zuazo commited on
Commit
54f3f3e
·
verified ·
1 Parent(s): a8bb11b

End of training

Browse files
README.md CHANGED
@@ -1,21 +1,24 @@
1
  ---
 
 
2
  license: apache-2.0
3
  base_model: openai/whisper-base
4
  tags:
 
5
  - generated_from_trainer
6
  datasets:
7
- - common_voice_13_0
8
  metrics:
9
  - wer
10
  model-index:
11
- - name: openai/whisper-base
12
  results:
13
  - task:
14
  name: Automatic Speech Recognition
15
  type: automatic-speech-recognition
16
  dataset:
17
- name: common_voice_13_0
18
- type: common_voice_13_0
19
  config: gl
20
  split: test
21
  args: gl
@@ -28,9 +31,9 @@ model-index:
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
29
  should probably proofread and complete it, then remove this comment. -->
30
 
31
- # openai/whisper-base
32
 
33
- This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the common_voice_13_0 dataset.
34
  It achieves the following results on the evaluation set:
35
  - Loss: 0.4360
36
  - Wer: 17.2910
 
1
  ---
2
+ language:
3
+ - gl
4
  license: apache-2.0
5
  base_model: openai/whisper-base
6
  tags:
7
+ - whisper-event
8
  - generated_from_trainer
9
  datasets:
10
+ - mozilla-foundation/common_voice_13_0
11
  metrics:
12
  - wer
13
  model-index:
14
+ - name: Whisper Base Galician
15
  results:
16
  - task:
17
  name: Automatic Speech Recognition
18
  type: automatic-speech-recognition
19
  dataset:
20
+ name: mozilla-foundation/common_voice_13_0 gl
21
+ type: mozilla-foundation/common_voice_13_0
22
  config: gl
23
  split: test
24
  args: gl
 
31
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
32
  should probably proofread and complete it, then remove this comment. -->
33
 
34
+ # Whisper Base Galician
35
 
36
+ This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the mozilla-foundation/common_voice_13_0 gl dataset.
37
  It achieves the following results on the evaluation set:
38
  - Loss: 0.4360
39
  - Wer: 17.2910
all_results.json CHANGED
@@ -1,12 +1,12 @@
1
  {
2
- "epoch": 49.02,
3
- "eval_loss": 0.4754403829574585,
4
- "eval_runtime": 720.2859,
5
- "eval_samples_per_second": 9.088,
6
- "eval_steps_per_second": 0.143,
7
- "eval_wer": 18.687913907284766,
8
- "train_loss": 0.054723825027979914,
9
- "train_runtime": 73614.7434,
10
- "train_samples_per_second": 8.694,
11
- "train_steps_per_second": 0.068
12
  }
 
1
  {
2
+ "epoch": 50.0,
3
+ "eval_loss": 0.43604204058647156,
4
+ "eval_runtime": 72.7164,
5
+ "eval_samples_per_second": 90.021,
6
+ "eval_steps_per_second": 1.416,
7
+ "eval_wer": 17.290976821192054,
8
+ "train_loss": 0.5045194145917893,
9
+ "train_runtime": 4484.1133,
10
+ "train_samples_per_second": 142.726,
11
+ "train_steps_per_second": 1.115
12
  }
eval_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
- "epoch": 49.02,
3
- "eval_loss": 0.4754403829574585,
4
- "eval_runtime": 720.2859,
5
- "eval_samples_per_second": 9.088,
6
- "eval_steps_per_second": 0.143,
7
- "eval_wer": 18.687913907284766
8
  }
 
1
  {
2
+ "epoch": 50.0,
3
+ "eval_loss": 0.43604204058647156,
4
+ "eval_runtime": 72.7164,
5
+ "eval_samples_per_second": 90.021,
6
+ "eval_steps_per_second": 1.416,
7
+ "eval_wer": 17.290976821192054
8
  }
runs/Jan02_16-03-37_hyperion-255/events.out.tfevents.1735834951.hyperion-255 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd61169583395c68cea3f272fabf9278de495cab646f27e71d7572b19011cd93
3
+ size 40
train_results.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "epoch": 49.02,
3
- "train_loss": 0.054723825027979914,
4
- "train_runtime": 73614.7434,
5
- "train_samples_per_second": 8.694,
6
- "train_steps_per_second": 0.068
7
  }
 
