
235
- "loss": 1.0876,
236
- "step": 380
237
- },
238
- {
239
- "epoch": 0.31,
240
- "learning_rate": 0.0001688995215311005,
241
- "loss": 1.0985,
242
- "step": 390
243
- },
244
- {
245
- "epoch": 0.32,
246
- "learning_rate": 0.00016810207336523126,
247
- "loss": 1.1361,
248
- "step": 400
249
- },
250
- {
251
- "epoch": 0.33,
252
- "learning_rate": 0.00016730462519936204,
253
- "loss": 1.1393,
254
- "step": 410
255
- },
256
- {
257
- "epoch": 0.33,
258
- "learning_rate": 0.00016650717703349283,
259
- "loss": 1.1464,
260
- "step": 420
261
- },
262
- {
263
- "epoch": 0.34,
264
- "learning_rate": 0.0001657097288676236,
265
- "loss": 1.1402,
266
- "step": 430
267
- },
268
- {
269
- "epoch": 0.35,
270
- "learning_rate": 0.0001649122807017544,
271
- "loss": 1.1232,
272
- "step": 440
273
- },
274
- {
275
- "epoch": 0.36,
276
- "learning_rate": 0.00016411483253588518,
277
- "loss": 1.1154,
278
- "step": 450
279
- },
280
- {
281
- "epoch": 0.37,
282
- "learning_rate": 0.00016331738437001597,
283
- "loss": 1.1609,
284
- "step": 460
285
- },
286
- {
287
- "epoch": 0.37,
288
- "learning_rate": 0.00016251993620414675,
289
- "loss": 1.1359,
290
- "step": 470
291
- },
292
- {
293
- "epoch": 0.38,
294
- "learning_rate": 0.0001617224880382775,
295
- "loss": 1.133,
296
- "step": 480
297
- },
298
- {
299
- "epoch": 0.39,
300
- "learning_rate": 0.0001609250398724083,
301
- "loss": 1.1198,
302
- "step": 490
303
- },
304
- {
305
- "epoch": 0.4,
306
- "learning_rate": 0.00016012759170653908,
307
- "loss": 1.137,
308
- "step": 500
309
- },
310
- {
311
- "epoch": 0.41,
312
- "learning_rate": 0.00015933014354066986,
313
- "loss": 1.0804,
314
- "step": 510
315
- },
316
- {
317
- "epoch": 0.41,
318
- "learning_rate": 0.00015853269537480065,
319
- "loss": 1.0814,
320
- "step": 520
321
- },
322
- {
323
- "epoch": 0.42,
324
- "learning_rate": 0.00015773524720893143,
325
- "loss": 1.1,
326
- "step": 530
327
- },
328
- {
329
- "epoch": 0.43,
330
- "learning_rate": 0.00015693779904306222,
331
- "loss": 1.1329,
332
- "step": 540
333
- },
334
- {
335
- "epoch": 0.44,
336
- "learning_rate": 0.00015614035087719297,
337
- "loss": 1.1124,
338
- "step": 550
339
- },
340
- {
341
- "epoch": 0.45,
342
- "learning_rate": 0.00015534290271132376,
343
- "loss": 1.0867,
344
- "step": 560
345
- },
346
- {
347
- "epoch": 0.45,
348
- "learning_rate": 0.00015454545454545454,
349
- "loss": 1.1013,
350
- "step": 570
351
- },
352
- {
353
- "epoch": 0.46,
354
- "learning_rate": 0.00015374800637958533,
355
- "loss": 1.1241,
356
- "step": 580
357
- },
358
- {
359
- "epoch": 0.47,
360
- "learning_rate": 0.0001529505582137161,
361
- "loss": 1.1047,
362
- "step": 590
363
- },
364
- {
365
- "epoch": 0.48,
366
- "learning_rate": 0.0001521531100478469,
367
- "loss": 1.136,
368
- "step": 600
369
- },
370
- {
371
- "epoch": 0.49,
372
- "learning_rate": 0.00015135566188197768,
373
- "loss": 1.1262,
374
- "step": 610
375
- },
376
- {
377
- "epoch": 0.49,
378
- "learning_rate": 0.00015055821371610844,
379
- "loss": 1.0916,
