Hugging Face's logo

gary109
/
wav2vec2-base-mirst500

Audio Classification
Transformers
PyTorch
TensorBoard
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
gary109 commited on
Commit
1440fd4
•
1 Parent(s): 5c830e4

End of training

Browse files
Files changed (7) hide show
  1. all_results.json +10 -10
  2. eval_results.json +6 -6
  3. pytorch_model.bin +1 -1
  4. runs/Apr12_04-18-26_e9bbb09c2704/events.out.tfevents.1649737182.e9bbb09c2704.24461.0 +2 -2
  5. runs/Apr12_04-18-26_e9bbb09c2704/events.out.tfevents.1649742227.e9bbb09c2704.24461.2 +3 -0
  6. train_results.json +5 -5
  7. trainer_state.json +3446 -278
all_results.json CHANGED
@@ -1,12 +1,12 @@
1
  {
2
- "epoch": 1.0,
3
- "eval_accuracy": 0.5816544091224876,
4
- "eval_loss": 1.1218500137329102,
5
- "eval_runtime": 299.459,
6
- "eval_samples_per_second": 139.395,
7
- "eval_steps_per_second": 69.699,
8
- "train_loss": 1.4751509523099184,
9
- "train_runtime": 962.063,
10
- "train_samples_per_second": 173.553,
11
- "train_steps_per_second": 1.355
12
  }
 
1
  {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.7016984883693075,
4
+ "eval_loss": 0.8677756190299988,
5
+ "eval_runtime": 297.1896,
6
+ "eval_samples_per_second": 140.459,
7
+ "eval_steps_per_second": 70.231,
8
+ "train_loss": 1.11384382686732,
9
+ "train_runtime": 4746.5723,
10
+ "train_samples_per_second": 175.884,
11
+ "train_steps_per_second": 1.374
12
  }
eval_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
- "epoch": 1.0,
3
- "eval_accuracy": 0.5816544091224876,
4
- "eval_loss": 1.1218500137329102,
5
- "eval_runtime": 299.459,
6
- "eval_samples_per_second": 139.395,
7
- "eval_steps_per_second": 69.699
8
  }
 
1
  {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.7016984883693075,
4
+ "eval_loss": 0.8677756190299988,
5
+ "eval_runtime": 297.1896,
6
+ "eval_samples_per_second": 140.459,
7
+ "eval_steps_per_second": 70.231
8
  }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9a55a4140e888ef84a55a7e903b79588dc2fd6f39cd620e7a65a5e302999d480
3
  size 378367891
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:85a3db02d061d77926da04f24fd89df13263ebf0d7f0a042788a7023ff9ccf10
3
  size 378367891
runs/Apr12_04-18-26_e9bbb09c2704/events.out.tfevents.1649737182.e9bbb09c2704.24461.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:56865c96948a55f600e48b71cbe894c4ee701533944ab0dca3e13e057a351da2
3
- size 26009
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:002ab880f320d71df3b9ad6d2af74685e89ea723ab9db9e01d67b0254ab32346
3
+ size 109609
runs/Apr12_04-18-26_e9bbb09c2704/events.out.tfevents.1649742227.e9bbb09c2704.24461.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6510c8597cf87ca808b8ae2b657dd92ca071ac0860043f1edea4b5de6937faec
3
+ size 363
train_results.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "epoch": 1.0,
3
- "train_loss": 1.4751509523099184,
4
- "train_runtime": 962.063,
5
- "train_samples_per_second": 173.553,
6
- "train_steps_per_second": 1.355
7
  }
 
1
  {
2
+ "epoch": 5.0,
3
+ "train_loss": 1.11384382686732,
4
+ "train_runtime": 4746.5723,
5
+ "train_samples_per_second": 175.884,
6
+ "train_steps_per_second": 1.374
7
  }
trainer_state.json CHANGED
@@ -1,814 +1,3982 @@
1
  {
2
- "best_metric": 0.5816544091224876,
3
- "best_model_checkpoint": "wav2vec2-base-mirst500/checkpoint-1304",
4
- "epoch": 0.9996167113836719,
5
- "global_step": 1304,
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": 2.2900763358778625e-06,
13
- "loss": 2.082,
14
  "step": 10
15
  },
16
  {
17
  "epoch": 0.02,
18
- "learning_rate": 4.580152671755725e-06,
19
- "loss": 2.0799,
20
  "step": 20
21
  },
22
  {
23
  "epoch": 0.02,
24
- "learning_rate": 6.870229007633588e-06,
25
- "loss": 2.0746,
26
  "step": 30
27
  },
28
  {
29
  "epoch": 0.03,
30
- "learning_rate": 9.16030534351145e-06,
31
- "loss": 2.0638,
32
  "step": 40
33
  },
34
  {
35
  "epoch": 0.04,
36
- "learning_rate": 1.1450381679389314e-05,
37
- "loss": 2.0536,
38
  "step": 50
39
  },
40
  {
41
  "epoch": 0.05,
42
- "learning_rate": 1.3740458015267176e-05,
43
- "loss": 2.0224,
44
  "step": 60
45
  },
46
  {
47
  "epoch": 0.05,
48
- "learning_rate": 1.6030534351145038e-05,
49
- "loss": 2.0083,
50
  "step": 70
51
  },
52
  {
53
  "epoch": 0.06,
54
- "learning_rate": 1.83206106870229e-05,
55
- "loss": 1.9666,
56
  "step": 80
57
  },
58
  {
59
  "epoch": 0.07,
60
- "learning_rate": 2.038167938931298e-05,
61
- "loss": 1.9604,
62
  "step": 90
63
  },
64
  {
65
  "epoch": 0.08,
66
- "learning_rate": 2.267175572519084e-05,
67
- "loss": 1.9438,
68
  "step": 100
69
  },
70
  {
71
  "epoch": 0.08,
72
- "learning_rate": 2.4961832061068703e-05,
73
- "loss": 1.951,
74
  "step": 110
75
  },
76
  {
77
  "epoch": 0.09,
78
- "learning_rate": 2.7251908396946566e-05,
79
- "loss": 1.9237,
80
  "step": 120
81
  },
82
  {
83
  "epoch": 0.1,
84
- "learning_rate": 2.9541984732824428e-05,
85
- "loss": 1.9137,
86
  "step": 130
87
  },
88
  {
89
  "epoch": 0.11,
90
- "learning_rate": 2.979539641943734e-05,
91
- "loss": 1.9285,
92
  "step": 140
93
  },
94
  {
95
  "epoch": 0.11,
96
- "learning_rate": 2.9539641943734016e-05,
97
- "loss": 1.8948,
98
  "step": 150
99
  },
100
  {
101
  "epoch": 0.12,
102
- "learning_rate": 2.928388746803069e-05,
103
- "loss": 1.9013,
104
  "step": 160
105
  },
106
  {
107
  "epoch": 0.13,
108
- "learning_rate": 2.9028132992327367e-05,
109
- "loss": 1.9093,
110
  "step": 170
111
  },
112
  {
113
  "epoch": 0.14,
114
- "learning_rate": 2.8772378516624042e-05,
115
- "loss": 1.8813,
116
  "step": 180
117
  },
118
  {
119
  "epoch": 0.15,
120
- "learning_rate": 2.8516624040920717e-05,
121
- "loss": 1.8655,
122
  "step": 190
123
  },
124
  {
125
  "epoch": 0.15,
126
- "learning_rate": 2.8260869565217392e-05,
127
- "loss": 1.8605,
128
  "step": 200
129
  },
130
  {
131
  "epoch": 0.16,
132
- "learning_rate": 2.8005115089514067e-05,
133
- "loss": 1.8136,
134
  "step": 210
135
  },
136
  {
137
  "epoch": 0.17,
138
- "learning_rate": 2.7749360613810743e-05,
139
- "loss": 1.8576,
140
  "step": 220
141
  },
142
  {
143
  "epoch": 0.18,
144
- "learning_rate": 2.7493606138107418e-05,
145
- "loss": 1.7765,
146
  "step": 230
147
  },
148
  {
149
  "epoch": 0.18,
150
- "learning_rate": 2.7237851662404093e-05,
151
- "loss": 1.7673,
152
  "step": 240
153
  },
154
  {
155
  "epoch": 0.19,
156
- "learning_rate": 2.6982097186700768e-05,
157
- "loss": 1.7767,
158
  "step": 250
159
  },
160
  {
161
  "epoch": 0.2,
162
- "learning_rate": 2.6726342710997444e-05,
163
- "loss": 1.789,
164
  "step": 260
165
  },
166
  {
167
  "epoch": 0.21,
168
- "learning_rate": 2.647058823529412e-05,
169
- "loss": 1.7593,
170
  "step": 270
171
  },
172
  {
173
  "epoch": 0.21,
174
- "learning_rate": 2.6214833759590794e-05,
175
- "loss": 1.7742,
176
  "step": 280
177
  },
178
  {
179
  "epoch": 0.22,
180
- "learning_rate": 2.595907928388747e-05,
181
- "loss": 1.7237,
182
  "step": 290
183
  },
184
  {
185
  "epoch": 0.23,
186
- "learning_rate": 2.5703324808184144e-05,
187
- "loss": 1.7631,
188
  "step": 300
189
  },
190
  {
191
  "epoch": 0.24,
192
- "learning_rate": 2.544757033248082e-05,
193
- "loss": 1.7139,
194
  "step": 310
195
  },
196
  {
197
  "epoch": 0.25,
198
- "learning_rate": 2.5191815856777495e-05,
199
- "loss": 1.7306,
200
  "step": 320
201
  },
202
  {
203
  "epoch": 0.25,
204
- "learning_rate": 2.493606138107417e-05,
205
- "loss": 1.6697,
206
  "step": 330
207
  },
208
  {
209
  "epoch": 0.26,
210
- "learning_rate": 2.4680306905370845e-05,
211
- "loss": 1.7109,
212
  "step": 340
213
  },
214
  {
215
  "epoch": 0.27,
216
- "learning_rate": 2.442455242966752e-05,
217
- "loss": 1.6932,
218
  "step": 350
219
  },
220
  {
221
  "epoch": 0.28,
222
- "learning_rate": 2.4168797953964196e-05,
223
- "loss": 1.6518,
224
  "step": 360
225
  },
226
  {
227
  "epoch": 0.28,
228
- "learning_rate": 2.391304347826087e-05,
229
- "loss": 1.6458,
230
  "step": 370
231
  },
232
  {
233
  "epoch": 0.29,
234
- "learning_rate": 2.3657289002557546e-05,
235
- "loss": 1.6027,
236
  "step": 380
237
  },
238
  {
239
  "epoch": 0.3,
240
- "learning_rate": 2.340153452685422e-05,
241
- "loss": 1.611,
242
  "step": 390
243
  },
244
  {
245
  "epoch": 0.31,
246
- "learning_rate": 2.3145780051150897e-05,
247
- "loss": 1.6442,
248
  "step": 400
249
  },
250
  {
251
  "epoch": 0.31,
252
- "learning_rate": 2.2890025575447572e-05,
253
- "loss": 1.5848,
254
  "step": 410
255
  },
256
  {
257
  "epoch": 0.32,
258
- "learning_rate": 2.2634271099744247e-05,
259
- "loss": 1.5634,
260
  "step": 420
261
  },
262
  {
263
  "epoch": 0.33,
264
- "learning_rate": 2.2378516624040922e-05,
265
- "loss": 1.5943,
266
  "step": 430
267
  },
268
  {
269
  "epoch": 0.34,
270
- "learning_rate": 2.2122762148337597e-05,
271
- "loss": 1.5414,
272
  "step": 440
273
  },
274
  {
275
  "epoch": 0.34,
276
- "learning_rate": 2.1867007672634273e-05,
277
- "loss": 1.5293,
278
  "step": 450
279
  },
280
  {
281
  "epoch": 0.35,
282
- "learning_rate": 2.1611253196930948e-05,
283
- "loss": 1.5084,
284
  "step": 460
285
  },
286
  {
287
  "epoch": 0.36,
288
- "learning_rate": 2.1355498721227623e-05,
289
- "loss": 1.5021,
290
  "step": 470
291
  },
292
  {
293
  "epoch": 0.37,
294
- "learning_rate": 2.1099744245524298e-05,
295
- "loss": 1.4932,
296
  "step": 480
297
  },
298
  {
299
  "epoch": 0.38,
300
- "learning_rate": 2.0843989769820973e-05,
301
- "loss": 1.4269,
302
  "step": 490
303
  },
304
  {
305
  "epoch": 0.38,
306
- "learning_rate": 2.058823529411765e-05,
307
- "loss": 1.5041,
308
  "step": 500
309
  },
310
  {
311
  "epoch": 0.39,
312
- "learning_rate": 2.0332480818414324e-05,
313
- "loss": 1.4384,
314
  "step": 510
315
  },
316
  {
317
  "epoch": 0.4,
318
- "learning_rate": 2.0076726342711e-05,
319
- "loss": 1.4581,
320
  "step": 520
321
  },
322
  {
323
  "epoch": 0.41,
324
- "learning_rate": 1.9820971867007674e-05,
325
- "loss": 1.4105,
326
  "step": 530
327
  },
328
  {
329
  "epoch": 0.41,
330
- "learning_rate": 1.956521739130435e-05,
331
- "loss": 1.4033,
332
  "step": 540
333
  },
334
  {
335
  "epoch": 0.42,
336
- "learning_rate": 1.9309462915601025e-05,
337
- "loss": 1.4241,
338
  "step": 550
339
  },
340
  {
341
  "epoch": 0.43,
342
- "learning_rate": 1.90537084398977e-05,
343
- "loss": 1.4365,
344
  "step": 560
345
  },
346
  {
347
  "epoch": 0.44,
348
- "learning_rate": 1.8797953964194375e-05,
349
- "loss": 1.4566,
350
  "step": 570
351
  },
352
  {
353
  "epoch": 0.44,
354
- "learning_rate": 1.854219948849105e-05,
355
- "loss": 1.3788,
356
  "step": 580
357
  },
358
  {
359
  "epoch": 0.45,
360
- "learning_rate": 1.8286445012787726e-05,
361
- "loss": 1.4141,
362
  "step": 590
363
  },
364
  {
365
  "epoch": 0.46,
366
- "learning_rate": 1.80306905370844e-05,
367
- "loss": 1.3489,
368
  "step": 600
369
  },
370
  {
371
  "epoch": 0.47,
372
- "learning_rate": 1.7774936061381076e-05,
373
- "loss": 1.3909,
374
  "step": 610
375
  },
376
  {
377
  "epoch": 0.48,
378
- "learning_rate": 1.751918158567775e-05,
379
- "loss": 1.3576,
380
  "step": 620
381
  },
382
  {
383
  "epoch": 0.48,
384
- "learning_rate": 1.7263427109974426e-05,
385
- "loss": 1.3591,
386
  "step": 630
387
  },
388
  {
389
  "epoch": 0.49,
390
- "learning_rate": 1.70076726342711e-05,
391
- "loss": 1.3649,
392
  "step": 640
393
  },
394
  {
395
  "epoch": 0.5,
396
- "learning_rate": 1.6751918158567777e-05,
397
- "loss": 1.3327,
398
  "step": 650
399
  },
400
  {
401
  "epoch": 0.51,
402
- "learning_rate": 1.6496163682864452e-05,
403
- "loss": 1.3663,
404
  "step": 660
405
  },
406
  {
407
  "epoch": 0.51,
408
- "learning_rate": 1.6240409207161127e-05,
409
- "loss": 1.3544,
410
  "step": 670
411
  },
412
  {
413
  "epoch": 0.52,
414
- "learning_rate": 1.5984654731457802e-05,
415
- "loss": 1.3469,
416
  "step": 680
417
  },
418
  {
419
  "epoch": 0.53,
420
- "learning_rate": 1.5728900255754478e-05,
421
- "loss": 1.3086,
422
  "step": 690
423
  },
424
  {
425
  "epoch": 0.54,
426
- "learning_rate": 1.5473145780051153e-05,
427
- "loss": 1.3193,
428
  "step": 700
429
  },
430
  {
431
  "epoch": 0.54,
432
- "learning_rate": 1.5217391304347828e-05,
433
- "loss": 1.3347,
434
  "step": 710
435
  },
436
  {
437
  "epoch": 0.55,
438
- "learning_rate": 1.4961636828644502e-05,
439
- "loss": 1.346,
440
  "step": 720
441
  },
442
  {
443
  "epoch": 0.56,
444
- "learning_rate": 1.4705882352941177e-05,
445
- "loss": 1.2925,
446
  "step": 730
447
  },
448
  {
449
  "epoch": 0.57,
450
- "learning_rate": 1.4450127877237852e-05,
451
- "loss": 1.308,
452
  "step": 740
453
  },
454
  {
455
  "epoch": 0.57,
456
- "learning_rate": 1.4194373401534527e-05,
457
- "loss": 1.286,
458
  "step": 750
459
  },
460
  {
461
  "epoch": 0.58,
462
- "learning_rate": 1.3938618925831203e-05,
463
- "loss": 1.3567,
464
  "step": 760
465
  },
466
  {
467
  "epoch": 0.59,
468
- "learning_rate": 1.3682864450127878e-05,
469
- "loss": 1.2899,
470
  "step": 770
471
  },
472
  {
473
  "epoch": 0.6,
474
- "learning_rate": 1.3427109974424553e-05,
475
- "loss": 1.3039,
476
  "step": 780
477
  },
478
  {
479
  "epoch": 0.61,
480
- "learning_rate": 1.3171355498721228e-05,
481
- "loss": 1.2917,
482
  "step": 790
483
  },
484
  {
485
  "epoch": 0.61,
486
- "learning_rate": 1.2915601023017903e-05,
487
- "loss": 1.3053,
488
  "step": 800
489
  },
490
  {
491
  "epoch": 0.62,
492
- "learning_rate": 1.2659846547314579e-05,
493
