Yurii Paniv commited on
Commit
fa5d080
1 Parent(s): 70372e7

Delete redundant files

Browse files
added_tokens.json DELETED
@@ -1,4 +0,0 @@
1
- {
2
- "</s>": 38,
3
- "<s>": 37
4
- }
 
 
 
 
 
preprocessor_config.json DELETED
@@ -1,10 +0,0 @@
1
- {
2
- "do_normalize": true,
3
- "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
- "feature_size": 1,
5
- "padding_side": "right",
6
- "padding_value": 0.0,
7
- "processor_class": "Wav2Vec2Processor",
8
- "return_attention_mask": true,
9
- "sampling_rate": 16000
10
- }
 
 
 
 
 
 
 
 
 
 
 
special_tokens_map.json DELETED
@@ -1,22 +0,0 @@
1
- {
2
- "additional_special_tokens": [
3
- {
4
- "content": "<s>",
5
- "lstrip": false,
6
- "normalized": true,
7
- "rstrip": false,
8
- "single_word": false
9
- },
10
- {
11
- "content": "</s>",
12
- "lstrip": false,
13
- "normalized": true,
14
- "rstrip": false,
15
- "single_word": false
16
- }
17
- ],
18
- "bos_token": "<s>",
19
- "eos_token": "</s>",
20
- "pad_token": "[PAD]",
21
- "unk_token": "[UNK]"
22
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
tokenizer_config.json DELETED
@@ -1,14 +0,0 @@
1
- {
2
- "bos_token": "<s>",
3
- "do_lower_case": false,
4
- "eos_token": "</s>",
5
- "model_max_length": 1000000000000000019884624838656,
6
- "name_or_path": "./",
7
- "pad_token": "[PAD]",
8
- "processor_class": "Wav2Vec2Processor",
9
- "replace_word_delimiter_char": " ",
10
- "special_tokens_map_file": null,
11
- "tokenizer_class": "Wav2Vec2CTCTokenizer",
12
- "unk_token": "[UNK]",
13
- "word_delimiter_token": "|"
14
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
trainer_state.json DELETED
@@ -1,864 +0,0 @@
1
- {
2
- "best_metric": null,
3
- "best_model_checkpoint": null,
4
- "epoch": 493.02296211251434,
5
- "global_step": 21200,
6
- "is_hyper_param_search": false,
7
- "is_local_process_zero": true,
8
- "is_world_process_zero": true,
9
- "log_history": [
10
- {
11
- "epoch": 9.3,
12
- "learning_rate": 0.00023999999999999998,
13
- "loss": 4.3982,
14
- "step": 400
15
- },
16
- {
17
- "epoch": 9.3,
18
- "eval_cer": 0.14366206687135194,
19
- "eval_loss": 0.5217852592468262,
20
- "eval_runtime": 117.8175,
21
- "eval_samples_per_second": 26.414,
22
- "eval_steps_per_second": 3.302,
23
- "eval_wer": 0.6506676906011348,
24
- "step": 400
25
- },
26
- {
27
- "epoch": 18.6,
28
- "learning_rate": 0.00029461077844311373,
29
- "loss": 0.229,
30
- "step": 800
31
- },
32
- {
33
- "epoch": 18.6,
34
- "eval_cer": 0.08482742332737273,
35
- "eval_loss": 0.36793914437294006,
36
- "eval_runtime": 118.0263,
37
- "eval_samples_per_second": 26.367,
38
- "eval_steps_per_second": 3.296,
39
- "eval_wer": 0.40479542642604205,
40
- "step": 800
41
- },
42
- {
43
- "epoch": 27.9,
44
- "learning_rate": 0.00028742514970059877,
45
- "loss": 0.1054,
46
- "step": 1200
47
- },
48
- {
49
- "epoch": 27.9,
50
- "eval_cer": 0.07780964270049597,
51
- "eval_loss": 0.38127270340919495,
52
- "eval_runtime": 116.0512,
53
- "eval_samples_per_second": 26.816,
54
- "eval_steps_per_second": 3.352,
55
