ChiyuSONG commited on
Commit
aa2c59c
1 Parent(s): d15f4c2

Delete trainer_state.json

Browse files
Files changed (1) hide show
  1. trainer_state.json +0 -3606
trainer_state.json DELETED
@@ -1,3606 +0,0 @@
1
- {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 4.0,
5
- "global_step": 5930,
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": 9.995951417004049e-06,
13
- "loss": 4.0911,
14
- "step": 10
15
- },
16
- {
17
- "epoch": 0.01,
18
- "learning_rate": 9.991902834008098e-06,
19
- "loss": 2.8012,
20
- "step": 20
21
- },
22
- {
23
- "epoch": 0.02,
24
- "learning_rate": 9.987854251012146e-06,
25
- "loss": 2.389,
26
- "step": 30
27
- },
28
- {
29
- "epoch": 0.03,
30
- "learning_rate": 9.983805668016194e-06,
31
- "loss": 2.1658,
32
- "step": 40
33
- },
34
- {
35
- "epoch": 0.03,
36
- "learning_rate": 9.979757085020244e-06,
37
- "loss": 2.026,
38
- "step": 50
39
- },
40
- {
41
- "epoch": 0.04,
42
- "learning_rate": 9.975708502024292e-06,
43
- "loss": 1.7846,
44
- "step": 60
45
- },
46
- {
47
- "epoch": 0.05,
48
- "learning_rate": 9.971659919028341e-06,
49
- "loss": 1.703,
50
- "step": 70
51
- },
52
- {
53
- "epoch": 0.05,
54
- "learning_rate": 9.96761133603239e-06,
55
- "loss": 1.5691,
56
- "step": 80
57
- },
58
- {
59
- "epoch": 0.06,
60
- "learning_rate": 9.963562753036437e-06,
61
- "loss": 1.4297,
62
- "step": 90
63
- },
64
- {
65
- "epoch": 0.07,
66
- "learning_rate": 9.959514170040487e-06,
67
- "loss": 1.4341,
68
- "step": 100
69
- },
70
- {
71
- "epoch": 0.07,
72
- "learning_rate": 9.955465587044535e-06,
73
- "loss": 1.3304,
74
- "step": 110
75
- },
76
- {
77
- "epoch": 0.08,
78
- "learning_rate": 9.951417004048583e-06,
79
- "loss": 1.3988,
80
- "step": 120
81
- },
82
- {
83
- "epoch": 0.09,
84
- "learning_rate": 9.947368421052632e-06,
85
- "loss": 1.3272,
86
- "step": 130
87
- },
88
- {
89
- "epoch": 0.09,
90
- "learning_rate": 9.943319838056682e-06,
91
- "loss": 1.2999,
92
- "step": 140
93
- },
94
- {
95
- "epoch": 0.1,
96
- "learning_rate": 9.93927125506073e-06,
97
- "loss": 1.2997,
98
- "step": 150
99
- },
100
- {
101
- "epoch": 0.11,
102
- "learning_rate": 9.935222672064778e-06,
103
- "loss": 1.3704,
104
- "step": 160
105
- },
106
- {
107
- "epoch": 0.11,
108
- "learning_rate": 9.931174089068828e-06,
109
- "loss": 1.3313,
110
- "step": 170
111
- },
112
- {
113
- "epoch": 0.12,
114
- "learning_rate": 9.927125506072876e-06,
115
- "loss": 1.3554,
116
- "step": 180
117
- },
118
- {
119
- "epoch": 0.13,
120
- "learning_rate": 9.923076923076923e-06,
121
- "loss": 1.41,
122
- "step": 190
123
- },
124
- {
125
- "epoch": 0.13,
126
- "learning_rate": 9.919028340080973e-06,
127
- "loss": 1.3005,
128
- "step": 200
129
- },
130
- {
131
- "epoch": 0.14,
132
- "learning_rate": 9.914979757085021e-06,
133
- "loss": 1.2311,
134
- "step": 210
135
- },
136
- {
137
- "epoch": 0.15,
138
- "learning_rate": 9.910931174089069e-06,
139
- "loss": 1.3162,
140
- "step": 220
141
- },
142
- {
143
- "epoch": 0.16,
144
- "learning_rate": 9.906882591093119e-06,
145
- "loss": 1.2895,
146
- "step": 230
147
- },
148
- {
149
- "epoch": 0.16,
150
- "learning_rate": 9.902834008097167e-06,
151
- "loss": 1.2812,
152
- "step": 240
153
- },
154
- {
155
- "epoch": 0.17,
156
- "learning_rate": 9.898785425101216e-06,
157
- "loss": 1.2682,
158
- "step": 250
159
- },
160
- {
161
- "epoch": 0.18,
162
- "learning_rate": 9.894736842105264e-06,
163
- "loss": 1.2992,
164
- "step": 260
165
- },
166
- {
167
- "epoch": 0.18,
168
- "learning_rate": 9.890688259109312e-06,
169
- "loss": 1.2832,
170
- "step": 270
171
- },
172
- {
173
- "epoch": 0.19,
174
- "learning_rate": 9.886639676113362e-06,
175
- "loss": 1.2912,
176
- "step": 280
177
- },
178
- {
179
- "epoch": 0.2,
180
- "learning_rate": 9.88259109311741e-06,
181
- "loss": 1.2949,
182
- "step": 290
183
- },
184
- {
185
- "epoch": 0.2,
186
- "learning_rate": 9.878542510121458e-06,
187
- "loss": 1.2557,
188
- "step": 300
189
- },
190
- {
191
- "epoch": 0.21,
192
- "learning_rate": 9.874493927125507e-06,
193
- "loss": 1.3119,
194
- "step": 310
195
- },
196
- {
197
- "epoch": 0.22,
198
- "learning_rate": 9.870445344129555e-06,
199
- "loss": 1.3092,
200
- "step": 320
201
- },
202
- {
203
- "epoch": 0.22,
204
- "learning_rate": 9.866396761133603e-06,
205
- "loss": 1.2441,
206
- "step": 330
207
- },
208
- {
209
- "epoch": 0.23,
210
- "learning_rate": 9.862348178137653e-06,
211
- "loss": 1.3288,
212
- "step": 340
213
- },
214
- {
215
- "epoch": 0.24,
216
- "learning_rate": 9.8582995951417e-06,
217
- "loss": 1.2382,
218
- "step": 350
219
- },
220
- {
221
- "epoch": 0.24,
222
- "learning_rate": 9.85425101214575e-06,
223
- "loss": 1.2819,
224
- "step": 360
225
- },
226
- {
227
- "epoch": 0.25,
228
- "learning_rate": 9.850202429149798e-06,
229
- "loss": 1.3011,
230
- "step": 370
231
- },
232
- {
233
- "epoch": 0.26,
234
- "learning_rate": 9.846153846153846e-06,
235
- "loss": 1.2421,
236
- "step": 380
237
- },
238
- {
239
- "epoch": 0.26,
240
- "learning_rate": 9.842105263157896e-06,
241
- "loss": 1.2071,
242
- "step": 390
243
- },
244
- {
245
- "epoch": 0.27,
246
- "learning_rate": 9.838056680161945e-06,
247
- "loss": 1.2838,
248
- "step": 400
249
- },
250
- {
251
- "epoch": 0.28,
252
- "learning_rate": 9.834008097165993e-06,
253
- "loss": 1.298,
254
- "step": 410
255
- },
256
- {
257
- "epoch": 0.28,
258
- "learning_rate": 9.829959514170041e-06,
259
- "loss": 1.2867,
260
- "step": 420
261
- },
262
- {
263
- "epoch": 0.29,
264
- "learning_rate": 9.825910931174091e-06,
265
- "loss": 1.1868,
266
- "step": 430
267
- },
268
- {
269
- "epoch": 0.3,
270
- "learning_rate": 9.821862348178139e-06,
271
- "loss": 1.2263,
272
- "step": 440
273
- },
274
- {
275
- "epoch": 0.3,
276
- "learning_rate": 9.817813765182187e-06,
277
- "loss": 1.2701,
278
- "step": 450
279
- },
280
- {
281
- "epoch": 0.31,
282
- "learning_rate": 9.813765182186236e-06,
283
- "loss": 1.2542,
284
- "step": 460
285
- },
286
- {
287
- "epoch": 0.32,
288
- "learning_rate": 9.809716599190284e-06,
289
- "loss": 1.2833,
290
- "step": 470
291
- },
292
- {
293
- "epoch": 0.32,
294
- "learning_rate": 9.805668016194332e-06,
295
- "loss": 1.2109,
296
- "step": 480
297
- },
298
- {
299
- "epoch": 0.33,
300
- "learning_rate": 9.801619433198382e-06,
301
- "loss": 1.2657,
302
- "step": 490
303
- },
304
- {
305
- "epoch": 0.34,
306
- "learning_rate": 9.79757085020243e-06,
307
- "loss": 1.2587,
308
- "step": 500
309
- },
310
- {
311
- "epoch": 0.34,
312
- "learning_rate": 9.79352226720648e-06,
313
- "loss": 1.2647,
314
- "step": 510
315
- },
316
- {
317
- "epoch": 0.35,
318
- "learning_rate": 9.789473684210527e-06,
319
- "loss": 1.1917,
320
- "step": 520
321
- },
322
- {
323
- "epoch": 0.36,
324
- "learning_rate": 9.785425101214575e-06,
325
- "loss": 1.2281,
326
- "step": 530
327
- },
328
- {
329
- "epoch": 0.36,
330
- "learning_rate": 9.781376518218625e-06,
331
- "loss": 1.2677,
332
- "step": 540
333
- },
334
- {
335
- "epoch": 0.37,
336
- "learning_rate": 9.777327935222673e-06,
337
- "loss": 1.2299,
338
- "step": 550
339
- },
340
- {
341
- "epoch": 0.38,
342
- "learning_rate": 9.77327935222672e-06,
343
- "loss": 1.2524,
344
- "step": 560
345
- },
346
- {
347
- "epoch": 0.38,
348
- "learning_rate": 9.76923076923077e-06,
349
- "loss": 1.2902,
350
- "step": 570
351
- },
352
- {
353
- "epoch": 0.39,
354
- "learning_rate": 9.765182186234818e-06,
355
- "loss": 1.2603,
356
- "step": 580
357
- },
358
- {
359
- "epoch": 0.4,
360
- "learning_rate": 9.761133603238866e-06,
361
- "loss": 1.2572,
362
- "step": 590
363
- },
364
- {
365
- "epoch": 0.4,
366
- "learning_rate": 9.757085020242916e-06,
367
- "loss": 1.2223,
368
- "step": 600
369
- },
370
- {
371
- "epoch": 0.41,
372
- "learning_rate": 9.753036437246964e-06,
373
- "loss": 1.2284,
374
- "step": 610
375
- },
376
- {
377
- "epoch": 0.42,
378
- "learning_rate": 9.748987854251012e-06,
379
- "loss": 1.2331,
380
- "step": 620
381
- },
382
- {
383
- "epoch": 0.42,
384
- "learning_rate": 9.744939271255061e-06,
385
- "loss": 1.2601,
386
- "step": 630
387
- },
388
- {
389
- "epoch": 0.43,
390
- "learning_rate": 9.74089068825911e-06,
391
- "loss": 1.195,
392
- "step": 640
393
- },
394
- {
395
- "epoch": 0.44,
396
- "learning_rate": 9.736842105263159e-06,
397
- "loss": 1.2176,
398
- "step": 650
399
- },
400
- {
401
- "epoch": 0.45,
402
- "learning_rate": 9.732793522267207e-06,
403
- "loss": 1.2684,
404
- "step": 660
405
- },
406
- {
407
- "epoch": 0.45,
408
- "learning_rate": 9.728744939271257e-06,
409
- "loss": 1.2575,
410
- "step": 670
411
- },
412
- {
413
- "epoch": 0.46,
414
- "learning_rate": 9.724696356275305e-06,
415
- "loss": 1.2187,
416
- "step": 680
417
- },
418
- {
419
- "epoch": 0.47,
420
- "learning_rate": 9.720647773279354e-06,
421
- "loss": 1.2184,
422
- "step": 690
423
- },
424
- {
425
- "epoch": 0.47,
426
- "learning_rate": 9.716599190283402e-06,
427
- "loss": 1.282,
428
- "step": 700
429
- },
430
- {
431
- "epoch": 0.48,
432
- "learning_rate": 9.71255060728745e-06,
433
- "loss": 1.2464,
434
- "step": 710
435
- },
436
- {
437
- "epoch": 0.49,
438
- "learning_rate": 9.7085020242915e-06,
439
- "loss": 1.2041,
440
- "step": 720
441
- },
442
- {
443
- "epoch": 0.49,
444
- "learning_rate": 9.704453441295548e-06,
445
- "loss": 1.26,
446
- "step": 730
447
- },
448
- {
449
- "epoch": 0.5,
450
- "learning_rate": 9.700404858299596e-06,
451
- "loss": 1.2401,
452
- "step": 740
453
- },
454
- {
455
- "epoch": 0.51,
456
- "learning_rate": 9.696356275303645e-06,
457
- "loss": 1.2289,
458
- "step": 750
459
- },
460
- {
461
- "epoch": 0.51,
462
- "learning_rate": 9.692307692307693e-06,
463
- "loss": 1.2587,
464
- "step": 760
465
- },
466
- {
467
- "epoch": 0.52,
468
- "learning_rate": 9.688259109311741e-06,
469
- "loss": 1.2078,
470