- "loss": 1.3558,
494
  "step": 810
495
  },
496
  {
497
  "epoch": 0.63,
498
- "learning_rate": 1.2404092071611254e-05,
499
- "loss": 1.27,
500
  "step": 820
501
  },
502
  {
503
  "epoch": 0.64,
504
- "learning_rate": 1.2148337595907929e-05,
505
- "loss": 1.278,
506
  "step": 830
507
  },
508
  {
509
  "epoch": 0.64,
510
- "learning_rate": 1.1892583120204604e-05,
511
- "loss": 1.2846,
512
  "step": 840
513
  },
514
  {
515
  "epoch": 0.65,
516
- "learning_rate": 1.163682864450128e-05,
517
- "loss": 1.301,
518
  "step": 850
519
  },
520
  {
521
  "epoch": 0.66,
522
- "learning_rate": 1.1381074168797955e-05,
523
- "loss": 1.2658,
524
  "step": 860
525
  },
526
  {
527
  "epoch": 0.67,
528
- "learning_rate": 1.112531969309463e-05,
529
- "loss": 1.3158,
530
  "step": 870
531
  },
532
  {
533
  "epoch": 0.67,
534
- "learning_rate": 1.0869565217391305e-05,
535
- "loss": 1.2507,
536
  "step": 880
537
  },
538
  {
539
  "epoch": 0.68,
540
- "learning_rate": 1.061381074168798e-05,
541
- "loss": 1.3014,
542
  "step": 890
543
  },
544
  {
545
  "epoch": 0.69,
546
- "learning_rate": 1.0358056265984656e-05,
547
- "loss": 1.2537,
548
  "step": 900
549
  },
550
  {
551
  "epoch": 0.7,
552
- "learning_rate": 1.010230179028133e-05,
553
- "loss": 1.3258,
554
  "step": 910
555
  },
556
  {
557
  "epoch": 0.71,
558
- "learning_rate": 9.846547314578006e-06,
559
- "loss": 1.2994,
560
  "step": 920
561
  },
562
  {
563
  "epoch": 0.71,
564
- "learning_rate": 9.590792838874681e-06,
565
- "loss": 1.2522,
566
  "step": 930
567
  },
568
  {
569
  "epoch": 0.72,
570
- "learning_rate": 9.335038363171356e-06,
571
- "loss": 1.2733,
572
  "step": 940
573
  },
574
  {
575
  "epoch": 0.73,
576
- "learning_rate": 9.079283887468032e-06,
577
- "loss": 1.2925,
578
  "step": 950
579
  },
580
  {
581
  "epoch": 0.74,
582
- "learning_rate": 8.823529411764707e-06,
583
- "loss": 1.2794,
584
  "step": 960
585
  },
586
  {
587
  "epoch": 0.74,
588
- "learning_rate": 8.567774936061382e-06,
589
- "loss": 1.2821,
590
  "step": 970
591
  },
592
  {
593
  "epoch": 0.75,
594
- "learning_rate": 8.312020460358057e-06,
595
- "loss": 1.2203,
596
  "step": 980
597
  },
598
  {
599
  "epoch": 0.76,
600
- "learning_rate": 8.056265984654732e-06,
601
- "loss": 1.2205,
602
  "step": 990
603
  },
604
  {
605
  "epoch": 0.77,
606
- "learning_rate": 7.800511508951408e-06,
607
- "loss": 1.2216,
608
  "step": 1000
609
  },
610
  {
611
  "epoch": 0.77,
612
- "learning_rate": 7.544757033248083e-06,
613
- "loss": 1.2165,
614
  "step": 1010
615
  },
616
  {
617
  "epoch": 0.78,
618
- "learning_rate": 7.289002557544757e-06,
619
- "loss": 1.2198,
620
  "step": 1020
621
  },
622
  {
623
  "epoch": 0.79,
624
- "learning_rate": 7.0332480818414325e-06,
625
- "loss": 1.2465,
626
  "step": 1030
627
  },
628
  {
629
  "epoch": 0.8,
630
- "learning_rate": 6.777493606138108e-06,
631
- "loss": 1.1886,
632
  "step": 1040
633
  },
634
  {
635
  "epoch": 0.8,
636
- "learning_rate": 6.521739130434783e-06,
637
- "loss": 1.2223,
638
  "step": 1050
639
  },
640
  {
641
  "epoch": 0.81,
642
- "learning_rate": 6.265984654731458e-06,
643
- "loss": 1.2033,
644
  "step": 1060
645
  },
646
  {
647
  "epoch": 0.82,
648
- "learning_rate": 6.010230179028133e-06,
649
- "loss": 1.229,
650
  "step": 1070
651
  },
652
  {
653
  "epoch": 0.83,
654
- "learning_rate": 5.7544757033248085e-06,
655
- "loss": 1.205,
656
  "step": 1080
657
  },
658
  {
659
  "epoch": 0.84,
660
- "learning_rate": 5.498721227621484e-06,
661
- "loss": 1.204,
662
  "step": 1090
663
  },
664
  {
665
  "epoch": 0.84,
666
- "learning_rate": 5.242966751918159e-06,
667
- "loss": 1.2272,
668
  "step": 1100
669
  },
670
  {
671
  "epoch": 0.85,
672
- "learning_rate": 4.987212276214834e-06,
673
- "loss": 1.2212,
674
  "step": 1110
675
  },
676
  {
677
  "epoch": 0.86,
678
- "learning_rate": 4.731457800511509e-06,
679
- "loss": 1.245,
680
  "step": 1120
681
  },
682
  {
683
  "epoch": 0.87,
684
- "learning_rate": 4.475703324808185e-06,
685
- "loss": 1.211,
686
  "step": 1130
687
  },
688
  {
689
  "epoch": 0.87,
690
- "learning_rate": 4.21994884910486e-06,
691
- "loss": 1.2395,
692
  "step": 1140
693
  },
694
  {
695
  "epoch": 0.88,
696
- "learning_rate": 3.964194373401535e-06,
697
- "loss": 1.2348,
698
  "step": 1150
699
  },
700
  {
701
  "epoch": 0.89,
702
- "learning_rate": 3.70843989769821e-06,
703
- "loss": 1.1843,
704
  "step": 1160
705
  },
706
  {
707
  "epoch": 0.9,
708
- "learning_rate": 3.452685421994885e-06,
709
- "loss": 1.1898,
710
  "step": 1170
711
  },
712
  {
713
  "epoch": 0.9,
714
- "learning_rate": 3.1969309462915602e-06,
715
- "loss": 1.1997,
716
  "step": 1180
717
  },
718
  {
719
  "epoch": 0.91,
720
- "learning_rate": 2.9411764705882355e-06,
721
- "loss": 1.1799,
722
  "step": 1190
723
  },
724
  {
725
  "epoch": 0.92,
726
- "learning_rate": 2.6854219948849107e-06,
727
- "loss": 1.1979,
728
  "step": 1200
729
  },
730
  {
731
  "epoch": 0.93,
732
- "learning_rate": 2.429667519181586e-06,
733
- "loss": 1.251,
734
  "step": 1210
735
  },
736
  {
737
  "epoch": 0.94,
738
- "learning_rate": 2.173913043478261e-06,
739
- "loss": 1.1851,
740
  "step": 1220
741
  },
742
  {
743
  "epoch": 0.94,
744
- "learning_rate": 1.9181585677749363e-06,
745
- "loss": 1.2571,
746
  "step": 1230
747
  },
748
  {
749
  "epoch": 0.95,
750
- "learning_rate": 1.6624040920716113e-06,
751
- "loss": 1.2452,
752
  "step": 1240
753
  },
754
  {
755
  "epoch": 0.96,
756
- "learning_rate": 1.4066496163682865e-06,
757
- "loss": 1.262,
758
  "step": 1250
759
  },
760
  {
761
  "epoch": 0.97,
762
- "learning_rate": 1.1508951406649617e-06,
763
- "loss": 1.2206,
764
  "step": 1260
765
  },
766
  {
767
  "epoch": 0.97,
768
- "learning_rate": 8.951406649616369e-07,
769
- "loss": 1.231,
770
  "step": 1270
771
  },
772
  {
773
  "epoch": 0.98,
774
- "learning_rate": 6.393861892583121e-07,
775
- "loss": 1.2134,
776
  "step": 1280
777
  },
778
  {
779
  "epoch": 0.99,
780
- "learning_rate": 3.8363171355498723e-07,
781
- "loss": 1.2325,
782
  "step": 1290
783
  },
784
  {
785
  "epoch": 1.0,
786
- "learning_rate": 1.2787723785166242e-07,
787
- "loss": 1.1779,
788
  "step": 1300
789
  },
790
  {
791
  "epoch": 1.0,
792
- "eval_accuracy": 0.5816544091224876,
793
- "eval_loss": 1.1218500137329102,
794
- "eval_runtime": 299.8253,
795
- "eval_samples_per_second": 139.224,
796
- "eval_steps_per_second": 69.614,
797
  "step": 1304
798
  },
799
  {
800
  "epoch": 1.0,
801
- "step": 1304,
802
- "total_flos": 2.2877841631614075e+18,
803
- "train_loss": 1.4751509523099184,
804
- "train_runtime": 962.063,
805
- "train_samples_per_second": 173.553,
806
- "train_steps_per_second": 1.355
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
807
  }
808
  ],
809
- "max_steps": 1304,
810
- "num_train_epochs": 1,
811
- "total_flos": 2.2877841631614075e+18,
812
  "trial_name": null,
813
  "trial_params": null
814
  }
 
1
  {
2
+ "best_metric": 0.7016984883693075,
3
+ "best_model_checkpoint": "wav2vec2-base-mirst500/checkpoint-6520",
4
+ "epoch": 4.999616711383672,
5
+ "global_step": 6520,
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": 4.6012269938650306e-07,
13
+ "loss": 2.0823,
14
  "step": 10
15
  },
16
  {
17
  "epoch": 0.02,
18
+ "learning_rate": 9.202453987730061e-07,
19
+ "loss": 2.0814,
20
  "step": 20
21
  },
22
  {
23
  "epoch": 0.02,
24
+ "learning_rate": 1.3803680981595093e-06,
25
+ "loss": 2.0803,
26
  "step": 30
27
  },
28
  {
29
  "epoch": 0.03,
30
+ "learning_rate": 1.8404907975460122e-06,
31
+ "loss": 2.0788,
32
  "step": 40
33
  },
34
  {
35
  "epoch": 0.04,
36
+ "learning_rate": 2.3006134969325154e-06,
37
+ "loss": 2.0776,
38
  "step": 50
39
  },
40
  {
41
  "epoch": 0.05,
42
+ "learning_rate": 2.7607361963190186e-06,
43
+ "loss": 2.0731,
44
  "step": 60
45
  },
46
  {
47
  "epoch": 0.05,
48
+ "learning_rate": 3.2208588957055217e-06,
49
+ "loss": 2.0728,
50
  "step": 70
51
  },
52
  {
53
  "epoch": 0.06,
54
+ "learning_rate": 3.6809815950920245e-06,
55
+ "loss": 2.0645,
56
  "step": 80
57
  },
58
  {
59
  "epoch": 0.07,
60
+ "learning_rate": 4.141104294478528e-06,
61
+ "loss": 2.0645,
62
  "step": 90
63
  },
64
  {
65
  "epoch": 0.08,
66
+ "learning_rate": 4.601226993865031e-06,
67
+ "loss": 2.0514,
68
  "step": 100
69
  },
70
  {
71
  "epoch": 0.08,
72
+ "learning_rate": 5.061349693251534e-06,
73
+ "loss": 2.0403,
74
  "step": 110
75
  },
76
  {
77
  "epoch": 0.09,
78
+ "learning_rate": 5.521472392638037e-06,
79
+ "loss": 2.0268,
80
  "step": 120
81
  },
82
  {
83
  "epoch": 0.1,
84
+ "learning_rate": 5.981595092024539e-06,
85
+ "loss": 2.0036,
86
  "step": 130
87
  },
88
  {
89
  "epoch": 0.11,
90
+ "learning_rate": 6.4417177914110434e-06,
91
+ "loss": 2.0021,
92
  "step": 140
93
  },
94
  {
95
  "epoch": 0.11,
96
+ "learning_rate": 6.901840490797547e-06,
97
+ "loss": 1.9849,
98
  "step": 150
99
  },
100
  {
101
  "epoch": 0.12,
102
+ "learning_rate": 7.361963190184049e-06,
103
+ "loss": 1.9574,
104
  "step": 160
105
  },
106
  {
107
  "epoch": 0.13,
108
+ "learning_rate": 7.822085889570552e-06,
109
+ "loss": 1.9671,
110
  "step": 170
111
  },
112
  {
113
  "epoch": 0.14,
114
+ "learning_rate": 8.282208588957055e-06,
115
+ "loss": 1.9527,
116
  "step": 180
117
  },
118
  {
119
  "epoch": 0.15,
120
+ "learning_rate": 8.742331288343558e-06,
121
+ "loss": 1.9551,
122
  "step": 190
123
  },
124
  {
125
  "epoch": 0.15,
126
+ "learning_rate": 9.202453987730062e-06,
127
+ "loss": 1.9561,
128
  "step": 200
129
  },
130
  {
131
  "epoch": 0.16,
132
+ "learning_rate": 9.662576687116563e-06,
133
+ "loss": 1.9289,
134
  "step": 210
135
  },
136
  {
137
  "epoch": 0.17,
138
+ "learning_rate": 1.0122699386503068e-05,
139
+ "loss": 1.9388,
140
  "step": 220
141
  },
142
  {
143
  "epoch": 0.18,
144
+ "learning_rate": 1.0582822085889571e-05,
145
+ "loss": 1.9122,
146
  "step": 230
147
  },
148
  {
149
  "epoch": 0.18,
150
+ "learning_rate": 1.1042944785276074e-05,
151
+ "loss": 1.8831,
152
  "step": 240
153
  },
154
  {
155
  "epoch": 0.19,
156
+ "learning_rate": 1.1503067484662577e-05,
157
+ "loss": 1.9166,
158
  "step": 250
159
  },
160
  {
161
  "epoch": 0.2,
162
+ "learning_rate": 1.1963190184049079e-05,
163
+ "loss": 1.8841,
164
  "step": 260
165
  },
166
  {
167
  "epoch": 0.21,
168
+ "learning_rate": 1.2423312883435584e-05,
169
+ "loss": 1.8977,
170
  "step": 270
171
  },
172
  {
173
  "epoch": 0.21,
174
+ "learning_rate": 1.2883435582822087e-05,
175
+ "loss": 1.9011,
176
  "step": 280
177
  },
178
  {
179
  "epoch": 0.22,
180
+ "learning_rate": 1.3343558282208588e-05,
181
+ "loss": 1.8778,
182
  "step": 290
183
  },
184
  {
185
  "epoch": 0.23,
186
+ "learning_rate": 1.3803680981595093e-05,
187
+ "loss": 1.8782,
188
  "step": 300
189
  },
190
  {
191
  "epoch": 0.24,
192
+ "learning_rate": 1.4263803680981595e-05,
193
+ "loss": 1.8568,
194
  "step": 310
195
  },
196
  {
197
  "epoch": 0.25,
198
+ "learning_rate": 1.4723926380368098e-05,
199
+ "loss": 1.8845,
200
  "step": 320
201
  },
202
  {
203
  "epoch": 0.25,
204
+ "learning_rate": 1.5184049079754601e-05,
205
+ "loss": 1.852,
206
  "step": 330
207
  },
208
  {
209
  "epoch": 0.26,
210
+ "learning_rate": 1.5598159509202456e-05,
211
+ "loss": 1.8479,
212
  "step": 340
213
  },
214
  {
215
  "epoch": 0.27,
216
+ "learning_rate": 1.6058282208588957e-05,
217
+ "loss": 1.8345,
218
  "step": 350
219
  },
220
  {
221
  "epoch": 0.28,
222
+ "learning_rate": 1.6518404907975462e-05,
223
+ "loss": 1.8518,
224
  "step": 360
225
  },
226
  {
227
  "epoch": 0.28,
228
+ "learning_rate": 1.6978527607361963e-05,
229
+ "loss": 1.8653,
230
  "step": 370
231
  },
232
  {
233
  "epoch": 0.29,
234
+ "learning_rate": 1.7438650306748465e-05,
235
+ "loss": 1.837,
236
  "step": 380
237
  },
238
  {
239
  "epoch": 0.3,
240
+ "learning_rate": 1.789877300613497e-05,
241
+ "loss": 1.8058,
242
  "step": 390
243
  },
244
  {
245
  "epoch": 0.31,
246
+ "learning_rate": 1.8358895705521475e-05,
247
+ "loss": 1.8301,
248
  "step": 400
249
  },
250
  {
251
  "epoch": 0.31,
252
+ "learning_rate": 1.8819018404907976e-05,
253
+ "loss": 1.8287,
254
  "step": 410
255
  },
256
  {
257
  "epoch": 0.32,
258
+ "learning_rate": 1.9279141104294478e-05,
259
+ "loss": 1.7848,
260
  "step": 420
261
  },
262
  {
263
  "epoch": 0.33,
264
+ "learning_rate": 1.9739263803680982e-05,
265
+ "loss": 1.7821,
266
  "step": 430
267
  },
268
  {
269
  "epoch": 0.34,
270
+ "learning_rate": 2.0199386503067484e-05,
271
+ "loss": 1.7985,
272
  "step": 440
273
  },
274
  {
275
  "epoch": 0.34,
276
+ "learning_rate": 2.065950920245399e-05,
277
+ "loss": 1.762,
278
  "step": 450
279
  },
280
  {
281
  "epoch": 0.35,
282
+ "learning_rate": 2.1119631901840494e-05,
283
+ "loss": 1.7812,
284
  "step": 460
285
  },
286
  {
287
  "epoch": 0.36,
288
+ "learning_rate": 2.1579754601226992e-05,
289
+ "loss": 1.7677,
290
  "step": 470
291
  },
292
  {
293
  "epoch": 0.37,
294
+ "learning_rate": 2.2039877300613497e-05,
295
+ "loss": 1.7733,
296
  "step": 480
297
  },
298
  {
299
  "epoch": 0.38,
300
+ "learning_rate": 2.25e-05,
301
+ "loss": 1.7699,
302
  "step": 490
303
  },
304
  {
305
  "epoch": 0.38,
306
+ "learning_rate": 2.2960122699386503e-05,
307
+ "loss": 1.7429,
308
  "step": 500
309
  },
310
  {
311
  "epoch": 0.39,
312
+ "learning_rate": 2.3420245398773008e-05,
313
+ "loss": 1.6911,
314
  "step": 510
315
  },
316
  {
317