- "eval_wer": 0.367037842911387,
56
- "step": 1200
57
- },
58
- {
59
- "epoch": 37.21,
60
- "learning_rate": 0.0002802395209580838,
61
- "loss": 0.0784,
62
- "step": 1600
63
- },
64
- {
65
- "epoch": 37.21,
66
- "eval_cer": 0.07465164141840143,
67
- "eval_loss": 0.3839242458343506,
68
- "eval_runtime": 116.2982,
69
- "eval_samples_per_second": 26.759,
70
- "eval_steps_per_second": 3.345,
71
- "eval_wer": 0.35496394897393235,
72
- "step": 1600
73
- },
74
- {
75
- "epoch": 46.51,
76
- "learning_rate": 0.00027305389221556883,
77
- "loss": 0.066,
78
- "step": 2000
79
- },
80
- {
81
- "epoch": 46.51,
82
- "eval_cer": 0.07363946152029421,
83
- "eval_loss": 0.3969549238681793,
84
- "eval_runtime": 118.9877,
85
- "eval_samples_per_second": 26.154,
86
- "eval_steps_per_second": 3.269,
87
- "eval_wer": 0.3443406288664192,
88
- "step": 2000
89
- },
90
- {
91
- "epoch": 55.8,
92
- "learning_rate": 0.00026586826347305386,
93
- "loss": 0.0603,
94
- "step": 2400
95
- },
96
- {
97
- "epoch": 55.8,
98
- "eval_cer": 0.07218192246701981,
99
- "eval_loss": 0.3702129125595093,
100
- "eval_runtime": 118.9492,
101
- "eval_samples_per_second": 26.162,
102
- "eval_steps_per_second": 3.27,
103
- "eval_wer": 0.3393489483339733,
104
- "step": 2400
105
- },
106
- {
107
- "epoch": 65.11,
108
- "learning_rate": 0.0002586826347305389,
109
- "loss": 0.0539,
110
- "step": 2800
111
- },
112
- {
113
- "epoch": 65.11,
114
- "eval_cer": 0.07241809777657816,
115
- "eval_loss": 0.3762107491493225,
116
- "eval_runtime": 117.3723,
117
- "eval_samples_per_second": 26.514,
118
- "eval_steps_per_second": 3.314,
119
- "eval_wer": 0.33875165322752676,
120
- "step": 2800
121
- },
122
- {
123
- "epoch": 74.41,
124
- "learning_rate": 0.00025149700598802393,
125
- "loss": 0.0497,
126
- "step": 3200
127
- },
128
- {
129
- "epoch": 74.41,
130
- "eval_cer": 0.07128445629069807,
131
- "eval_loss": 0.36228740215301514,
132
- "eval_runtime": 116.3914,
133
- "eval_samples_per_second": 26.737,
134
- "eval_steps_per_second": 3.342,
135
- "eval_wer": 0.3413541533341866,
136
- "step": 3200
137
- },
138
- {
139
- "epoch": 83.71,
140
- "learning_rate": 0.00024431137724550896,
141
- "loss": 0.0432,
142
- "step": 3600
143
- },
144
- {
145
- "epoch": 83.71,
146
- "eval_cer": 0.07248557643645197,
147
- "eval_loss": 0.3847475051879883,
148
- "eval_runtime": 116.3314,
149
- "eval_samples_per_second": 26.751,
150
- "eval_steps_per_second": 3.344,
151
- "eval_wer": 0.3346132514185759,
152
- "step": 3600
153
- },
154
- {
155
- "epoch": 93.02,
156
- "learning_rate": 0.000237125748502994,
157
- "loss": 0.0438,
158
- "step": 4000
159
- },
160
- {
161
- "epoch": 93.02,
162
- "eval_cer": 0.07503626977968217,
163
- "eval_loss": 0.4057835340499878,
164
- "eval_runtime": 116.616,
165
- "eval_samples_per_second": 26.686,
166
- "eval_steps_per_second": 3.336,
167
- "eval_wer": 0.3393489483339733,
168
- "step": 4000
169
- },
170
- {
171
- "epoch": 102.32,
172
- "learning_rate": 0.00022994011976047902,
173
- "loss": 0.0413,
174
- "step": 4400
175
- },
176