- "step": 770
471
- },
472
- {
473
- "epoch": 0.53,
474
- "learning_rate": 9.68421052631579e-06,
475
- "loss": 1.192,
476
- "step": 780
477
- },
478
- {
479
- "epoch": 0.53,
480
- "learning_rate": 9.680161943319839e-06,
481
- "loss": 1.2981,
482
- "step": 790
483
- },
484
- {
485
- "epoch": 0.54,
486
- "learning_rate": 9.676113360323888e-06,
487
- "loss": 1.1942,
488
- "step": 800
489
- },
490
- {
491
- "epoch": 0.55,
492
- "learning_rate": 9.672064777327936e-06,
493
- "loss": 1.247,
494
- "step": 810
495
- },
496
- {
497
- "epoch": 0.55,
498
- "learning_rate": 9.668016194331984e-06,
499
- "loss": 1.1902,
500
- "step": 820
501
- },
502
- {
503
- "epoch": 0.56,
504
- "learning_rate": 9.663967611336034e-06,
505
- "loss": 1.2271,
506
- "step": 830
507
- },
508
- {
509
- "epoch": 0.57,
510
- "learning_rate": 9.659919028340082e-06,
511
- "loss": 1.2407,
512
- "step": 840
513
- },
514
- {
515
- "epoch": 0.57,
516
- "learning_rate": 9.65587044534413e-06,
517
- "loss": 1.2067,
518
- "step": 850
519
- },
520
- {
521
- "epoch": 0.58,
522
- "learning_rate": 9.65182186234818e-06,
523
- "loss": 1.2284,
524
- "step": 860
525
- },
526
- {
527
- "epoch": 0.59,
528
- "learning_rate": 9.647773279352227e-06,
529
- "loss": 1.2676,
530
- "step": 870
531
- },
532
- {
533
- "epoch": 0.59,
534
- "learning_rate": 9.643724696356275e-06,
535
- "loss": 1.2542,
536
- "step": 880
537
- },
538
- {
539
- "epoch": 0.6,
540
- "learning_rate": 9.639676113360325e-06,
541
- "loss": 1.2364,
542
- "step": 890
543
- },
544
- {
545
- "epoch": 0.61,
546
- "learning_rate": 9.635627530364373e-06,
547
- "loss": 1.3608,
548
- "step": 900
549
- },
550
- {
551
- "epoch": 0.61,
552
- "learning_rate": 9.63157894736842e-06,
553
- "loss": 1.2739,
554
- "step": 910
555
- },
556
- {
557
- "epoch": 0.62,
558
- "learning_rate": 9.62753036437247e-06,
559
- "loss": 1.1984,
560
- "step": 920
561
- },
562
- {
563
- "epoch": 0.63,
564
- "learning_rate": 9.62348178137652e-06,
565
- "loss": 1.1875,
566
- "step": 930
567
- },
568
- {
569
- "epoch": 0.63,
570
- "learning_rate": 9.619433198380568e-06,
571
- "loss": 1.2447,
572
- "step": 940
573
- },
574
- {
575
- "epoch": 0.64,
576
- "learning_rate": 9.615384615384616e-06,
577
- "loss": 1.26,
578
- "step": 950
579
- },
580
- {
581
- "epoch": 0.65,
582
- "learning_rate": 9.611336032388665e-06,
583
- "loss": 1.2059,
584
- "step": 960
585
- },
586
- {
587
- "epoch": 0.65,
588
- "learning_rate": 9.607287449392713e-06,
589
- "loss": 1.233,
590
- "step": 970
591
- },
592
- {
593
- "epoch": 0.66,
594
- "learning_rate": 9.603238866396763e-06,
595
- "loss": 1.244,
596
- "step": 980
597
- },
598
- {
599
- "epoch": 0.67,
600
- "learning_rate": 9.59919028340081e-06,
601
- "loss": 1.2613,
602
- "step": 990
603
- },
604
- {
605
- "epoch": 0.67,
606
- "learning_rate": 9.595141700404859e-06,
607
- "loss": 1.2274,
608
- "step": 1000
609
- },
610
- {
611
- "epoch": 0.68,
612
- "learning_rate": 9.591093117408908e-06,
613
- "loss": 1.2727,
614
- "step": 1010
615
- },
616
- {
617
- "epoch": 0.69,
618
- "learning_rate": 9.587044534412956e-06,
619
- "loss": 1.2834,
620
- "step": 1020
621
- },
622
- {
623
- "epoch": 0.69,
624
- "learning_rate": 9.582995951417004e-06,
625
- "loss": 1.259,
626
- "step": 1030
627
- },
628
- {
629
- "epoch": 0.7,
630
- "learning_rate": 9.578947368421054e-06,
631
- "loss": 1.2742,
632
- "step": 1040
633
- },
634
- {
635
- "epoch": 0.71,
636
- "learning_rate": 9.574898785425102e-06,
637
- "loss": 1.2658,
638
- "step": 1050
639
- },
640
- {
641
- "epoch": 0.72,
642
- "learning_rate": 9.57085020242915e-06,
643
- "loss": 1.2203,
644
- "step": 1060
645
- },
646
- {
647
- "epoch": 0.72,
648
- "learning_rate": 9.5668016194332e-06,
649
- "loss": 1.1757,
650
- "step": 1070
651
- },
652
- {
653
- "epoch": 0.73,
654
- "learning_rate": 9.562753036437247e-06,
655
- "loss": 1.2296,
656
- "step": 1080
657
- },
658
- {
659
- "epoch": 0.74,
660
- "learning_rate": 9.558704453441297e-06,
661
- "loss": 1.2714,
662
- "step": 1090
663
- },
664
- {
665
- "epoch": 0.74,
666
- "learning_rate": 9.554655870445345e-06,
667
- "loss": 1.2265,
668
- "step": 1100
669
- },
670
- {
671
- "epoch": 0.75,
672
- "learning_rate": 9.550607287449393e-06,
673
- "loss": 1.1984,
674
- "step": 1110
675
- },
676
- {
677
- "epoch": 0.76,
678
- "learning_rate": 9.546558704453442e-06,
679
- "loss": 1.2706,
680
- "step": 1120
681
- },
682
- {
683
- "epoch": 0.76,
684
- "learning_rate": 9.54251012145749e-06,
685
- "loss": 1.2611,
686
- "step": 1130
687
- },
688
- {
689
- "epoch": 0.77,
690
- "learning_rate": 9.538461538461538e-06,
691
- "loss": 1.2584,
692
- "step": 1140
693
- },
694
- {
695
- "epoch": 0.78,
696
- "learning_rate": 9.534412955465588e-06,
697
- "loss": 1.2849,
698
- "step": 1150
699
- },
700
- {
701
- "epoch": 0.78,
702
- "learning_rate": 9.530364372469636e-06,
703
- "loss": 1.2254,
704
- "step": 1160
705
- },
706
- {
707
- "epoch": 0.79,
708
- "learning_rate": 9.526315789473686e-06,
709
- "loss": 1.2698,
710
- "step": 1170
711
- },
712
- {
713
- "epoch": 0.8,
714
- "learning_rate": 9.522267206477733e-06,
715
- "loss": 1.2037,
716
- "step": 1180
717
- },
718
- {
719
- "epoch": 0.8,
720
- "learning_rate": 9.518218623481783e-06,
721
- "loss": 1.2408,
722
- "step": 1190
723
- },
724
- {
725
- "epoch": 0.81,
726
- "learning_rate": 9.514170040485831e-06,
727
- "loss": 1.2116,
728
- "step": 1200
729
- },
730
- {
731
- "epoch": 0.82,
732
- "learning_rate": 9.510121457489879e-06,
733
- "loss": 1.2393,
734
- "step": 1210
735
- },
736
- {
737
- "epoch": 0.82,
738
- "learning_rate": 9.506072874493929e-06,
739
- "loss": 1.2177,
740
- "step": 1220
741
- },
742
- {
743
- "epoch": 0.83,
744
- "learning_rate": 9.502024291497977e-06,
745
- "loss": 1.23,
746
- "step": 1230
747
- },
748
- {
749
- "epoch": 0.84,
750
- "learning_rate": 9.497975708502026e-06,
751
- "loss": 1.2245,
752
- "step": 1240
753
- },
754
- {
755
- "epoch": 0.84,
756
- "learning_rate": 9.493927125506074e-06,
757
- "loss": 1.2602,
758
- "step": 1250
759
- },
760
- {
761
- "epoch": 0.85,
762
- "learning_rate": 9.489878542510122e-06,
763
- "loss": 1.2017,
764
- "step": 1260
765
- },
766
- {
767
- "epoch": 0.86,
768
- "learning_rate": 9.485829959514172e-06,
769
- "loss": 1.2281,
770
- "step": 1270
771
- },
772
- {
773
- "epoch": 0.86,
774
- "learning_rate": 9.48178137651822e-06,
775
- "loss": 1.2808,
776
- "step": 1280
777
- },
778
- {
779
- "epoch": 0.87,
780
- "learning_rate": 9.477732793522268e-06,
781
- "loss": 1.2017,
782
- "step": 1290
783
- },
784
- {
785
- "epoch": 0.88,
786
- "learning_rate": 9.473684210526317e-06,
787
- "loss": 1.2557,
788
- "step": 1300
789
- },
790
- {
791
- "epoch": 0.88,
792
- "learning_rate": 9.469635627530365e-06,
793
- "loss": 1.2952,
794
- "step": 1310
795
- },
796
- {
797
- "epoch": 0.89,
798
- "learning_rate": 9.465587044534413e-06,
799
- "loss": 1.2477,
800
- "step": 1320
801
- },
802
- {
803
- "epoch": 0.9,
804
- "learning_rate": 9.461538461538463e-06,
805
- "loss": 1.2397,
806
- "step": 1330
807
- },
808
- {
809
- "epoch": 0.9,
810
- "learning_rate": 9.45748987854251e-06,
811
- "loss": 1.1556,
812
- "step": 1340
813
- },
814
- {
815
- "epoch": 0.91,
816
- "learning_rate": 9.453441295546559e-06,
817
- "loss": 1.217,
818
- "step": 1350
819
- },
820
- {
821
- "epoch": 0.92,
822
- "learning_rate": 9.449392712550608e-06,
823
- "loss": 1.2725,
824
- "step": 1360
825
- },
826
- {
827
- "epoch": 0.92,
828
- "learning_rate": 9.445344129554656e-06,
829
- "loss": 1.2079,
830
- "step": 1370
831
- },
832
- {
833
- "epoch": 0.93,
834
- "learning_rate": 9.441295546558706e-06,
835
- "loss": 1.2646,
836
- "step": 1380
837
- },
838
- {
839
- "epoch": 0.94,
840
- "learning_rate": 9.437246963562754e-06,
841
- "loss": 1.2034,
842
- "step": 1390
843
- },
844
- {
845
- "epoch": 0.94,
846
- "learning_rate": 9.433198380566802e-06,
847
- "loss": 1.2442,
848
- "step": 1400
849
- },
850
- {
851
- "epoch": 0.95,
852
- "learning_rate": 9.429149797570851e-06,
853
- "loss": 1.2463,
854
- "step": 1410
855
- },
856
- {
857
- "epoch": 0.96,
858
- "learning_rate": 9.4251012145749e-06,
859
- "loss": 1.2257,
860
- "step": 1420
861
- },
862
- {
863
- "epoch": 0.96,
864
- "learning_rate": 9.421052631578947e-06,
865
- "loss": 1.2969,
866
- "step": 1430
867
- },
868
- {
869
- "epoch": 0.97,
870
- "learning_rate": 9.417004048582997e-06,
871
- "loss": 1.2288,
872
- "step": 1440
873
- },
874
- {
875
- "epoch": 0.98,
876
- "learning_rate": 9.412955465587045e-06,
877
- "loss": 1.2307,
878
- "step": 1450
879
- },
880
- {
881
- "epoch": 0.98,
882
- "learning_rate": 9.408906882591094e-06,
883
- "loss": 1.2101,
884
- "step": 1460
885
- },
886
- {
887
- "epoch": 0.99,
888
- "learning_rate": 9.404858299595142e-06,
889
- "loss": 1.2009,
890
- "step": 1470
891
- },
892
- {
893
- "epoch": 1.0,
894
- "learning_rate": 9.400809716599192e-06,
895
- "loss": 1.2698,
896
- "step": 1480
897
- },
898
- {
899
- "epoch": 1.0,
900
- "eval_loss": 1.1532880067825317,
901
- "eval_runtime": 11.4194,
902
- "eval_samples_per_second": 96.327,
903
- "eval_steps_per_second": 12.085,
904
- "step": 1482
905
- },
906
- {
907
- "epoch": 1.01,
908
- "learning_rate": 9.39676113360324e-06,
909
- "loss": 1.1808,
910
- "step": 1490
911
- },
912
- {
913
- "epoch": 1.01,
914
- "learning_rate": 9.392712550607288e-06,
915
- "loss": 1.1629,
916
- "step": 1500
917
- },
918
- {
919
- "epoch": 1.02,
920
- "learning_rate": 9.388663967611337e-06,
921
- "loss": 1.1656,
922
- "step": 1510
923
- },
924
- {
925
- "epoch": 1.03,
926
- "learning_rate": 9.384615384615385e-06,
927
- "loss": 1.1163,
928
- "step": 1520
929
- },
930
- {
931
- "epoch": 1.03,
932
- "learning_rate": 9.380566801619435e-06,
933
- "loss": 1.1403,
934
- "step": 1530
935
- },
936
- {
937