380
- "step": 620
381
- },
382
- {
383
- "epoch": 0.5,
384
- "learning_rate": 0.00014976076555023922,
385
- "loss": 1.114,
386
- "step": 630
387
- },
388
- {
389
- "epoch": 0.51,
390
- "learning_rate": 0.00014896331738437,
391
- "loss": 1.0845,
392
- "step": 640
393
- },
394
- {
395
- "epoch": 0.52,
396
- "learning_rate": 0.0001481658692185008,
397
- "loss": 1.1224,
398
- "step": 650
399
- },
400
- {
401
- "epoch": 0.53,
402
- "learning_rate": 0.00014736842105263158,
403
- "loss": 1.1181,
404
- "step": 660
405
- },
406
- {
407
- "epoch": 0.53,
408
- "learning_rate": 0.00014657097288676236,
409
- "loss": 1.1082,
410
- "step": 670
411
- },
412
- {
413
- "epoch": 0.54,
414
- "learning_rate": 0.00014577352472089315,
415
- "loss": 1.0977,
416
- "step": 680
417
- },
418
- {
419
- "epoch": 0.55,
420
- "learning_rate": 0.00014497607655502393,
421
- "loss": 1.1359,
422
- "step": 690
423
- },
424
- {
425
- "epoch": 0.56,
426
- "learning_rate": 0.0001441786283891547,
427
- "loss": 1.1276,
428
- "step": 700
429
- },
430
- {
431
- "epoch": 0.57,
432
- "learning_rate": 0.00014338118022328547,
433
- "loss": 1.1239,
434
- "step": 710
435
- },
436
- {
437
- "epoch": 0.57,
438
- "learning_rate": 0.00014258373205741626,
439
- "loss": 1.096,
440
- "step": 720
441
- },
442
- {
443
- "epoch": 0.58,
444
- "learning_rate": 0.00014178628389154707,
445
- "loss": 1.1071,
446
- "step": 730
447
- },
448
- {
449
- "epoch": 0.59,
450
- "learning_rate": 0.00014098883572567786,
451
- "loss": 1.113,
452
- "step": 740
453
- },
454
- {
455
- "epoch": 0.6,
456
- "learning_rate": 0.00014019138755980864,
457
- "loss": 1.0914,
458
- "step": 750
459
- },
460
- {
461
- "epoch": 0.61,
462
- "learning_rate": 0.0001393939393939394,
463
- "loss": 1.1012,
464
- "step": 760
465
- },
466
- {
467
- "epoch": 0.61,
468
- "learning_rate": 0.00013859649122807018,
469
- "loss": 1.1075,
470
- "step": 770
471
- },
472
- {
473
- "epoch": 0.62,
474
- "learning_rate": 0.00013779904306220097,
475
- "loss": 1.1001,
476
- "step": 780
477
- },
478
- {
479
- "epoch": 0.63,
480
- "learning_rate": 0.00013700159489633175,
481
- "loss": 1.1321,
482
- "step": 790
483
- },
484
- {
485
- "epoch": 0.64,
486
- "learning_rate": 0.00013620414673046254,
487
- "loss": 1.1067,
488
- "step": 800
489
- },
490
- {
491
- "epoch": 0.65,
492
- "learning_rate": 0.00013540669856459332,
493
- "loss": 1.137,
494
- "step": 810
495
- },
496
- {
497
- "epoch": 0.65,
498
- "learning_rate": 0.0001346092503987241,
499
- "loss": 1.0714,
500
- "step": 820
501
- },
502
- {
503
- "epoch": 0.66,
504
- "learning_rate": 0.00013381180223285486,
505
- "loss": 1.1269,
506
- "step": 830
507
- },
508
- {
509
- "epoch": 0.67,
510
- "learning_rate": 0.00013301435406698565,
511
- "loss": 1.1011,
512
- "step": 840
513
- },
514
- {
515
- "epoch": 0.68,
516
- "learning_rate": 0.00013221690590111643,
517
- "loss": 1.0957,
518
- "step": 850
519
- },
520
- {
521
- "epoch": 0.69,
522
- "learning_rate": 0.00013141945773524722,
523
- "loss": 1.0886,
524
- "step": 860