- {
177
- "epoch": 102.32,
178
- "eval_cer": 0.07270825601403556,
179
- "eval_loss": 0.3957100510597229,
180
- "eval_runtime": 116.9156,
181
- "eval_samples_per_second": 26.617,
182
- "eval_steps_per_second": 3.327,
183
- "eval_wer": 0.3362771449293912,
184
- "step": 4400
185
- },
186
- {
187
- "epoch": 111.62,
188
- "learning_rate": 0.00022275449101796406,
189
- "loss": 0.039,
190
- "step": 4800
191
- },
192
- {
193
- "epoch": 111.62,
194
- "eval_cer": 0.07179729410573905,
195
- "eval_loss": 0.386459618806839,
196
- "eval_runtime": 119.0004,
197
- "eval_samples_per_second": 26.151,
198
- "eval_steps_per_second": 3.269,
199
- "eval_wer": 0.3330346857801101,
200
- "step": 4800
201
- },
202
- {
203
- "epoch": 120.92,
204
- "learning_rate": 0.0002155688622754491,
205
- "loss": 0.0356,
206
- "step": 5200
207
- },
208
- {
209
- "epoch": 120.92,
210
- "eval_cer": 0.07110226390903876,
211
- "eval_loss": 0.38599926233291626,
212
- "eval_runtime": 118.9882,
213
- "eval_samples_per_second": 26.154,
214
- "eval_steps_per_second": 3.269,
215
- "eval_wer": 0.33192542343956655,
216
- "step": 5200
217
- },
218
- {
219
- "epoch": 130.23,
220
- "learning_rate": 0.00020838323353293412,
221
- "loss": 0.0336,
222
- "step": 5600
223
- },
224
- {
225
- "epoch": 130.23,
226
- "eval_cer": 0.07001585748507035,
227
- "eval_loss": 0.3902195692062378,
228
- "eval_runtime": 116.3404,
229
- "eval_samples_per_second": 26.749,
230
- "eval_steps_per_second": 3.344,
231
- "eval_wer": 0.3241605870557618,
232
- "step": 5600
233
- },
234
- {
235
- "epoch": 139.53,
236
- "learning_rate": 0.00020119760479041913,
237
- "loss": 0.034,
238
- "step": 6000
239
- },
240
- {
241
- "epoch": 139.53,
242
- "eval_cer": 0.07322109382907656,
243
- "eval_loss": 0.39304569363594055,
244
- "eval_runtime": 116.7151,
245
- "eval_samples_per_second": 26.663,
246
- "eval_steps_per_second": 3.333,
247
- "eval_wer": 0.3337173087589061,
248
- "step": 6000
249
- },
250
- {
251
- "epoch": 148.83,
252
- "learning_rate": 0.00019402994011976046,
253
- "loss": 0.0273,
254
- "step": 6400
255
- },
256
- {
257
- "epoch": 148.83,
258
- "eval_cer": 0.07479334660413645,
259
- "eval_loss": 0.39119070768356323,
260
- "eval_runtime": 116.4756,
261
- "eval_samples_per_second": 26.718,
262
- "eval_steps_per_second": 3.34,
263
- "eval_wer": 0.33747173514228423,
264
- "step": 6400
265
- },
266
- {
267
- "epoch": 158.14,
268
- "learning_rate": 0.0001868443113772455,
269
- "loss": 0.027,
270
- "step": 6800
271
- },
272
- {
273
- "epoch": 158.14,
274
- "eval_cer": 0.07523870575930362,
275
- "eval_loss": 0.42656052112579346,
276
- "eval_runtime": 116.5029,
277
- "eval_samples_per_second": 26.712,
278
- "eval_steps_per_second": 3.339,
279
- "eval_wer": 0.34344468620674945,
280
- "step": 6800
281
- },
282
- {
283
- "epoch": 167.44,
284
- "learning_rate": 0.0001796586826347305,
285
- "loss": 0.028,
286
- "step": 7200
287
- },
288
- {
289
- "epoch": 167.44,
290
- "eval_cer": 0.0707648706096697,
291
- "eval_loss": 0.38949263095855713,
292