- "epoch": 1.04,
938
- "learning_rate": 9.376518218623483e-06,
939
- "loss": 1.122,
940
- "step": 1540
941
- },
942
- {
943
- "epoch": 1.05,
944
- "learning_rate": 9.37246963562753e-06,
945
- "loss": 1.1426,
946
- "step": 1550
947
- },
948
- {
949
- "epoch": 1.05,
950
- "learning_rate": 9.36842105263158e-06,
951
- "loss": 1.1142,
952
- "step": 1560
953
- },
954
- {
955
- "epoch": 1.06,
956
- "learning_rate": 9.364372469635628e-06,
957
- "loss": 1.1143,
958
- "step": 1570
959
- },
960
- {
961
- "epoch": 1.07,
962
- "learning_rate": 9.360323886639676e-06,
963
- "loss": 1.1798,
964
- "step": 1580
965
- },
966
- {
967
- "epoch": 1.07,
968
- "learning_rate": 9.356275303643726e-06,
969
- "loss": 1.1575,
970
- "step": 1590
971
- },
972
- {
973
- "epoch": 1.08,
974
- "learning_rate": 9.352226720647774e-06,
975
- "loss": 1.1459,
976
- "step": 1600
977
- },
978
- {
979
- "epoch": 1.09,
980
- "learning_rate": 9.348178137651822e-06,
981
- "loss": 1.199,
982
- "step": 1610
983
- },
984
- {
985
- "epoch": 1.09,
986
- "learning_rate": 9.344129554655871e-06,
987
- "loss": 1.1568,
988
- "step": 1620
989
- },
990
- {
991
- "epoch": 1.1,
992
- "learning_rate": 9.34008097165992e-06,
993
- "loss": 1.1577,
994
- "step": 1630
995
- },
996
- {
997
- "epoch": 1.11,
998
- "learning_rate": 9.336032388663967e-06,
999
- "loss": 1.0549,
1000
- "step": 1640
1001
- },
1002
- {
1003
- "epoch": 1.11,
1004
- "learning_rate": 9.331983805668017e-06,
1005
- "loss": 1.1889,
1006
- "step": 1650
1007
- },
1008
- {
1009
- "epoch": 1.12,
1010
- "learning_rate": 9.327935222672065e-06,
1011
- "loss": 1.1355,
1012
- "step": 1660
1013
- },
1014
- {
1015
- "epoch": 1.13,
1016
- "learning_rate": 9.323886639676115e-06,
1017
- "loss": 1.1725,
1018
- "step": 1670
1019
- },
1020
- {
1021
- "epoch": 1.13,
1022
- "learning_rate": 9.319838056680162e-06,
1023
- "loss": 1.1096,
1024
- "step": 1680
1025
- },
1026
- {
1027
- "epoch": 1.14,
1028
- "learning_rate": 9.315789473684212e-06,
1029
- "loss": 1.2392,
1030
- "step": 1690
1031
- },
1032
- {
1033
- "epoch": 1.15,
1034
- "learning_rate": 9.31174089068826e-06,
1035
- "loss": 1.1508,
1036
- "step": 1700
1037
- },
1038
- {
1039
- "epoch": 1.15,
1040
- "learning_rate": 9.30769230769231e-06,
1041
- "loss": 1.1437,
1042
- "step": 1710
1043
- },
1044
- {
1045
- "epoch": 1.16,
1046
- "learning_rate": 9.303643724696358e-06,
1047
- "loss": 1.204,
1048
- "step": 1720
1049
- },
1050
- {
1051
- "epoch": 1.17,
1052
- "learning_rate": 9.299595141700406e-06,
1053
- "loss": 1.1795,
1054
- "step": 1730
1055
- },
1056
- {
1057
- "epoch": 1.17,
1058
- "learning_rate": 9.295546558704455e-06,
1059
- "loss": 1.112,
1060
- "step": 1740
1061
- },
1062
- {
1063
- "epoch": 1.18,
1064
- "learning_rate": 9.291497975708503e-06,
1065
- "loss": 1.2072,
1066
- "step": 1750
1067
- },
1068
- {
1069
- "epoch": 1.19,
1070
- "learning_rate": 9.287449392712551e-06,
1071
- "loss": 1.1518,
1072
- "step": 1760
1073
- },
1074
- {
1075
- "epoch": 1.19,
1076
- "learning_rate": 9.2834008097166e-06,
1077
- "loss": 1.1479,
1078
- "step": 1770
1079
- },
1080
- {
1081
- "epoch": 1.2,
1082
- "learning_rate": 9.279352226720649e-06,
1083
- "loss": 1.1183,
1084
- "step": 1780
1085
- },
1086
- {
1087
- "epoch": 1.21,
1088
- "learning_rate": 9.275303643724697e-06,
1089
- "loss": 1.1493,
1090
- "step": 1790
1091
- },
1092
- {
1093
- "epoch": 1.21,
1094
- "learning_rate": 9.271255060728746e-06,
1095
- "loss": 1.1603,
1096
- "step": 1800
1097
- },
1098
- {
1099
- "epoch": 1.22,
1100
- "learning_rate": 9.267206477732794e-06,
1101
- "loss": 1.186,
1102
- "step": 1810
1103
- },
1104
- {
1105
- "epoch": 1.23,
1106
- "learning_rate": 9.263157894736844e-06,
1107
- "loss": 1.1857,
1108
- "step": 1820
1109
- },
1110
- {
1111
- "epoch": 1.23,
1112
- "learning_rate": 9.259109311740892e-06,
1113
- "loss": 1.1957,
1114
- "step": 1830
1115
- },
1116
- {
1117
- "epoch": 1.24,
1118
- "learning_rate": 9.25506072874494e-06,
1119
- "loss": 1.1102,
1120
- "step": 1840
1121
- },
1122
- {
1123
- "epoch": 1.25,
1124
- "learning_rate": 9.25101214574899e-06,
1125
- "loss": 1.1439,
1126
- "step": 1850
1127
- },
1128
- {
1129
- "epoch": 1.25,
1130
- "learning_rate": 9.246963562753037e-06,
1131
- "loss": 1.1252,
1132
- "step": 1860
1133
- },
1134
- {
1135
- "epoch": 1.26,
1136
- "learning_rate": 9.242914979757085e-06,
1137
- "loss": 1.1238,
1138
- "step": 1870
1139
- },
1140
- {
1141
- "epoch": 1.27,
1142
- "learning_rate": 9.238866396761135e-06,
1143
- "loss": 1.135,
1144
- "step": 1880
1145
- },
1146
- {
1147
- "epoch": 1.27,
1148
- "learning_rate": 9.234817813765183e-06,
1149
- "loss": 1.1542,
1150
- "step": 1890
1151
- },
1152
- {
1153
- "epoch": 1.28,
1154
- "learning_rate": 9.23076923076923e-06,
1155
- "loss": 1.1481,
1156
- "step": 1900
1157
- },
1158
- {
1159
- "epoch": 1.29,
1160
- "learning_rate": 9.22672064777328e-06,
1161
- "loss": 1.0846,
1162
- "step": 1910
1163
- },
1164
- {
1165
- "epoch": 1.3,
1166
- "learning_rate": 9.222672064777328e-06,
1167
- "loss": 1.1346,
1168
- "step": 1920
1169
- },
1170
- {
1171
- "epoch": 1.3,
1172
- "learning_rate": 9.218623481781376e-06,
1173
- "loss": 1.0831,
1174
- "step": 1930
1175
- },
1176
- {
1177
- "epoch": 1.31,
1178
- "learning_rate": 9.214574898785426e-06,
1179
- "loss": 1.2031,
1180
- "step": 1940
1181
- },
1182
- {
1183
- "epoch": 1.32,
1184
- "learning_rate": 9.210526315789474e-06,
1185
- "loss": 1.0651,
1186
- "step": 1950
1187
- },
1188
- {
1189
- "epoch": 1.32,
1190
- "learning_rate": 9.206477732793523e-06,
1191
- "loss": 1.1237,
1192
- "step": 1960
1193
- },
1194
- {
1195
- "epoch": 1.33,
1196
- "learning_rate": 9.202429149797571e-06,
1197
- "loss": 1.0612,
1198
- "step": 1970
1199
- },
1200
- {
1201
- "epoch": 1.34,
1202
- "learning_rate": 9.198380566801621e-06,
1203
- "loss": 1.1377,
1204
- "step": 1980
1205
- },
1206
- {
1207
- "epoch": 1.34,
1208
- "learning_rate": 9.194331983805669e-06,
1209
- "loss": 1.1692,
1210
- "step": 1990
1211
- },
1212
- {
1213
- "epoch": 1.35,
1214
- "learning_rate": 9.190283400809718e-06,
1215
- "loss": 1.1136,
1216
- "step": 2000
1217
- },
1218
- {
1219
- "epoch": 1.36,
1220
- "learning_rate": 9.186234817813766e-06,
1221
- "loss": 1.1361,
1222
- "step": 2010
1223
- },
1224
- {
1225
- "epoch": 1.36,
1226
- "learning_rate": 9.182186234817814e-06,
1227
- "loss": 1.1663,
1228
- "step": 2020
1229
- },
1230
- {
1231
- "epoch": 1.37,
1232
- "learning_rate": 9.178137651821864e-06,
1233
- "loss": 1.1054,
1234
- "step": 2030
1235
- },
1236
- {
1237
- "epoch": 1.38,
1238
- "learning_rate": 9.174089068825912e-06,
1239
- "loss": 1.1427,
1240
- "step": 2040
1241
- },
1242
- {
1243
- "epoch": 1.38,
1244
- "learning_rate": 9.17004048582996e-06,
1245
- "loss": 1.1392,
1246
- "step": 2050
1247
- },
1248
- {
1249
- "epoch": 1.39,
1250
- "learning_rate": 9.16599190283401e-06,
1251
- "loss": 1.191,
1252
- "step": 2060
1253
- },
1254
- {
1255
- "epoch": 1.4,
1256
- "learning_rate": 9.161943319838057e-06,
1257
- "loss": 1.1006,
1258
- "step": 2070
1259
- },
1260
- {
1261
- "epoch": 1.4,
1262
- "learning_rate": 9.157894736842105e-06,
1263
- "loss": 1.1372,
1264
- "step": 2080
1265
- },
1266
- {
1267
- "epoch": 1.41,
1268
- "learning_rate": 9.153846153846155e-06,
1269
- "loss": 1.1428,
1270
- "step": 2090
1271
- },
1272
- {
1273
- "epoch": 1.42,
1274
- "learning_rate": 9.149797570850203e-06,
1275
- "loss": 1.1433,
1276
- "step": 2100
1277
- },
1278
- {
1279
- "epoch": 1.42,
1280
- "learning_rate": 9.145748987854253e-06,
1281
- "loss": 1.0997,
1282
- "step": 2110
1283
- },
1284
- {
1285
- "epoch": 1.43,
1286
- "learning_rate": 9.1417004048583e-06,
1287
- "loss": 1.1478,
1288
- "step": 2120
1289
- },
1290
- {
1291
- "epoch": 1.44,
1292
- "learning_rate": 9.137651821862348e-06,
1293
- "loss": 1.0928,
1294
- "step": 2130
1295
- },
1296
- {
1297
- "epoch": 1.44,
1298
- "learning_rate": 9.133603238866398e-06,
1299
- "loss": 1.0993,
1300
- "step": 2140
1301
- },
1302
- {
1303
- "epoch": 1.45,
1304
- "learning_rate": 9.129554655870446e-06,
1305
- "loss": 1.1306,
1306
- "step": 2150
1307
- },
1308
- {
1309
- "epoch": 1.46,
1310
- "learning_rate": 9.125506072874494e-06,
1311
- "loss": 1.1527,
1312
- "step": 2160
1313
- },
1314
- {
1315
- "epoch": 1.46,
1316
- "learning_rate": 9.121457489878544e-06,
1317
- "loss": 1.1394,
1318
- "step": 2170
1319
- },
1320
- {
1321
- "epoch": 1.47,
1322
- "learning_rate": 9.117408906882591e-06,
1323
- "loss": 1.1787,
1324
- "step": 2180
1325
- },
1326
- {
1327
- "epoch": 1.48,
1328
- "learning_rate": 9.11336032388664e-06,
1329
- "loss": 1.1295,
1330
- "step": 2190
1331
- },
1332
- {
1333
- "epoch": 1.48,
1334
- "learning_rate": 9.109311740890689e-06,
1335
- "loss": 1.1843,
1336
- "step": 2200
1337
- },
1338
- {
1339
- "epoch": 1.49,
1340
- "learning_rate": 9.105263157894739e-06,
1341
- "loss": 1.1243,
1342
- "step": 2210
1343
- },
1344
- {
1345
- "epoch": 1.5,
1346
- "learning_rate": 9.101214574898785e-06,
1347
- "loss": 1.1748,
1348
- "step": 2220
1349
- },
1350
- {
1351
- "epoch": 1.5,
1352
- "learning_rate": 9.097165991902835e-06,
1353
- "loss": 1.1282,
1354
- "step": 2230
1355
- },
1356
- {
1357
- "epoch": 1.51,
1358
- "learning_rate": 9.093117408906884e-06,
1359
- "loss": 1.0969,
1360
- "step": 2240
1361
- },
1362
- {
1363
- "epoch": 1.52,
1364
- "learning_rate": 9.089068825910932e-06,
1365
- "loss": 1.158,
1366
- "step": 2250
1367
- },
1368
- {
1369
- "epoch": 1.52,
1370
- "learning_rate": 9.085020242914982e-06,
1371
- "loss": 1.1417,
1372
- "step": 2260
1373
- },
1374
- {
1375
- "epoch": 1.53,
1376
- "learning_rate": 9.08097165991903e-06,
1377
- "loss": 1.1397,
1378
- "step": 2270
1379
- },
1380
- {
1381
- "epoch": 1.54,
1382
- "learning_rate": 9.076923076923078e-06,
1383
- "loss": 1.1622,
1384
- "step": 2280
