Bhoomika33 commited on
Commit
f319b63
·
verified ·
1 Parent(s): 72afeca

Upload trainer_state.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. trainer_state.json +760 -0
trainer_state.json ADDED
@@ -0,0 +1,760 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.11637073463718609,
5
+ "global_step": 3100,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.0,
12
+ "learning_rate": 0.0002,
13
+ "loss": 3.5545,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.0,
18
+ "learning_rate": 0.0002,
19
+ "loss": 3.3808,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.0,
24
+ "learning_rate": 0.0002,
25
+ "loss": 3.2177,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.0,
30
+ "learning_rate": 0.0002,
31
+ "loss": 3.14,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.0,
36
+ "learning_rate": 0.0002,
37
+ "loss": 3.135,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.01,
42
+ "learning_rate": 0.0002,
43
+ "loss": 3.0592,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.01,
48
+ "learning_rate": 0.0002,
49
+ "loss": 3.083,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.01,
54
+ "learning_rate": 0.0002,
55
+ "loss": 3.0834,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.01,
60
+ "learning_rate": 0.0002,
61
+ "loss": 3.1127,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.01,
66
+ "learning_rate": 0.0002,
67
+ "loss": 2.9984,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.01,
72
+ "learning_rate": 0.0002,
73
+ "loss": 3.0828,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.01,
78
+ "learning_rate": 0.0002,
79
+ "loss": 2.9751,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 0.01,
84
+ "learning_rate": 0.0002,
85
+ "loss": 3.0941,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 0.01,
90
+ "learning_rate": 0.0002,
91
+ "loss": 2.9631,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 0.01,
96
+ "learning_rate": 0.0002,
97
+ "loss": 3.0859,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 0.02,
102
+ "learning_rate": 0.0002,
103
+ "loss": 2.9828,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 0.02,
108
+ "learning_rate": 0.0002,
109
+ "loss": 3.0143,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 0.02,
114
+ "learning_rate": 0.0002,
115
+ "loss": 2.9813,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 0.02,
120
+ "learning_rate": 0.0002,
121
+ "loss": 3.0787,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 0.02,
126
+ "learning_rate": 0.0002,
127
+ "loss": 3.0272,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 0.02,
132
+ "learning_rate": 0.0002,
133
+ "loss": 3.0479,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 0.02,
138
+ "learning_rate": 0.0002,
139
+ "loss": 2.9969,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 0.02,
144
+ "learning_rate": 0.0002,
145
+ "loss": 3.0361,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 0.02,
150
+ "learning_rate": 0.0002,
151
+ "loss": 2.9057,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 0.02,
156
+ "learning_rate": 0.0002,
157
+ "loss": 2.965,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 0.02,
162
+ "learning_rate": 0.0002,
163
+ "loss": 2.9288,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 0.03,
168
+ "learning_rate": 0.0002,
169
+ "loss": 2.9974,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 0.03,
174
+ "learning_rate": 0.0002,
175
+ "loss": 2.945,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 0.03,
180
+ "learning_rate": 0.0002,
181
+ "loss": 3.045,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 0.03,