- "eval_runtime": 116.3646,
293
- "eval_samples_per_second": 26.744,
294
- "eval_steps_per_second": 3.343,
295
- "eval_wer": 0.32266734928964547,
296
- "step": 7200
297
- },
298
- {
299
- "epoch": 176.73,
300
- "learning_rate": 0.00017247305389221556,
301
- "loss": 0.0241,
302
- "step": 7600
303
- },
304
- {
305
- "epoch": 176.73,
306
- "eval_cer": 0.07274874320995985,
307
- "eval_loss": 0.3967472016811371,
308
- "eval_runtime": 116.4729,
309
- "eval_samples_per_second": 26.719,
310
- "eval_steps_per_second": 3.34,
311
- "eval_wer": 0.32936558726908144,
312
- "step": 7600
313
- },
314
- {
315
- "epoch": 186.05,
316
- "learning_rate": 0.00016530538922155687,
317
- "loss": 0.0241,
318
- "step": 8000
319
- },
320
- {
321
- "epoch": 186.05,
322
- "eval_cer": 0.07122372549681164,
323
- "eval_loss": 0.4058191776275635,
324
- "eval_runtime": 116.3498,
325
- "eval_samples_per_second": 26.747,
326
- "eval_steps_per_second": 3.343,
327
- "eval_wer": 0.32548316907717906,
328
- "step": 8000
329
- },
330
- {
331
- "epoch": 195.34,
332
- "learning_rate": 0.0001581197604790419,
333
- "loss": 0.0209,
334
- "step": 8400
335
- },
336
- {
337
- "epoch": 195.34,
338
- "eval_cer": 0.07019804986672964,
339
- "eval_loss": 0.4101807177066803,
340
- "eval_runtime": 116.6051,
341
- "eval_samples_per_second": 26.688,
342
- "eval_steps_per_second": 3.336,
343
- "eval_wer": 0.3233499722684415,
344
- "step": 8400
345
- },
346
- {
347
- "epoch": 204.64,
348
- "learning_rate": 0.00015093413173652694,
349
- "loss": 0.0206,
350
- "step": 8800
351
- },
352
- {
353
- "epoch": 204.64,
354
- "eval_cer": 0.06987415229933533,
355
- "eval_loss": 0.40751102566719055,
356
- "eval_runtime": 117.508,
357
- "eval_samples_per_second": 26.483,
358
- "eval_steps_per_second": 3.31,
359
- "eval_wer": 0.3193822262041896,
360
- "step": 8800
361
- },
362
- {
363
- "epoch": 213.94,
364
- "learning_rate": 0.00014376646706586825,
365
- "loss": 0.0172,
366
- "step": 9200
367
- },
368
- {
369
- "epoch": 213.94,
370
- "eval_cer": 0.06948952393805459,
371
- "eval_loss": 0.42218008637428284,
372
- "eval_runtime": 116.7394,
373
- "eval_samples_per_second": 26.658,
374
- "eval_steps_per_second": 3.332,
375
- "eval_wer": 0.31912624258714106,
376
- "step": 9200
377
- },
378
- {
379
- "epoch": 223.25,
380
- "learning_rate": 0.00013658083832335328,
381
- "loss": 0.0166,
382
- "step": 9600
383
- },
384
- {
385
- "epoch": 223.25,
386
- "eval_cer": 0.06777556597725969,
387
- "eval_loss": 0.38604938983917236,
388
- "eval_runtime": 116.6232,
389
- "eval_samples_per_second": 26.684,
390
- "eval_steps_per_second": 3.336,
391
- "eval_wer": 0.31345193907589913,
392
- "step": 9600
393
- },
394
- {
395
- "epoch": 232.55,
396
- "learning_rate": 0.0001293952095808383,
397
- "loss": 0.0156,
398
- "step": 10000
399
- },
400
- {
401
- "epoch": 232.55,
402
- "eval_cer": 0.0677013394513985,
403
- "eval_loss": 0.40345117449760437,
404
- "eval_runtime": 117.639,
405
- "eval_samples_per_second": 26.454,
406
- "eval_steps_per_second": 3.307,
407