  "epoch": 0.4,
318
+ "learning_rate": 2.388036809815951e-05,
319
+ "loss": 1.6975,
320
  "step": 520
321
  },
322
  {
323
  "epoch": 0.41,
324
+ "learning_rate": 2.434049079754601e-05,
325
+ "loss": 1.7233,
326
  "step": 530
327
  },
328
  {
329
  "epoch": 0.41,
330
+ "learning_rate": 2.4800613496932516e-05,
331
+ "loss": 1.6838,
332
  "step": 540
333
  },
334
  {
335
  "epoch": 0.42,
336
+ "learning_rate": 2.526073619631902e-05,
337
+ "loss": 1.6709,
338
  "step": 550
339
  },
340
  {
341
  "epoch": 0.43,
342
+ "learning_rate": 2.5720858895705522e-05,
343
+ "loss": 1.7078,
344
  "step": 560
345
  },
346
  {
347
  "epoch": 0.44,
348
+ "learning_rate": 2.6180981595092023e-05,
349
+ "loss": 1.6882,
350
  "step": 570
351
  },
352
  {
353
  "epoch": 0.44,
354
+ "learning_rate": 2.6641104294478528e-05,
355
+ "loss": 1.6386,
356
  "step": 580
357
  },
358
  {
359
  "epoch": 0.45,
360
+ "learning_rate": 2.7101226993865033e-05,
361
+ "loss": 1.6315,
362
  "step": 590
363
  },
364
  {
365
  "epoch": 0.46,
366
+ "learning_rate": 2.7561349693251535e-05,
367
+ "loss": 1.5773,
368
  "step": 600
369
  },
370
  {
371
  "epoch": 0.47,
372
+ "learning_rate": 2.802147239263804e-05,
373
+ "loss": 1.6123,
374
  "step": 610
375
  },
376
  {
377
  "epoch": 0.48,
378
+ "learning_rate": 2.848159509202454e-05,
379
+ "loss": 1.6244,
380
  "step": 620
381
  },
382
  {
383
  "epoch": 0.48,
384
+ "learning_rate": 2.8941717791411042e-05,
385
+ "loss": 1.5473,
386
  "step": 630
387
  },
388
  {
389
  "epoch": 0.49,
390
+ "learning_rate": 2.9401840490797547e-05,
391
+ "loss": 1.5417,
392
  "step": 640
393
  },
394
  {
395
  "epoch": 0.5,
396
+ "learning_rate": 2.9861963190184052e-05,
397
+ "loss": 1.5088,
398
  "step": 650
399
  },
400
  {
401
  "epoch": 0.51,
402
+ "learning_rate": 2.9964212678936606e-05,
403
+ "loss": 1.51,
404
  "step": 660
405
  },
406
  {
407
  "epoch": 0.51,
408
+ "learning_rate": 2.9913087934560325e-05,
409
+ "loss": 1.5141,
410
  "step": 670
411
  },
412
  {
413
  "epoch": 0.52,
414
+ "learning_rate": 2.9861963190184052e-05,
415
+ "loss": 1.4878,
416
  "step": 680
417
  },
418
  {
419
  "epoch": 0.53,
420
+ "learning_rate": 2.9810838445807772e-05,
421
+ "loss": 1.4722,
422
  "step": 690
423
  },
424
  {
425
  "epoch": 0.54,
426
+ "learning_rate": 2.9759713701431495e-05,
427
+ "loss": 1.4969,
428
  "step": 700
429
  },
430
  {
431
  "epoch": 0.54,
432
+ "learning_rate": 2.9708588957055215e-05,
433
+ "loss": 1.4691,
434
  "step": 710
435
  },
436
  {
437
  "epoch": 0.55,
438
+ "learning_rate": 2.9657464212678938e-05,
439
+ "loss": 1.4537,
440
  "step": 720
441
  },
442
  {
443
  "epoch": 0.56,
444
+ "learning_rate": 2.9606339468302658e-05,
445
+ "loss": 1.4395,
446
  "step": 730
447
  },
448
  {
449
  "epoch": 0.57,
450
+ "learning_rate": 2.955521472392638e-05,
451
+ "loss": 1.4219,
452
  "step": 740
453
  },
454
  {
455
  "epoch": 0.57,
456
+ "learning_rate": 2.9504089979550104e-05,
457
+ "loss": 1.421,
458
  "step": 750
459
  },
460
  {
461
  "epoch": 0.58,
462
+ "learning_rate": 2.9452965235173824e-05,
463
+ "loss": 1.4472,
464
  "step": 760
465
  },
466
  {
467
  "epoch": 0.59,
468
+ "learning_rate": 2.9401840490797547e-05,
469
+ "loss": 1.404,
470
  "step": 770
471
  },
472
  {
473
  "epoch": 0.6,
474
+ "learning_rate": 2.9350715746421267e-05,
475
+ "loss": 1.3762,
476
  "step": 780
477
  },
478
  {
479
  "epoch": 0.61,
480
+ "learning_rate": 2.929959100204499e-05,
481
+ "loss": 1.3499,
482
  "step": 790
483
  },
484
  {
485
  "epoch": 0.61,
486
+ "learning_rate": 2.924846625766871e-05,
487
+ "loss": 1.387,
488
  "step": 800
489
  },
490
  {
491
  "epoch": 0.62,
492
+ "learning_rate": 2.9197341513292437e-05,
493
+ "loss": 1.3872,
494
  "step": 810
495
  },
496
  {
497
  "epoch": 0.63,
498
+ "learning_rate": 2.9146216768916156e-05,
499
+ "loss": 1.3594,
500
  "step": 820
501
  },
502
  {
503
  "epoch": 0.64,
504
+ "learning_rate": 2.909509202453988e-05,
505
+ "loss": 1.3386,
506
  "step": 830
507
  },
508
  {
509
  "epoch": 0.64,
510
+ "learning_rate": 2.90439672801636e-05,
511
+ "loss": 1.362,
512
  "step": 840
513
  },
514
  {
515
  "epoch": 0.65,
516
+ "learning_rate": 2.8992842535787323e-05,
517
+ "loss": 1.4005,
518
  "step": 850
519
  },
520
  {
521
  "epoch": 0.66,
522
+ "learning_rate": 2.8941717791411042e-05,
523
+ "loss": 1.2967,
524
  "step": 860
525
  },
526
  {
527
  "epoch": 0.67,
528
+ "learning_rate": 2.8890593047034765e-05,
529
+ "loss": 1.3309,
530
  "step": 870
531
  },
532
  {
533
  "epoch": 0.67,
534
+ "learning_rate": 2.883946830265849e-05,
535
+ "loss": 1.3563,
536
  "step": 880
537
  },
538
  {
539
  "epoch": 0.68,
540
+ "learning_rate": 2.8788343558282212e-05,
541
+ "loss": 1.3469,
542
  "step": 890
543
  },
544
  {
545
  "epoch": 0.69,
546
+ "learning_rate": 2.873721881390593e-05,
547
+ "loss": 1.3561,
548
  "step": 900
549
  },
550
  {
551
  "epoch": 0.7,
552
+ "learning_rate": 2.8686094069529655e-05,
553
+ "loss": 1.3207,
554
  "step": 910
555
  },
556
  {
557
  "epoch": 0.71,
558
+ "learning_rate": 2.8634969325153375e-05,
559
+ "loss": 1.327,
560
  "step": 920
561
  },
562
  {
563
  "epoch": 0.71,
564
+ "learning_rate": 2.8583844580777094e-05,
565
+ "loss": 1.2817,
566
  "step": 930
567
  },
568
  {
569
  "epoch": 0.72,
570
+ "learning_rate": 2.8532719836400818e-05,
571
+ "loss": 1.2823,
572
  "step": 940
573
  },
574
  {
575
  "epoch": 0.73,
576
+ "learning_rate": 2.848159509202454e-05,
577
+ "loss": 1.2915,
578
  "step": 950
579
  },
580
  {
581
  "epoch": 0.74,
582
+ "learning_rate": 2.8430470347648264e-05,
583
+ "loss": 1.2361,
584
  "step": 960
585
  },
586
  {
587
  "epoch": 0.74,
588
+ "learning_rate": 2.8379345603271984e-05,
589
+ "loss": 1.2647,
590
  "step": 970
591
  },
592
  {
593
  "epoch": 0.75,
594
+ "learning_rate": 2.8328220858895707e-05,
595
+ "loss": 1.2677,
596
  "step": 980
597
  },
598
  {
599
  "epoch": 0.76,
600
+ "learning_rate": 2.8277096114519427e-05,
601
+ "loss": 1.2611,
602
  "step": 990
603
  },
604
  {
605
  "epoch": 0.77,
606
+ "learning_rate": 2.822597137014315e-05,
607
+ "loss": 1.293,
608
  "step": 1000
609
  },
610
  {
611
  "epoch": 0.77,
612
+ "learning_rate": 2.817484662576687e-05,
613
+ "loss": 1.2216,
614
  "step": 1010
615
  },
616
  {
617
  "epoch": 0.78,
618
+ "learning_rate": 2.8123721881390596e-05,
619
+ "loss": 1.3012,
620
  "step": 1020
621
  },
622
  {
623
  "epoch": 0.79,
624
+ "learning_rate": 2.8072597137014316e-05,
625
+ "loss": 1.2432,
626
  "step": 1030
627
  },
628
  {
629
  "epoch": 0.8,
630
+ "learning_rate": 2.802147239263804e-05,
631
+ "loss": 1.2935,
632
  "step": 1040
633
  },
634
  {
635
  "epoch": 0.8,
636
+ "learning_rate": 2.797034764826176e-05,
637
+ "loss": 1.2375,
638
  "step": 1050
639
  },
640
  {
641
  "epoch": 0.81,
642
+ "learning_rate": 2.7919222903885482e-05,
643
+ "loss": 1.2617,
644
  "step": 1060
645
  },
646
  {
647
  "epoch": 0.82,
648
+ "learning_rate": 2.7868098159509202e-05,
649
+ "loss": 1.2519,
650
  "step": 1070
651
  },
652
  {
653
  "epoch": 0.83,
654
+ "learning_rate": 2.7816973415132922e-05,
655
+ "loss": 1.1971,
656
  "step": 1080
657
  },
658
  {
659
  "epoch": 0.84,
660
+ "learning_rate": 2.776584867075665e-05,
661
+ "loss": 1.2598,
662
  "step": 1090
663
  },
664
  {
665
  "epoch": 0.84,
666
+ "learning_rate": 2.771472392638037e-05,
667
+ "loss": 1.2187,
668
  "step": 1100
669
  },
670
  {
671
  "epoch": 0.85,
672
+ "learning_rate": 2.766359918200409e-05,
673
+ "loss": 1.2202,
674
  "step": 1110
675
  },
676
  {
677
  "epoch": 0.86,
678
+ "learning_rate": 2.761247443762781e-05,
679
+ "loss": 1.2007,
680
  "step": 1120
681
  },
682
  {
683
  "epoch": 0.87,
684
+ "learning_rate": 2.7561349693251535e-05,
685
+ "loss": 1.2117,
686
  "step": 1130
687
  },
688
  {
689
  "epoch": 0.87,
690
+ "learning_rate": 2.7510224948875254e-05,
691
+ "loss": 1.2165,
692
  "step": 1140
693
  },
694
  {
695
  "epoch": 0.88,
696
+ "learning_rate": 2.7459100204498977e-05,
697
+ "loss": 1.2141,
698
  "step": 1150
699
  },
700
  {
701
  "epoch": 0.89,
702
+ "learning_rate": 2.74079754601227e-05,
703
+ "loss": 1.2257,
704
  "step": 1160
705
  },
706
  {
707
  "epoch": 0.9,
708
+ "learning_rate": 2.7356850715746424e-05,
709
+ "loss": 1.1844,
710
  "step": 1170
711
  },
712
  {
713
  "epoch": 0.9,
714
+ "learning_rate": 2.7305725971370144e-05,
715
+ "loss": 1.229,
716
  "step": 1180
717
  },
718
  {
719
  "epoch": 0.91,
720
+ "learning_rate": 2.7254601226993867e-05,
721
+ "loss": 1.2189,
722
  "step": 1190
723
  },
724
  {
725
  "epoch": 0.92,
726
+ "learning_rate": 2.7203476482617587e-05,
727
+ "loss": 1.2027,
728
  "step": 1200
729
  },
730
  {
731
  "epoch": 0.93,
732
+ "learning_rate": 2.715235173824131e-05,
733
+ "loss": 1.2362,
734
  "step": 1210
735
  },
736
  {
737
  "epoch": 0.94,
738
+ "learning_rate": 2.7101226993865033e-05,
739
+ "loss": 1.1787,
740
  "step": 1220
741
  },
742
  {
743
  "epoch": 0.94,
744
+ "learning_rate": 2.7050102249488753e-05,
745
+ "loss": 1.2087,
746
  "step": 1230
747
  },
748
  {
749
  "epoch": 0.95,
750
+ "learning_rate": 2.6998977505112476e-05,
751
+ "loss": 1.1646,
752
  "step": 1240
753
  },
754
  {
755
  "epoch": 0.96,
756
+ "learning_rate": 2.6947852760736196e-05,
757
+ "loss": 1.2548,
758
  "step": 1250
759
  },
760
  {
761
  "epoch": 0.97,
762
+ "learning_rate": 2.689672801635992e-05,
763
+ "loss": 1.2269,
764
  "step": 1260
765
  },
766
  {
767
  "epoch": 0.97,
768
+ "learning_rate": 2.684560327198364e-05,
769
+ "loss": 1.1421,
770
  "step": 1270
771
  },
772
  {
773
  "epoch": 0.98,
774
+ "learning_rate": 2.6794478527607362e-05,
775
+ "loss": 1.2349,
776
  "step": 1280
777
  },
778
  {
779
  "epoch": 0.99,
780
+ "learning_rate": 2.6743353783231085e-05,
781
+ "loss": 1.1553,
782
  "step": 1290
783
  },
784
  {
785
  "epoch": 1.0,
786
+ "learning_rate": 2.669222903885481e-05,
787
+ "loss": 1.1999,
788
  "step": 1300
789
  },
790
  {
791
  "epoch": 1.0,
792
+ "eval_accuracy": 0.5877392616726158,
793
+ "eval_loss": 1.102908730506897,
794
+ "eval_runtime": 301.1056,
795
+ "eval_samples_per_second": 138.632,
796
+ "eval_steps_per_second": 69.318,
797
  "step": 1304
798
  },
799
  {
800
  "epoch": 1.0,
801
+ "learning_rate": 2.6641104294478528e-05,
802
+ "loss": 1.2812,
803
+ "step": 1310
804
+ },
805
+ {
806
+ "epoch": 1.01,
807
+ "learning_rate": 2.658997955010225e-05,
808
+ "loss": 1.1977,
809
+ "step": 1320
810
+ },
811
+ {
812
+ "epoch": 1.02,
813
+ "learning_rate": 2.653885480572597e-05,
814
+ "loss": 1.1985,
815
+ "step": 1330
816
+ },
817
+ {
818
+ "epoch": 1.03,
819
+ "learning_rate": 2.6487730061349694e-05,
820
+ "loss": 1.1709,
821
+ "step": 1340
822
+ },
823
+ {
824
+ "epoch": 1.04,
825
+ "learning_rate": 2.6441717791411043e-05,
826
+ "loss": 1.2057,
827
+ "step": 1350
828
+ },
829
+ {
830
+ "epoch": 1.04,
831
+ "learning_rate": 2.6390593047034766e-05,
832
+ "loss": 1.1377,
833
+ "step": 1360
834
+ },
835
+ {
836
+ "epoch": 1.05,
837
+ "learning_rate": 2.6339468302658486e-05,
838
+ "loss": 1.1278,
839
+ "step": 1370
840
+ },
841
+ {
842
+ "epoch": 1.06,
843
+ "learning_rate": 2.628834355828221e-05,
844
+ "loss": 1.1997,
845
+ "step": 1380
846
+ },
847
+ {
848
+ "epoch": 1.07,
849
+ "learning_rate": 2.623721881390593e-05,
850
+ "loss": 1.121,
851
+ "step": 1390
852
+ },
853
+ {
854
+ "epoch": 1.07,
855
+ "learning_rate": 2.6186094069529655e-05,
856
+ "loss": 1.2115,
857
+ "step": 1400
858
+ },
859
+ {
860
+ "epoch": 1.08,
861
+ "learning_rate": 2.6134969325153375e-05,
862
+ "loss": 1.18,
863
+ "step": 1410
864
+ },
865
+ {
866
+ "epoch": 1.09,
867
+ "learning_rate": 2.6083844580777098e-05,
868
+ "loss": 1.1797,
869
+ "step": 1420
870
+ },
871
+ {
872
+ "epoch": 1.1,
873
+ "learning_rate": 2.6032719836400818e-05,
874
+ "loss": 1.1343,
875
+ "step": 1430
876
+ },
877
+ {
878
+ "epoch": 1.1,
879
+ "learning_rate": 2.598159509202454e-05,
880
+ "loss": 1.1437,
881
+ "step": 1440
882
+ },
883
+ {
884
+ "epoch": 1.11,
885
+ "learning_rate": 2.593047034764826e-05,
886
+ "loss": 1.1184,
887
+ "step": 1450
888
+ },
889
+ {
890
+ "epoch": 1.12,
891
+ "learning_rate": 2.5879345603271984e-05,
892
+ "loss": 1.1675,
893
+ "step": 1460
894
+ },
895
+ {
896
+ "epoch": 1.13,
897
+ "learning_rate": 2.5828220858895707e-05,
898
+ "loss": 1.171,
899
+ "step": 1470
900
+ },
901
+ {
902
+ "epoch": 1.13,
903
+ "learning_rate": 2.577709611451943e-05,
904
+ "loss": 1.1481,
905
+ "step": 1480
906
+ },
907
+ {
908
+ "epoch": 1.14,
909
+ "learning_rate": 2.572597137014315e-05,
910
+ "loss": 1.1237,
911
+ "step": 1490
912
+ },
913
+ {
914
+ "epoch": 1.15,
915
+ "learning_rate": 2.5674846625766873e-05,
916
+ "loss": 1.1696,
917
+ "step": 1500
918
+ },
919
+ {
920
+ "epoch": 1.16,
921
+ "learning_rate": 2.5623721881390593e-05,
922
+ "loss": 1.1376,
923
+ "step": 1510
924
+ },
925
+ {
926
+ "epoch": 1.17,
927
+ "learning_rate": 2.5572597137014313e-05,
928
+ "loss": 1.126,
929
+ "step": 1520
930
+ },
931
+ {
932
+ "epoch": 1.17,
933
+ "learning_rate": 2.5521472392638036e-05,
934
+ "loss": 1.1429,
935
+ "step": 1530
936
+ },
937
+ {
938
+ "epoch": 1.18,
939
+ "learning_rate": 2.547034764826176e-05,
940
+ "loss": 1.1271,
941
+ "step": 1540
942
+ },
943
+ {
944
+ "epoch": 1.19,
945
+ "learning_rate": 2.5419222903885483e-05,
946
+ "loss": 1.1538,
947
+ "step": 1550
948
+ },
949
+ {
950
+ "epoch": 1.2,
951
+ "learning_rate": 2.5368098159509202e-05,
952