186
+ "learning_rate": 0.0002,
187
+ "loss": 2.9497,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 0.03,
192
+ "learning_rate": 0.0002,
193
+ "loss": 3.011,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 0.03,
198
+ "learning_rate": 0.0002,
199
+ "loss": 2.9471,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 0.03,
204
+ "learning_rate": 0.0002,
205
+ "loss": 3.0071,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 0.03,
210
+ "learning_rate": 0.0002,
211
+ "loss": 2.8933,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 0.03,
216
+ "learning_rate": 0.0002,
217
+ "loss": 3.012,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 0.03,
222
+ "learning_rate": 0.0002,
223
+ "loss": 2.922,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 0.03,
228
+ "learning_rate": 0.0002,
229
+ "loss": 3.0512,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 0.04,
234
+ "learning_rate": 0.0002,
235
+ "loss": 2.9355,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 0.04,
240
+ "learning_rate": 0.0002,
241
+ "loss": 3.0477,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 0.04,
246
+ "learning_rate": 0.0002,
247
+ "loss": 2.9325,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 0.04,
252
+ "learning_rate": 0.0002,
253
+ "loss": 2.9595,
254
+ "step": 1025
255
+ },
256
+ {
257
+ "epoch": 0.04,
258
+ "learning_rate": 0.0002,
259
+ "loss": 2.8729,
260
+ "step": 1050
261
+ },
262
+ {
263
+ "epoch": 0.04,
264
+ "learning_rate": 0.0002,
265
+ "loss": 3.0294,
266
+ "step": 1075
267
+ },
268
+ {
269
+ "epoch": 0.04,
270
+ "learning_rate": 0.0002,
271
+ "loss": 2.9156,
272
+ "step": 1100
273
+ },
274
+ {
275
+ "epoch": 0.04,
276
+ "learning_rate": 0.0002,
277
+ "loss": 2.9877,
278
+ "step": 1125
279
+ },
280
+ {
281
+ "epoch": 0.04,
282
+ "learning_rate": 0.0002,
283
+ "loss": 2.9953,
284
+ "step": 1150
285
+ },
286
+ {
287
+ "epoch": 0.04,
288
+ "learning_rate": 0.0002,
289
+ "loss": 2.9605,
290
+ "step": 1175
291
+ },
292
+ {
293
+ "epoch": 0.05,
294
+ "learning_rate": 0.0002,
295
+ "loss": 2.8817,
296
+ "step": 1200
297
+ },
298
+ {
299
+ "epoch": 0.05,
300
+ "learning_rate": 0.0002,
301
+ "loss": 2.9458,
302
+ "step": 1225
303
+ },
304
+ {
305
+ "epoch": 0.05,
306
+ "learning_rate": 0.0002,
307
+ "loss": 2.8464,
308
+ "step": 1250
309
+ },
310
+ {
311
+ "epoch": 0.05,
312
+ "learning_rate": 0.0002,
313
+ "loss": 2.9767,
314
+ "step": 1275
315
+ },
316
+ {
317
+ "epoch": 0.05,
318
+ "learning_rate": 0.0002,
319
+ "loss": 2.9091,
320
+ "step": 1300
321
+ },
322
+ {
323
+ "epoch": 0.05,
324
+ "learning_rate": 0.0002,
325
+ "loss": 2.9615,
326
+ "step": 1325
327
+ },
328
+ {
329
+ "epoch": 0.05,
330
+ "learning_rate": 0.0002,
331
+ "loss": 2.8904,
332
+ "step": 1350
333
+ },
334
+ {
335
+ "epoch": 0.05,
336
+ "learning_rate": 0.0002,
337
+ "loss": 2.9142,
338
+ "step": 1375
339
+ },
340
+ {
341
+ "epoch": 0.05,
342
+ "learning_rate": 0.0002,
343
+ "loss": 2.8625,
344
+ "step": 1400
345
+ },
346
+ {
347
+ "epoch": 0.05,
348
+ "learning_rate": 0.0002,
349
+ "loss": 3.0317,
350
+ "step": 1425
351
+ },
352
+ {
353
+ "epoch": 0.05,
354
+ "learning_rate": 0.0002,
355
+ "loss": 2.8543,
356
+ "step": 1450
357
+ },
358
+ {
359
+ "epoch": 0.06,
360
+ "learning_rate": 0.0002,
361
+ "loss": 2.9401,
362
+ "step": 1475
363
+ },
364
+ {
365
+ "epoch": 0.06,
366
+ "learning_rate": 0.0002,
367
+ "loss": 2.8642,
368
+ "step": 1500
369
+ },
370
+ {
371
+ "epoch": 0.06,
372
+ "learning_rate": 0.0002,
373
+ "loss": 2.9703,
374
+ "step": 1525
375
+ },
376
+ {