- "eval_wer": 0.31170271769273433,
408
- "step": 10000
409
- },
410
- {
411
- "epoch": 241.85,
412
- "learning_rate": 0.00012220958083832334,
413
- "loss": 0.0149,
414
- "step": 10400
415
- },
416
- {
417
- "epoch": 241.85,
418
- "eval_cer": 0.0677013394513985,
419
- "eval_loss": 0.39512303471565247,
420
- "eval_runtime": 120.4059,
421
- "eval_samples_per_second": 25.846,
422
- "eval_steps_per_second": 3.231,
423
- "eval_wer": 0.30867357822432695,
424
- "step": 10400
425
- },
426
- {
427
- "epoch": 251.16,
428
- "learning_rate": 0.00011502395209580837,
429
- "loss": 0.0142,
430
- "step": 10800
431
- },
432
- {
433
- "epoch": 251.16,
434
- "eval_cer": 0.06735045042005466,
435
- "eval_loss": 0.3971852958202362,
436
- "eval_runtime": 118.3102,
437
- "eval_samples_per_second": 26.304,
438
- "eval_steps_per_second": 3.288,
439
- "eval_wer": 0.309697512692521,
440
- "step": 10800
441
- },
442
- {
443
- "epoch": 260.46,
444
- "learning_rate": 0.0001078383233532934,
445
- "loss": 0.0134,
446
- "step": 11200
447
- },
448
- {
449
- "epoch": 260.46,
450
- "eval_cer": 0.06749215560578967,
451
- "eval_loss": 0.40693503618240356,
452
- "eval_runtime": 120.6637,
453
- "eval_samples_per_second": 25.791,
454
- "eval_steps_per_second": 3.224,
455
- "eval_wer": 0.31114808652246256,
456
- "step": 11200
457
- },
458
- {
459
- "epoch": 269.76,
460
- "learning_rate": 0.00010065269461077844,
461
- "loss": 0.0116,
462
- "step": 11600
463
- },
464
- {
465
- "epoch": 269.76,
466
- "eval_cer": 0.06968521205168865,
467
- "eval_loss": 0.41885173320770264,
468
- "eval_runtime": 118.3717,
469
- "eval_samples_per_second": 26.29,
470
- "eval_steps_per_second": 3.286,
471
- "eval_wer": 0.31609710311873374,
472
- "step": 11600
473
- },
474
- {
475
- "epoch": 279.07,
476
- "learning_rate": 9.346706586826346e-05,
477
- "loss": 0.0119,
478
- "step": 12000
479
- },
480
- {
481
- "epoch": 279.07,
482
- "eval_cer": 0.0648200006747866,
483
- "eval_loss": 0.3901657462120056,
484
- "eval_runtime": 119.8759,
485
- "eval_samples_per_second": 25.96,
486
- "eval_steps_per_second": 3.245,
487
- "eval_wer": 0.3008234139681727,
488
- "step": 12000
489
- },
490
- {
491
- "epoch": 288.37,
492
- "learning_rate": 8.62814371257485e-05,
493
- "loss": 0.0098,
494
- "step": 12400
495
- },
496
- {
497
- "epoch": 288.37,
498
- "eval_cer": 0.06515064610816829,
499
- "eval_loss": 0.40946489572525024,
500
- "eval_runtime": 120.8583,
501
- "eval_samples_per_second": 25.749,
502
- "eval_steps_per_second": 3.219,
503
- "eval_wer": 0.30018345492555143,
504
- "step": 12400
505
- },
506
- {
507
- "epoch": 297.67,
508
- "learning_rate": 7.909580838323352e-05,
509
- "loss": 0.0091,
510
- "step": 12800
511
- },
512
- {
513
- "epoch": 297.67,
514
- "eval_cer": 0.06441512871554371,
515
- "eval_loss": 0.3892023265361786,
516
- "eval_runtime": 118.1231,
517
- "eval_samples_per_second": 26.345,
518
- "eval_steps_per_second": 3.293,
519
- "eval_wer": 0.2989888647126584,
520
- "step": 12800
521
- },
522
- {
523
- "epoch": 306.96,
524