1385
- },
1386
- {
1387
- "epoch": 1.54,
1388
- "learning_rate": 9.072874493927127e-06,
1389
- "loss": 1.1217,
1390
- "step": 2290
1391
- },
1392
- {
1393
- "epoch": 1.55,
1394
- "learning_rate": 9.068825910931175e-06,
1395
- "loss": 1.1527,
1396
- "step": 2300
1397
- },
1398
- {
1399
- "epoch": 1.56,
1400
- "learning_rate": 9.064777327935223e-06,
1401
- "loss": 1.1927,
1402
- "step": 2310
1403
- },
1404
- {
1405
- "epoch": 1.56,
1406
- "learning_rate": 9.060728744939273e-06,
1407
- "loss": 1.1505,
1408
- "step": 2320
1409
- },
1410
- {
1411
- "epoch": 1.57,
1412
- "learning_rate": 9.05668016194332e-06,
1413
- "loss": 1.123,
1414
- "step": 2330
1415
- },
1416
- {
1417
- "epoch": 1.58,
1418
- "learning_rate": 9.052631578947369e-06,
1419
- "loss": 1.1453,
1420
- "step": 2340
1421
- },
1422
- {
1423
- "epoch": 1.59,
1424
- "learning_rate": 9.048582995951418e-06,
1425
- "loss": 1.1903,
1426
- "step": 2350
1427
- },
1428
- {
1429
- "epoch": 1.59,
1430
- "learning_rate": 9.044534412955466e-06,
1431
- "loss": 1.144,
1432
- "step": 2360
1433
- },
1434
- {
1435
- "epoch": 1.6,
1436
- "learning_rate": 9.040485829959514e-06,
1437
- "loss": 1.1469,
1438
- "step": 2370
1439
- },
1440
- {
1441
- "epoch": 1.61,
1442
- "learning_rate": 9.036437246963564e-06,
1443
- "loss": 1.1463,
1444
- "step": 2380
1445
- },
1446
- {
1447
- "epoch": 1.61,
1448
- "learning_rate": 9.032388663967612e-06,
1449
- "loss": 1.1702,
1450
- "step": 2390
1451
- },
1452
- {
1453
- "epoch": 1.62,
1454
- "learning_rate": 9.028340080971661e-06,
1455
- "loss": 1.078,
1456
- "step": 2400
1457
- },
1458
- {
1459
- "epoch": 1.63,
1460
- "learning_rate": 9.02429149797571e-06,
1461
- "loss": 1.0907,
1462
- "step": 2410
1463
- },
1464
- {
1465
- "epoch": 1.63,
1466
- "learning_rate": 9.020242914979757e-06,
1467
- "loss": 1.1621,
1468
- "step": 2420
1469
- },
1470
- {
1471
- "epoch": 1.64,
1472
- "learning_rate": 9.016194331983807e-06,
1473
- "loss": 1.0947,
1474
- "step": 2430
1475
- },
1476
- {
1477
- "epoch": 1.65,
1478
- "learning_rate": 9.012145748987855e-06,
1479
- "loss": 1.1301,
1480
- "step": 2440
1481
- },
1482
- {
1483
- "epoch": 1.65,
1484
- "learning_rate": 9.008097165991903e-06,
1485
- "loss": 1.1232,
1486
- "step": 2450
1487
- },
1488
- {
1489
- "epoch": 1.66,
1490
- "learning_rate": 9.004048582995952e-06,
1491
- "loss": 1.1737,
1492
- "step": 2460
1493
- },
1494
- {
1495
- "epoch": 1.67,
1496
- "learning_rate": 9e-06,
1497
- "loss": 1.1475,
1498
- "step": 2470
1499
- },
1500
- {
1501
- "epoch": 1.67,
1502
- "learning_rate": 8.99595141700405e-06,
1503
- "loss": 1.1242,
1504
- "step": 2480
1505
- },
1506
- {
1507
- "epoch": 1.68,
1508
- "learning_rate": 8.991902834008098e-06,
1509
- "loss": 1.1321,
1510
- "step": 2490
1511
- },
1512
- {
1513
- "epoch": 1.69,
1514
- "learning_rate": 8.987854251012147e-06,
1515
- "loss": 1.1131,
1516
- "step": 2500
1517
- },
1518
- {
1519
- "epoch": 1.69,
1520
- "learning_rate": 8.983805668016195e-06,
1521
- "loss": 1.1456,
1522
- "step": 2510
1523
- },
1524
- {
1525
- "epoch": 1.7,
1526
- "learning_rate": 8.979757085020243e-06,
1527
- "loss": 1.0983,
1528
- "step": 2520
1529
- },
1530
- {
1531
- "epoch": 1.71,
1532
- "learning_rate": 8.975708502024293e-06,
1533
- "loss": 1.1598,
1534
- "step": 2530
1535
- },
1536
- {
1537
- "epoch": 1.71,
1538
- "learning_rate": 8.971659919028341e-06,
1539
- "loss": 1.1861,
1540
- "step": 2540
1541
- },
1542
- {
1543
- "epoch": 1.72,
1544
- "learning_rate": 8.96761133603239e-06,
1545
- "loss": 1.1334,
1546
- "step": 2550
1547
- },
1548
- {
1549
- "epoch": 1.73,
1550
- "learning_rate": 8.963562753036438e-06,
1551
- "loss": 1.1605,
1552
- "step": 2560
1553
- },
1554
- {
1555
- "epoch": 1.73,
1556
- "learning_rate": 8.959514170040486e-06,
1557
- "loss": 1.136,
1558
- "step": 2570
1559
- },
1560
- {
1561
- "epoch": 1.74,
1562
- "learning_rate": 8.955465587044536e-06,
1563
- "loss": 1.1162,
1564
- "step": 2580
1565
- },
1566
- {
1567
- "epoch": 1.75,
1568
- "learning_rate": 8.951417004048584e-06,
1569
- "loss": 1.1115,
1570
- "step": 2590
1571
- },
1572
- {
1573
- "epoch": 1.75,
1574
- "learning_rate": 8.947368421052632e-06,
1575
- "loss": 1.1025,
1576
- "step": 2600
1577
- },
1578
- {
1579
- "epoch": 1.76,
1580
- "learning_rate": 8.943319838056681e-06,
1581
- "loss": 1.1616,
1582
- "step": 2610
1583
- },
1584
- {
1585
- "epoch": 1.77,
1586
- "learning_rate": 8.93927125506073e-06,
1587
- "loss": 1.1407,
1588
- "step": 2620
1589
- },
1590
- {
1591
- "epoch": 1.77,
1592
- "learning_rate": 8.935222672064777e-06,
1593
- "loss": 1.1016,
1594
- "step": 2630
1595
- },
1596
- {
1597
- "epoch": 1.78,
1598
- "learning_rate": 8.931174089068827e-06,
1599
- "loss": 1.111,
1600
- "step": 2640
1601
- },
1602
- {
1603
- "epoch": 1.79,
1604
- "learning_rate": 8.927125506072875e-06,
1605
- "loss": 1.1226,
1606
- "step": 2650
1607
- },
1608
- {
1609
- "epoch": 1.79,
1610
- "learning_rate": 8.923076923076923e-06,
1611
- "loss": 1.1105,
1612
- "step": 2660
1613
- },
1614
- {
1615
- "epoch": 1.8,
1616
- "learning_rate": 8.919028340080972e-06,
1617
- "loss": 1.1227,
1618
- "step": 2670
1619
- },
1620
- {
1621
- "epoch": 1.81,
1622
- "learning_rate": 8.91497975708502e-06,
1623
- "loss": 1.1462,
1624
- "step": 2680
1625
- },
1626
- {
1627
- "epoch": 1.81,
1628
- "learning_rate": 8.91093117408907e-06,
1629
- "loss": 1.1119,
1630
- "step": 2690
1631
- },
1632
- {
1633
- "epoch": 1.82,
1634
- "learning_rate": 8.906882591093118e-06,
1635
- "loss": 1.1567,
1636
- "step": 2700
1637
- },
1638
- {
1639
- "epoch": 1.83,
1640
- "learning_rate": 8.902834008097166e-06,
1641
- "loss": 1.1565,
1642
- "step": 2710
1643
- },
1644
- {
1645
- "epoch": 1.83,
1646
- "learning_rate": 8.898785425101216e-06,
1647
- "loss": 1.1116,
1648
- "step": 2720
1649
- },
1650
- {
1651
- "epoch": 1.84,
1652
- "learning_rate": 8.894736842105265e-06,
1653
- "loss": 1.1139,
1654
- "step": 2730
1655
- },
1656
- {
1657
- "epoch": 1.85,
1658
- "learning_rate": 8.890688259109311e-06,
1659
- "loss": 1.1454,
1660
- "step": 2740
1661
- },
1662
- {
1663
- "epoch": 1.85,
1664
- "learning_rate": 8.886639676113361e-06,
1665
- "loss": 1.1084,
1666
- "step": 2750
1667
- },
1668
- {
1669
- "epoch": 1.86,
1670
- "learning_rate": 8.882591093117409e-06,
1671
- "loss": 1.1829,
1672
- "step": 2760
1673
- },
1674
- {
1675
- "epoch": 1.87,
1676
- "learning_rate": 8.878542510121459e-06,
1677
- "loss": 1.1262,
1678
- "step": 2770
1679
- },
1680
- {
1681
- "epoch": 1.88,
1682
- "learning_rate": 8.874493927125507e-06,
1683
- "loss": 1.1754,
1684
- "step": 2780
1685
- },
1686
- {
1687
- "epoch": 1.88,
1688
- "learning_rate": 8.870445344129556e-06,
1689
- "loss": 1.1269,
1690
- "step": 2790
1691
- },
1692
- {
1693
- "epoch": 1.89,
1694
- "learning_rate": 8.866396761133604e-06,
1695
- "loss": 1.1663,
1696
- "step": 2800
1697
- },
1698
- {
1699
- "epoch": 1.9,
1700
- "learning_rate": 8.862348178137652e-06,
1701
- "loss": 1.1457,
1702
- "step": 2810
1703
- },
1704
- {
1705
- "epoch": 1.9,
1706
- "learning_rate": 8.858299595141702e-06,
1707
- "loss": 1.1143,
1708
- "step": 2820
1709
- },
1710
- {
1711
- "epoch": 1.91,
1712
- "learning_rate": 8.85425101214575e-06,
1713
- "loss": 1.1956,
1714
- "step": 2830
1715
- },
1716
- {
1717
- "epoch": 1.92,
1718
- "learning_rate": 8.8502024291498e-06,
1719
- "loss": 1.152,
1720
- "step": 2840
1721
- },
1722
- {
1723
- "epoch": 1.92,
1724
- "learning_rate": 8.846153846153847e-06,
1725
- "loss": 1.1365,
1726
- "step": 2850
1727
- },
1728
- {
1729
- "epoch": 1.93,
1730
- "learning_rate": 8.842105263157895e-06,
1731
- "loss": 1.1519,
1732
- "step": 2860
1733
- },
1734
- {
1735
- "epoch": 1.94,
1736
- "learning_rate": 8.838056680161945e-06,
1737
- "loss": 1.126,
1738
- "step": 2870
1739
- },
1740
- {
1741
- "epoch": 1.94,
1742
- "learning_rate": 8.834008097165993e-06,
1743
- "loss": 1.1408,
1744
- "step": 2880
1745
- },
1746
- {
1747
- "epoch": 1.95,
1748
- "learning_rate": 8.82995951417004e-06,
1749
- "loss": 1.1522,
1750
- "step": 2890
1751
- },
1752
- {
1753
- "epoch": 1.96,
1754
- "learning_rate": 8.82591093117409e-06,
1755
- "loss": 1.1236,
1756
- "step": 2900
1757
- },
1758
- {
1759
- "epoch": 1.96,
1760
- "learning_rate": 8.821862348178138e-06,
1761
- "loss": 1.0927,
1762
- "step": 2910
1763
- },
1764
- {
1765
- "epoch": 1.97,
1766
- "learning_rate": 8.817813765182186e-06,
1767
- "loss": 1.21,
1768
- "step": 2920
1769
- },
1770
- {
1771
- "epoch": 1.98,
1772
- "learning_rate": 8.813765182186236e-06,
1773
- "loss": 1.1119,
1774
- "step": 2930
1775
- },
1776
- {
1777
- "epoch": 1.98,
1778
- "learning_rate": 8.809716599190284e-06,
1779
- "loss": 1.1383,
1780
- "step": 2940
1781
- },
1782
- {
1783
- "epoch": 1.99,
1784
- "learning_rate": 8.805668016194333e-06,
1785
- "loss": 1.1438,
1786
- "step": 2950
1787
- },
1788
- {
1789
- "epoch": 2.0,
1790
- "learning_rate": 8.801619433198381e-06,
1791
- "loss": 1.1747,
1792
- "step": 2960
1793
- },
1794
- {
1795
- "epoch": 2.0,
1796
- "eval_loss": 1.1016342639923096,
1797
- "eval_runtime": 10.6294,
1798
- "eval_samples_per_second": 103.487,
1799
- "eval_steps_per_second": 12.983,
1800
- "step": 2965
1801
- },
1802
- {
1803
- "epoch": 2.0,
1804
- "learning_rate": 8.79757085020243e-06,
1805
- "loss": 1.083,
1806
- "step": 2970
1807
- },
1808
- {
1809
- "epoch": 2.01,
1810
- "learning_rate": 8.793522267206479e-06,
1811
- "loss": 1.0542,
1812
- "step": 2980
1813
- },
1814
- {
1815
- "epoch": 2.02,
1816
- "learning_rate": 8.789473684210527e-06,
1817
- "loss": 1.0974,
1818
- "step": 2990
1819
- },
1820
- {
1821
- "epoch": 2.02,
1822
- "learning_rate": 8.785425101214576e-06,
1823
- "loss": 1.0074,
1824
- "step": 3000
1825
- },
1826
- {
1827
- "epoch": 2.03,
1828
- "learning_rate": 8.781376518218624e-06,