377
+ "epoch": 0.06,
378
+ "learning_rate": 0.0002,
379
+ "loss": 2.8528,
380
+ "step": 1550
381
+ },
382
+ {
383
+ "epoch": 0.06,
384
+ "learning_rate": 0.0002,
385
+ "loss": 2.9115,
386
+ "step": 1575
387
+ },
388
+ {
389
+ "epoch": 0.06,
390
+ "learning_rate": 0.0002,
391
+ "loss": 2.773,
392
+ "step": 1600
393
+ },
394
+ {
395
+ "epoch": 0.06,
396
+ "learning_rate": 0.0002,
397
+ "loss": 2.9731,
398
+ "step": 1625
399
+ },
400
+ {
401
+ "epoch": 0.06,
402
+ "learning_rate": 0.0002,
403
+ "loss": 2.8812,
404
+ "step": 1650
405
+ },
406
+ {
407
+ "epoch": 0.06,
408
+ "learning_rate": 0.0002,
409
+ "loss": 2.9683,
410
+ "step": 1675
411
+ },
412
+ {
413
+ "epoch": 0.06,
414
+ "learning_rate": 0.0002,
415
+ "loss": 2.8224,
416
+ "step": 1700
417
+ },
418
+ {
419
+ "epoch": 0.06,
420
+ "learning_rate": 0.0002,
421
+ "loss": 2.9658,
422
+ "step": 1725
423
+ },
424
+ {
425
+ "epoch": 0.07,
426
+ "learning_rate": 0.0002,
427
+ "loss": 2.8459,
428
+ "step": 1750
429
+ },
430
+ {
431
+ "epoch": 0.07,
432
+ "learning_rate": 0.0002,
433
+ "loss": 2.9334,
434
+ "step": 1775
435
+ },
436
+ {
437
+ "epoch": 0.07,
438
+ "learning_rate": 0.0002,
439
+ "loss": 2.8489,
440
+ "step": 1800
441
+ },
442
+ {
443
+ "epoch": 0.07,
444
+ "learning_rate": 0.0002,
445
+ "loss": 2.9709,
446
+ "step": 1825
447
+ },
448
+ {
449
+ "epoch": 0.07,
450
+ "learning_rate": 0.0002,
451
+ "loss": 2.8089,
452
+ "step": 1850
453
+ },
454
+ {
455
+ "epoch": 0.07,
456
+ "learning_rate": 0.0002,
457
+ "loss": 2.9309,
458
+ "step": 1875
459
+ },
460
+ {
461
+ "epoch": 0.07,
462
+ "learning_rate": 0.0002,
463
+ "loss": 2.7292,
464
+ "step": 1900
465
+ },
466
+ {
467
+ "epoch": 0.07,
468
+ "learning_rate": 0.0002,
469
+ "loss": 2.8867,
470
+ "step": 1925
471
+ },
472
+ {
473
+ "epoch": 0.07,
474
+ "learning_rate": 0.0002,
475
+ "loss": 2.8445,
476
+ "step": 1950
477
+ },
478
+ {
479
+ "epoch": 0.07,
480
+ "learning_rate": 0.0002,
481
+ "loss": 2.9522,
482
+ "step": 1975
483
+ },
484
+ {
485
+ "epoch": 0.08,
486
+ "learning_rate": 0.0002,
487
+ "loss": 2.8645,
488
+ "step": 2000
489
+ },
490
+ {
491
+ "epoch": 0.08,
492
+ "learning_rate": 0.0002,
493
+ "loss": 2.9292,
494
+ "step": 2025
495
+ },
496
+ {
497
+ "epoch": 0.08,
498
+ "learning_rate": 0.0002,
499
+ "loss": 2.8197,
500
+ "step": 2050
501
+ },
502
+ {
503
+ "epoch": 0.08,
504
+ "learning_rate": 0.0002,
505
+ "loss": 2.9131,
506
+ "step": 2075
507
+ },
508
+ {
509
+ "epoch": 0.08,
510
+ "learning_rate": 0.0002,
511
+ "loss": 2.7743,
512
+ "step": 2100
513
+ },
514
+ {
515
+ "epoch": 0.08,
516
+ "learning_rate": 0.0002,
517
+ "loss": 2.8766,
518
+ "step": 2125
519
+ },
520
+ {
521
+ "epoch": 0.08,
522
+ "learning_rate": 0.0002,
523
+ "loss": 2.8322,
524
+ "step": 2150
525
+ },
526
+ {
527
+ "epoch": 0.08,
528
+ "learning_rate": 0.0002,
529
+ "loss": 2.8837,
530
+ "step": 2175
531
+ },
532
+ {
533
+ "epoch": 0.08,
534
+ "learning_rate": 0.0002,
535
+ "loss": 2.8804,
536
+ "step": 2200
537
+ },
538
+ {
539
+ "epoch": 0.08,
540
+ "learning_rate": 0.0002,
541
+ "loss": 2.9015,
542
+ "step": 2225
543
+ },
544
+ {
545
+ "epoch": 0.08,
546
+ "learning_rate": 0.0002,
547
+ "loss": 2.8675,
548
+ "step": 2250
549
+ },
550
+ {
551
+ "epoch": 0.09,
552
+ "learning_rate": 0.0002,
553
+ "loss": 2.8653,
554
+ "step": 2275
555
+ },
556
+ {
557
+ "epoch": 0.09,
558
+ "learning_rate": 0.0002,
559
+ "loss": 2.7899,
560
+ "step": 2300
561
+ },
562
+ {
563
+ "epoch": 0.09,
564
+ "learning_rate": 0.0002,
565
+ "loss": 2.9419,
566
+ "step": 2325
567
+ },
568
+ {
569
+ "epoch": 0.09,