525
- },
526
- {
527
- "epoch": 0.69,
528
- "learning_rate": 0.000130622009569378,
529
- "loss": 1.0875,
530
- "step": 870
531
- },
532
- {
533
- "epoch": 0.7,
534
- "learning_rate": 0.0001298245614035088,
535
- "loss": 1.1221,
536
- "step": 880
537
- },
538
- {
539
- "epoch": 0.71,
540
- "learning_rate": 0.00012902711323763957,
541
- "loss": 1.0864,
542
- "step": 890
543
- },
544
- {
545
- "epoch": 0.72,
546
- "learning_rate": 0.00012822966507177036,
547
- "loss": 1.1336,
548
- "step": 900
549
- },
550
- {
551
- "epoch": 0.73,
552
- "learning_rate": 0.0001274322169059011,
553
- "loss": 1.0748,
554
- "step": 910
555
- },
556
- {
557
- "epoch": 0.73,
558
- "learning_rate": 0.0001266347687400319,
559
- "loss": 1.1184,
560
- "step": 920
561
- },
562
- {
563
- "epoch": 0.74,
564
- "learning_rate": 0.00012583732057416268,
565
- "loss": 1.0824,
566
- "step": 930
567
- },
568
- {
569
- "epoch": 0.75,
570
- "learning_rate": 0.00012503987240829347,
571
- "loss": 1.1031,
572
- "step": 940
573
- },
574
- {
575
- "epoch": 0.76,
576
- "learning_rate": 0.00012424242424242425,
577
- "loss": 1.0578,
578
- "step": 950
579
- },
580
- {
581
- "epoch": 0.77,
582
- "learning_rate": 0.00012344497607655504,
583
- "loss": 1.1143,
584
- "step": 960
585
- },
586
- {
587
- "epoch": 0.77,
588
- "learning_rate": 0.00012264752791068582,
589
- "loss": 1.0915,
590
- "step": 970
591
- },
592
- {
593
- "epoch": 0.78,
594
- "learning_rate": 0.00012185007974481659,
595
- "loss": 1.1074,
596
- "step": 980
597
- },
598
- {
599
- "epoch": 0.79,
600
- "learning_rate": 0.00012105263157894738,
601
- "loss": 1.043,
602
- "step": 990
603
- },
604
- {
605
- "epoch": 0.8,
606
- "learning_rate": 0.00012025518341307815,
607
- "loss": 1.0994,
608
- "step": 1000
609
- },
610
- {
611
- "epoch": 0.81,
612
- "learning_rate": 0.00011945773524720893,
613
- "loss": 1.1137,
614
- "step": 1010
615
- },
616
- {
617
- "epoch": 0.81,
618
- "learning_rate": 0.00011866028708133972,
619
- "loss": 1.0959,
620
- "step": 1020
621
- },
622
- {
623
- "epoch": 0.82,
624
- "learning_rate": 0.0001178628389154705,
625
- "loss": 1.1064,
626
- "step": 1030
627
- },
628
- {
629
- "epoch": 0.83,
630
- "learning_rate": 0.00011706539074960127,
631
- "loss": 1.0502,
632
- "step": 1040
633
- },
634
- {
635
- "epoch": 0.84,
636
- "learning_rate": 0.00011626794258373206,
637
- "loss": 1.0904,
638
- "step": 1050
639
- },
640
- {
641
- "epoch": 0.85,
642
- "learning_rate": 0.00011547049441786284,
643
- "loss": 1.1022,
644
- "step": 1060
645
- },
646
- {
647
- "epoch": 0.85,
648
- "learning_rate": 0.00011467304625199363,
649
- "loss": 1.0933,
650
- "step": 1070
651
- },
652
- {
653
- "epoch": 0.86,
654
- "learning_rate": 0.0001138755980861244,
655
- "loss": 1.1365,
656
- "step": 1080
657
- },
658
- {
659
- "epoch": 0.87,
660
- "learning_rate": 0.00011307814992025518,
661
- "loss": 1.0681,
662
- "step": 1090
663
- },
664
- {
665
- "epoch": 0.88,
666
- "learning_rate": 0.00011228070175438597,
667
- "loss": 1.0998,
668
- "step": 1100