1829
- "loss": 1.0825,
1830
- "step": 3010
1831
- },
1832
- {
1833
- "epoch": 2.04,
1834
- "learning_rate": 8.777327935222674e-06,
1835
- "loss": 1.065,
1836
- "step": 3020
1837
- },
1838
- {
1839
- "epoch": 2.04,
1840
- "learning_rate": 8.77327935222672e-06,
1841
- "loss": 1.0996,
1842
- "step": 3030
1843
- },
1844
- {
1845
- "epoch": 2.05,
1846
- "learning_rate": 8.76923076923077e-06,
1847
- "loss": 1.1079,
1848
- "step": 3040
1849
- },
1850
- {
1851
- "epoch": 2.06,
1852
- "learning_rate": 8.76518218623482e-06,
1853
- "loss": 1.0803,
1854
- "step": 3050
1855
- },
1856
- {
1857
- "epoch": 2.06,
1858
- "learning_rate": 8.761133603238867e-06,
1859
- "loss": 1.0923,
1860
- "step": 3060
1861
- },
1862
- {
1863
- "epoch": 2.07,
1864
- "learning_rate": 8.757085020242915e-06,
1865
- "loss": 1.1171,
1866
- "step": 3070
1867
- },
1868
- {
1869
- "epoch": 2.08,
1870
- "learning_rate": 8.753036437246965e-06,
1871
- "loss": 1.0404,
1872
- "step": 3080
1873
- },
1874
- {
1875
- "epoch": 2.08,
1876
- "learning_rate": 8.748987854251013e-06,
1877
- "loss": 1.0881,
1878
- "step": 3090
1879
- },
1880
- {
1881
- "epoch": 2.09,
1882
- "learning_rate": 8.744939271255061e-06,
1883
- "loss": 1.1536,
1884
- "step": 3100
1885
- },
1886
- {
1887
- "epoch": 2.1,
1888
- "learning_rate": 8.74089068825911e-06,
1889
- "loss": 1.0868,
1890
- "step": 3110
1891
- },
1892
- {
1893
- "epoch": 2.1,
1894
- "learning_rate": 8.736842105263158e-06,
1895
- "loss": 1.1422,
1896
- "step": 3120
1897
- },
1898
- {
1899
- "epoch": 2.11,
1900
- "learning_rate": 8.732793522267208e-06,
1901
- "loss": 1.1048,
1902
- "step": 3130
1903
- },
1904
- {
1905
- "epoch": 2.12,
1906
- "learning_rate": 8.728744939271256e-06,
1907
- "loss": 1.136,
1908
- "step": 3140
1909
- },
1910
- {
1911
- "epoch": 2.12,
1912
- "learning_rate": 8.724696356275304e-06,
1913
- "loss": 1.07,
1914
- "step": 3150
1915
- },
1916
- {
1917
- "epoch": 2.13,
1918
- "learning_rate": 8.720647773279354e-06,
1919
- "loss": 1.1336,
1920
- "step": 3160
1921
- },
1922
- {
1923
- "epoch": 2.14,
1924
- "learning_rate": 8.716599190283401e-06,
1925
- "loss": 1.0946,
1926
- "step": 3170
1927
- },
1928
- {
1929
- "epoch": 2.15,
1930
- "learning_rate": 8.71255060728745e-06,
1931
- "loss": 1.061,
1932
- "step": 3180
1933
- },
1934
- {
1935
- "epoch": 2.15,
1936
- "learning_rate": 8.708502024291499e-06,
1937
- "loss": 1.0275,
1938
- "step": 3190
1939
- },
1940
- {
1941
- "epoch": 2.16,
1942
- "learning_rate": 8.704453441295547e-06,
1943
- "loss": 1.0665,
1944
- "step": 3200
1945
- },
1946
- {
1947
- "epoch": 2.17,
1948
- "learning_rate": 8.700404858299595e-06,
1949
- "loss": 1.0563,
1950
- "step": 3210
1951
- },
1952
- {
1953
- "epoch": 2.17,
1954
- "learning_rate": 8.696356275303645e-06,
1955
- "loss": 1.0756,
1956
- "step": 3220
1957
- },
1958
- {
1959
- "epoch": 2.18,
1960
- "learning_rate": 8.692307692307692e-06,
1961
- "loss": 1.0981,
1962
- "step": 3230
1963
- },
1964
- {
1965
- "epoch": 2.19,
1966
- "learning_rate": 8.688259109311742e-06,
1967
- "loss": 1.1138,
1968
- "step": 3240
1969
- },
1970
- {
1971
- "epoch": 2.19,
1972
- "learning_rate": 8.68421052631579e-06,
1973
- "loss": 1.0838,
1974
- "step": 3250
1975
- },
1976
- {
1977
- "epoch": 2.2,
1978
- "learning_rate": 8.680161943319838e-06,
1979
- "loss": 1.1053,
1980
- "step": 3260
1981
- },
1982
- {
1983
- "epoch": 2.21,
1984
- "learning_rate": 8.676113360323888e-06,
1985
- "loss": 1.1099,
1986
- "step": 3270
1987
- },
1988
- {
1989
- "epoch": 2.21,
1990
- "learning_rate": 8.672064777327936e-06,
1991
- "loss": 1.0884,
1992
- "step": 3280
1993
- },
1994
- {
1995
- "epoch": 2.22,
1996
- "learning_rate": 8.668016194331985e-06,
1997
- "loss": 1.0857,
1998
- "step": 3290
1999
- },
2000
- {
2001
- "epoch": 2.23,
2002
- "learning_rate": 8.663967611336033e-06,
2003
- "loss": 1.0629,
2004
- "step": 3300
2005
- },
2006
- {
2007
- "epoch": 2.23,
2008
- "learning_rate": 8.659919028340083e-06,
2009
- "loss": 1.0753,
2010
- "step": 3310
2011
- },
2012
- {
2013
- "epoch": 2.24,
2014
- "learning_rate": 8.65587044534413e-06,
2015
- "loss": 1.1481,
2016
- "step": 3320
2017
- },
2018
- {
2019
- "epoch": 2.25,
2020
- "learning_rate": 8.651821862348179e-06,
2021
- "loss": 1.1127,
2022
- "step": 3330
2023
- },
2024
- {
2025
- "epoch": 2.25,
2026
- "learning_rate": 8.647773279352228e-06,
2027
- "loss": 1.1666,
2028
- "step": 3340
2029
- },
2030
- {
2031
- "epoch": 2.26,
2032
- "learning_rate": 8.643724696356276e-06,
2033
- "loss": 1.1086,
2034
- "step": 3350
2035
- },
2036
- {
2037
- "epoch": 2.27,
2038
- "learning_rate": 8.639676113360324e-06,
2039
- "loss": 1.1102,
2040
- "step": 3360
2041
- },
2042
- {
2043
- "epoch": 2.27,
2044
- "learning_rate": 8.635627530364374e-06,
2045
- "loss": 1.1194,
2046
- "step": 3370
2047
- },
2048
- {
2049
- "epoch": 2.28,
2050
- "learning_rate": 8.631578947368422e-06,
2051
- "loss": 1.145,
2052
- "step": 3380
2053
- },
2054
- {
2055
- "epoch": 2.29,
2056
- "learning_rate": 8.62753036437247e-06,
2057
- "loss": 1.1103,
2058
- "step": 3390
2059
- },
2060
- {
2061
- "epoch": 2.29,
2062
- "learning_rate": 8.62348178137652e-06,
2063
- "loss": 1.0998,
2064
- "step": 3400
2065
- },
2066
- {
2067
- "epoch": 2.3,
2068
- "learning_rate": 8.619433198380567e-06,
2069
- "loss": 1.1551,
2070
- "step": 3410
2071
- },
2072
- {
2073
- "epoch": 2.31,
2074
- "learning_rate": 8.615384615384615e-06,
2075
- "loss": 1.0687,
2076
- "step": 3420
2077
- },
2078
- {
2079
- "epoch": 2.31,
2080
- "learning_rate": 8.611336032388665e-06,
2081
- "loss": 1.0959,
2082
- "step": 3430
2083
- },
2084
- {
2085
- "epoch": 2.32,
2086
- "learning_rate": 8.607287449392713e-06,
2087
- "loss": 1.0799,
2088
- "step": 3440
2089
- },
2090
- {
2091
- "epoch": 2.33,
2092
- "learning_rate": 8.60323886639676e-06,
2093
- "loss": 1.0673,
2094
- "step": 3450
2095
- },
2096
- {
2097
- "epoch": 2.33,
2098
- "learning_rate": 8.59919028340081e-06,
2099
- "loss": 1.0873,
2100
- "step": 3460
2101
- },
2102
- {
2103
- "epoch": 2.34,
2104
- "learning_rate": 8.595141700404858e-06,
2105
- "loss": 1.1119,
2106
- "step": 3470
2107
- },
2108
- {
2109
- "epoch": 2.35,
2110
- "learning_rate": 8.591093117408906e-06,
2111
- "loss": 1.1356,
2112
- "step": 3480
2113
- },
2114
- {
2115
- "epoch": 2.35,
2116
- "learning_rate": 8.587044534412956e-06,
2117
- "loss": 1.0924,
2118
- "step": 3490
2119
- },
2120
- {
2121
- "epoch": 2.36,
2122
- "learning_rate": 8.582995951417005e-06,
2123
- "loss": 1.0818,
2124
- "step": 3500
2125
- },
2126
- {
2127
- "epoch": 2.37,
2128
- "learning_rate": 8.578947368421053e-06,
2129
- "loss": 1.1007,
2130
- "step": 3510
2131
- },
2132
- {
2133
- "epoch": 2.37,
2134
- "learning_rate": 8.574898785425103e-06,
2135
- "loss": 1.0591,
2136
- "step": 3520
2137
- },
2138
- {
2139
- "epoch": 2.38,
2140
- "learning_rate": 8.57085020242915e-06,
2141
- "loss": 1.1069,
2142
- "step": 3530
2143
- },
2144
- {
2145
- "epoch": 2.39,
2146
- "learning_rate": 8.566801619433199e-06,
2147
- "loss": 1.1191,
2148
- "step": 3540
2149
- },
2150
- {
2151
- "epoch": 2.39,
2152
- "learning_rate": 8.562753036437247e-06,
2153
- "loss": 1.077,
2154
- "step": 3550
2155
- },
2156
- {
2157
- "epoch": 2.4,
2158
- "learning_rate": 8.558704453441296e-06,
2159
- "loss": 1.104,
2160
- "step": 3560
2161
- },
2162
- {
2163
- "epoch": 2.41,
2164
- "learning_rate": 8.554655870445344e-06,
2165
- "loss": 1.0758,
2166
- "step": 3570
2167
- },
2168
- {
2169
- "epoch": 2.41,
2170
- "learning_rate": 8.550607287449394e-06,
2171
- "loss": 1.0885,
2172
- "step": 3580
2173
- },
2174
- {
2175
- "epoch": 2.42,
2176
- "learning_rate": 8.546558704453442e-06,
2177
- "loss": 1.084,
2178
- "step": 3590
2179
- },
2180
- {
2181
- "epoch": 2.43,
2182
- "learning_rate": 8.54251012145749e-06,
2183
- "loss": 1.0869,
2184
- "step": 3600
2185
- },
2186
- {
2187
- "epoch": 2.44,
2188
- "learning_rate": 8.53846153846154e-06,
2189
- "loss": 1.073,
2190
- "step": 3610
2191
- },
2192
- {
2193
- "epoch": 2.44,
2194
- "learning_rate": 8.534412955465587e-06,
2195
- "loss": 1.0766,
2196
- "step": 3620
2197
- },
2198
- {
2199
- "epoch": 2.45,
2200
- "learning_rate": 8.530364372469635e-06,
2201
- "loss": 1.1085,
2202
- "step": 3630
2203
- },
2204
- {
2205
- "epoch": 2.46,
2206
- "learning_rate": 8.526315789473685e-06,
2207
- "loss": 1.097,
2208
- "step": 3640
2209
- },
2210
- {
2211
- "epoch": 2.46,
2212
- "learning_rate": 8.522267206477733e-06,
2213
- "loss": 1.0968,
2214
- "step": 3650
2215
- },
2216
- {
2217
- "epoch": 2.47,
2218
- "learning_rate": 8.518218623481783e-06,
2219
- "loss": 1.0917,
2220
- "step": 3660
2221
- },
2222
- {
2223
- "epoch": 2.48,
2224
- "learning_rate": 8.51417004048583e-06,
2225
- "loss": 1.047,
2226
- "step": 3670
2227
- },
2228
- {
2229
- "epoch": 2.48,
2230
- "learning_rate": 8.510121457489878e-06,
2231
- "loss": 1.0469,
2232
- "step": 3680
2233
- },
2234
- {
2235
- "epoch": 2.49,
2236
- "learning_rate": 8.506072874493928e-06,
2237
- "loss": 1.1114,
2238
- "step": 3690
2239
- },
2240
- {
2241
- "epoch": 2.5,
2242
- "learning_rate": 8.502024291497976e-06,
2243
- "loss": 1.0319,
2244
- "step": 3700
2245
- },
2246
- {
2247
- "epoch": 2.5,
2248
- "learning_rate": 8.497975708502024e-06,
2249
- "loss": 1.0733,
2250
- "step": 3710
2251
- },
2252
- {
2253
- "epoch": 2.51,
2254
- "learning_rate": 8.493927125506074e-06,
2255
- "loss": 1.1237,
2256
- "step": 3720
2257
- },
2258
- {
2259
- "epoch": 2.52,
2260
- "learning_rate": 8.489878542510121e-06,
2261
- "loss": 1.0649,
2262
- "step": 3730
2263
- },
2264
- {
2265
- "epoch": 2.52,
2266
- "learning_rate": 8.48582995951417e-06,
2267
- "loss": 1.0903,
2268
- "step": 3740
2269
- },
2270
- {
2271
- "epoch": 2.53,
2272
- "learning_rate": 8.481781376518219e-06,
2273
- "loss": 1.0633,
2274
- "step": 3750