+ "loss": 1.1844,
953
+ "step": 1560
954
+ },
955
+ {
956
+ "epoch": 1.2,
957
+ "learning_rate": 2.5316973415132926e-05,
958
+ "loss": 1.1682,
959
+ "step": 1570
960
+ },
961
+ {
962
+ "epoch": 1.21,
963
+ "learning_rate": 2.5265848670756645e-05,
964
+ "loss": 1.1432,
965
+ "step": 1580
966
+ },
967
+ {
968
+ "epoch": 1.22,
969
+ "learning_rate": 2.521472392638037e-05,
970
+ "loss": 1.1812,
971
+ "step": 1590
972
+ },
973
+ {
974
+ "epoch": 1.23,
975
+ "learning_rate": 2.516359918200409e-05,
976
+ "loss": 1.1812,
977
+ "step": 1600
978
+ },
979
+ {
980
+ "epoch": 1.23,
981
+ "learning_rate": 2.5112474437627815e-05,
982
+ "loss": 1.1207,
983
+ "step": 1610
984
+ },
985
+ {
986
+ "epoch": 1.24,
987
+ "learning_rate": 2.5061349693251535e-05,
988
+ "loss": 1.105,
989
+ "step": 1620
990
+ },
991
+ {
992
+ "epoch": 1.25,
993
+ "learning_rate": 2.5010224948875258e-05,
994
+ "loss": 1.1223,
995
+ "step": 1630
996
+ },
997
+ {
998
+ "epoch": 1.26,
999
+ "learning_rate": 2.4959100204498978e-05,
1000
+ "loss": 1.0973,
1001
+ "step": 1640
1002
+ },
1003
+ {
1004
+ "epoch": 1.27,
1005
+ "learning_rate": 2.49079754601227e-05,
1006
+ "loss": 1.131,
1007
+ "step": 1650
1008
+ },
1009
+ {
1010
+ "epoch": 1.27,
1011
+ "learning_rate": 2.485685071574642e-05,
1012
+ "loss": 1.1277,
1013
+ "step": 1660
1014
+ },
1015
+ {
1016
+ "epoch": 1.28,
1017
+ "learning_rate": 2.4805725971370144e-05,
1018
+ "loss": 1.0796,
1019
+ "step": 1670
1020
+ },
1021
+ {
1022
+ "epoch": 1.29,
1023
+ "learning_rate": 2.4754601226993867e-05,
1024
+ "loss": 1.1054,
1025
+ "step": 1680
1026
+ },
1027
+ {
1028
+ "epoch": 1.3,
1029
+ "learning_rate": 2.4703476482617587e-05,
1030
+ "loss": 1.1159,
1031
+ "step": 1690
1032
+ },
1033
+ {
1034
+ "epoch": 1.3,
1035
+ "learning_rate": 2.465235173824131e-05,
1036
+ "loss": 1.0675,
1037
+ "step": 1700
1038
+ },
1039
+ {
1040
+ "epoch": 1.31,
1041
+ "learning_rate": 2.460122699386503e-05,
1042
+ "loss": 1.1283,
1043
+ "step": 1710
1044
+ },
1045
+ {
1046
+ "epoch": 1.32,
1047
+ "learning_rate": 2.4550102249488753e-05,
1048
+ "loss": 1.1507,
1049
+ "step": 1720
1050
+ },
1051
+ {
1052
+ "epoch": 1.33,
1053
+ "learning_rate": 2.4498977505112473e-05,
1054
+ "loss": 1.1587,
1055
+ "step": 1730
1056
+ },
1057
+ {
1058
+ "epoch": 1.33,
1059
+ "learning_rate": 2.44478527607362e-05,
1060
+ "loss": 1.1495,
1061
+ "step": 1740
1062
+ },
1063
+ {
1064
+ "epoch": 1.34,
1065
+ "learning_rate": 2.439672801635992e-05,
1066
+ "loss": 1.1109,
1067
+ "step": 1750
1068
+ },
1069
+ {
1070
+ "epoch": 1.35,
1071
+ "learning_rate": 2.4345603271983642e-05,
1072
+ "loss": 1.124,
1073
+ "step": 1760
1074
+ },
1075
+ {
1076
+ "epoch": 1.36,
1077
+ "learning_rate": 2.4294478527607362e-05,
1078
+ "loss": 1.1369,
1079
+ "step": 1770
1080
+ },
1081
+ {
1082
+ "epoch": 1.36,
1083
+ "learning_rate": 2.4243353783231085e-05,
1084
+ "loss": 1.1066,
1085
+ "step": 1780
1086
+ },
1087
+ {
1088
+ "epoch": 1.37,
1089
+ "learning_rate": 2.4192229038854805e-05,
1090
+ "loss": 1.1119,
1091
+ "step": 1790
1092
+ },
1093
+ {
1094
+ "epoch": 1.38,
1095
+ "learning_rate": 2.414110429447853e-05,
1096
+ "loss": 1.0885,
1097
+ "step": 1800
1098
+ },
1099
+ {
1100
+ "epoch": 1.39,
1101
+ "learning_rate": 2.408997955010225e-05,
1102
+ "loss": 1.1023,
1103
+ "step": 1810
1104
+ },
1105
+ {
1106
+ "epoch": 1.4,
1107
+ "learning_rate": 2.403885480572597e-05,
1108
+ "loss": 1.1624,
1109
+ "step": 1820
1110
+ },
1111
+ {
1112
+ "epoch": 1.4,
1113
+ "learning_rate": 2.3987730061349695e-05,
1114
+ "loss": 1.089,
1115
+ "step": 1830
1116
+ },
1117
+ {
1118
+ "epoch": 1.41,
1119
+ "learning_rate": 2.3936605316973414e-05,
1120
+ "loss": 1.1186,
1121
+ "step": 1840
1122
+ },
1123
+ {
1124
+ "epoch": 1.42,
1125
+ "learning_rate": 2.3885480572597138e-05,
1126
+ "loss": 1.0953,
1127
+ "step": 1850
1128
+ },
1129
+ {
1130
+ "epoch": 1.43,
1131
+ "learning_rate": 2.3834355828220857e-05,
1132
+ "loss": 1.1036,
1133
+ "step": 1860
1134
+ },
1135
+ {
1136
+ "epoch": 1.43,
1137
+ "learning_rate": 2.378323108384458e-05,
1138
+ "loss": 1.0797,
1139
+ "step": 1870
1140
+ },
1141
+ {
1142
+ "epoch": 1.44,
1143
+ "learning_rate": 2.3732106339468304e-05,
1144
+ "loss": 1.1168,
1145
+ "step": 1880
1146
+ },
1147
+ {
1148
+ "epoch": 1.45,
1149
+ "learning_rate": 2.3680981595092027e-05,
1150
+ "loss": 1.1152,
1151
+ "step": 1890
1152
+ },
1153
+ {
1154
+ "epoch": 1.46,
1155
+ "learning_rate": 2.3629856850715747e-05,
1156
+ "loss": 1.0968,
1157
+ "step": 1900
1158
+ },
1159
+ {
1160
+ "epoch": 1.46,
1161
+ "learning_rate": 2.357873210633947e-05,
1162
+ "loss": 1.1493,
1163
+ "step": 1910
1164
+ },
1165
+ {
1166
+ "epoch": 1.47,
1167
+ "learning_rate": 2.352760736196319e-05,
1168
+ "loss": 1.0944,
1169
+ "step": 1920
1170
+ },
1171
+ {
1172
+ "epoch": 1.48,
1173
+ "learning_rate": 2.3476482617586913e-05,
1174
+ "loss": 1.0999,
1175
+ "step": 1930
1176
+ },
1177
+ {
1178
+ "epoch": 1.49,
1179
+ "learning_rate": 2.3425357873210633e-05,
1180
+ "loss": 1.072,
1181
+ "step": 1940
1182
+ },
1183
+ {
1184
+ "epoch": 1.5,
1185
+ "learning_rate": 2.337423312883436e-05,
1186
+ "loss": 1.1158,
1187
+ "step": 1950
1188
+ },
1189
+ {
1190
+ "epoch": 1.5,
1191
+ "learning_rate": 2.332310838445808e-05,
1192
+ "loss": 1.082,
1193
+ "step": 1960
1194
+ },
1195
+ {
1196
+ "epoch": 1.51,
1197
+ "learning_rate": 2.32719836400818e-05,
1198
+ "loss": 1.058,
1199
+ "step": 1970
1200
+ },
1201
+ {
1202
+ "epoch": 1.52,
1203
+ "learning_rate": 2.3220858895705522e-05,
1204
+ "loss": 1.1094,
1205
+ "step": 1980
1206
+ },
1207
+ {
1208
+ "epoch": 1.53,
1209
+ "learning_rate": 2.3169734151329242e-05,
1210
+ "loss": 1.0677,
1211
+ "step": 1990
1212
+ },
1213
+ {
1214
+ "epoch": 1.53,
1215
+ "learning_rate": 2.3118609406952965e-05,
1216
+ "loss": 1.0579,
1217
+ "step": 2000
1218
+ },
1219
+ {
1220
+ "epoch": 1.54,
1221
+ "learning_rate": 2.3067484662576688e-05,
1222
+ "loss": 1.0428,
1223
+ "step": 2010
1224
+ },
1225
+ {
1226
+ "epoch": 1.55,
1227
+ "learning_rate": 2.301635991820041e-05,
1228
+ "loss": 1.0322,
1229
+ "step": 2020
1230
+ },
1231
+ {
1232
+ "epoch": 1.56,
1233
+ "learning_rate": 2.296523517382413e-05,
1234
+ "loss": 1.083,
1235
+ "step": 2030
1236
+ },
1237
+ {
1238
+ "epoch": 1.56,
1239
+ "learning_rate": 2.2914110429447854e-05,
1240
+ "loss": 1.0952,
1241
+ "step": 2040
1242
+ },
1243
+ {
1244
+ "epoch": 1.57,
1245
+ "learning_rate": 2.2862985685071574e-05,
1246
+ "loss": 1.0667,
1247
+ "step": 2050
1248
+ },
1249
+ {
1250
+ "epoch": 1.58,
1251
+ "learning_rate": 2.2811860940695297e-05,
1252
+ "loss": 1.0549,
1253
+ "step": 2060
1254
+ },
1255
+ {
1256
+ "epoch": 1.59,
1257
+ "learning_rate": 2.2760736196319017e-05,
1258
+ "loss": 1.0916,
1259
+ "step": 2070
1260
+ },
1261
+ {
1262
+ "epoch": 1.59,
1263
+ "learning_rate": 2.2709611451942744e-05,
1264
+ "loss": 1.0978,
1265
+ "step": 2080
1266
+ },
1267
+ {
1268
+ "epoch": 1.6,
1269
+ "learning_rate": 2.2658486707566464e-05,
1270
+ "loss": 1.058,
1271
+ "step": 2090
1272
+ },
1273
+ {
1274
+ "epoch": 1.61,
1275
+ "learning_rate": 2.2607361963190187e-05,
1276
+ "loss": 1.1053,
1277
+ "step": 2100
1278
+ },
1279
+ {
1280
+ "epoch": 1.62,
1281
+ "learning_rate": 2.2556237218813907e-05,
1282
+ "loss": 1.1061,
1283
+ "step": 2110
1284
+ },
1285
+ {
1286
+ "epoch": 1.63,
1287
+ "learning_rate": 2.2505112474437626e-05,
1288
+ "loss": 1.0919,
1289
+ "step": 2120
1290
+ },
1291
+ {
1292
+ "epoch": 1.63,
1293
+ "learning_rate": 2.245398773006135e-05,
1294
+ "loss": 1.0499,
1295
+ "step": 2130
1296
+ },
1297
+ {
1298
+ "epoch": 1.64,
1299
+ "learning_rate": 2.240286298568507e-05,
1300
+ "loss": 1.0577,
1301
+ "step": 2140
1302
+ },
1303
+ {
1304
+ "epoch": 1.65,
1305
+ "learning_rate": 2.2351738241308796e-05,
1306
+ "loss": 1.0188,
1307
+ "step": 2150
1308
+ },
1309
+ {
1310
+ "epoch": 1.66,
1311
+ "learning_rate": 2.2300613496932516e-05,
1312
+ "loss": 1.0922,
1313
+ "step": 2160
1314
+ },
1315
+ {
1316
+ "epoch": 1.66,
1317
+ "learning_rate": 2.224948875255624e-05,
1318
+ "loss": 1.0659,
1319
+ "step": 2170
1320
+ },
1321
+ {
1322
+ "epoch": 1.67,
1323
+ "learning_rate": 2.219836400817996e-05,
1324
+ "loss": 1.0656,
1325
+ "step": 2180
1326
+ },
1327
+ {
1328
+ "epoch": 1.68,
1329
+ "learning_rate": 2.2147239263803682e-05,
1330
+ "loss": 1.0858,
1331
+ "step": 2190
1332
+ },
1333
+ {
1334
+ "epoch": 1.69,
1335
+ "learning_rate": 2.2096114519427402e-05,
1336
+ "loss": 1.0432,
1337
+ "step": 2200
1338
+ },
1339
+ {
1340
+ "epoch": 1.69,
1341
+ "learning_rate": 2.2044989775051125e-05,
1342
+ "loss": 1.0386,
1343
+ "step": 2210
1344
+ },
1345
+ {
1346
+ "epoch": 1.7,
1347
+ "learning_rate": 2.1993865030674848e-05,
1348
+ "loss": 1.0843,
1349
+ "step": 2220
1350
+ },
1351
+ {
1352
+ "epoch": 1.71,
1353
+ "learning_rate": 2.194274028629857e-05,
1354
+ "loss": 1.1272,
1355
+ "step": 2230
1356
+ },
1357
+ {
1358
+ "epoch": 1.72,
1359
+ "learning_rate": 2.189161554192229e-05,
1360
+ "loss": 1.0712,
1361
+ "step": 2240
1362
+ },
1363
+ {
1364
+ "epoch": 1.73,
1365
+ "learning_rate": 2.1840490797546014e-05,
1366
+ "loss": 0.9976,
1367
+ "step": 2250
1368
+ },
1369
+ {
1370
+ "epoch": 1.73,
1371
+ "learning_rate": 2.1789366053169734e-05,
1372
+ "loss": 1.0925,
1373
+ "step": 2260
1374
+ },
1375
+ {
1376
+ "epoch": 1.74,
1377
+ "learning_rate": 2.1738241308793454e-05,
1378
+ "loss": 1.0628,
1379
+ "step": 2270
1380
+ },
1381
+ {
1382
+ "epoch": 1.75,
1383
+ "learning_rate": 2.1687116564417177e-05,
1384
+ "loss": 1.0742,
1385
+ "step": 2280
1386
+ },
1387
+ {
1388
+ "epoch": 1.76,
1389
+ "learning_rate": 2.16359918200409e-05,
1390
+ "loss": 1.058,
1391
+ "step": 2290
1392
+ },
1393
+ {
1394
+ "epoch": 1.76,
1395
+ "learning_rate": 2.1584867075664623e-05,
1396
+ "loss": 1.033,
1397
+ "step": 2300
1398
+ },
1399
+ {
1400
+ "epoch": 1.77,
1401
+ "learning_rate": 2.1533742331288343e-05,
1402
+ "loss": 1.0682,
1403
+ "step": 2310
1404
+ },
1405
+ {
1406
+ "epoch": 1.78,
1407
+ "learning_rate": 2.1482617586912066e-05,
1408
+ "loss": 1.1028,
1409
+ "step": 2320
1410
+ },
1411
+ {
1412
+ "epoch": 1.79,
1413
+ "learning_rate": 2.1431492842535786e-05,
1414
+ "loss": 1.0428,
1415
+ "step": 2330
1416
+ },
1417
+ {
1418
+ "epoch": 1.79,
1419
+ "learning_rate": 2.138036809815951e-05,
1420
+ "loss": 1.0432,
1421
+ "step": 2340
1422
+ },
1423
+ {
1424
+ "epoch": 1.8,
1425
+ "learning_rate": 2.1329243353783233e-05,
1426
+ "loss": 1.0666,
1427
+ "step": 2350
1428
+ },
1429
+ {
1430
+ "epoch": 1.81,
1431
+ "learning_rate": 2.1278118609406956e-05,
1432
+ "loss": 1.0205,
1433
+ "step": 2360
1434
+ },
1435
+ {
1436
+ "epoch": 1.82,
1437
+ "learning_rate": 2.1226993865030676e-05,
1438
+ "loss": 1.0398,
1439
+ "step": 2370
1440
+ },
1441
+ {
1442
+ "epoch": 1.82,
1443
+ "learning_rate": 2.11758691206544e-05,
1444
+ "loss": 1.0386,
1445
+ "step": 2380
1446
+ },
1447
+ {
1448
+ "epoch": 1.83,
1449
+ "learning_rate": 2.112474437627812e-05,
1450
+ "loss": 1.0642,
1451
+ "step": 2390
1452
+ },
1453
+ {
1454
+ "epoch": 1.84,
1455
+ "learning_rate": 2.1073619631901842e-05,
1456
+ "loss": 1.0635,
1457
+ "step": 2400
1458
+ },
1459
+ {
1460
+ "epoch": 1.85,
1461
+ "learning_rate": 2.102249488752556e-05,
1462
+ "loss": 1.0366,
1463
+ "step": 2410
1464
+ },
1465
+ {
1466
+ "epoch": 1.86,
1467
+ "learning_rate": 2.0971370143149285e-05,
1468
+ "loss": 1.0517,
1469
+ "step": 2420
1470
+ },
1471
+ {
1472
+ "epoch": 1.86,
1473
+ "learning_rate": 2.0920245398773008e-05,
1474
+ "loss": 1.0518,
1475
+ "step": 2430
1476
+ },
1477
+ {
1478
+ "epoch": 1.87,
1479
+ "learning_rate": 2.0869120654396728e-05,
1480
+ "loss": 1.0715,
1481
+ "step": 2440
1482
+ },
1483
+ {
1484
+ "epoch": 1.88,
1485
+ "learning_rate": 2.081799591002045e-05,
1486
+ "loss": 1.0363,
1487
+ "step": 2450
1488
+ },
1489
+ {
1490
+ "epoch": 1.89,
1491
+ "learning_rate": 2.076687116564417e-05,
1492
+ "loss": 1.0384,
1493
+ "step": 2460
1494
+ },
1495
+ {
1496
+ "epoch": 1.89,
1497
+ "learning_rate": 2.0715746421267894e-05,
1498
+ "loss": 1.0585,
1499
+ "step": 2470
1500
+ },
1501
+ {
1502
+ "epoch": 1.9,
1503
+ "learning_rate": 2.0664621676891614e-05,
1504
+ "loss": 1.0972,
1505
+ "step": 2480
1506
+ },
1507
+ {
1508
+ "epoch": 1.91,
1509
+ "learning_rate": 2.061349693251534e-05,
1510
+ "loss": 1.0248,
1511
+ "step": 2490
1512
+ },
1513
+ {
1514
+ "epoch": 1.92,
1515
+ "learning_rate": 2.056237218813906e-05,
1516
+ "loss": 1.0279,
1517
+ "step": 2500
1518
+ },
1519
+ {
1520
+ "epoch": 1.92,
1521
+ "learning_rate": 2.0511247443762783e-05,
1522
+ "loss": 1.0053,
1523
+ "step": 2510
1524
+ },
1525
+ {
1526
+ "epoch": 1.93,
1527
+ "learning_rate": 2.0460122699386503e-05,
1528
+ "loss": 1.0685,
1529
+ "step": 2520
1530
+ },
1531
+ {
1532
+ "epoch": 1.94,
1533
+ "learning_rate": 2.0408997955010226e-05,
1534
+ "loss": 1.0059,
1535
+ "step": 2530
1536
+ },
1537
+ {
1538
+ "epoch": 1.95,
1539
+ "learning_rate": 2.0357873210633946e-05,
1540
+ "loss": 1.029,
1541
+ "step": 2540
1542
+ },
1543
+ {
1544
+ "epoch": 1.96,
1545
+ "learning_rate": 2.030674846625767e-05,
1546
+ "loss": 1.0916,
1547
+ "step": 2550
1548
+ },
1549
+ {
1550
+ "epoch": 1.96,
1551
+ "learning_rate": 2.0255623721881392e-05,
1552
+ "loss": 1.0296,
1553
+ "step": 2560
1554
+ },
1555
+ {
1556
+ "epoch": 1.97,
1557
+ "learning_rate": 2.0204498977505112e-05,
1558
+ "loss": 1.0179,
1559
+ "step": 2570
1560
+ },
1561
+ {
1562
+ "epoch": 1.98,
1563