669
- },
670
- {
671
- "epoch": 0.89,
672
- "learning_rate": 0.00011148325358851674,
673
- "loss": 1.1021,
674
- "step": 1110
675
- },
676
- {
677
- "epoch": 0.89,
678
- "learning_rate": 0.00011068580542264752,
679
- "loss": 1.0832,
680
- "step": 1120
681
- },
682
- {
683
- "epoch": 0.9,
684
- "learning_rate": 0.00010988835725677831,
685
- "loss": 1.1063,
686
- "step": 1130
687
- },
688
- {
689
- "epoch": 0.91,
690
- "learning_rate": 0.00010909090909090909,
691
- "loss": 1.0963,
692
- "step": 1140
693
- },
694
- {
695
- "epoch": 0.92,
696
- "learning_rate": 0.00010829346092503986,
697
- "loss": 1.0916,
698
- "step": 1150
699
- },
700
- {
701
- "epoch": 0.93,
702
- "learning_rate": 0.00010749601275917065,
703
- "loss": 1.1188,
704
- "step": 1160
705
- },
706
- {
707
- "epoch": 0.93,
708
- "learning_rate": 0.00010669856459330143,
709
- "loss": 1.1009,
710
- "step": 1170
711
- },
712
- {
713
- "epoch": 0.94,
714
- "learning_rate": 0.00010590111642743222,
715
- "loss": 1.0814,
716
- "step": 1180
717
- },
718
- {
719
- "epoch": 0.95,
720
- "learning_rate": 0.00010510366826156299,
721
- "loss": 1.1109,
722
- "step": 1190
723
- },
724
- {
725
- "epoch": 0.96,
726
- "learning_rate": 0.00010430622009569377,
727
- "loss": 1.0631,
728
- "step": 1200
729
- },
730
- {
731
- "epoch": 0.96,
732
- "learning_rate": 0.00010350877192982457,
733
- "loss": 1.1277,
734
- "step": 1210
735
- },
736
- {
737
- "epoch": 0.97,
738
- "learning_rate": 0.00010271132376395536,
739
- "loss": 1.0947,
740
- "step": 1220
741
- },
742
- {
743
- "epoch": 0.98,
744
- "learning_rate": 0.00010191387559808614,
745
- "loss": 1.0699,
746
- "step": 1230
747
- },
748
- {
749
- "epoch": 0.99,
750
- "learning_rate": 0.00010111642743221693,
751
- "loss": 1.0619,
752
- "step": 1240
753
- },
754
- {
755
- "epoch": 1.0,
756
- "learning_rate": 0.0001003189792663477,
757
- "loss": 1.0974,
758
- "step": 1250
759
- },
760
- {
761
- "epoch": 1.0,
762
- "learning_rate": 9.952153110047847e-05,
763
- "loss": 1.0663,
764
- "step": 1260
765
- },
766
- {
767
- "epoch": 1.01,
768
- "learning_rate": 9.872408293460925e-05,
769
- "loss": 1.03,
770
- "step": 1270
771
- },
772
- {
773
- "epoch": 1.02,
774
- "learning_rate": 9.792663476874004e-05,
775
- "loss": 1.0108,
776
- "step": 1280
777
- },
778
- {
779
- "epoch": 1.03,
780
- "learning_rate": 9.712918660287081e-05,
781
- "loss": 0.9865,
782
- "step": 1290
783
- },
784
- {
785
- "epoch": 1.04,
786
- "learning_rate": 9.633173843700159e-05,
787
- "loss": 1.0006,
788
- "step": 1300
789
- },
790
- {
791
- "epoch": 1.04,
792
- "learning_rate": 9.553429027113238e-05,
793
- "loss": 1.029,
794
- "step": 1310
795
- },
796
- {
797
- "epoch": 1.05,
798
- "learning_rate": 9.473684210526316e-05,
799
- "loss": 1.0177,
800
- "step": 1320
801
- },
802
- {
803
- "epoch": 1.06,
804
- "learning_rate": 9.393939393939395e-05,
805
- "loss": 1.004,
806
- "step": 1330
807
- },
808
- {
809
- "epoch": 1.07,
810
- "learning_rate": 9.314194577352473e-05,
811
- "loss": 1.0121,
812
- "step": 1340