2275
- },
2276
- {
2277
- "epoch": 2.54,
2278
- "learning_rate": 8.477732793522267e-06,
2279
- "loss": 1.1259,
2280
- "step": 3760
2281
- },
2282
- {
2283
- "epoch": 2.54,
2284
- "learning_rate": 8.473684210526317e-06,
2285
- "loss": 1.0978,
2286
- "step": 3770
2287
- },
2288
- {
2289
- "epoch": 2.55,
2290
- "learning_rate": 8.469635627530365e-06,
2291
- "loss": 1.1362,
2292
- "step": 3780
2293
- },
2294
- {
2295
- "epoch": 2.56,
2296
- "learning_rate": 8.465587044534414e-06,
2297
- "loss": 1.1487,
2298
- "step": 3790
2299
- },
2300
- {
2301
- "epoch": 2.56,
2302
- "learning_rate": 8.461538461538462e-06,
2303
- "loss": 1.0887,
2304
- "step": 3800
2305
- },
2306
- {
2307
- "epoch": 2.57,
2308
- "learning_rate": 8.457489878542512e-06,
2309
- "loss": 1.1039,
2310
- "step": 3810
2311
- },
2312
- {
2313
- "epoch": 2.58,
2314
- "learning_rate": 8.453441295546558e-06,
2315
- "loss": 1.0906,
2316
- "step": 3820
2317
- },
2318
- {
2319
- "epoch": 2.58,
2320
- "learning_rate": 8.449392712550608e-06,
2321
- "loss": 1.1039,
2322
- "step": 3830
2323
- },
2324
- {
2325
- "epoch": 2.59,
2326
- "learning_rate": 8.445344129554657e-06,
2327
- "loss": 1.0432,
2328
- "step": 3840
2329
- },
2330
- {
2331
- "epoch": 2.6,
2332
- "learning_rate": 8.441295546558705e-06,
2333
- "loss": 1.1565,
2334
- "step": 3850
2335
- },
2336
- {
2337
- "epoch": 2.6,
2338
- "learning_rate": 8.437246963562753e-06,
2339
- "loss": 1.1786,
2340
- "step": 3860
2341
- },
2342
- {
2343
- "epoch": 2.61,
2344
- "learning_rate": 8.433198380566803e-06,
2345
- "loss": 1.0724,
2346
- "step": 3870
2347
- },
2348
- {
2349
- "epoch": 2.62,
2350
- "learning_rate": 8.42914979757085e-06,
2351
- "loss": 1.0863,
2352
- "step": 3880
2353
- },
2354
- {
2355
- "epoch": 2.62,
2356
- "learning_rate": 8.425101214574899e-06,
2357
- "loss": 1.0825,
2358
- "step": 3890
2359
- },
2360
- {
2361
- "epoch": 2.63,
2362
- "learning_rate": 8.421052631578948e-06,
2363
- "loss": 1.0831,
2364
- "step": 3900
2365
- },
2366
- {
2367
- "epoch": 2.64,
2368
- "learning_rate": 8.417004048582996e-06,
2369
- "loss": 1.0681,
2370
- "step": 3910
2371
- },
2372
- {
2373
- "epoch": 2.64,
2374
- "learning_rate": 8.412955465587044e-06,
2375
- "loss": 1.0871,
2376
- "step": 3920
2377
- },
2378
- {
2379
- "epoch": 2.65,
2380
- "learning_rate": 8.408906882591094e-06,
2381
- "loss": 1.0531,
2382
- "step": 3930
2383
- },
2384
- {
2385
- "epoch": 2.66,
2386
- "learning_rate": 8.404858299595142e-06,
2387
- "loss": 1.1174,
2388
- "step": 3940
2389
- },
2390
- {
2391
- "epoch": 2.66,
2392
- "learning_rate": 8.400809716599191e-06,
2393
- "loss": 1.117,
2394
- "step": 3950
2395
- },
2396
- {
2397
- "epoch": 2.67,
2398
- "learning_rate": 8.39676113360324e-06,
2399
- "loss": 1.0938,
2400
- "step": 3960
2401
- },
2402
- {
2403
- "epoch": 2.68,
2404
- "learning_rate": 8.392712550607287e-06,
2405
- "loss": 1.0736,
2406
- "step": 3970
2407
- },
2408
- {
2409
- "epoch": 2.68,
2410
- "learning_rate": 8.388663967611337e-06,
2411
- "loss": 1.0872,
2412
- "step": 3980
2413
- },
2414
- {
2415
- "epoch": 2.69,
2416
- "learning_rate": 8.384615384615385e-06,
2417
- "loss": 1.1,
2418
- "step": 3990
2419
- },
2420
- {
2421
- "epoch": 2.7,
2422
- "learning_rate": 8.380566801619433e-06,
2423
- "loss": 1.1201,
2424
- "step": 4000
2425
- },
2426
- {
2427
- "epoch": 2.7,
2428
- "learning_rate": 8.376518218623482e-06,
2429
- "loss": 1.0755,
2430
- "step": 4010
2431
- },
2432
- {
2433
- "epoch": 2.71,
2434
- "learning_rate": 8.37246963562753e-06,
2435
- "loss": 1.0917,
2436
- "step": 4020
2437
- },
2438
- {
2439
- "epoch": 2.72,
2440
- "learning_rate": 8.368421052631578e-06,
2441
- "loss": 1.0873,
2442
- "step": 4030
2443
- },
2444
- {
2445
- "epoch": 2.73,
2446
- "learning_rate": 8.364372469635628e-06,
2447
- "loss": 1.0858,
2448
- "step": 4040
2449
- },
2450
- {
2451
- "epoch": 2.73,
2452
- "learning_rate": 8.360323886639676e-06,
2453
- "loss": 1.0721,
2454
- "step": 4050
2455
- },
2456
- {
2457
- "epoch": 2.74,
2458
- "learning_rate": 8.356275303643725e-06,
2459
- "loss": 1.138,
2460
- "step": 4060
2461
- },
2462
- {
2463
- "epoch": 2.75,
2464
- "learning_rate": 8.352226720647773e-06,
2465
- "loss": 1.1194,
2466
- "step": 4070
2467
- },
2468
- {
2469
- "epoch": 2.75,
2470
- "learning_rate": 8.348178137651823e-06,
2471
- "loss": 1.1364,
2472
- "step": 4080
2473
- },
2474
- {
2475
- "epoch": 2.76,
2476
- "learning_rate": 8.344129554655871e-06,
2477
- "loss": 1.1201,
2478
- "step": 4090
2479
- },
2480
- {
2481
- "epoch": 2.77,
2482
- "learning_rate": 8.34008097165992e-06,
2483
- "loss": 1.0585,
2484
- "step": 4100
2485
- },
2486
- {
2487
- "epoch": 2.77,
2488
- "learning_rate": 8.336032388663968e-06,
2489
- "loss": 1.0613,
2490
- "step": 4110
2491
- },
2492
- {
2493
- "epoch": 2.78,
2494
- "learning_rate": 8.331983805668016e-06,
2495
- "loss": 1.1091,
2496
- "step": 4120
2497
- },
2498
- {
2499
- "epoch": 2.79,
2500
- "learning_rate": 8.327935222672066e-06,
2501
- "loss": 1.1446,
2502
- "step": 4130
2503
- },
2504
- {
2505
- "epoch": 2.79,
2506
- "learning_rate": 8.323886639676114e-06,
2507
- "loss": 1.0413,
2508
- "step": 4140
2509
- },
2510
- {
2511
- "epoch": 2.8,
2512
- "learning_rate": 8.319838056680162e-06,
2513
- "loss": 1.1225,
2514
- "step": 4150
2515
- },
2516
- {
2517
- "epoch": 2.81,
2518
- "learning_rate": 8.315789473684212e-06,
2519
- "loss": 1.1121,
2520
- "step": 4160
2521
- },
2522
- {
2523
- "epoch": 2.81,
2524
- "learning_rate": 8.31174089068826e-06,
2525
- "loss": 1.0534,
2526
- "step": 4170
2527
- },
2528
- {
2529
- "epoch": 2.82,
2530
- "learning_rate": 8.307692307692307e-06,
2531
- "loss": 1.1097,
2532
- "step": 4180
2533
- },
2534
- {
2535
- "epoch": 2.83,
2536
- "learning_rate": 8.303643724696357e-06,
2537
- "loss": 1.1147,
2538
- "step": 4190
2539
- },
2540
- {
2541
- "epoch": 2.83,
2542
- "learning_rate": 8.299595141700405e-06,
2543
- "loss": 1.1009,
2544
- "step": 4200
2545
- },
2546
- {
2547
- "epoch": 2.84,
2548
- "learning_rate": 8.295546558704453e-06,
2549
- "loss": 1.0584,
2550
- "step": 4210
2551
- },
2552
- {
2553
- "epoch": 2.85,
2554
- "learning_rate": 8.291497975708503e-06,
2555
- "loss": 1.1045,
2556
- "step": 4220
2557
- },
2558
- {
2559
- "epoch": 2.85,
2560
- "learning_rate": 8.28744939271255e-06,
2561
- "loss": 1.0559,
2562
- "step": 4230
2563
- },
2564
- {
2565
- "epoch": 2.86,
2566
- "learning_rate": 8.2834008097166e-06,
2567
- "loss": 1.063,
2568
- "step": 4240
2569
- },
2570
- {
2571
- "epoch": 2.87,
2572
- "learning_rate": 8.279352226720648e-06,
2573
- "loss": 1.1671,
2574
- "step": 4250
2575
- },
2576
- {
2577
- "epoch": 2.87,
2578
- "learning_rate": 8.275303643724696e-06,
2579
- "loss": 1.077,
2580
- "step": 4260
2581
- },
2582
- {
2583
- "epoch": 2.88,
2584
- "learning_rate": 8.271255060728746e-06,
2585
- "loss": 1.1119,
2586
- "step": 4270
2587
- },
2588
- {
2589
- "epoch": 2.89,
2590
- "learning_rate": 8.267206477732794e-06,
2591
- "loss": 1.0773,
2592
- "step": 4280
2593
- },
2594
- {
2595
- "epoch": 2.89,
2596
- "learning_rate": 8.263157894736843e-06,
2597
- "loss": 1.0874,
2598
- "step": 4290
2599
- },
2600
- {
2601
- "epoch": 2.9,
2602
- "learning_rate": 8.259109311740891e-06,
2603
- "loss": 1.0793,
2604
- "step": 4300
2605
- },
2606
- {
2607
- "epoch": 2.91,
2608
- "learning_rate": 8.25506072874494e-06,
2609
- "loss": 1.1011,
2610
- "step": 4310
2611
- },
2612
- {
2613
- "epoch": 2.91,
2614
- "learning_rate": 8.251012145748987e-06,
2615
- "loss": 1.1232,
2616
- "step": 4320
2617
- },
2618
- {
2619
- "epoch": 2.92,
2620
- "learning_rate": 8.246963562753037e-06,
2621
- "loss": 1.0953,
2622
- "step": 4330
2623
- },
2624
- {
2625
- "epoch": 2.93,
2626
- "learning_rate": 8.242914979757085e-06,
2627
- "loss": 1.0841,
2628
- "step": 4340
2629
- },
2630
- {
2631
- "epoch": 2.93,
2632
- "learning_rate": 8.238866396761134e-06,
2633
- "loss": 1.1036,
2634
- "step": 4350
2635
- },
2636
- {
2637
- "epoch": 2.94,
2638
- "learning_rate": 8.234817813765182e-06,
2639
- "loss": 1.1207,
2640
- "step": 4360
2641
- },
2642
- {
2643
- "epoch": 2.95,
2644
- "learning_rate": 8.230769230769232e-06,
2645
- "loss": 1.1033,
2646
- "step": 4370
2647
- },
2648
- {
2649
- "epoch": 2.95,
2650
- "learning_rate": 8.22672064777328e-06,
2651
- "loss": 1.0719,
2652
- "step": 4380
2653
- },
2654
- {
2655
- "epoch": 2.96,
2656
- "learning_rate": 8.22267206477733e-06,
2657
- "loss": 1.0539,
2658
- "step": 4390
2659
- },
2660
- {
2661
- "epoch": 2.97,
2662
- "learning_rate": 8.218623481781377e-06,
2663
- "loss": 1.0818,
2664
- "step": 4400
2665
- },
2666
- {
2667
- "epoch": 2.97,
2668
- "learning_rate": 8.214574898785425e-06,
2669
- "loss": 1.1149,
2670
- "step": 4410
2671
- },
2672
- {
2673
- "epoch": 2.98,
2674
- "learning_rate": 8.210526315789475e-06,
2675
- "loss": 1.0624,
2676
- "step": 4420
2677
- },
2678
- {
2679
- "epoch": 2.99,
2680
- "learning_rate": 8.206477732793523e-06,
2681
- "loss": 1.1197,
2682
- "step": 4430
2683
- },
2684
- {
2685
- "epoch": 2.99,
2686
- "learning_rate": 8.20242914979757e-06,
2687
- "loss": 1.0567,
2688
- "step": 4440
2689
- },
2690
- {
2691
- "epoch": 3.0,
2692
- "eval_loss": 1.0876251459121704,
2693
- "eval_runtime": 10.6357,
2694
- "eval_samples_per_second": 103.425,
2695
- "eval_steps_per_second": 12.975,
2696
- "step": 4447
2697
- },
2698
- {
2699
- "epoch": 3.0,
2700
- "learning_rate": 8.19838056680162e-06,
2701
- "loss": 1.0768,
2702
- "step": 4450
2703
- },
2704
- {
2705
- "epoch": 3.01,
2706
- "learning_rate": 8.194331983805668e-06,
2707
- "loss": 1.0687,
2708
- "step": 4460
2709
- },
2710
- {
2711
- "epoch": 3.02,
2712
- "learning_rate": 8.190283400809716e-06,
2713
- "loss": 1.1029,
2714
- "step": 4470
2715
- },
2716
- {
2717
- "epoch": 3.02,
2718
- "learning_rate": 8.186234817813766e-06,