- "learning_rate": 7.191017964071855e-05,
525
- "loss": 0.0094,
526
- "step": 13200
527
- },
528
- {
529
- "epoch": 306.96,
530
- "eval_cer": 0.06469853908701373,
531
- "eval_loss": 0.40261197090148926,
532
- "eval_runtime": 118.2919,
533
- "eval_samples_per_second": 26.308,
534
- "eval_steps_per_second": 3.288,
535
- "eval_wer": 0.29834890567003713,
536
- "step": 13200
537
- },
538
- {
539
- "epoch": 316.28,
540
- "learning_rate": 6.474251497005988e-05,
541
- "loss": 0.0081,
542
- "step": 13600
543
- },
544
- {
545
- "epoch": 316.28,
546
- "eval_cer": 0.06462431256115253,
547
- "eval_loss": 0.4302999675273895,
548
- "eval_runtime": 118.0293,
549
- "eval_samples_per_second": 26.366,
550
- "eval_steps_per_second": 3.296,
551
- "eval_wer": 0.29779427449976537,
552
- "step": 13600
553
- },
554
- {
555
- "epoch": 325.57,
556
- "learning_rate": 5.7556886227544904e-05,
557
- "loss": 0.0079,
558
- "step": 14000
559
- },
560
- {
561
- "epoch": 325.57,
562
- "eval_cer": 0.06431391072573299,
563
- "eval_loss": 0.40440893173217773,
564
- "eval_runtime": 118.2184,
565
- "eval_samples_per_second": 26.324,
566
- "eval_steps_per_second": 3.291,
567
- "eval_wer": 0.29796493024446435,
568
- "step": 14000
569
- },
570
- {
571
- "epoch": 334.87,
572
- "learning_rate": 5.038922155688622e-05,
573
- "loss": 0.0072,
574
- "step": 14400
575
- },
576
- {
577
- "epoch": 334.87,
578
- "eval_cer": 0.06546779580957522,
579
- "eval_loss": 0.382755845785141,
580
- "eval_runtime": 118.6873,
581
- "eval_samples_per_second": 26.22,
582
- "eval_steps_per_second": 3.278,
583
- "eval_wer": 0.29992747130850295,
584
- "step": 14400
585
- },
586
- {
587
- "epoch": 344.18,
588
- "learning_rate": 9.578571428571428e-05,
589
- "loss": 0.0081,
590
- "step": 14800
591
- },
592
- {
593
- "epoch": 344.18,
594
- "eval_cer": 0.06676338607915247,
595
- "eval_loss": 0.4108315706253052,
596
- "eval_runtime": 114.661,
597
- "eval_samples_per_second": 27.141,
598
- "eval_steps_per_second": 3.393,
599
- "eval_wer": 0.30457784035155083,
600
- "step": 14800
601
- },
602
- {
603
- "epoch": 353.48,
604
- "learning_rate": 9.007142857142856e-05,
605
- "loss": 0.0088,
606
- "step": 15200
607
- },
608
- {
609
- "epoch": 353.48,
610
- "eval_cer": 0.06539356928371402,
611
- "eval_loss": 0.40191251039505005,
612
- "eval_runtime": 117.0774,
613
- "eval_samples_per_second": 26.581,
614
- "eval_steps_per_second": 3.323,
615
- "eval_wer": 0.2993301762020564,
616
- "step": 15200
617
- },
618
- {
619
- "epoch": 362.78,
620
- "learning_rate": 8.435714285714286e-05,
621
- "loss": 0.0088,
622
- "step": 15600
623
- },
624
- {
625
- "epoch": 362.78,
626
- "eval_cer": 0.06814669860656568,
627
- "eval_loss": 0.4072999954223633,
628
- "eval_runtime": 114.628,
629
- "eval_samples_per_second": 27.149,
630
- "eval_steps_per_second": 3.394,
631
- "eval_wer": 0.3091002175860745,
632
- "step": 15600
633
- },
634
- {
635
- "epoch": 372.09,
636
- "learning_rate": 7.864285714285714e-05,
637
- "loss": 0.0079,
638
- "step": 16000
639
- },
640
- {
641
- "epoch": 372.09,