813
- },
814
- {
815
- "epoch": 1.08,
816
- "learning_rate": 9.234449760765552e-05,
817
- "loss": 0.9881,
818
- "step": 1350
819
- },
820
- {
821
- "epoch": 1.08,
822
- "learning_rate": 9.154704944178629e-05,
823
- "loss": 0.9858,
824
- "step": 1360
825
- },
826
- {
827
- "epoch": 1.09,
828
- "learning_rate": 9.074960127591707e-05,
829
- "loss": 1.0383,
830
- "step": 1370
831
- },
832
- {
833
- "epoch": 1.1,
834
- "learning_rate": 8.995215311004786e-05,
835
- "loss": 0.9925,
836
- "step": 1380
837
- },
838
- {
839
- "epoch": 1.11,
840
- "learning_rate": 8.915470494417864e-05,
841
- "loss": 1.0,
842
- "step": 1390
843
- },
844
- {
845
- "epoch": 1.12,
846
- "learning_rate": 8.835725677830941e-05,
847
- "loss": 1.0067,
848
- "step": 1400
849
- },
850
- {
851
- "epoch": 1.12,
852
- "learning_rate": 8.75598086124402e-05,
853
- "loss": 0.9852,
854
- "step": 1410
855
- },
856
- {
857
- "epoch": 1.13,
858
- "learning_rate": 8.676236044657098e-05,
859
- "loss": 1.0226,
860
- "step": 1420
861
- },
862
- {
863
- "epoch": 1.14,
864
- "learning_rate": 8.596491228070177e-05,
865
- "loss": 0.9996,
866
- "step": 1430
867
- },
868
- {
869
- "epoch": 1.15,
870
- "learning_rate": 8.516746411483254e-05,
871
- "loss": 0.9906,
872
- "step": 1440
873
- },
874
- {
875
- "epoch": 1.16,
876
- "learning_rate": 8.437001594896332e-05,
877
- "loss": 1.0306,
878
- "step": 1450
879
- },
880
- {
881
- "epoch": 1.16,
882
- "learning_rate": 8.35725677830941e-05,
883
- "loss": 1.0275,
884
- "step": 1460
885
- },
886
- {
887
- "epoch": 1.17,
888
- "learning_rate": 8.277511961722488e-05,
889
- "loss": 1.01,
890
- "step": 1470
891
- },
892
- {
893
- "epoch": 1.18,
894
- "learning_rate": 8.197767145135566e-05,
895
- "loss": 1.0231,
896
- "step": 1480
897
- },
898
- {
899
- "epoch": 1.19,
900
- "learning_rate": 8.118022328548645e-05,
901
- "loss": 1.0173,
902
- "step": 1490
903
- },
904
- {
905
- "epoch": 1.2,
906
- "learning_rate": 8.038277511961723e-05,
907
- "loss": 0.9819,
908
- "step": 1500
909
- },
910
- {
911
- "epoch": 1.2,
912
- "learning_rate": 7.9585326953748e-05,
913
- "loss": 1.0046,
914
- "step": 1510
915
- },
916
- {
917
- "epoch": 1.21,
918
- "learning_rate": 7.878787878787879e-05,
919
- "loss": 1.0291,
920
- "step": 1520
921
- },
922
- {
923
- "epoch": 1.22,
924
- "learning_rate": 7.799043062200957e-05,
925
- "loss": 1.0095,
926
- "step": 1530
927
- },
928
- {
929
- "epoch": 1.23,
930
- "learning_rate": 7.719298245614036e-05,
931
- "loss": 1.0321,
932
- "step": 1540
933
- },
934
- {
935
- "epoch": 1.24,
936
- "learning_rate": 7.639553429027113e-05,
937
- "loss": 0.9853,
938
- "step": 1550
939
- },
940
- {
941
- "epoch": 1.24,
942
- "learning_rate": 7.559808612440191e-05,
943
- "loss": 0.9841,
944
- "step": 1560
945
- },
946
- {
947
- "epoch": 1.25,
948
- "learning_rate": 7.480063795853271e-05,
949
- "loss": 1.0013,
950
- "step": 1570
951
- },
952
- {
953
- "epoch": 1.26,
954
- "learning_rate": 7.400318979266348e-05,
955
- "loss": 0.9834,
956
- "step": 1580
957
- },
958