2719
- "loss": 1.0569,
2720
- "step": 4480
2721
- },
2722
- {
2723
- "epoch": 3.03,
2724
- "learning_rate": 8.182186234817814e-06,
2725
- "loss": 1.0151,
2726
- "step": 4490
2727
- },
2728
- {
2729
- "epoch": 3.04,
2730
- "learning_rate": 8.178137651821862e-06,
2731
- "loss": 1.0405,
2732
- "step": 4500
2733
- },
2734
- {
2735
- "epoch": 3.04,
2736
- "learning_rate": 8.174089068825911e-06,
2737
- "loss": 1.07,
2738
- "step": 4510
2739
- },
2740
- {
2741
- "epoch": 3.05,
2742
- "learning_rate": 8.17004048582996e-06,
2743
- "loss": 1.0643,
2744
- "step": 4520
2745
- },
2746
- {
2747
- "epoch": 3.06,
2748
- "learning_rate": 8.165991902834009e-06,
2749
- "loss": 1.0859,
2750
- "step": 4530
2751
- },
2752
- {
2753
- "epoch": 3.06,
2754
- "learning_rate": 8.161943319838057e-06,
2755
- "loss": 1.0384,
2756
- "step": 4540
2757
- },
2758
- {
2759
- "epoch": 3.07,
2760
- "learning_rate": 8.157894736842105e-06,
2761
- "loss": 1.0873,
2762
- "step": 4550
2763
- },
2764
- {
2765
- "epoch": 3.08,
2766
- "learning_rate": 8.153846153846154e-06,
2767
- "loss": 1.0847,
2768
- "step": 4560
2769
- },
2770
- {
2771
- "epoch": 3.08,
2772
- "learning_rate": 8.149797570850202e-06,
2773
- "loss": 1.0105,
2774
- "step": 4570
2775
- },
2776
- {
2777
- "epoch": 3.09,
2778
- "learning_rate": 8.145748987854252e-06,
2779
- "loss": 1.0407,
2780
- "step": 4580
2781
- },
2782
- {
2783
- "epoch": 3.1,
2784
- "learning_rate": 8.1417004048583e-06,
2785
- "loss": 1.0437,
2786
- "step": 4590
2787
- },
2788
- {
2789
- "epoch": 3.1,
2790
- "learning_rate": 8.13765182186235e-06,
2791
- "loss": 1.0641,
2792
- "step": 4600
2793
- },
2794
- {
2795
- "epoch": 3.11,
2796
- "learning_rate": 8.133603238866397e-06,
2797
- "loss": 1.0662,
2798
- "step": 4610
2799
- },
2800
- {
2801
- "epoch": 3.12,
2802
- "learning_rate": 8.129554655870445e-06,
2803
- "loss": 1.0608,
2804
- "step": 4620
2805
- },
2806
- {
2807
- "epoch": 3.12,
2808
- "learning_rate": 8.125506072874495e-06,
2809
- "loss": 1.0728,
2810
- "step": 4630
2811
- },
2812
- {
2813
- "epoch": 3.13,
2814
- "learning_rate": 8.121457489878543e-06,
2815
- "loss": 1.039,
2816
- "step": 4640
2817
- },
2818
- {
2819
- "epoch": 3.14,
2820
- "learning_rate": 8.117408906882591e-06,
2821
- "loss": 1.0898,
2822
- "step": 4650
2823
- },
2824
- {
2825
- "epoch": 3.14,
2826
- "learning_rate": 8.11336032388664e-06,
2827
- "loss": 1.107,
2828
- "step": 4660
2829
- },
2830
- {
2831
- "epoch": 3.15,
2832
- "learning_rate": 8.109311740890688e-06,
2833
- "loss": 1.1001,
2834
- "step": 4670
2835
- },
2836
- {
2837
- "epoch": 3.16,
2838
- "learning_rate": 8.105263157894738e-06,
2839
- "loss": 1.0371,
2840
- "step": 4680
2841
- },
2842
- {
2843
- "epoch": 3.16,
2844
- "learning_rate": 8.101214574898786e-06,
2845
- "loss": 1.0945,
2846
- "step": 4690
2847
- },
2848
- {
2849
- "epoch": 3.17,
2850
- "learning_rate": 8.097165991902834e-06,
2851
- "loss": 1.0509,
2852
- "step": 4700
2853
- },
2854
- {
2855
- "epoch": 3.18,
2856
- "learning_rate": 8.093117408906884e-06,
2857
- "loss": 1.0694,
2858
- "step": 4710
2859
- },
2860
- {
2861
- "epoch": 3.18,
2862
- "learning_rate": 8.089068825910931e-06,
2863
- "loss": 1.0153,
2864
- "step": 4720
2865
- },
2866
- {
2867
- "epoch": 3.19,
2868
- "learning_rate": 8.08502024291498e-06,
2869
- "loss": 1.0373,
2870
- "step": 4730
2871
- },
2872
- {
2873
- "epoch": 3.2,
2874
- "learning_rate": 8.080971659919029e-06,
2875
- "loss": 1.0583,
2876
- "step": 4740
2877
- },
2878
- {
2879
- "epoch": 3.2,
2880
- "learning_rate": 8.076923076923077e-06,
2881
- "loss": 1.0213,
2882
- "step": 4750
2883
- },
2884
- {
2885
- "epoch": 3.21,
2886
- "learning_rate": 8.072874493927125e-06,
2887
- "loss": 1.0707,
2888
- "step": 4760
2889
- },
2890
- {
2891
- "epoch": 3.22,
2892
- "learning_rate": 8.068825910931175e-06,
2893
- "loss": 1.0801,
2894
- "step": 4770
2895
- },
2896
- {
2897
- "epoch": 3.22,
2898
- "learning_rate": 8.064777327935222e-06,
2899
- "loss": 1.0336,
2900
- "step": 4780
2901
- },
2902
- {
2903
- "epoch": 3.23,
2904
- "learning_rate": 8.060728744939272e-06,
2905
- "loss": 1.0829,
2906
- "step": 4790
2907
- },
2908
- {
2909
- "epoch": 3.24,
2910
- "learning_rate": 8.05668016194332e-06,
2911
- "loss": 1.0853,
2912
- "step": 4800
2913
- },
2914
- {
2915
- "epoch": 3.24,
2916
- "learning_rate": 8.05263157894737e-06,
2917
- "loss": 1.074,
2918
- "step": 4810
2919
- },
2920
- {
2921
- "epoch": 3.25,
2922
- "learning_rate": 8.048582995951418e-06,
2923
- "loss": 1.0805,
2924
- "step": 4820
2925
- },
2926
- {
2927
- "epoch": 3.26,
2928
- "learning_rate": 8.044534412955467e-06,
2929
- "loss": 1.0644,
2930
- "step": 4830
2931
- },
2932
- {
2933
- "epoch": 3.26,
2934
- "learning_rate": 8.040485829959514e-06,
2935
- "loss": 1.039,
2936
- "step": 4840
2937
- },
2938
- {
2939
- "epoch": 3.27,
2940
- "learning_rate": 8.036437246963563e-06,
2941
- "loss": 1.0326,
2942
- "step": 4850
2943
- },
2944
- {
2945
- "epoch": 3.28,
2946
- "learning_rate": 8.032388663967611e-06,
2947
- "loss": 1.0776,
2948
- "step": 4860
2949
- },
2950
- {
2951
- "epoch": 3.28,
2952
- "learning_rate": 8.02834008097166e-06,
2953
- "loss": 1.043,
2954
- "step": 4870
2955
- },
2956
- {
2957
- "epoch": 3.29,
2958
- "learning_rate": 8.024291497975709e-06,
2959
- "loss": 1.0784,
2960
- "step": 4880
2961
- },
2962
- {
2963
- "epoch": 3.3,
2964
- "learning_rate": 8.020242914979758e-06,
2965
- "loss": 1.0783,
2966
- "step": 4890
2967
- },
2968
- {
2969
- "epoch": 3.31,
2970
- "learning_rate": 8.016194331983806e-06,
2971
- "loss": 1.0987,
2972
- "step": 4900
2973
- },
2974
- {
2975
- "epoch": 3.31,
2976
- "learning_rate": 8.012145748987854e-06,
2977
- "loss": 1.0515,
2978
- "step": 4910
2979
- },
2980
- {
2981
- "epoch": 3.32,
2982
- "learning_rate": 8.008097165991904e-06,
2983
- "loss": 1.0667,
2984
- "step": 4920
2985
- },
2986
- {
2987
- "epoch": 3.33,
2988
- "learning_rate": 8.004048582995952e-06,
2989
- "loss": 1.1095,
2990
- "step": 4930
2991
- },
2992
- {
2993
- "epoch": 3.33,
2994
- "learning_rate": 8e-06,
2995
- "loss": 1.0245,
2996
- "step": 4940
2997
- },
2998
- {
2999
- "epoch": 3.34,
3000
- "learning_rate": 7.99595141700405e-06,
3001
- "loss": 1.0753,
3002
- "step": 4950
3003
- },
3004
- {
3005
- "epoch": 3.35,
3006
- "learning_rate": 7.991902834008097e-06,
3007
- "loss": 1.0522,
3008
- "step": 4960
3009
- },
3010
- {
3011
- "epoch": 3.35,
3012
- "learning_rate": 7.987854251012147e-06,
3013
- "loss": 1.0106,
3014
- "step": 4970
3015
- },
3016
- {
3017
- "epoch": 3.36,
3018
- "learning_rate": 7.983805668016195e-06,
3019
- "loss": 1.0948,
3020
- "step": 4980
3021
- },
3022
- {
3023
- "epoch": 3.37,
3024
- "learning_rate": 7.979757085020243e-06,
3025
- "loss": 1.0649,
3026
- "step": 4990
3027
- },
3028
- {
3029
- "epoch": 3.37,
3030
- "learning_rate": 7.975708502024292e-06,
3031
- "loss": 1.0096,
3032
- "step": 5000
3033
- },
3034
- {
3035
- "epoch": 3.38,
3036
- "learning_rate": 7.97165991902834e-06,
3037
- "loss": 1.0076,
3038
- "step": 5010
3039
- },
3040
- {
3041
- "epoch": 3.39,
3042
- "learning_rate": 7.967611336032388e-06,
3043
- "loss": 1.0605,
3044
- "step": 5020
3045
- },
3046
- {
3047
- "epoch": 3.39,
3048
- "learning_rate": 7.963562753036438e-06,
3049
- "loss": 1.0544,
3050
- "step": 5030
3051
- },
3052
- {
3053
- "epoch": 3.4,
3054
- "learning_rate": 7.959514170040486e-06,
3055
- "loss": 1.0638,
3056
- "step": 5040
3057
- },
3058
- {
3059
- "epoch": 3.41,
3060
- "learning_rate": 7.955465587044534e-06,
3061
- "loss": 1.021,
3062
- "step": 5050
3063
- },
3064
- {
3065
- "epoch": 3.41,
3066
- "learning_rate": 7.951417004048583e-06,
3067
- "loss": 1.0622,
3068
- "step": 5060
3069
- },
3070
- {
3071
- "epoch": 3.42,
3072
- "learning_rate": 7.947368421052631e-06,
3073
- "loss": 1.0452,
3074
- "step": 5070
3075
- },
3076
- {
3077
- "epoch": 3.43,
3078
- "learning_rate": 7.943319838056681e-06,
3079
- "loss": 1.0678,
3080
- "step": 5080
3081
- },
3082
- {
3083
- "epoch": 3.43,
3084
- "learning_rate": 7.939271255060729e-06,
3085
- "loss": 0.9841,
3086
- "step": 5090
3087
- },
3088
- {
3089
- "epoch": 3.44,
3090
- "learning_rate": 7.935222672064778e-06,
3091
- "loss": 1.0684,
3092
- "step": 5100
3093
- },
3094
- {
3095
- "epoch": 3.45,
3096
- "learning_rate": 7.931174089068826e-06,
3097
- "loss": 1.0393,
3098
- "step": 5110
3099
- },
3100
- {
3101
- "epoch": 3.45,
3102
- "learning_rate": 7.927125506072876e-06,
3103
- "loss": 1.0816,
3104
- "step": 5120
3105
- },
3106
- {
3107
- "epoch": 3.46,
3108
- "learning_rate": 7.923076923076922e-06,
3109
- "loss": 1.0744,
3110
- "step": 5130
3111
- },
3112
- {
3113
- "epoch": 3.47,
3114
- "learning_rate": 7.919028340080972e-06,
3115
- "loss": 1.0575,
3116
- "step": 5140
3117
- },
3118
- {
3119
- "epoch": 3.47,
3120
- "learning_rate": 7.914979757085022e-06,
3121
- "loss": 1.119,
3122
- "step": 5150
3123
- },
3124
- {
3125
- "epoch": 3.48,
3126
- "learning_rate": 7.91093117408907e-06,
3127
- "loss": 1.0652,
3128
- "step": 5160
3129
- },
3130
- {
3131
- "epoch": 3.49,
3132
- "learning_rate": 7.906882591093117e-06,
3133
- "loss": 0.9888,
3134
- "step": 5170
3135
- },
3136
- {
3137
- "epoch": 3.49,
3138
- "learning_rate": 7.902834008097167e-06,
3139
- "loss": 1.0415,
3140
- "step": 5180
3141
- },
3142
- {
3143
- "epoch": 3.5,
3144
- "learning_rate": 7.898785425101215e-06,
3145
- "loss": 1.0266,
3146
- "step": 5190
3147
- },
3148
- {
3149
- "epoch": 3.51,
3150
- "learning_rate": 7.894736842105263e-06,
3151
- "loss": 1.0204,
3152
- "step": 5200
3153
- },
3154
- {
3155
- "epoch": 3.51,
3156
- "learning_rate": 7.890688259109313e-06,
3157
- "loss": 1.0709,
3158
- "step": 5210
3159
- },
3160
- {
3161
- "epoch": 3.52,
3162
- "learning_rate": 7.88663967611336e-06,
3163
- "loss": 1.0342,
3164