+ "learning_rate": 2.0153374233128835e-05,
1564
+ "loss": 1.0126,
1565
+ "step": 2580
1566
+ },
1567
+ {
1568
+ "epoch": 1.99,
1569
+ "learning_rate": 2.0102249488752555e-05,
1570
+ "loss": 1.0458,
1571
+ "step": 2590
1572
+ },
1573
+ {
1574
+ "epoch": 1.99,
1575
+ "learning_rate": 2.005112474437628e-05,
1576
+ "loss": 1.0779,
1577
+ "step": 2600
1578
+ },
1579
+ {
1580
+ "epoch": 2.0,
1581
+ "eval_accuracy": 0.6555350597705004,
1582
+ "eval_loss": 0.9455426335334778,
1583
+ "eval_runtime": 303.4096,
1584
+ "eval_samples_per_second": 137.58,
1585
+ "eval_steps_per_second": 68.792,
1586
+ "step": 2608
1587
+ },
1588
+ {
1589
+ "epoch": 2.0,
1590
+ "learning_rate": 1.9999999999999998e-05,
1591
+ "loss": 1.1476,
1592
+ "step": 2610
1593
+ },
1594
+ {
1595
+ "epoch": 2.01,
1596
+ "learning_rate": 1.994887525562372e-05,
1597
+ "loss": 1.0322,
1598
+ "step": 2620
1599
+ },
1600
+ {
1601
+ "epoch": 2.02,
1602
+ "learning_rate": 1.9897750511247445e-05,
1603
+ "loss": 1.0394,
1604
+ "step": 2630
1605
+ },
1606
+ {
1607
+ "epoch": 2.02,
1608
+ "learning_rate": 1.9846625766871168e-05,
1609
+ "loss": 1.0319,
1610
+ "step": 2640
1611
+ },
1612
+ {
1613
+ "epoch": 2.03,
1614
+ "learning_rate": 1.9795501022494888e-05,
1615
+ "loss": 1.0061,
1616
+ "step": 2650
1617
+ },
1618
+ {
1619
+ "epoch": 2.04,
1620
+ "learning_rate": 1.974437627811861e-05,
1621
+ "loss": 1.0337,
1622
+ "step": 2660
1623
+ },
1624
+ {
1625
+ "epoch": 2.05,
1626
+ "learning_rate": 1.969325153374233e-05,
1627
+ "loss": 1.0778,
1628
+ "step": 2670
1629
+ },
1630
+ {
1631
+ "epoch": 2.06,
1632
+ "learning_rate": 1.9642126789366054e-05,
1633
+ "loss": 1.0107,
1634
+ "step": 2680
1635
+ },
1636
+ {
1637
+ "epoch": 2.06,
1638
+ "learning_rate": 1.9591002044989777e-05,
1639
+ "loss": 1.0841,
1640
+ "step": 2690
1641
+ },
1642
+ {
1643
+ "epoch": 2.07,
1644
+ "learning_rate": 1.95398773006135e-05,
1645
+ "loss": 1.0172,
1646
+ "step": 2700
1647
+ },
1648
+ {
1649
+ "epoch": 2.08,
1650
+ "learning_rate": 1.948875255623722e-05,
1651
+ "loss": 1.0077,
1652
+ "step": 2710
1653
+ },
1654
+ {
1655
+ "epoch": 2.09,
1656
+ "learning_rate": 1.943762781186094e-05,
1657
+ "loss": 1.0349,
1658
+ "step": 2720
1659
+ },
1660
+ {
1661
+ "epoch": 2.09,
1662
+ "learning_rate": 1.9386503067484663e-05,
1663
+ "loss": 0.9735,
1664
+ "step": 2730
1665
+ },
1666
+ {
1667
+ "epoch": 2.1,
1668
+ "learning_rate": 1.9335378323108383e-05,
1669
+ "loss": 1.0829,
1670
+ "step": 2740
1671
+ },
1672
+ {
1673
+ "epoch": 2.11,
1674
+ "learning_rate": 1.9284253578732106e-05,
1675
+ "loss": 1.0084,
1676
+ "step": 2750
1677
+ },
1678
+ {
1679
+ "epoch": 2.12,
1680
+ "learning_rate": 1.923312883435583e-05,
1681
+ "loss": 1.1453,
1682
+ "step": 2760
1683
+ },
1684
+ {
1685
+ "epoch": 2.12,
1686
+ "learning_rate": 1.9182004089979552e-05,
1687
+ "loss": 1.0449,
1688
+ "step": 2770
1689
+ },
1690
+ {
1691
+ "epoch": 2.13,
1692
+ "learning_rate": 1.9130879345603272e-05,
1693
+ "loss": 1.0324,
1694
+ "step": 2780
1695
+ },
1696
+ {
1697
+ "epoch": 2.14,
1698
+ "learning_rate": 1.9079754601226995e-05,
1699
+ "loss": 1.0473,
1700
+ "step": 2790
1701
+ },
1702
+ {
1703
+ "epoch": 2.15,
1704
+ "learning_rate": 1.9028629856850715e-05,
1705
+ "loss": 1.0417,
1706
+ "step": 2800
1707
+ },
1708
+ {
1709
+ "epoch": 2.15,
1710
+ "learning_rate": 1.8977505112474438e-05,
1711
+ "loss": 1.0275,
1712
+ "step": 2810
1713
+ },
1714
+ {
1715
+ "epoch": 2.16,
1716
+ "learning_rate": 1.8926380368098158e-05,
1717
+ "loss": 0.9963,
1718
+ "step": 2820
1719
+ },
1720
+ {
1721
+ "epoch": 2.17,
1722
+ "learning_rate": 1.8875255623721885e-05,
1723
+ "loss": 1.0089,
1724
+ "step": 2830
1725
+ },
1726
+ {
1727
+ "epoch": 2.18,
1728
+ "learning_rate": 1.8824130879345604e-05,
1729
+ "loss": 1.0995,
1730
+ "step": 2840
1731
+ },
1732
+ {
1733
+ "epoch": 2.19,
1734
+ "learning_rate": 1.8773006134969328e-05,
1735
+ "loss": 1.0087,
1736
+ "step": 2850
1737
+ },
1738
+ {
1739
+ "epoch": 2.19,
1740
+ "learning_rate": 1.8721881390593047e-05,
1741
+ "loss": 1.0147,
1742
+ "step": 2860
1743
+ },
1744
+ {
1745
+ "epoch": 2.2,
1746
+ "learning_rate": 1.8670756646216767e-05,
1747
+ "loss": 1.0342,
1748
+ "step": 2870
1749
+ },
1750
+ {
1751
+ "epoch": 2.21,
1752
+ "learning_rate": 1.861963190184049e-05,
1753
+ "loss": 1.0367,
1754
+ "step": 2880
1755
+ },
1756
+ {
1757
+ "epoch": 2.22,
1758
+ "learning_rate": 1.856850715746421e-05,
1759
+ "loss": 1.0202,
1760
+ "step": 2890
1761
+ },
1762
+ {
1763
+ "epoch": 2.22,
1764
+ "learning_rate": 1.8517382413087937e-05,
1765
+ "loss": 1.0607,
1766
+ "step": 2900
1767
+ },
1768
+ {
1769
+ "epoch": 2.23,
1770
+ "learning_rate": 1.8466257668711657e-05,
1771
+ "loss": 1.0357,
1772
+ "step": 2910
1773
+ },
1774
+ {
1775
+ "epoch": 2.24,
1776
+ "learning_rate": 1.841513292433538e-05,
1777
+ "loss": 1.0233,
1778
+ "step": 2920
1779
+ },
1780
+ {
1781
+ "epoch": 2.25,
1782
+ "learning_rate": 1.83640081799591e-05,
1783
+ "loss": 1.0346,
1784
+ "step": 2930
1785
+ },
1786
+ {
1787
+ "epoch": 2.25,
1788
+ "learning_rate": 1.8312883435582823e-05,
1789
+ "loss": 1.03,
1790
+ "step": 2940
1791
+ },
1792
+ {
1793
+ "epoch": 2.26,
1794
+ "learning_rate": 1.8261758691206543e-05,
1795
+ "loss": 0.9792,
1796
+ "step": 2950
1797
+ },
1798
+ {
1799
+ "epoch": 2.27,
1800
+ "learning_rate": 1.8210633946830266e-05,
1801
+ "loss": 1.0475,
1802
+ "step": 2960
1803
+ },
1804
+ {
1805
+ "epoch": 2.28,
1806
+ "learning_rate": 1.815950920245399e-05,
1807
+ "loss": 1.0535,
1808
+ "step": 2970
1809
+ },
1810
+ {
1811
+ "epoch": 2.29,
1812
+ "learning_rate": 1.8108384458077712e-05,
1813
+ "loss": 1.0164,
1814
+ "step": 2980
1815
+ },
1816
+ {
1817
+ "epoch": 2.29,
1818
+ "learning_rate": 1.8057259713701432e-05,
1819
+ "loss": 1.0234,
1820
+ "step": 2990
1821
+ },
1822
+ {
1823
+ "epoch": 2.3,
1824
+ "learning_rate": 1.8006134969325155e-05,
1825
+ "loss": 1.0178,
1826
+ "step": 3000
1827
+ },
1828
+ {
1829
+ "epoch": 2.31,
1830
+ "learning_rate": 1.7955010224948875e-05,
1831
+ "loss": 1.0093,
1832
+ "step": 3010
1833
+ },
1834
+ {
1835
+ "epoch": 2.32,
1836
+ "learning_rate": 1.7903885480572595e-05,
1837
+ "loss": 1.0081,
1838
+ "step": 3020
1839
+ },
1840
+ {
1841
+ "epoch": 2.32,
1842
+ "learning_rate": 1.785276073619632e-05,
1843
+ "loss": 0.9826,
1844
+ "step": 3030
1845
+ },
1846
+ {
1847
+ "epoch": 2.33,
1848
+ "learning_rate": 1.780163599182004e-05,
1849
+ "loss": 0.964,
1850
+ "step": 3040
1851
+ },
1852
+ {
1853
+ "epoch": 2.34,
1854
+ "learning_rate": 1.7750511247443764e-05,
1855
+ "loss": 0.9796,
1856
+ "step": 3050
1857
+ },
1858
+ {
1859
+ "epoch": 2.35,
1860
+ "learning_rate": 1.7699386503067484e-05,
1861
+ "loss": 1.0112,
1862
+ "step": 3060
1863
+ },
1864
+ {
1865
+ "epoch": 2.35,
1866
+ "learning_rate": 1.7648261758691207e-05,
1867
+ "loss": 0.9958,
1868
+ "step": 3070
1869
+ },
1870
+ {
1871
+ "epoch": 2.36,
1872
+ "learning_rate": 1.7597137014314927e-05,
1873
+ "loss": 1.0626,
1874
+ "step": 3080
1875
+ },
1876
+ {
1877
+ "epoch": 2.37,
1878
+ "learning_rate": 1.754601226993865e-05,
1879
+ "loss": 0.949,
1880
+ "step": 3090
1881
+ },
1882
+ {
1883
+ "epoch": 2.38,
1884
+ "learning_rate": 1.7494887525562373e-05,
1885
+ "loss": 0.9996,
1886
+ "step": 3100
1887
+ },
1888
+ {
1889
+ "epoch": 2.38,
1890
+ "learning_rate": 1.7443762781186097e-05,
1891
+ "loss": 0.923,
1892
+ "step": 3110
1893
+ },
1894
+ {
1895
+ "epoch": 2.39,
1896
+ "learning_rate": 1.7392638036809816e-05,
1897
+ "loss": 1.0195,
1898
+ "step": 3120
1899
+ },
1900
+ {
1901
+ "epoch": 2.4,
1902
+ "learning_rate": 1.734151329243354e-05,
1903
+ "loss": 1.0198,
1904
+ "step": 3130
1905
+ },
1906
+ {
1907
+ "epoch": 2.41,
1908
+ "learning_rate": 1.729038854805726e-05,
1909
+ "loss": 1.0267,
1910
+ "step": 3140
1911
+ },
1912
+ {
1913
+ "epoch": 2.42,
1914
+ "learning_rate": 1.7239263803680983e-05,
1915
+ "loss": 0.9649,
1916
+ "step": 3150
1917
+ },
1918
+ {
1919
+ "epoch": 2.42,
1920
+ "learning_rate": 1.7188139059304702e-05,
1921
+ "loss": 1.0171,
1922
+ "step": 3160
1923
+ },
1924
+ {
1925
+ "epoch": 2.43,
1926
+ "learning_rate": 1.7137014314928426e-05,
1927
+ "loss": 1.0343,
1928
+ "step": 3170
1929
+ },
1930
+ {
1931
+ "epoch": 2.44,
1932
+ "learning_rate": 1.708588957055215e-05,
1933
+ "loss": 1.0254,
1934
+ "step": 3180
1935
+ },
1936
+ {
1937
+ "epoch": 2.45,
1938
+ "learning_rate": 1.703476482617587e-05,
1939
+ "loss": 1.0452,
1940
+ "step": 3190
1941
+ },
1942
+ {
1943
+ "epoch": 2.45,
1944
+ "learning_rate": 1.6983640081799592e-05,
1945
+ "loss": 1.0527,
1946
+ "step": 3200
1947
+ },
1948
+ {
1949
+ "epoch": 2.46,
1950
+ "learning_rate": 1.693251533742331e-05,
1951
+ "loss": 0.9567,
1952
+ "step": 3210
1953
+ },
1954
+ {
1955
+ "epoch": 2.47,
1956
+ "learning_rate": 1.6881390593047035e-05,
1957
+ "loss": 0.9941,
1958
+ "step": 3220
1959
+ },
1960
+ {
1961
+ "epoch": 2.48,
1962
+ "learning_rate": 1.6830265848670755e-05,
1963
+ "loss": 1.0011,
1964
+ "step": 3230
1965
+ },
1966
+ {
1967
+ "epoch": 2.48,
1968
+ "learning_rate": 1.677914110429448e-05,
1969
+ "loss": 0.9983,
1970
+ "step": 3240
1971
+ },
1972
+ {
1973
+ "epoch": 2.49,
1974
+ "learning_rate": 1.67280163599182e-05,
1975
+ "loss": 1.0388,
1976
+ "step": 3250
1977
+ },
1978
+ {
1979
+ "epoch": 2.5,
1980
+ "learning_rate": 1.6676891615541924e-05,
1981
+ "loss": 0.9829,
1982
+ "step": 3260
1983
+ },
1984
+ {
1985
+ "epoch": 2.51,
1986
+ "learning_rate": 1.6625766871165644e-05,
1987
+ "loss": 1.047,
1988
+ "step": 3270
1989
+ },
1990
+ {
1991
+ "epoch": 2.52,
1992
+ "learning_rate": 1.6574642126789367e-05,
1993
+ "loss": 0.9815,
1994
+ "step": 3280
1995
+ },
1996
+ {
1997
+ "epoch": 2.52,
1998
+ "learning_rate": 1.6523517382413087e-05,
1999
+ "loss": 0.9917,
2000
+ "step": 3290
2001
+ },
2002
+ {
2003
+ "epoch": 2.53,
2004
+ "learning_rate": 1.647239263803681e-05,
2005
+ "loss": 1.0505,
2006
+ "step": 3300
2007
+ },
2008
+ {
2009
+ "epoch": 2.54,
2010
+ "learning_rate": 1.6421267893660533e-05,
2011
+ "loss": 0.9786,
2012
+ "step": 3310
2013
+ },
2014
+ {
2015
+ "epoch": 2.55,
2016
+ "learning_rate": 1.6370143149284253e-05,
2017
+ "loss": 0.9499,
2018
+ "step": 3320
2019
+ },
2020
+ {
2021
+ "epoch": 2.55,
2022
+ "learning_rate": 1.6319018404907976e-05,
2023
+ "loss": 1.0274,
2024
+ "step": 3330
2025
+ },
2026
+ {
2027
+ "epoch": 2.56,
2028
+ "learning_rate": 1.6267893660531696e-05,
2029
+ "loss": 1.0012,
2030
+ "step": 3340
2031
+ },
2032
+ {
2033
+ "epoch": 2.57,
2034
+ "learning_rate": 1.621676891615542e-05,
2035
+ "loss": 1.0759,
2036
+ "step": 3350
2037
+ },
2038
+ {
2039
+ "epoch": 2.58,
2040
+ "learning_rate": 1.616564417177914e-05,
2041
+ "loss": 1.009,
2042
+ "step": 3360
2043
+ },
2044
+ {
2045
+ "epoch": 2.58,
2046
+ "learning_rate": 1.6114519427402862e-05,
2047
+ "loss": 0.9257,
2048
+ "step": 3370
2049
+ },
2050
+ {
2051
+ "epoch": 2.59,
2052
+ "learning_rate": 1.6063394683026586e-05,
2053
+ "loss": 1.0042,
2054
+ "step": 3380
2055
+ },
2056
+ {
2057
+ "epoch": 2.6,
2058
+ "learning_rate": 1.601226993865031e-05,
2059
+ "loss": 1.0145,
2060
+ "step": 3390
2061
+ },
2062
+ {
2063
+ "epoch": 2.61,
2064
+ "learning_rate": 1.596114519427403e-05,
2065
+ "loss": 0.9835,
2066
+ "step": 3400
2067
+ },
2068
+ {
2069
+ "epoch": 2.61,
2070
+ "learning_rate": 1.591002044989775e-05,
2071
+ "loss": 1.0112,
2072
+ "step": 3410
2073
+ },
2074
+ {
2075
+ "epoch": 2.62,
2076
+ "learning_rate": 1.585889570552147e-05,
2077
+ "loss": 1.0209,
2078
+ "step": 3420
2079
+ },
2080
+ {
2081
+ "epoch": 2.63,
2082
+ "learning_rate": 1.5807770961145195e-05,
2083
+ "loss": 1.1021,
2084
+ "step": 3430
2085
+ },
2086
+ {
2087
+ "epoch": 2.64,
2088
+ "learning_rate": 1.5756646216768918e-05,
2089
+ "loss": 0.9853,
2090
+ "step": 3440
2091
+ },
2092
+ {
2093
+ "epoch": 2.65,
2094
+ "learning_rate": 1.570552147239264e-05,
2095
+ "loss": 0.9676,
2096
+ "step": 3450
2097
+ },
2098
+ {
2099
+ "epoch": 2.65,
2100
+ "learning_rate": 1.565439672801636e-05,
2101
+ "loss": 1.0247,
2102
+ "step": 3460
2103
+ },
2104
+ {
2105
+ "epoch": 2.66,
2106
+ "learning_rate": 1.560327198364008e-05,
2107
+ "loss": 0.9681,
2108
+ "step": 3470
2109
+ },
2110
+ {
2111
+ "epoch": 2.67,
2112
+ "learning_rate": 1.5552147239263804e-05,
2113
+ "loss": 0.9989,
2114
+ "step": 3480
2115
+ },
2116
+ {
2117
+ "epoch": 2.68,
2118
+ "learning_rate": 1.5501022494887524e-05,
2119
+ "loss": 1.0103,
2120
+ "step": 3490
2121
+ },
2122
+ {
2123
+ "epoch": 2.68,
2124
+ "learning_rate": 1.5449897750511247e-05,
2125
+ "loss": 0.9726,
2126
+ "step": 3500
2127
+ },
2128
+ {
2129
+ "epoch": 2.69,
2130
+ "learning_rate": 1.539877300613497e-05,
2131
+ "loss": 0.9575,
2132
+ "step": 3510
2133
+ },
2134
+ {
2135
+ "epoch": 2.7,
2136
+ "learning_rate": 1.5347648261758693e-05,
2137
+ "loss": 1.0127,
2138
+ "step": 3520
2139
+ },
2140
+ {
2141
+ "epoch": 2.71,
2142
+ "learning_rate": 1.5296523517382413e-05,
2143
+ "loss": 0.9866,
2144
+ "step": 3530
2145
+ },
2146
+ {
2147
+ "epoch": 2.71,
2148
+ "learning_rate": 1.5245398773006136e-05,
2149
+ "loss": 0.9727,
2150
+ "step": 3540
2151
+ },
2152
+ {
2153
+ "epoch": 2.72,
2154
+ "learning_rate": 1.5194274028629856e-05,
2155
+ "loss": 1.0087,
2156
+ "step": 3550
2157
+ },
2158
+ {
2159
+ "epoch": 2.73,
2160
+ "learning_rate": 1.5143149284253581e-05,
2161
+ "loss": 1.0048,
2162
+ "step": 3560
2163
+ },
2164
+ {
2165
+ "epoch": 2.74,
2166
+ "learning_rate": 1.50920245398773e-05,
2167
+ "loss": 1.0397,