- {
959
- "epoch": 1.27,
960
- "learning_rate": 7.320574162679427e-05,
961
- "loss": 1.0124,
962
- "step": 1590
963
- },
964
- {
965
- "epoch": 1.28,
966
- "learning_rate": 7.240829346092505e-05,
967
- "loss": 0.9983,
968
- "step": 1600
969
- },
970
- {
971
- "epoch": 1.28,
972
- "learning_rate": 7.161084529505583e-05,
973
- "loss": 1.0449,
974
- "step": 1610
975
- },
976
- {
977
- "epoch": 1.29,
978
- "learning_rate": 7.08133971291866e-05,
979
- "loss": 1.0002,
980
- "step": 1620
981
- },
982
- {
983
- "epoch": 1.3,
984
- "learning_rate": 7.001594896331739e-05,
985
- "loss": 0.9899,
986
- "step": 1630
987
- },
988
- {
989
- "epoch": 1.31,
990
- "learning_rate": 6.921850079744818e-05,
991
- "loss": 1.0273,
992
- "step": 1640
993
- },
994
- {
995
- "epoch": 1.32,
996
- "learning_rate": 6.842105263157895e-05,
997
- "loss": 1.0238,
998
- "step": 1650
999
- },
1000
- {
1001
- "epoch": 1.32,
1002
- "learning_rate": 6.762360446570973e-05,
1003
- "loss": 0.9946,
1004
- "step": 1660
1005
- },
1006
- {
1007
- "epoch": 1.33,
1008
- "learning_rate": 6.682615629984052e-05,
1009
- "loss": 0.9924,
1010
- "step": 1670
1011
- },
1012
- {
1013
- "epoch": 1.34,
1014
- "learning_rate": 6.60287081339713e-05,
1015
- "loss": 0.9789,
1016
- "step": 1680
1017
- },
1018
- {
1019
- "epoch": 1.35,
1020
- "learning_rate": 6.523125996810207e-05,
1021
- "loss": 1.0082,
1022
- "step": 1690
1023
- },
1024
- {
1025
- "epoch": 1.36,
1026
- "learning_rate": 6.443381180223286e-05,
1027
- "loss": 0.9767,
1028
- "step": 1700
1029
- },
1030
- {
1031
- "epoch": 1.36,
1032
- "learning_rate": 6.363636363636364e-05,
1033
- "loss": 1.0172,
1034
- "step": 1710
1035
- },
1036
- {
1037
- "epoch": 1.37,
1038
- "learning_rate": 6.283891547049443e-05,
1039
- "loss": 1.0048,
1040
- "step": 1720
1041
- },
1042
- {
1043
- "epoch": 1.38,
1044
- "learning_rate": 6.20414673046252e-05,
1045
- "loss": 1.0004,
1046
- "step": 1730
1047
- },
1048
- {
1049
- "epoch": 1.39,
1050
- "learning_rate": 6.124401913875598e-05,
1051
- "loss": 1.0146,
1052
- "step": 1740
1053
- },
1054
- {
1055
- "epoch": 1.4,
1056
- "learning_rate": 6.0446570972886766e-05,
1057
- "loss": 1.0125,
1058
- "step": 1750
1059
- },
1060
- {
1061
- "epoch": 1.4,
1062
- "learning_rate": 5.9649122807017544e-05,
1063
- "loss": 0.9935,
1064
- "step": 1760
1065
- },
1066
- {
1067
- "epoch": 1.41,
1068
- "learning_rate": 5.885167464114832e-05,
1069
- "loss": 0.972,
1070
- "step": 1770
1071
- },
1072
- {
1073
- "epoch": 1.42,
1074
- "learning_rate": 5.8054226475279106e-05,
1075
- "loss": 1.0154,
1076
- "step": 1780
1077
- },
1078
- {
1079
- "epoch": 1.43,
1080
- "learning_rate": 5.7256778309409884e-05,
1081
- "loss": 0.9627,
1082
- "step": 1790
1083
- },
1084
- {
1085
- "epoch": 1.44,
1086
- "learning_rate": 5.645933014354067e-05,
1087
- "loss": 0.9957,
1088
- "step": 1800
1089
- }
1090
- ],
1091
- "max_steps": 2508,
1092
- "num_train_epochs": 2,
1093
- "total_flos": 1.525857663320064e+17,
1094
- "trial_name": null,
1095
- "trial_params": null
1096
- }