642
- "eval_cer": 0.0667161510172408,
643
- "eval_loss": 0.42044562101364136,
644
- "eval_runtime": 115.1974,
645
- "eval_samples_per_second": 27.014,
646
- "eval_steps_per_second": 3.377,
647
- "eval_wer": 0.30547378301122063,
648
- "step": 16000
649
- },
650
- {
651
- "epoch": 381.39,
652
- "learning_rate": 7.292857142857142e-05,
653
- "loss": 0.0072,
654
- "step": 16400
655
- },
656
- {
657
- "epoch": 381.39,
658
- "eval_cer": 0.06564998819123452,
659
- "eval_loss": 0.40300747752189636,
660
- "eval_runtime": 114.5196,
661
- "eval_samples_per_second": 27.174,
662
- "eval_steps_per_second": 3.397,
663
- "eval_wer": 0.3027859550322113,
664
- "step": 16400
665
- },
666
- {
667
- "epoch": 390.69,
668
- "learning_rate": 6.721428571428571e-05,
669
- "loss": 0.0073,
670
- "step": 16800
671
- },
672
- {
673
- "epoch": 390.69,
674
- "eval_cer": 0.0677350787813354,
675
- "eval_loss": 0.4031626284122467,
676
- "eval_runtime": 114.7831,
677
- "eval_samples_per_second": 27.112,
678
- "eval_steps_per_second": 3.389,
679
- "eval_wer": 0.30807628311788043,
680
- "step": 16800
681
- },
682
- {
683
- "epoch": 399.99,
684
- "learning_rate": 6.151428571428571e-05,
685
- "loss": 0.0069,
686
- "step": 17200
687
- },
688
- {
689
- "epoch": 399.99,
690
- "eval_cer": 0.06693208272883701,
691
- "eval_loss": 0.41302183270454407,
692
- "eval_runtime": 114.8457,
693
- "eval_samples_per_second": 27.097,
694
- "eval_steps_per_second": 3.387,
695
- "eval_wer": 0.30214599598959,
696
- "step": 17200
697
- },
698
- {
699
- "epoch": 409.3,
700
- "learning_rate": 5.5799999999999994e-05,
701
- "loss": 0.0063,
702
- "step": 17600
703
- },
704
- {
705
- "epoch": 409.3,
706
- "eval_cer": 0.06513715037619353,
707
- "eval_loss": 0.4071926772594452,
708
- "eval_runtime": 114.6356,
709
- "eval_samples_per_second": 27.147,
710
- "eval_steps_per_second": 3.393,
711
- "eval_wer": 0.2979222663082896,
712
- "step": 17600
713
- },
714
- {
715
- "epoch": 418.6,
716
- "learning_rate": 5.008571428571428e-05,
717
- "loss": 0.0059,
718
- "step": 18000
719
- },
720
- {
721
- "epoch": 418.6,
722
- "eval_cer": 0.06403724822025035,
723
- "eval_loss": 0.41102761030197144,
724
- "eval_runtime": 116.5576,
725
- "eval_samples_per_second": 26.699,
726
- "eval_steps_per_second": 3.337,
727
- "eval_wer": 0.2969409957762703,
728
- "step": 18000
729
- },
730
- {
731
- "epoch": 427.9,
732
- "learning_rate": 4.437142857142857e-05,
733
- "loss": 0.0056,
734
- "step": 18400
735
- },
736
- {
737
- "epoch": 427.9,
738
- "eval_cer": 0.06465805189108945,
739
- "eval_loss": 0.4228787422180176,
740
- "eval_runtime": 114.7096,
741
- "eval_samples_per_second": 27.129,
742
- "eval_steps_per_second": 3.391,
743
- "eval_wer": 0.29945816801058067,
744
- "step": 18400
745
- },
746
- {
747
- "epoch": 437.21,
748
- "learning_rate": 3.8657142857142856e-05,
749
- "loss": 0.005,
750
- "step": 18800
751
- },
752
- {
753
- "epoch": 437.21,
754
- "eval_cer": 0.062370525321367117,
755
- "eval_loss": 0.41175001859664917,
756
- "eval_runtime": 115.5821,