2168
+ "step": 3570
2169
+ },
2170
+ {
2171
+ "epoch": 2.75,
2172
+ "learning_rate": 1.5040899795501024e-05,
2173
+ "loss": 0.9903,
2174
+ "step": 3580
2175
+ },
2176
+ {
2177
+ "epoch": 2.75,
2178
+ "learning_rate": 1.4989775051124744e-05,
2179
+ "loss": 1.0429,
2180
+ "step": 3590
2181
+ },
2182
+ {
2183
+ "epoch": 2.76,
2184
+ "learning_rate": 1.4938650306748467e-05,
2185
+ "loss": 1.0008,
2186
+ "step": 3600
2187
+ },
2188
+ {
2189
+ "epoch": 2.77,
2190
+ "learning_rate": 1.4887525562372188e-05,
2191
+ "loss": 0.9799,
2192
+ "step": 3610
2193
+ },
2194
+ {
2195
+ "epoch": 2.78,
2196
+ "learning_rate": 1.483640081799591e-05,
2197
+ "loss": 0.9668,
2198
+ "step": 3620
2199
+ },
2200
+ {
2201
+ "epoch": 2.78,
2202
+ "learning_rate": 1.4785276073619633e-05,
2203
+ "loss": 1.0046,
2204
+ "step": 3630
2205
+ },
2206
+ {
2207
+ "epoch": 2.79,
2208
+ "learning_rate": 1.4734151329243355e-05,
2209
+ "loss": 1.0146,
2210
+ "step": 3640
2211
+ },
2212
+ {
2213
+ "epoch": 2.8,
2214
+ "learning_rate": 1.4683026584867076e-05,
2215
+ "loss": 0.9645,
2216
+ "step": 3650
2217
+ },
2218
+ {
2219
+ "epoch": 2.81,
2220
+ "learning_rate": 1.46319018404908e-05,
2221
+ "loss": 1.0449,
2222
+ "step": 3660
2223
+ },
2224
+ {
2225
+ "epoch": 2.81,
2226
+ "learning_rate": 1.458077709611452e-05,
2227
+ "loss": 1.0,
2228
+ "step": 3670
2229
+ },
2230
+ {
2231
+ "epoch": 2.82,
2232
+ "learning_rate": 1.4529652351738242e-05,
2233
+ "loss": 0.9965,
2234
+ "step": 3680
2235
+ },
2236
+ {
2237
+ "epoch": 2.83,
2238
+ "learning_rate": 1.4478527607361964e-05,
2239
+ "loss": 1.0086,
2240
+ "step": 3690
2241
+ },
2242
+ {
2243
+ "epoch": 2.84,
2244
+ "learning_rate": 1.4427402862985685e-05,
2245
+ "loss": 1.0889,
2246
+ "step": 3700
2247
+ },
2248
+ {
2249
+ "epoch": 2.84,
2250
+ "learning_rate": 1.4376278118609407e-05,
2251
+ "loss": 1.0154,
2252
+ "step": 3710
2253
+ },
2254
+ {
2255
+ "epoch": 2.85,
2256
+ "learning_rate": 1.4325153374233128e-05,
2257
+ "loss": 1.017,
2258
+ "step": 3720
2259
+ },
2260
+ {
2261
+ "epoch": 2.86,
2262
+ "learning_rate": 1.4274028629856851e-05,
2263
+ "loss": 0.9491,
2264
+ "step": 3730
2265
+ },
2266
+ {
2267
+ "epoch": 2.87,
2268
+ "learning_rate": 1.4222903885480573e-05,
2269
+ "loss": 0.9901,
2270
+ "step": 3740
2271
+ },
2272
+ {
2273
+ "epoch": 2.88,
2274
+ "learning_rate": 1.4171779141104294e-05,
2275
+ "loss": 1.0308,
2276
+ "step": 3750
2277
+ },
2278
+ {
2279
+ "epoch": 2.88,
2280
+ "learning_rate": 1.4120654396728016e-05,
2281
+ "loss": 0.9649,
2282
+ "step": 3760
2283
+ },
2284
+ {
2285
+ "epoch": 2.89,
2286
+ "learning_rate": 1.4069529652351739e-05,
2287
+ "loss": 0.9903,
2288
+ "step": 3770
2289
+ },
2290
+ {
2291
+ "epoch": 2.9,
2292
+ "learning_rate": 1.401840490797546e-05,
2293
+ "loss": 0.9283,
2294
+ "step": 3780
2295
+ },
2296
+ {
2297
+ "epoch": 2.91,
2298
+ "learning_rate": 1.3967280163599182e-05,
2299
+ "loss": 0.9714,
2300
+ "step": 3790
2301
+ },
2302
+ {
2303
+ "epoch": 2.91,
2304
+ "learning_rate": 1.3916155419222905e-05,
2305
+ "loss": 0.9989,
2306
+ "step": 3800
2307
+ },
2308
+ {
2309
+ "epoch": 2.92,
2310
+ "learning_rate": 1.3865030674846627e-05,
2311
+ "loss": 0.9711,
2312
+ "step": 3810
2313
+ },
2314
+ {
2315
+ "epoch": 2.93,
2316
+ "learning_rate": 1.3813905930470348e-05,
2317
+ "loss": 1.0136,
2318
+ "step": 3820
2319
+ },
2320
+ {
2321
+ "epoch": 2.94,
2322
+ "learning_rate": 1.3762781186094071e-05,
2323
+ "loss": 0.9706,
2324
+ "step": 3830
2325
+ },
2326
+ {
2327
+ "epoch": 2.94,
2328
+ "learning_rate": 1.3711656441717793e-05,
2329
+ "loss": 1.0298,
2330
+ "step": 3840
2331
+ },
2332
+ {
2333
+ "epoch": 2.95,
2334
+ "learning_rate": 1.3660531697341513e-05,
2335
+ "loss": 0.9527,
2336
+ "step": 3850
2337
+ },
2338
+ {
2339
+ "epoch": 2.96,
2340
+ "learning_rate": 1.3609406952965234e-05,
2341
+ "loss": 0.9867,
2342
+ "step": 3860
2343
+ },
2344
+ {
2345
+ "epoch": 2.97,
2346
+ "learning_rate": 1.3558282208588957e-05,
2347
+ "loss": 0.9672,
2348
+ "step": 3870
2349
+ },
2350
+ {
2351
+ "epoch": 2.98,
2352
+ "learning_rate": 1.3507157464212679e-05,
2353
+ "loss": 0.9654,
2354
+ "step": 3880
2355
+ },
2356
+ {
2357
+ "epoch": 2.98,
2358
+ "learning_rate": 1.34560327198364e-05,
2359
+ "loss": 0.9646,
2360
+ "step": 3890
2361
+ },
2362
+ {
2363
+ "epoch": 2.99,
2364
+ "learning_rate": 1.3404907975460124e-05,
2365
+ "loss": 0.9597,
2366
+ "step": 3900
2367
+ },
2368
+ {
2369
+ "epoch": 3.0,
2370
+ "learning_rate": 1.3353783231083845e-05,
2371
+ "loss": 0.9775,
2372
+ "step": 3910
2373
+ },
2374
+ {
2375
+ "epoch": 3.0,
2376
+ "eval_accuracy": 0.6522530723714156,
2377
+ "eval_loss": 0.9669816493988037,
2378
+ "eval_runtime": 294.3351,
2379
+ "eval_samples_per_second": 141.821,
2380
+ "eval_steps_per_second": 70.912,
2381
+ "step": 3912
2382
+ },
2383
+ {
2384
+ "epoch": 3.01,
2385
+ "learning_rate": 1.3302658486707567e-05,
2386
+ "loss": 1.0613,
2387
+ "step": 3920
2388
+ },
2389
+ {
2390
+ "epoch": 3.01,
2391
+ "learning_rate": 1.3251533742331288e-05,
2392
+ "loss": 0.9874,
2393
+ "step": 3930
2394
+ },
2395
+ {
2396
+ "epoch": 3.02,
2397
+ "learning_rate": 1.3200408997955011e-05,
2398
+ "loss": 0.9708,
2399
+ "step": 3940
2400
+ },
2401
+ {
2402
+ "epoch": 3.03,
2403
+ "learning_rate": 1.3149284253578733e-05,
2404
+ "loss": 0.9617,
2405
+ "step": 3950
2406
+ },
2407
+ {
2408
+ "epoch": 3.04,
2409
+ "learning_rate": 1.3098159509202454e-05,
2410
+ "loss": 0.968,
2411
+ "step": 3960
2412
+ },
2413
+ {
2414
+ "epoch": 3.04,
2415
+ "learning_rate": 1.3047034764826177e-05,
2416
+ "loss": 0.9683,
2417
+ "step": 3970
2418
+ },
2419
+ {
2420
+ "epoch": 3.05,
2421
+ "learning_rate": 1.2995910020449899e-05,
2422
+ "loss": 0.9496,
2423
+ "step": 3980
2424
+ },
2425
+ {
2426
+ "epoch": 3.06,
2427
+ "learning_rate": 1.294478527607362e-05,
2428
+ "loss": 0.919,
2429
+ "step": 3990
2430
+ },
2431
+ {
2432
+ "epoch": 3.07,
2433
+ "learning_rate": 1.2893660531697342e-05,
2434
+ "loss": 0.9563,
2435
+ "step": 4000
2436
+ },
2437
+ {
2438
+ "epoch": 3.08,
2439
+ "learning_rate": 1.2842535787321063e-05,
2440
+ "loss": 0.9115,
2441
+ "step": 4010
2442
+ },
2443
+ {
2444
+ "epoch": 3.08,
2445
+ "learning_rate": 1.2791411042944785e-05,
2446
+ "loss": 0.9078,
2447
+ "step": 4020
2448
+ },
2449
+ {
2450
+ "epoch": 3.09,
2451
+ "learning_rate": 1.2740286298568506e-05,
2452
+ "loss": 0.9824,
2453
+ "step": 4030
2454
+ },
2455
+ {
2456
+ "epoch": 3.1,
2457
+ "learning_rate": 1.268916155419223e-05,
2458
+ "loss": 0.9953,
2459
+ "step": 4040
2460
+ },
2461
+ {
2462
+ "epoch": 3.11,
2463
+ "learning_rate": 1.2638036809815951e-05,
2464
+ "loss": 1.004,
2465
+ "step": 4050
2466
+ },
2467
+ {
2468
+ "epoch": 3.11,
2469
+ "learning_rate": 1.2586912065439673e-05,
2470
+ "loss": 0.9191,
2471
+ "step": 4060
2472
+ },
2473
+ {
2474
+ "epoch": 3.12,
2475
+ "learning_rate": 1.2535787321063396e-05,
2476
+ "loss": 0.9923,
2477
+ "step": 4070
2478
+ },
2479
+ {
2480
+ "epoch": 3.13,
2481
+ "learning_rate": 1.2484662576687117e-05,
2482
+ "loss": 0.9668,
2483
+ "step": 4080
2484
+ },
2485
+ {
2486
+ "epoch": 3.14,
2487
+ "learning_rate": 1.2433537832310839e-05,
2488
+ "loss": 1.0262,
2489
+ "step": 4090
2490
+ },
2491
+ {
2492
+ "epoch": 3.14,
2493
+ "learning_rate": 1.238241308793456e-05,
2494
+ "loss": 1.0177,
2495
+ "step": 4100
2496
+ },
2497
+ {
2498
+ "epoch": 3.15,
2499
+ "learning_rate": 1.2331288343558283e-05,
2500
+ "loss": 0.9461,
2501
+ "step": 4110
2502
+ },
2503
+ {
2504
+ "epoch": 3.16,
2505
+ "learning_rate": 1.2280163599182005e-05,
2506
+ "loss": 1.0468,
2507
+ "step": 4120
2508
+ },
2509
+ {
2510
+ "epoch": 3.17,
2511
+ "learning_rate": 1.2229038854805726e-05,
2512
+ "loss": 0.9567,
2513
+ "step": 4130
2514
+ },
2515
+ {
2516
+ "epoch": 3.17,
2517
+ "learning_rate": 1.217791411042945e-05,
2518
+ "loss": 0.936,
2519
+ "step": 4140
2520
+ },
2521
+ {
2522
+ "epoch": 3.18,
2523
+ "learning_rate": 1.212678936605317e-05,
2524
+ "loss": 0.9586,
2525
+ "step": 4150
2526
+ },
2527
+ {
2528
+ "epoch": 3.19,
2529
+ "learning_rate": 1.2075664621676891e-05,
2530
+ "loss": 0.964,
2531
+ "step": 4160
2532
+ },
2533
+ {
2534
+ "epoch": 3.2,
2535
+ "learning_rate": 1.2024539877300614e-05,
2536
+ "loss": 0.9221,
2537
+ "step": 4170
2538
+ },
2539
+ {
2540
+ "epoch": 3.21,
2541
+ "learning_rate": 1.1973415132924336e-05,
2542
+ "loss": 0.9604,
2543
+ "step": 4180
2544
+ },
2545
+ {
2546
+ "epoch": 3.21,
2547
+ "learning_rate": 1.1922290388548057e-05,
2548
+ "loss": 1.0227,
2549
+ "step": 4190
2550
+ },
2551
+ {
2552
+ "epoch": 3.22,
2553
+ "learning_rate": 1.1871165644171779e-05,
2554
+ "loss": 0.9817,
2555
+ "step": 4200
2556
+ },
2557
+ {
2558
+ "epoch": 3.23,
2559
+ "learning_rate": 1.1820040899795502e-05,
2560
+ "loss": 0.9773,
2561
+ "step": 4210
2562
+ },
2563
+ {
2564
+ "epoch": 3.24,
2565
+ "learning_rate": 1.1768916155419223e-05,
2566
+ "loss": 0.9468,
2567
+ "step": 4220
2568
+ },
2569
+ {
2570
+ "epoch": 3.24,
2571
+ "learning_rate": 1.1717791411042945e-05,
2572
+ "loss": 0.9356,
2573
+ "step": 4230
2574
+ },
2575
+ {
2576
+ "epoch": 3.25,
2577
+ "learning_rate": 1.1666666666666668e-05,
2578
+ "loss": 0.9818,
2579
+ "step": 4240
2580
+ },
2581
+ {
2582
+ "epoch": 3.26,
2583
+ "learning_rate": 1.161554192229039e-05,
2584
+ "loss": 0.9984,
2585
+ "step": 4250
2586
+ },
2587
+ {
2588
+ "epoch": 3.27,
2589
+ "learning_rate": 1.1564417177914111e-05,
2590
+ "loss": 1.0297,
2591
+ "step": 4260
2592
+ },
2593
+ {
2594
+ "epoch": 3.27,
2595
+ "learning_rate": 1.1513292433537832e-05,
2596
+ "loss": 0.9418,
2597
+ "step": 4270
2598
+ },
2599
+ {
2600
+ "epoch": 3.28,
2601
+ "learning_rate": 1.1462167689161556e-05,
2602
+ "loss": 0.9486,
2603
+ "step": 4280
2604
+ },
2605
+ {
2606
+ "epoch": 3.29,
2607
+ "learning_rate": 1.1411042944785277e-05,
2608
+ "loss": 0.9456,
2609
+ "step": 4290
2610
+ },
2611
+ {
2612
+ "epoch": 3.3,
2613
+ "learning_rate": 1.1359918200408997e-05,
2614
+ "loss": 0.9821,
2615
+ "step": 4300
2616
+ },
2617
+ {
2618
+ "epoch": 3.31,
2619
+ "learning_rate": 1.130879345603272e-05,
2620
+ "loss": 0.9672,
2621
+ "step": 4310
2622
+ },
2623
+ {
2624
+ "epoch": 3.31,
2625
+ "learning_rate": 1.1257668711656442e-05,
2626
+ "loss": 0.9769,
2627
+ "step": 4320
2628
+ },
2629
+ {
2630
+ "epoch": 3.32,
2631
+ "learning_rate": 1.1206543967280163e-05,
2632
+ "loss": 0.9806,
2633
+ "step": 4330
2634
+ },
2635
+ {
2636
+ "epoch": 3.33,
2637
+ "learning_rate": 1.1155419222903885e-05,
2638
+ "loss": 0.9755,
2639
+ "step": 4340
2640
+ },
2641
+ {
2642
+ "epoch": 3.34,
2643
+ "learning_rate": 1.1104294478527608e-05,
2644
+ "loss": 0.9019,
2645
+ "step": 4350
2646
+ },
2647
+ {
2648
+ "epoch": 3.34,
2649
+ "learning_rate": 1.105316973415133e-05,
2650
+ "loss": 0.9812,
2651
+ "step": 4360
2652
+ },
2653
+ {
2654
+ "epoch": 3.35,
2655
+ "learning_rate": 1.100204498977505e-05,
2656
+ "loss": 0.9808,
2657
+ "step": 4370
2658
+ },
2659
+ {
2660
+ "epoch": 3.36,
2661
+ "learning_rate": 1.0950920245398774e-05,
2662
+ "loss": 1.0226,
2663
+ "step": 4380
2664
+ },
2665
+ {
2666
+ "epoch": 3.37,
2667
+ "learning_rate": 1.0899795501022495e-05,
2668
+ "loss": 0.9444,
2669
+ "step": 4390
2670
+ },
2671
+ {
2672
+ "epoch": 3.37,
2673
+ "learning_rate": 1.0848670756646217e-05,
2674
+ "loss": 0.9383,
2675
+ "step": 4400
2676
+ },
2677
+ {
2678
+ "epoch": 3.38,
2679
+ "learning_rate": 1.079754601226994e-05,
2680
+ "loss": 0.8953,
2681
+ "step": 4410
2682
+ },
2683
+ {
2684
+ "epoch": 3.39,
2685
+ "learning_rate": 1.0746421267893662e-05,
2686
+ "loss": 0.9296,
2687
+ "step": 4420
2688
+ },
2689
+ {
2690
+ "epoch": 3.4,
2691
+ "learning_rate": 1.0695296523517383e-05,
2692
+ "loss": 0.967,
2693
+ "step": 4430
2694
+ },
2695
+ {
2696
+ "epoch": 3.4,
2697
+ "learning_rate": 1.0644171779141105e-05,
2698
+ "loss": 0.9711,
2699
+ "step": 4440
2700
+ },
2701
+ {
2702
+ "epoch": 3.41,
2703
+ "learning_rate": 1.0593047034764828e-05,
2704
+ "loss": 0.9028,
2705
+ "step": 4450
2706
+ },
2707
+ {
2708
+ "epoch": 3.42,
2709
+ "learning_rate": 1.0541922290388548e-05,
2710
+ "loss": 0.972,
2711
+ "step": 4460
2712
+ },
2713
+ {
2714
+ "epoch": 3.43,
2715
+ "learning_rate": 1.0490797546012269e-05,
2716
+ "loss": 0.9936,
2717
+ "step": 4470
2718
+ },
2719
+ {
2720
+ "epoch": 3.44,
2721
+ "learning_rate": 1.0439672801635992e-05,
2722
+ "loss": 1.0213,
2723
+ "step": 4480
2724
+ },
2725
+ {
2726
+ "epoch": 3.44,
2727
+ "learning_rate": 1.0388548057259714e-05,
2728
+ "loss": 0.9247,
2729
+ "step": 4490
2730
+ },
2731
+ {
2732
+ "epoch": 3.45,
2733
+ "learning_rate": 1.0337423312883435e-05,
2734
+ "loss": 0.9832,
2735
+ "step": 4500
2736
+ },
2737
+ {
2738
+ "epoch": 3.46,
2739
+ "learning_rate": 1.0286298568507157e-05,
2740
+ "loss": 0.9604,
2741
+ "step": 4510
2742
+ },
2743
+ {
2744
+ "epoch": 3.47,
2745
+ "learning_rate": 1.023517382413088e-05,
2746
+ "loss": 0.9799,
2747
+ "step": 4520
2748
+ },
2749
+ {
2750
+ "epoch": 3.47,
2751
+ "learning_rate": 1.0184049079754601e-05,
2752
+ "loss": 0.9318,
2753
+ "step": 4530
2754
+ },
2755
+ {
2756
+ "epoch": 3.48,
2757
+ "learning_rate": 1.0132924335378323e-05,
2758
+ "loss": 0.9727,
2759
+ "step": 4540
2760
+ },
2761
+ {
2762
+ "epoch": 3.49,
2763
+ "learning_rate": 1.0081799591002046e-05,
2764
+ "loss": 0.9929,
2765
+ "step": 4550
2766
+ },
2767
+ {
2768
+ "epoch": 3.5,
2769
+ "learning_rate": 1.0030674846625768e-05,
2770
+ "loss": 0.9488,
2771
+ "step": 4560