570
+ "learning_rate": 0.0002,
571
+ "loss": 2.8745,
572
+ "step": 2350
573
+ },
574
+ {
575
+ "epoch": 0.09,
576
+ "learning_rate": 0.0002,
577
+ "loss": 2.939,
578
+ "step": 2375
579
+ },
580
+ {
581
+ "epoch": 0.09,
582
+ "learning_rate": 0.0002,
583
+ "loss": 2.8301,
584
+ "step": 2400
585
+ },
586
+ {
587
+ "epoch": 0.09,
588
+ "learning_rate": 0.0002,
589
+ "loss": 2.9117,
590
+ "step": 2425
591
+ },
592
+ {
593
+ "epoch": 0.09,
594
+ "learning_rate": 0.0002,
595
+ "loss": 2.8191,
596
+ "step": 2450
597
+ },
598
+ {
599
+ "epoch": 0.09,
600
+ "learning_rate": 0.0002,
601
+ "loss": 2.8762,
602
+ "step": 2475
603
+ },
604
+ {
605
+ "epoch": 0.09,
606
+ "learning_rate": 0.0002,
607
+ "loss": 2.8237,
608
+ "step": 2500
609
+ },
610
+ {
611
+ "epoch": 0.09,
612
+ "learning_rate": 0.0002,
613
+ "loss": 2.8568,
614
+ "step": 2525
615
+ },
616
+ {
617
+ "epoch": 0.1,
618
+ "learning_rate": 0.0002,
619
+ "loss": 2.8007,
620
+ "step": 2550
621
+ },
622
+ {
623
+ "epoch": 0.1,
624
+ "learning_rate": 0.0002,
625
+ "loss": 2.917,
626
+ "step": 2575
627
+ },
628
+ {
629
+ "epoch": 0.1,
630
+ "learning_rate": 0.0002,
631
+ "loss": 2.8378,
632
+ "step": 2600
633
+ },
634
+ {
635
+ "epoch": 0.1,
636
+ "learning_rate": 0.0002,
637
+ "loss": 2.931,
638
+ "step": 2625
639
+ },
640
+ {
641
+ "epoch": 0.1,
642
+ "learning_rate": 0.0002,
643
+ "loss": 2.7476,
644
+ "step": 2650
645
+ },
646
+ {
647
+ "epoch": 0.1,
648
+ "learning_rate": 0.0002,
649
+ "loss": 2.8783,
650
+ "step": 2675
651
+ },
652
+ {
653
+ "epoch": 0.1,
654
+ "learning_rate": 0.0002,
655
+ "loss": 2.7566,
656
+ "step": 2700
657
+ },
658
+ {
659
+ "epoch": 0.1,
660
+ "learning_rate": 0.0002,
661
+ "loss": 2.8552,
662
+ "step": 2725
663
+ },
664
+ {
665
+ "epoch": 0.1,
666
+ "learning_rate": 0.0002,
667
+ "loss": 2.762,
668
+ "step": 2750
669
+ },
670
+ {
671
+ "epoch": 0.1,
672
+ "learning_rate": 0.0002,
673
+ "loss": 2.8504,
674
+ "step": 2775
675
+ },
676
+ {
677
+ "epoch": 0.11,
678
+ "learning_rate": 0.0002,
679
+ "loss": 2.7906,
680
+ "step": 2800
681
+ },
682
+ {
683
+ "epoch": 0.11,
684
+ "learning_rate": 0.0002,
685
+ "loss": 2.9305,
686
+ "step": 2825
687
+ },
688
+ {
689
+ "epoch": 0.11,
690
+ "learning_rate": 0.0002,
691
+ "loss": 2.7938,
692
+ "step": 2850
693
+ },
694
+ {
695
+ "epoch": 0.11,
696
+ "learning_rate": 0.0002,
697
+ "loss": 2.8926,
698
+ "step": 2875
699
+ },
700
+ {
701
+ "epoch": 0.11,
702
+ "learning_rate": 0.0002,
703
+ "loss": 2.8061,
704
+ "step": 2900
705
+ },
706
+ {
707
+ "epoch": 0.11,
708
+ "learning_rate": 0.0002,
709
+ "loss": 2.8761,
710
+ "step": 2925
711
+ },
712
+ {
713
+ "epoch": 0.11,
714
+ "learning_rate": 0.0002,
715
+ "loss": 2.7909,
716
+ "step": 2950
717
+ },
718
+ {
719
+ "epoch": 0.11,
720
+ "learning_rate": 0.0002,
721
+ "loss": 2.8898,
722
+ "step": 2975
723
+ },
724
+ {
725
+ "epoch": 0.11,
726
+ "learning_rate": 0.0002,
727
+ "loss": 2.7425,
728
+ "step": 3000
729
+ },
730
+ {
731
+ "epoch": 0.11,
732
+ "learning_rate": 0.0002,
733
+ "loss": 2.9364,
734
+ "step": 3025
735
+ },
736
+ {
737
+ "epoch": 0.11,
738
+ "learning_rate": 0.0002,
739
+ "loss": 2.7867,
740
+ "step": 3050
741
+ },
742
+ {
743
+ "epoch": 0.12,
744
+ "learning_rate": 0.0002,
745
+ "loss": 2.8904,
746
+ "step": 3075
747
+ },
748
+ {
749
+ "epoch": 0.12,
750
+ "learning_rate": 0.0002,
751
+ "loss": 2.8158,
752
+ "step": 3100
753
+ }
754
+ ],
755
+ "max_steps": 26639,
756
+ "num_train_epochs": 1,
757
+ "total_flos": 1.183808995344384e+16,
758
+ "trial_name": null,
759
+ "trial_params": null
760
+ }