757
- "eval_samples_per_second": 26.925,
758
- "eval_steps_per_second": 3.366,
759
- "eval_wer": 0.2884508724774948,
760
- "step": 18800
761
- },
762
- {
763
- "epoch": 446.51,
764
- "learning_rate": 3.294285714285714e-05,
765
- "loss": 0.0046,
766
- "step": 19200
767
- },
768
- {
769
- "epoch": 446.51,
770
- "eval_cer": 0.06147305914504538,
771
- "eval_loss": 0.41112595796585083,
772
- "eval_runtime": 115.5282,
773
- "eval_samples_per_second": 26.937,
774
- "eval_steps_per_second": 3.367,
775
- "eval_wer": 0.28409915098767013,
776
- "step": 19200
777
- },
778
- {
779
- "epoch": 455.8,
780
- "learning_rate": 2.7228571428571427e-05,
781
- "loss": 0.0043,
782
- "step": 19600
783
- },
784
- {
785
- "epoch": 455.8,
786
- "eval_cer": 0.06160126859880563,
787
- "eval_loss": 0.40707847476005554,
788
- "eval_runtime": 117.681,
789
- "eval_samples_per_second": 26.444,
790
- "eval_steps_per_second": 3.306,
791
- "eval_wer": 0.28495242971116513,
792
- "step": 19600
793
- },
794
- {
795
- "epoch": 465.11,
796
- "learning_rate": 2.1514285714285714e-05,
797
- "loss": 0.0038,
798
- "step": 20000
799
- },
800
- {
801
- "epoch": 465.11,
802
- "eval_cer": 0.062363777455379736,
803
- "eval_loss": 0.4267757534980774,
804
- "eval_runtime": 115.119,
805
- "eval_samples_per_second": 27.033,
806
- "eval_steps_per_second": 3.379,
807
- "eval_wer": 0.28670165109433,
808
- "step": 20000
809
- },
810
- {
811
- "epoch": 474.41,
812
- "learning_rate": 1.5799999999999998e-05,
813
- "loss": 0.0035,
814
- "step": 20400
815
- },
816
- {
817
- "epoch": 474.41,
818
- "eval_cer": 0.06053510577279935,
819
- "eval_loss": 0.4116959869861603,
820
- "eval_runtime": 115.3416,
821
- "eval_samples_per_second": 26.981,
822
- "eval_steps_per_second": 3.373,
823
- "eval_wer": 0.2820086181151073,
824
- "step": 20400
825
- },
826
- {
827
- "epoch": 483.71,
828
- "learning_rate": 1.0085714285714285e-05,
829
- "loss": 0.0035,
830
- "step": 20800
831
- },
832
- {
833
- "epoch": 483.71,
834
- "eval_cer": 0.060238199669354564,
835
- "eval_loss": 0.4154604375362396,
836
- "eval_runtime": 115.2471,
837
- "eval_samples_per_second": 27.003,
838
- "eval_steps_per_second": 3.375,
839
- "eval_wer": 0.2819232902427578,
840
- "step": 20800
841
- },
842
- {
843
- "epoch": 493.02,
844
- "learning_rate": 4.371428571428571e-06,
845
- "loss": 0.0034,
846
- "step": 21200
847
- },
848
- {
849
- "epoch": 493.02,
850
- "eval_cer": 0.06007625088565741,
851
- "eval_loss": 0.41654759645462036,
852
- "eval_runtime": 115.3904,
853
- "eval_samples_per_second": 26.969,
854
- "eval_steps_per_second": 3.371,
855
- "eval_wer": 0.27991808524254447,
856
- "step": 21200
857
- }
858
- ],
859
- "max_steps": 21500,
860
- "num_train_epochs": 500,
861
- "total_flos": 5.302246573116527e+20,
862
- "trial_name": null,
863
- "trial_params": null
864
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
training_args.bin DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:9a9f2d55e2d30c3efb297191f66a757c1a52ea11bcbc472646289172e2a0d182
3
- size 2927