2772
+ },
2773
+ {
2774
+ "epoch": 3.5,
2775
+ "learning_rate": 9.979550102249489e-06,
2776
+ "loss": 0.97,
2777
+ "step": 4570
2778
+ },
2779
+ {
2780
+ "epoch": 3.51,
2781
+ "learning_rate": 9.928425357873212e-06,
2782
+ "loss": 0.953,
2783
+ "step": 4580
2784
+ },
2785
+ {
2786
+ "epoch": 3.52,
2787
+ "learning_rate": 9.877300613496934e-06,
2788
+ "loss": 0.9396,
2789
+ "step": 4590
2790
+ },
2791
+ {
2792
+ "epoch": 3.53,
2793
+ "learning_rate": 9.826175869120655e-06,
2794
+ "loss": 1.0,
2795
+ "step": 4600
2796
+ },
2797
+ {
2798
+ "epoch": 3.54,
2799
+ "learning_rate": 9.775051124744375e-06,
2800
+ "loss": 0.9872,
2801
+ "step": 4610
2802
+ },
2803
+ {
2804
+ "epoch": 3.54,
2805
+ "learning_rate": 9.723926380368098e-06,
2806
+ "loss": 0.9254,
2807
+ "step": 4620
2808
+ },
2809
+ {
2810
+ "epoch": 3.55,
2811
+ "learning_rate": 9.67280163599182e-06,
2812
+ "loss": 0.989,
2813
+ "step": 4630
2814
+ },
2815
+ {
2816
+ "epoch": 3.56,
2817
+ "learning_rate": 9.621676891615541e-06,
2818
+ "loss": 0.8786,
2819
+ "step": 4640
2820
+ },
2821
+ {
2822
+ "epoch": 3.57,
2823
+ "learning_rate": 9.570552147239264e-06,
2824
+ "loss": 0.9236,
2825
+ "step": 4650
2826
+ },
2827
+ {
2828
+ "epoch": 3.57,
2829
+ "learning_rate": 9.519427402862986e-06,
2830
+ "loss": 0.9162,
2831
+ "step": 4660
2832
+ },
2833
+ {
2834
+ "epoch": 3.58,
2835
+ "learning_rate": 9.468302658486707e-06,
2836
+ "loss": 0.9444,
2837
+ "step": 4670
2838
+ },
2839
+ {
2840
+ "epoch": 3.59,
2841
+ "learning_rate": 9.417177914110429e-06,
2842
+ "loss": 0.9698,
2843
+ "step": 4680
2844
+ },
2845
+ {
2846
+ "epoch": 3.6,
2847
+ "learning_rate": 9.366053169734152e-06,
2848
+ "loss": 1.0498,
2849
+ "step": 4690
2850
+ },
2851
+ {
2852
+ "epoch": 3.6,
2853
+ "learning_rate": 9.314928425357874e-06,
2854
+ "loss": 0.9529,
2855
+ "step": 4700
2856
+ },
2857
+ {
2858
+ "epoch": 3.61,
2859
+ "learning_rate": 9.263803680981595e-06,
2860
+ "loss": 0.9147,
2861
+ "step": 4710
2862
+ },
2863
+ {
2864
+ "epoch": 3.62,
2865
+ "learning_rate": 9.212678936605318e-06,
2866
+ "loss": 0.9755,
2867
+ "step": 4720
2868
+ },
2869
+ {
2870
+ "epoch": 3.63,
2871
+ "learning_rate": 9.16155419222904e-06,
2872
+ "loss": 1.007,
2873
+ "step": 4730
2874
+ },
2875
+ {
2876
+ "epoch": 3.63,
2877
+ "learning_rate": 9.110429447852761e-06,
2878
+ "loss": 0.9338,
2879
+ "step": 4740
2880
+ },
2881
+ {
2882
+ "epoch": 3.64,
2883
+ "learning_rate": 9.059304703476484e-06,
2884
+ "loss": 0.9356,
2885
+ "step": 4750
2886
+ },
2887
+ {
2888
+ "epoch": 3.65,
2889
+ "learning_rate": 9.008179959100204e-06,
2890
+ "loss": 0.9711,
2891
+ "step": 4760
2892
+ },
2893
+ {
2894
+ "epoch": 3.66,
2895
+ "learning_rate": 8.957055214723926e-06,
2896
+ "loss": 0.9662,
2897
+ "step": 4770
2898
+ },
2899
+ {
2900
+ "epoch": 3.67,
2901
+ "learning_rate": 8.905930470347647e-06,
2902
+ "loss": 0.9794,
2903
+ "step": 4780
2904
+ },
2905
+ {
2906
+ "epoch": 3.67,
2907
+ "learning_rate": 8.85480572597137e-06,
2908
+ "loss": 1.0016,
2909
+ "step": 4790
2910
+ },
2911
+ {
2912
+ "epoch": 3.68,
2913
+ "learning_rate": 8.803680981595092e-06,
2914
+ "loss": 0.9433,
2915
+ "step": 4800
2916
+ },
2917
+ {
2918
+ "epoch": 3.69,
2919
+ "learning_rate": 8.752556237218813e-06,
2920
+ "loss": 0.9584,
2921
+ "step": 4810
2922
+ },
2923
+ {
2924
+ "epoch": 3.7,
2925
+ "learning_rate": 8.701431492842537e-06,
2926
+ "loss": 0.9664,
2927
+ "step": 4820
2928
+ },
2929
+ {
2930
+ "epoch": 3.7,
2931
+ "learning_rate": 8.650306748466258e-06,
2932
+ "loss": 0.9164,
2933
+ "step": 4830
2934
+ },
2935
+ {
2936
+ "epoch": 3.71,
2937
+ "learning_rate": 8.59918200408998e-06,
2938
+ "loss": 0.9198,
2939
+ "step": 4840
2940
+ },
2941
+ {
2942
+ "epoch": 3.72,
2943
+ "learning_rate": 8.548057259713701e-06,
2944
+ "loss": 0.9766,
2945
+ "step": 4850
2946
+ },
2947
+ {
2948
+ "epoch": 3.73,
2949
+ "learning_rate": 8.496932515337424e-06,
2950
+ "loss": 0.9515,
2951
+ "step": 4860
2952
+ },
2953
+ {
2954
+ "epoch": 3.73,
2955
+ "learning_rate": 8.445807770961146e-06,
2956
+ "loss": 0.9538,
2957
+ "step": 4870
2958
+ },
2959
+ {
2960
+ "epoch": 3.74,
2961
+ "learning_rate": 8.394683026584867e-06,
2962
+ "loss": 0.9635,
2963
+ "step": 4880
2964
+ },
2965
+ {
2966
+ "epoch": 3.75,
2967
+ "learning_rate": 8.34355828220859e-06,
2968
+ "loss": 0.9385,
2969
+ "step": 4890
2970
+ },
2971
+ {
2972
+ "epoch": 3.76,
2973
+ "learning_rate": 8.292433537832312e-06,
2974
+ "loss": 0.9669,
2975
+ "step": 4900
2976
+ },
2977
+ {
2978
+ "epoch": 3.77,
2979
+ "learning_rate": 8.241308793456032e-06,
2980
+ "loss": 0.9672,
2981
+ "step": 4910
2982
+ },
2983
+ {
2984
+ "epoch": 3.77,
2985
+ "learning_rate": 8.190184049079755e-06,
2986
+ "loss": 0.9436,
2987
+ "step": 4920
2988
+ },
2989
+ {
2990
+ "epoch": 3.78,
2991
+ "learning_rate": 8.139059304703476e-06,
2992
+ "loss": 0.936,
2993
+ "step": 4930
2994
+ },
2995
+ {
2996
+ "epoch": 3.79,
2997
+ "learning_rate": 8.087934560327198e-06,
2998
+ "loss": 0.999,
2999
+ "step": 4940
3000
+ },
3001
+ {
3002
+ "epoch": 3.8,
3003
+ "learning_rate": 8.03680981595092e-06,
3004
+ "loss": 0.9724,
3005
+ "step": 4950
3006
+ },
3007
+ {
3008
+ "epoch": 3.8,
3009
+ "learning_rate": 7.985685071574643e-06,
3010
+ "loss": 0.9303,
3011
+ "step": 4960
3012
+ },
3013
+ {
3014
+ "epoch": 3.81,
3015
+ "learning_rate": 7.934560327198364e-06,
3016
+ "loss": 0.9705,
3017
+ "step": 4970
3018
+ },
3019
+ {
3020
+ "epoch": 3.82,
3021
+ "learning_rate": 7.883435582822086e-06,
3022
+ "loss": 0.9082,
3023
+ "step": 4980
3024
+ },
3025
+ {
3026
+ "epoch": 3.83,
3027
+ "learning_rate": 7.832310838445809e-06,
3028
+ "loss": 0.9417,
3029
+ "step": 4990
3030
+ },
3031
+ {
3032
+ "epoch": 3.83,
3033
+ "learning_rate": 7.78118609406953e-06,
3034
+ "loss": 0.9132,
3035
+ "step": 5000
3036
+ },
3037
+ {
3038
+ "epoch": 3.84,
3039
+ "learning_rate": 7.730061349693252e-06,
3040
+ "loss": 0.9844,
3041
+ "step": 5010
3042
+ },
3043
+ {
3044
+ "epoch": 3.85,
3045
+ "learning_rate": 7.678936605316973e-06,
3046
+ "loss": 0.9299,
3047
+ "step": 5020
3048
+ },
3049
+ {
3050
+ "epoch": 3.86,
3051
+ "learning_rate": 7.6278118609406965e-06,
3052
+ "loss": 0.9278,
3053
+ "step": 5030
3054
+ },
3055
+ {
3056
+ "epoch": 3.86,
3057
+ "learning_rate": 7.576687116564418e-06,
3058
+ "loss": 0.9335,
3059
+ "step": 5040
3060
+ },
3061
+ {
3062
+ "epoch": 3.87,
3063
+ "learning_rate": 7.52556237218814e-06,
3064
+ "loss": 0.9042,
3065
+ "step": 5050
3066
+ },
3067
+ {
3068
+ "epoch": 3.88,
3069
+ "learning_rate": 7.474437627811861e-06,
3070
+ "loss": 0.971,
3071
+ "step": 5060
3072
+ },
3073
+ {
3074
+ "epoch": 3.89,
3075
+ "learning_rate": 7.423312883435583e-06,
3076
+ "loss": 0.9237,
3077
+ "step": 5070
3078
+ },
3079
+ {
3080
+ "epoch": 3.9,
3081
+ "learning_rate": 7.372188139059305e-06,
3082
+ "loss": 0.9598,
3083
+ "step": 5080
3084
+ },
3085
+ {
3086
+ "epoch": 3.9,
3087
+ "learning_rate": 7.321063394683027e-06,
3088
+ "loss": 0.9941,
3089
+ "step": 5090
3090
+ },
3091
+ {
3092
+ "epoch": 3.91,
3093
+ "learning_rate": 7.269938650306749e-06,
3094
+ "loss": 0.9467,
3095
+ "step": 5100
3096
+ },
3097
+ {
3098
+ "epoch": 3.92,
3099
+ "learning_rate": 7.21881390593047e-06,
3100
+ "loss": 0.9411,
3101
+ "step": 5110
3102
+ },
3103
+ {
3104
+ "epoch": 3.93,
3105
+ "learning_rate": 7.1676891615541924e-06,
3106
+ "loss": 0.9672,
3107
+ "step": 5120
3108
+ },
3109
+ {
3110
+ "epoch": 3.93,
3111
+ "learning_rate": 7.116564417177914e-06,
3112
+ "loss": 0.8993,
3113
+ "step": 5130
3114
+ },
3115
+ {
3116
+ "epoch": 3.94,
3117
+ "learning_rate": 7.065439672801636e-06,
3118
+ "loss": 0.9833,
3119
+ "step": 5140
3120
+ },
3121
+ {
3122
+ "epoch": 3.95,
3123
+ "learning_rate": 7.014314928425359e-06,
3124
+ "loss": 0.9743,
3125
+ "step": 5150
3126
+ },
3127
+ {
3128
+ "epoch": 3.96,
3129
+ "learning_rate": 6.96319018404908e-06,
3130
+ "loss": 0.9413,
3131
+ "step": 5160
3132
+ },
3133
+ {
3134
+ "epoch": 3.96,
3135
+ "learning_rate": 6.912065439672802e-06,
3136
+ "loss": 0.9357,
3137
+ "step": 5170
3138
+ },
3139
+ {
3140
+ "epoch": 3.97,
3141
+ "learning_rate": 6.860940695296523e-06,
3142
+ "loss": 0.9368,
3143
+ "step": 5180
3144
+ },
3145
+ {
3146
+ "epoch": 3.98,
3147
+ "learning_rate": 6.8098159509202454e-06,
3148
+ "loss": 0.9175,
3149
+ "step": 5190
3150
+ },
3151
+ {
3152
+ "epoch": 3.99,
3153
+ "learning_rate": 6.758691206543968e-06,
3154
+ "loss": 0.9567,
3155
+ "step": 5200
3156
+ },
3157
+ {
3158
+ "epoch": 4.0,
3159
+ "learning_rate": 6.707566462167689e-06,
3160
+ "loss": 0.9542,
3161
+ "step": 5210
3162
+ },
3163
+ {
3164
+ "epoch": 4.0,
3165
+ "eval_accuracy": 0.6945835229858899,
3166
+ "eval_loss": 0.8809502124786377,
3167
+ "eval_runtime": 297.753,
3168
+ "eval_samples_per_second": 140.193,
3169
+ "eval_steps_per_second": 70.098,
3170
+ "step": 5216
3171
+ },
3172
+ {
3173
+ "epoch": 4.0,
3174
+ "learning_rate": 6.656441717791412e-06,
3175
+ "loss": 0.9247,
3176
+ "step": 5220
3177
+ },
3178
+ {
3179
+ "epoch": 4.01,
3180
+ "learning_rate": 6.605316973415133e-06,
3181
+ "loss": 0.9581,
3182
+ "step": 5230
3183
+ },
3184
+ {
3185
+ "epoch": 4.02,
3186
+ "learning_rate": 6.5541922290388555e-06,
3187
+ "loss": 0.9375,
3188
+ "step": 5240
3189
+ },
3190
+ {
3191
+ "epoch": 4.03,
3192
+ "learning_rate": 6.503067484662576e-06,
3193
+ "loss": 0.9146,
3194
+ "step": 5250
3195
+ },
3196
+ {
3197
+ "epoch": 4.03,
3198
+ "learning_rate": 6.4519427402862984e-06,
3199
+ "loss": 0.9437,
3200
+ "step": 5260
3201
+ },
3202
+ {
3203
+ "epoch": 4.04,
3204
+ "learning_rate": 6.400817995910021e-06,
3205
+ "loss": 0.9554,
3206
+ "step": 5270
3207
+ },
3208
+ {
3209
+ "epoch": 4.05,
3210
+ "learning_rate": 6.349693251533742e-06,
3211
+ "loss": 0.9519,
3212
+ "step": 5280
3213
+ },
3214
+ {
3215
+ "epoch": 4.06,
3216
+ "learning_rate": 6.298568507157465e-06,
3217
+ "loss": 0.9291,
3218
+ "step": 5290
3219
+ },
3220
+ {
3221
+ "epoch": 4.06,
3222
+ "learning_rate": 6.247443762781186e-06,
3223
+ "loss": 0.8958,
3224
+ "step": 5300
3225
+ },
3226
+ {
3227
+ "epoch": 4.07,
3228
+ "learning_rate": 6.1963190184049085e-06,
3229
+ "loss": 0.9188,
3230
+ "step": 5310
3231
+ },
3232
+ {
3233
+ "epoch": 4.08,
3234
+ "learning_rate": 6.14519427402863e-06,
3235
+ "loss": 0.949,
3236
+ "step": 5320
3237
+ },
3238
+ {
3239
+ "epoch": 4.09,
3240
+ "learning_rate": 6.0940695296523515e-06,
3241
+ "loss": 0.9866,
3242
+ "step": 5330
3243
+ },
3244
+ {
3245
+ "epoch": 4.1,
3246
+ "learning_rate": 6.042944785276074e-06,
3247
+ "loss": 0.9078,
3248
+ "step": 5340
3249
+ },
3250
+ {
3251
+ "epoch": 4.1,
3252
+ "learning_rate": 5.991820040899795e-06,
3253
+ "loss": 0.9243,
3254
+ "step": 5350
3255
+ },
3256
+ {
3257
+ "epoch": 4.11,
3258
+ "learning_rate": 5.940695296523518e-06,
3259
+ "loss": 0.8902,
3260
+ "step": 5360
3261
+ },
3262
+ {
3263
+ "epoch": 4.12,
3264
+ "learning_rate": 5.88957055214724e-06,
3265
+ "loss": 0.9754,
3266
+ "step": 5370
3267
+ },
3268
+ {
3269
+ "epoch": 4.13,
3270
+ "learning_rate": 5.8384458077709615e-06,
3271
+ "loss": 0.8929,
3272
+ "step": 5380
3273
+ },
3274
+ {
3275
+ "epoch": 4.13,
3276
+ "learning_rate": 5.787321063394684e-06,
3277
+ "loss": 0.9311,
3278
+ "step": 5390
3279
+ },
3280
+ {
3281
+ "epoch": 4.14,
3282
+ "learning_rate": 5.7361963190184045e-06,
3283
+ "loss": 0.9267,
3284
+ "step": 5400
3285
+ },
3286
+ {
3287
+ "epoch": 4.15,
3288
+ "learning_rate": 5.685071574642127e-06,
3289
+ "loss": 0.9586,
3290
+ "step": 5410
3291
+ },
3292
+ {
3293
+ "epoch": 4.16,
3294
+ "learning_rate": 5.633946830265848e-06,
3295
+ "loss": 0.9446,
3296
+ "step": 5420
3297
+ },
3298
+ {
3299
+ "epoch": 4.16,
3300
+ "learning_rate": 5.582822085889571e-06,
3301
+ "loss": 0.9085,
3302
+ "step": 5430
3303
+ },
3304
+ {
3305
+ "epoch": 4.17,
3306
+ "learning_rate": 5.531697341513293e-06,
3307
+ "loss": 0.9376,
3308
+ "step": 5440
3309
+ },
3310
+ {
3311
+ "epoch": 4.18,
3312
+ "learning_rate": 5.4805725971370145e-06,
3313
+ "loss": 0.9498,
3314
+ "step": 5450
3315
+ },
3316
+ {
3317
+ "epoch": 4.19,
3318
+ "learning_rate": 5.429447852760737e-06,
3319
+ "loss": 0.8981,
3320
+ "step": 5460
3321
+ },
3322
+ {
3323
+ "epoch": 4.19,
3324
+ "learning_rate": 5.378323108384458e-06,
3325
+ "loss": 0.961,
3326
+ "step": 5470
3327
+ },
3328
+ {
3329
+ "epoch": 4.2,
3330
+ "learning_rate": 5.32719836400818e-06,
3331
+ "loss": 0.9169,
3332
+ "step": 5480
3333
+ },
3334
+ {
3335
+ "epoch": 4.21,
3336
+ "learning_rate": 5.276073619631902e-06,
3337
+ "loss": 0.9219,
3338
+ "step": 5490
3339
+ },
3340
+ {
3341
+ "epoch": 4.22,
3342
+ "learning_rate": 5.224948875255624e-06,
3343
+ "loss": 0.9491,
3344
+ "step": 5500
3345
+ },
3346
+ {
3347
+ "epoch": 4.23,
3348
+ "learning_rate": 5.173824130879346e-06,
3349
+ "loss": 0.9504,
3350
+ "step": 5510
3351
+ },
3352
+ {
3353
+ "epoch": 4.23,
3354
+ "learning_rate": 5.1226993865030675e-06,
3355
+ "loss": 0.9376,
3356
+ "step": 5520
3357
+ },
3358
+ {
3359
+ "epoch": 4.24,
3360
+ "learning_rate": 5.07157464212679e-06,
3361
+ "loss": 0.9051,
3362
+ "step": 5530
3363
+ },
3364
+ {
3365
+ "epoch": 4.25,
3366
+ "learning_rate": 5.020449897750511e-06,
3367
+ "loss": 0.9578,
3368
+ "step": 5540
3369
+ },
3370
+ {
3371
+ "epoch": 4.26,
3372
+ "learning_rate": 4.969325153374233e-06,
3373
+ "loss": 0.9037,
3374
+ "step": 5550
3375
+ },
3376
+ {
3377
+ "epoch": 4.26,
3378
+ "learning_rate": 4.918200408997955e-06,
3379
+ "loss": 0.8804,
3380