- "step": 5220
3165
- },
3166
- {
3167
- "epoch": 3.53,
3168
- "learning_rate": 7.882591093117408e-06,
3169
- "loss": 1.0492,
3170
- "step": 5230
3171
- },
3172
- {
3173
- "epoch": 3.53,
3174
- "learning_rate": 7.878542510121458e-06,
3175
- "loss": 1.0375,
3176
- "step": 5240
3177
- },
3178
- {
3179
- "epoch": 3.54,
3180
- "learning_rate": 7.874493927125506e-06,
3181
- "loss": 1.0188,
3182
- "step": 5250
3183
- },
3184
- {
3185
- "epoch": 3.55,
3186
- "learning_rate": 7.870445344129556e-06,
3187
- "loss": 1.0951,
3188
- "step": 5260
3189
- },
3190
- {
3191
- "epoch": 3.55,
3192
- "learning_rate": 7.866396761133604e-06,
3193
- "loss": 1.0133,
3194
- "step": 5270
3195
- },
3196
- {
3197
- "epoch": 3.56,
3198
- "learning_rate": 7.862348178137651e-06,
3199
- "loss": 1.0414,
3200
- "step": 5280
3201
- },
3202
- {
3203
- "epoch": 3.57,
3204
- "learning_rate": 7.858299595141701e-06,
3205
- "loss": 1.0863,
3206
- "step": 5290
3207
- },
3208
- {
3209
- "epoch": 3.58,
3210
- "learning_rate": 7.854251012145749e-06,
3211
- "loss": 1.0998,
3212
- "step": 5300
3213
- },
3214
- {
3215
- "epoch": 3.58,
3216
- "learning_rate": 7.850202429149797e-06,
3217
- "loss": 1.0854,
3218
- "step": 5310
3219
- },
3220
- {
3221
- "epoch": 3.59,
3222
- "learning_rate": 7.846153846153847e-06,
3223
- "loss": 1.0594,
3224
- "step": 5320
3225
- },
3226
- {
3227
- "epoch": 3.6,
3228
- "learning_rate": 7.842105263157896e-06,
3229
- "loss": 1.0225,
3230
- "step": 5330
3231
- },
3232
- {
3233
- "epoch": 3.6,
3234
- "learning_rate": 7.838056680161942e-06,
3235
- "loss": 1.0728,
3236
- "step": 5340
3237
- },
3238
- {
3239
- "epoch": 3.61,
3240
- "learning_rate": 7.834008097165992e-06,
3241
- "loss": 1.0417,
3242
- "step": 5350
3243
- },
3244
- {
3245
- "epoch": 3.62,
3246
- "learning_rate": 7.82995951417004e-06,
3247
- "loss": 1.0226,
3248
- "step": 5360
3249
- },
3250
- {
3251
- "epoch": 3.62,
3252
- "learning_rate": 7.82591093117409e-06,
3253
- "loss": 1.0953,
3254
- "step": 5370
3255
- },
3256
- {
3257
- "epoch": 3.63,
3258
- "learning_rate": 7.821862348178138e-06,
3259
- "loss": 1.053,
3260
- "step": 5380
3261
- },
3262
- {
3263
- "epoch": 3.64,
3264
- "learning_rate": 7.817813765182187e-06,
3265
- "loss": 1.0523,
3266
- "step": 5390
3267
- },
3268
- {
3269
- "epoch": 3.64,
3270
- "learning_rate": 7.813765182186235e-06,
3271
- "loss": 1.1017,
3272
- "step": 5400
3273
- },
3274
- {
3275
- "epoch": 3.65,
3276
- "learning_rate": 7.809716599190285e-06,
3277
- "loss": 1.0032,
3278
- "step": 5410
3279
- },
3280
- {
3281
- "epoch": 3.66,
3282
- "learning_rate": 7.805668016194333e-06,
3283
- "loss": 1.0331,
3284
- "step": 5420
3285
- },
3286
- {
3287
- "epoch": 3.66,
3288
- "learning_rate": 7.80161943319838e-06,
3289
- "loss": 1.0483,
3290
- "step": 5430
3291
- },
3292
- {
3293
- "epoch": 3.67,
3294
- "learning_rate": 7.79757085020243e-06,
3295
- "loss": 1.0291,
3296
- "step": 5440
3297
- },
3298
- {
3299
- "epoch": 3.68,
3300
- "learning_rate": 7.793522267206478e-06,
3301
- "loss": 1.0113,
3302
- "step": 5450
3303
- },
3304
- {
3305
- "epoch": 3.68,
3306
- "learning_rate": 7.789473684210526e-06,
3307
- "loss": 1.0407,
3308
- "step": 5460
3309
- },
3310
- {
3311
- "epoch": 3.69,
3312
- "learning_rate": 7.785425101214576e-06,
3313
- "loss": 1.077,
3314
- "step": 5470
3315
- },
3316
- {
3317
- "epoch": 3.7,
3318
- "learning_rate": 7.781376518218624e-06,
3319
- "loss": 1.1063,
3320
- "step": 5480
3321
- },
3322
- {
3323
- "epoch": 3.7,
3324
- "learning_rate": 7.777327935222672e-06,
3325
- "loss": 1.0377,
3326
- "step": 5490
3327
- },
3328
- {
3329
- "epoch": 3.71,
3330
- "learning_rate": 7.773279352226721e-06,
3331
- "loss": 1.0841,
3332
- "step": 5500
3333
- },
3334
- {
3335
- "epoch": 3.72,
3336
- "learning_rate": 7.76923076923077e-06,
3337
- "loss": 1.0638,
3338
- "step": 5510
3339
- },
3340
- {
3341
- "epoch": 3.72,
3342
- "learning_rate": 7.765182186234819e-06,
3343
- "loss": 1.0525,
3344
- "step": 5520
3345
- },
3346
- {
3347
- "epoch": 3.73,
3348
- "learning_rate": 7.761133603238867e-06,
3349
- "loss": 1.0638,
3350
- "step": 5530
3351
- },
3352
- {
3353
- "epoch": 3.74,
3354
- "learning_rate": 7.757085020242915e-06,
3355
- "loss": 1.0478,
3356
- "step": 5540
3357
- },
3358
- {
3359
- "epoch": 3.74,
3360
- "learning_rate": 7.753036437246964e-06,
3361
- "loss": 1.0429,
3362
- "step": 5550
3363
- },
3364
- {
3365
- "epoch": 3.75,
3366
- "learning_rate": 7.748987854251014e-06,
3367
- "loss": 1.0615,
3368
- "step": 5560
3369
- },
3370
- {
3371
- "epoch": 3.76,
3372
- "learning_rate": 7.74493927125506e-06,
3373
- "loss": 1.0653,
3374
- "step": 5570
3375
- },
3376
- {
3377
- "epoch": 3.76,
3378
- "learning_rate": 7.74089068825911e-06,
3379
- "loss": 1.0755,
3380
- "step": 5580
3381
- },
3382
- {
3383
- "epoch": 3.77,
3384
- "learning_rate": 7.736842105263158e-06,
3385
- "loss": 1.0447,
3386
- "step": 5590
3387
- },
3388
- {
3389
- "epoch": 3.78,
3390
- "learning_rate": 7.732793522267207e-06,
3391
- "loss": 1.0372,
3392
- "step": 5600
3393
- },
3394
- {
3395
- "epoch": 3.78,
3396
- "learning_rate": 7.728744939271255e-06,
3397
- "loss": 1.1158,
3398
- "step": 5610
3399
- },
3400
- {
3401
- "epoch": 3.79,
3402
- "learning_rate": 7.724696356275305e-06,
3403
- "loss": 1.1195,
3404
- "step": 5620
3405
- },
3406
- {
3407
- "epoch": 3.8,
3408
- "learning_rate": 7.720647773279351e-06,
3409
- "loss": 1.0246,
3410
- "step": 5630
3411
- },
3412
- {
3413
- "epoch": 3.8,
3414
- "learning_rate": 7.716599190283401e-06,
3415
- "loss": 1.0339,
3416
- "step": 5640
3417
- },
3418
- {
3419
- "epoch": 3.81,
3420
- "learning_rate": 7.712550607287449e-06,
3421
- "loss": 1.0578,
3422
- "step": 5650
3423
- },
3424
- {
3425
- "epoch": 3.82,
3426
- "learning_rate": 7.708502024291498e-06,
3427
- "loss": 1.0747,
3428
- "step": 5660
3429
- },
3430
- {
3431
- "epoch": 3.82,
3432
- "learning_rate": 7.704453441295546e-06,
3433
- "loss": 1.0301,
3434
- "step": 5670
3435
- },
3436
- {
3437
- "epoch": 3.83,
3438
- "learning_rate": 7.700404858299596e-06,
3439
- "loss": 0.9873,
3440
- "step": 5680
3441
- },
3442
- {
3443
- "epoch": 3.84,
3444
- "learning_rate": 7.696356275303644e-06,
3445
- "loss": 1.0217,
3446
- "step": 5690
3447
- },
3448
- {
3449
- "epoch": 3.84,
3450
- "learning_rate": 7.692307692307694e-06,
3451
- "loss": 1.0941,
3452
- "step": 5700
3453
- },
3454
- {
3455
- "epoch": 3.85,
3456
- "learning_rate": 7.688259109311742e-06,
3457
- "loss": 1.0842,
3458
- "step": 5710
3459
- },
3460
- {
3461
- "epoch": 3.86,
3462
- "learning_rate": 7.68421052631579e-06,
3463
- "loss": 1.0163,
3464
- "step": 5720
3465
- },
3466
- {
3467
- "epoch": 3.87,
3468
- "learning_rate": 7.680161943319839e-06,
3469
- "loss": 1.0387,
3470
- "step": 5730
3471
- },
3472
- {
3473
- "epoch": 3.87,
3474
- "learning_rate": 7.676113360323887e-06,
3475
- "loss": 1.0719,
3476
- "step": 5740
3477
- },
3478
- {
3479
- "epoch": 3.88,
3480
- "learning_rate": 7.672064777327935e-06,
3481
- "loss": 1.0495,
3482
- "step": 5750
3483
- },
3484
- {
3485
- "epoch": 3.89,
3486
- "learning_rate": 7.668016194331985e-06,
3487
- "loss": 1.0669,
3488
- "step": 5760
3489
- },
3490
- {
3491
- "epoch": 3.89,
3492
- "learning_rate": 7.663967611336033e-06,
3493
- "loss": 1.1226,
3494
- "step": 5770
3495
- },
3496
- {
3497
- "epoch": 3.9,
3498
- "learning_rate": 7.65991902834008e-06,
3499
- "loss": 1.0147,
3500
- "step": 5780
3501
- },
3502
- {
3503
- "epoch": 3.91,
3504
- "learning_rate": 7.65587044534413e-06,
3505
- "loss": 1.0244,
3506
- "step": 5790
3507
- },
3508
- {
3509
- "epoch": 3.91,
3510
- "learning_rate": 7.651821862348178e-06,
3511
- "loss": 1.0134,
3512
- "step": 5800
3513
- },
3514
- {
3515
- "epoch": 3.92,
3516
- "learning_rate": 7.647773279352228e-06,
3517
- "loss": 1.0215,
3518
- "step": 5810
3519
- },
3520
- {
3521
- "epoch": 3.93,
3522
- "learning_rate": 7.643724696356276e-06,
3523
- "loss": 1.1177,
3524
- "step": 5820
3525
- },
3526
- {
3527
- "epoch": 3.93,
3528
- "learning_rate": 7.639676113360325e-06,
3529
- "loss": 1.0472,
3530
- "step": 5830
3531
- },
3532
- {
3533
- "epoch": 3.94,
3534
- "learning_rate": 7.635627530364373e-06,
3535
- "loss": 1.0335,
3536
- "step": 5840
3537
- },
3538
- {
3539
- "epoch": 3.95,
3540
- "learning_rate": 7.631578947368423e-06,
3541
- "loss": 1.0346,
3542
- "step": 5850
3543
- },
3544
- {
3545
- "epoch": 3.95,
3546
- "learning_rate": 7.627530364372468e-06,
3547
- "loss": 1.0354,
3548
- "step": 5860
3549
- },
3550
- {
3551
- "epoch": 3.96,
3552
- "learning_rate": 7.623481781376519e-06,
3553
- "loss": 1.0663,
3554
- "step": 5870
3555
- },
3556
- {
3557
- "epoch": 3.97,
3558
- "learning_rate": 7.619433198380567e-06,
3559
- "loss": 1.0602,
3560
- "step": 5880
3561
- },
3562
- {
3563
- "epoch": 3.97,
3564
- "learning_rate": 7.615384615384615e-06,
3565
- "loss": 1.0458,
3566
- "step": 5890
3567
- },
3568
- {
3569
- "epoch": 3.98,
3570
- "learning_rate": 7.611336032388663e-06,
3571
- "loss": 1.079,
3572
- "step": 5900
3573
- },
3574
- {
3575
- "epoch": 3.99,
3576
- "learning_rate": 7.607287449392713e-06,
3577
- "loss": 1.0888,
3578
- "step": 5910
3579
- },
3580
- {
3581
- "epoch": 3.99,
3582
- "learning_rate": 7.60323886639676e-06,
3583
- "loss": 1.0395,
3584
- "step": 5920
3585
- },
3586
- {
3587
- "epoch": 4.0,
3588
- "learning_rate": 7.5991902834008105e-06,
3589
- "loss": 1.1178,
3590
- "step": 5930
3591
- },
3592
- {
3593
- "epoch": 4.0,
3594
- "eval_loss": 1.0923843383789062,
3595
- "eval_runtime": 10.6309,
3596
- "eval_samples_per_second": 103.472,
3597
- "eval_steps_per_second": 12.981,
3598
- "step": 5930
3599
- }
3600
- ],
3601
- "max_steps": 22230,
3602
- "num_train_epochs": 15,
3603
- "total_flos": 6.084121621168128e+18,
3604
- "trial_name": null,
3605
- "trial_params": null
3606
- }