+ "step": 5560
3381
+ },
3382
+ {
3383
+ "epoch": 4.27,
3384
+ "learning_rate": 4.867075664621677e-06,
3385
+ "loss": 0.9226,
3386
+ "step": 5570
3387
+ },
3388
+ {
3389
+ "epoch": 4.28,
3390
+ "learning_rate": 4.815950920245399e-06,
3391
+ "loss": 0.8782,
3392
+ "step": 5580
3393
+ },
3394
+ {
3395
+ "epoch": 4.29,
3396
+ "learning_rate": 4.7648261758691205e-06,
3397
+ "loss": 0.8848,
3398
+ "step": 5590
3399
+ },
3400
+ {
3401
+ "epoch": 4.29,
3402
+ "learning_rate": 4.713701431492843e-06,
3403
+ "loss": 0.929,
3404
+ "step": 5600
3405
+ },
3406
+ {
3407
+ "epoch": 4.3,
3408
+ "learning_rate": 4.662576687116565e-06,
3409
+ "loss": 0.9304,
3410
+ "step": 5610
3411
+ },
3412
+ {
3413
+ "epoch": 4.31,
3414
+ "learning_rate": 4.611451942740287e-06,
3415
+ "loss": 0.8859,
3416
+ "step": 5620
3417
+ },
3418
+ {
3419
+ "epoch": 4.32,
3420
+ "learning_rate": 4.560327198364008e-06,
3421
+ "loss": 0.9106,
3422
+ "step": 5630
3423
+ },
3424
+ {
3425
+ "epoch": 4.33,
3426
+ "learning_rate": 4.50920245398773e-06,
3427
+ "loss": 0.9991,
3428
+ "step": 5640
3429
+ },
3430
+ {
3431
+ "epoch": 4.33,
3432
+ "learning_rate": 4.458077709611452e-06,
3433
+ "loss": 0.9229,
3434
+ "step": 5650
3435
+ },
3436
+ {
3437
+ "epoch": 4.34,
3438
+ "learning_rate": 4.406952965235174e-06,
3439
+ "loss": 0.9479,
3440
+ "step": 5660
3441
+ },
3442
+ {
3443
+ "epoch": 4.35,
3444
+ "learning_rate": 4.355828220858896e-06,
3445
+ "loss": 0.9028,
3446
+ "step": 5670
3447
+ },
3448
+ {
3449
+ "epoch": 4.36,
3450
+ "learning_rate": 4.304703476482618e-06,
3451
+ "loss": 0.9395,
3452
+ "step": 5680
3453
+ },
3454
+ {
3455
+ "epoch": 4.36,
3456
+ "learning_rate": 4.25357873210634e-06,
3457
+ "loss": 0.9212,
3458
+ "step": 5690
3459
+ },
3460
+ {
3461
+ "epoch": 4.37,
3462
+ "learning_rate": 4.202453987730061e-06,
3463
+ "loss": 0.9599,
3464
+ "step": 5700
3465
+ },
3466
+ {
3467
+ "epoch": 4.38,
3468
+ "learning_rate": 4.151329243353783e-06,
3469
+ "loss": 0.9109,
3470
+ "step": 5710
3471
+ },
3472
+ {
3473
+ "epoch": 4.39,
3474
+ "learning_rate": 4.100204498977505e-06,
3475
+ "loss": 0.9477,
3476
+ "step": 5720
3477
+ },
3478
+ {
3479
+ "epoch": 4.39,
3480
+ "learning_rate": 4.049079754601227e-06,
3481
+ "loss": 0.8823,
3482
+ "step": 5730
3483
+ },
3484
+ {
3485
+ "epoch": 4.4,
3486
+ "learning_rate": 3.997955010224949e-06,
3487
+ "loss": 0.9318,
3488
+ "step": 5740
3489
+ },
3490
+ {
3491
+ "epoch": 4.41,
3492
+ "learning_rate": 3.946830265848671e-06,
3493
+ "loss": 0.8762,
3494
+ "step": 5750
3495
+ },
3496
+ {
3497
+ "epoch": 4.42,
3498
+ "learning_rate": 3.895705521472393e-06,
3499
+ "loss": 0.935,
3500
+ "step": 5760
3501
+ },
3502
+ {
3503
+ "epoch": 4.42,
3504
+ "learning_rate": 3.844580777096115e-06,
3505
+ "loss": 0.9151,
3506
+ "step": 5770
3507
+ },
3508
+ {
3509
+ "epoch": 4.43,
3510
+ "learning_rate": 3.793456032719836e-06,
3511
+ "loss": 0.9244,
3512
+ "step": 5780
3513
+ },
3514
+ {
3515
+ "epoch": 4.44,
3516
+ "learning_rate": 3.7423312883435584e-06,
3517
+ "loss": 0.9223,
3518
+ "step": 5790
3519
+ },
3520
+ {
3521
+ "epoch": 4.45,
3522
+ "learning_rate": 3.6912065439672803e-06,
3523
+ "loss": 0.8812,
3524
+ "step": 5800
3525
+ },
3526
+ {
3527
+ "epoch": 4.46,
3528
+ "learning_rate": 3.6400817995910022e-06,
3529
+ "loss": 0.9402,
3530
+ "step": 5810
3531
+ },
3532
+ {
3533
+ "epoch": 4.46,
3534
+ "learning_rate": 3.588957055214724e-06,
3535
+ "loss": 0.9159,
3536
+ "step": 5820
3537
+ },
3538
+ {
3539
+ "epoch": 4.47,
3540
+ "learning_rate": 3.537832310838446e-06,
3541
+ "loss": 0.9543,
3542
+ "step": 5830
3543
+ },
3544
+ {
3545
+ "epoch": 4.48,
3546
+ "learning_rate": 3.4867075664621676e-06,
3547
+ "loss": 0.9197,
3548
+ "step": 5840
3549
+ },
3550
+ {
3551
+ "epoch": 4.49,
3552
+ "learning_rate": 3.4355828220858895e-06,
3553
+ "loss": 0.945,
3554
+ "step": 5850
3555
+ },
3556
+ {
3557
+ "epoch": 4.49,
3558
+ "learning_rate": 3.384458077709612e-06,
3559
+ "loss": 0.9491,
3560
+ "step": 5860
3561
+ },
3562
+ {
3563
+ "epoch": 4.5,
3564
+ "learning_rate": 3.3333333333333333e-06,
3565
+ "loss": 0.8835,
3566
+ "step": 5870
3567
+ },
3568
+ {
3569
+ "epoch": 4.51,
3570
+ "learning_rate": 3.2822085889570552e-06,
3571
+ "loss": 0.877,
3572
+ "step": 5880
3573
+ },
3574
+ {
3575
+ "epoch": 4.52,
3576
+ "learning_rate": 3.231083844580777e-06,
3577
+ "loss": 0.9487,
3578
+ "step": 5890
3579
+ },
3580
+ {
3581
+ "epoch": 4.52,
3582
+ "learning_rate": 3.179959100204499e-06,
3583
+ "loss": 0.9424,
3584
+ "step": 5900
3585
+ },
3586
+ {
3587
+ "epoch": 4.53,
3588
+ "learning_rate": 3.1288343558282206e-06,
3589
+ "loss": 0.9083,
3590
+ "step": 5910
3591
+ },
3592
+ {
3593
+ "epoch": 4.54,
3594
+ "learning_rate": 3.077709611451943e-06,
3595
+ "loss": 0.9283,
3596
+ "step": 5920
3597
+ },
3598
+ {
3599
+ "epoch": 4.55,
3600
+ "learning_rate": 3.026584867075665e-06,
3601
+ "loss": 0.9977,
3602
+ "step": 5930
3603
+ },
3604
+ {
3605
+ "epoch": 4.56,
3606
+ "learning_rate": 2.9754601226993867e-06,
3607
+ "loss": 0.9019,
3608
+ "step": 5940
3609
+ },
3610
+ {
3611
+ "epoch": 4.56,
3612
+ "learning_rate": 2.9243353783231082e-06,
3613
+ "loss": 0.9302,
3614
+ "step": 5950
3615
+ },
3616
+ {
3617
+ "epoch": 4.57,
3618
+ "learning_rate": 2.87321063394683e-06,
3619
+ "loss": 0.9267,
3620
+ "step": 5960
3621
+ },
3622
+ {
3623
+ "epoch": 4.58,
3624
+ "learning_rate": 2.8220858895705525e-06,
3625
+ "loss": 0.9545,
3626
+ "step": 5970
3627
+ },
3628
+ {
3629
+ "epoch": 4.59,
3630
+ "learning_rate": 2.7709611451942744e-06,
3631
+ "loss": 0.9349,
3632
+ "step": 5980
3633
+ },
3634
+ {
3635
+ "epoch": 4.59,
3636
+ "learning_rate": 2.724948875255624e-06,
3637
+ "loss": 0.8813,
3638
+ "step": 5990
3639
+ },
3640
+ {
3641
+ "epoch": 4.6,
3642
+ "learning_rate": 2.6738241308793458e-06,
3643
+ "loss": 0.9202,
3644
+ "step": 6000
3645
+ },
3646
+ {
3647
+ "epoch": 4.61,
3648
+ "learning_rate": 2.6226993865030673e-06,
3649
+ "loss": 0.9854,
3650
+ "step": 6010
3651
+ },
3652
+ {
3653
+ "epoch": 4.62,
3654
+ "learning_rate": 2.571574642126789e-06,
3655
+ "loss": 0.9226,
3656
+ "step": 6020
3657
+ },
3658
+ {
3659
+ "epoch": 4.62,
3660
+ "learning_rate": 2.5204498977505115e-06,
3661
+ "loss": 0.9729,
3662
+ "step": 6030
3663
+ },
3664
+ {
3665
+ "epoch": 4.63,
3666
+ "learning_rate": 2.4693251533742334e-06,
3667
+ "loss": 0.9218,
3668
+ "step": 6040
3669
+ },
3670
+ {
3671
+ "epoch": 4.64,
3672
+ "learning_rate": 2.418200408997955e-06,
3673
+ "loss": 0.917,
3674
+ "step": 6050
3675
+ },
3676
+ {
3677
+ "epoch": 4.65,
3678
+ "learning_rate": 2.367075664621677e-06,
3679
+ "loss": 0.9602,
3680
+ "step": 6060
3681
+ },
3682
+ {
3683
+ "epoch": 4.65,
3684
+ "learning_rate": 2.3159509202453988e-06,
3685
+ "loss": 0.8968,
3686
+ "step": 6070
3687
+ },
3688
+ {
3689
+ "epoch": 4.66,
3690
+ "learning_rate": 2.264826175869121e-06,
3691
+ "loss": 0.9104,
3692
+ "step": 6080
3693
+ },
3694
+ {
3695
+ "epoch": 4.67,
3696
+ "learning_rate": 2.2137014314928426e-06,
3697
+ "loss": 0.8855,
3698
+ "step": 6090
3699
+ },
3700
+ {
3701
+ "epoch": 4.68,
3702
+ "learning_rate": 2.1625766871165645e-06,
3703
+ "loss": 0.9322,
3704
+ "step": 6100
3705
+ },
3706
+ {
3707
+ "epoch": 4.69,
3708
+ "learning_rate": 2.1114519427402864e-06,
3709
+ "loss": 0.9361,
3710
+ "step": 6110
3711
+ },
3712
+ {
3713
+ "epoch": 4.69,
3714
+ "learning_rate": 2.060327198364008e-06,
3715
+ "loss": 0.917,
3716
+ "step": 6120
3717
+ },
3718
+ {
3719
+ "epoch": 4.7,
3720
+ "learning_rate": 2.00920245398773e-06,
3721
+ "loss": 0.9389,
3722
+ "step": 6130
3723
+ },
3724
+ {
3725
+ "epoch": 4.71,
3726
+ "learning_rate": 1.958077709611452e-06,
3727
+ "loss": 0.9851,
3728
+ "step": 6140
3729
+ },
3730
+ {
3731
+ "epoch": 4.72,
3732
+ "learning_rate": 1.9069529652351741e-06,
3733
+ "loss": 0.9481,
3734
+ "step": 6150
3735
+ },
3736
+ {
3737
+ "epoch": 4.72,
3738
+ "learning_rate": 1.8558282208588958e-06,
3739
+ "loss": 0.8991,
3740
+ "step": 6160
3741
+ },
3742
+ {
3743
+ "epoch": 4.73,
3744
+ "learning_rate": 1.8047034764826175e-06,
3745
+ "loss": 0.9207,
3746
+ "step": 6170
3747
+ },
3748
+ {
3749
+ "epoch": 4.74,
3750
+ "learning_rate": 1.7535787321063397e-06,
3751
+ "loss": 0.9402,
3752
+ "step": 6180
3753
+ },
3754
+ {
3755
+ "epoch": 4.75,
3756
+ "learning_rate": 1.7024539877300614e-06,
3757
+ "loss": 0.9713,
3758
+ "step": 6190
3759
+ },
3760
+ {
3761
+ "epoch": 4.75,
3762
+ "learning_rate": 1.6513292433537833e-06,
3763
+ "loss": 0.9374,
3764
+ "step": 6200
3765
+ },
3766
+ {
3767
+ "epoch": 4.76,
3768
+ "learning_rate": 1.6002044989775052e-06,
3769
+ "loss": 0.9352,
3770
+ "step": 6210
3771
+ },
3772
+ {
3773
+ "epoch": 4.77,
3774
+ "learning_rate": 1.5490797546012271e-06,
3775
+ "loss": 0.9246,
3776
+ "step": 6220
3777
+ },
3778
+ {
3779
+ "epoch": 4.78,
3780
+ "learning_rate": 1.4979550102249488e-06,
3781
+ "loss": 0.8953,
3782
+ "step": 6230
3783
+ },
3784
+ {
3785
+ "epoch": 4.78,
3786
+ "learning_rate": 1.446830265848671e-06,
3787
+ "loss": 0.9044,
3788
+ "step": 6240
3789
+ },
3790
+ {
3791
+ "epoch": 4.79,
3792
+ "learning_rate": 1.3957055214723927e-06,
3793
+ "loss": 0.927,
3794
+ "step": 6250
3795
+ },
3796
+ {
3797
+ "epoch": 4.8,
3798
+ "learning_rate": 1.3445807770961146e-06,
3799
+ "loss": 0.926,
3800
+ "step": 6260
3801
+ },
3802
+ {
3803
+ "epoch": 4.81,
3804
+ "learning_rate": 1.2934560327198365e-06,
3805
+ "loss": 0.9082,
3806
+ "step": 6270
3807
+ },
3808
+ {
3809
+ "epoch": 4.82,
3810
+ "learning_rate": 1.2423312883435582e-06,
3811
+ "loss": 0.9158,
3812
+ "step": 6280
3813
+ },
3814
+ {
3815
+ "epoch": 4.82,
3816
+ "learning_rate": 1.1912065439672801e-06,
3817
+ "loss": 0.9389,
3818
+ "step": 6290
3819
+ },
3820
+ {
3821
+ "epoch": 4.83,
3822
+ "learning_rate": 1.140081799591002e-06,
3823
+ "loss": 0.9288,
3824
+ "step": 6300
3825
+ },
3826
+ {
3827
+ "epoch": 4.84,
3828
+ "learning_rate": 1.088957055214724e-06,
3829
+ "loss": 0.93,
3830
+ "step": 6310
3831
+ },
3832
+ {
3833
+ "epoch": 4.85,
3834
+ "learning_rate": 1.0378323108384457e-06,
3835
+ "loss": 0.8904,
3836
+ "step": 6320
3837
+ },
3838
+ {
3839
+ "epoch": 4.85,
3840
+ "learning_rate": 9.867075664621678e-07,
3841
+ "loss": 0.9087,
3842
+ "step": 6330
3843
+ },
3844
+ {
3845
+ "epoch": 4.86,
3846
+ "learning_rate": 9.355828220858896e-07,
3847
+ "loss": 0.9703,
3848
+ "step": 6340
3849
+ },
3850
+ {
3851
+ "epoch": 4.87,
3852
+ "learning_rate": 8.844580777096115e-07,
3853
+ "loss": 0.8408,
3854
+ "step": 6350
3855
+ },
3856
+ {
3857
+ "epoch": 4.88,
3858
+ "learning_rate": 8.333333333333333e-07,
3859
+ "loss": 0.8819,
3860
+ "step": 6360
3861
+ },
3862
+ {
3863
+ "epoch": 4.88,
3864
+ "learning_rate": 7.822085889570551e-07,
3865
+ "loss": 0.8903,
3866
+ "step": 6370
3867
+ },
3868
+ {
3869
+ "epoch": 4.89,
3870
+ "learning_rate": 7.310838445807771e-07,
3871
+ "loss": 0.9052,
3872
+ "step": 6380
3873
+ },
3874
+ {
3875
+ "epoch": 4.9,
3876
+ "learning_rate": 6.79959100204499e-07,
3877
+ "loss": 0.9522,
3878
+ "step": 6390
3879
+ },
3880
+ {
3881
+ "epoch": 4.91,
3882
+ "learning_rate": 6.288343558282209e-07,
3883
+ "loss": 0.9269,
3884
+ "step": 6400
3885
+ },
3886
+ {
3887
+ "epoch": 4.92,
3888
+ "learning_rate": 5.777096114519427e-07,
3889
+ "loss": 0.9148,
3890
+ "step": 6410
3891
+ },
3892
+ {
3893
+ "epoch": 4.92,
3894
+ "learning_rate": 5.265848670756646e-07,
3895
+ "loss": 0.9424,
3896
+ "step": 6420
3897
+ },
3898
+ {
3899
+ "epoch": 4.93,
3900
+ "learning_rate": 4.754601226993865e-07,
3901
+ "loss": 0.9658,
3902
+ "step": 6430
3903
+ },
3904
+ {
3905
+ "epoch": 4.94,
3906
+ "learning_rate": 4.243353783231084e-07,
3907
+ "loss": 0.9388,
3908
+ "step": 6440
3909
+ },
3910
+ {
3911
+ "epoch": 4.95,
3912
+ "learning_rate": 3.732106339468303e-07,
3913
+ "loss": 0.9308,
3914
+ "step": 6450
3915
+ },
3916
+ {
3917
+ "epoch": 4.95,
3918
+ "learning_rate": 3.220858895705522e-07,
3919
+ "loss": 0.9122,
3920
+ "step": 6460
3921
+ },
3922
+ {
3923
+ "epoch": 4.96,
3924
+ "learning_rate": 2.7096114519427406e-07,
3925
+ "loss": 0.8978,
3926
+ "step": 6470
3927
+ },
3928
+ {
3929
+ "epoch": 4.97,
3930
+ "learning_rate": 2.1983640081799592e-07,
3931
+ "loss": 0.9449,
3932
+ "step": 6480
3933
+ },
3934
+ {
3935
+ "epoch": 4.98,
3936
+ "learning_rate": 1.6871165644171781e-07,
3937
+ "loss": 0.9018,
3938
+ "step": 6490
3939
+ },
3940
+ {
3941
+ "epoch": 4.98,
3942
+ "learning_rate": 1.1758691206543969e-07,
3943
+ "loss": 0.9552,
3944
+ "step": 6500
3945
+ },
3946
+ {
3947
+ "epoch": 4.99,
3948
+ "learning_rate": 6.646216768916156e-08,
3949
+ "loss": 0.9161,
3950
+ "step": 6510
3951
+ },
3952
+ {
3953
+ "epoch": 5.0,
3954
+ "learning_rate": 1.533742331288344e-08,
3955
+ "loss": 0.9403,
3956
+ "step": 6520
3957
+ },
3958
+ {
3959
+ "epoch": 5.0,
3960
+ "eval_accuracy": 0.7016984883693075,
3961
+ "eval_loss": 0.8677756190299988,
3962
+ "eval_runtime": 303.2543,
3963
+ "eval_samples_per_second": 137.65,
3964
+ "eval_steps_per_second": 68.827,
3965
+ "step": 6520
3966
+ },
3967
+ {
3968
+ "epoch": 5.0,
3969
+ "step": 6520,
3970
+ "total_flos": 1.1511308168038711e+19,
3971
+ "train_loss": 1.11384382686732,
3972
+ "train_runtime": 4746.5723,
3973
+ "train_samples_per_second": 175.884,
3974
+ "train_steps_per_second": 1.374
3975
  }
3976
  ],
3977
+ "max_steps": 6520,
3978
+ "num_train_epochs": 5,
3979
+ "total_flos": 1.1511308168038711e+19,
3980
  "trial_name": null,
3981
  "trial_params": null
3982
  }