1
  {
2
+ "epoch": 50.0,
3
+ "train_loss": 0.5045194145917893,
4
+ "train_runtime": 4484.1133,
5
+ "train_samples_per_second": 142.726,
6
+ "train_steps_per_second": 1.115
7
  }
trainer_state.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "best_metric": 18.687913907284766,
3
- "best_model_checkpoint": "./checkpoint-2000",
4
- "epoch": 49.02,
5
  "eval_steps": 1000,
6
  "global_step": 5000,
7
  "is_hyper_param_search": false,
@@ -9,1265 +9,1267 @@
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.01,
13
- "learning_rate": 1.1e-06,
14
- "loss": 2.0938,
15
  "step": 25
16
  },
17
  {
18
- "epoch": 0.01,
19
- "learning_rate": 2.35e-06,
20
- "loss": 1.5466,
21
  "step": 50
22
  },
23
  {
24
- "epoch": 0.01,
25
- "learning_rate": 3.6e-06,
26
- "loss": 0.9893,
27
  "step": 75
28
  },
29
  {
30
- "epoch": 0.02,
31
- "learning_rate": 4.85e-06,
32
- "loss": 0.6816,
33
  "step": 100
34
  },
35
  {
36
- "epoch": 1.0,
37
- "learning_rate": 6.1e-06,
38
- "loss": 0.6133,
39
  "step": 125
40
  },
41
  {
42
- "epoch": 1.01,
43
- "learning_rate": 7.35e-06,
44
- "loss": 0.5238,
45
  "step": 150
46
  },
47
  {
48
- "epoch": 1.01,
49
- "learning_rate": 8.599999999999999e-06,
50
- "loss": 0.46,
51
  "step": 175
52
  },
53
  {
54
- "epoch": 1.02,
55
- "learning_rate": 9.85e-06,
56
- "loss": 0.3833,
57
  "step": 200
58
  },
59
  {
60
- "epoch": 2.0,
61
- "learning_rate": 1.11e-05,
62
- "loss": 0.3859,
63
  "step": 225
64
  },
65
  {
66
- "epoch": 2.01,
67
- "learning_rate": 1.235e-05,
68
- "loss": 0.34,
69
  "step": 250
70
  },
71
  {
72
- "epoch": 2.02,
73
- "learning_rate": 1.3600000000000002e-05,
74
- "loss": 0.3041,
75
  "step": 275
76
  },
77
  {
78
- "epoch": 2.02,
79
- "learning_rate": 1.485e-05,
80
- "loss": 0.2583,
81
  "step": 300
82
  },
83
  {
84
- "epoch": 3.0,
85
- "learning_rate": 1.6100000000000002e-05,
86
- "loss": 0.2608,
87
  "step": 325
88
  },
89
  {
90
- "epoch": 3.01,
91
- "learning_rate": 1.7349999999999998e-05,
92
- "loss": 0.2305,
93
  "step": 350
94
  },
95
  {
96
- "epoch": 3.02,
97
- "learning_rate": 1.86e-05,
98
- "loss": 0.2024,
99
  "step": 375
100
  },
101
  {
102
- "epoch": 3.02,
103
- "learning_rate": 1.985e-05,
104
- "loss": 0.1694,
105
  "step": 400
106
  },
107
  {
108
- "epoch": 4.0,
109
- "learning_rate": 2.11e-05,
110
- "loss": 0.1707,
111
  "step": 425
112
  },
113
  {
114
- "epoch": 4.01,
115
- "learning_rate": 2.235e-05,
116
- "loss": 0.154,
117
  "step": 450
118
  },
119
  {
120
- "epoch": 4.01,
121
- "learning_rate": 2.36e-05,
122
- "loss": 0.1236,
123
  "step": 475
124
  },
125
  {
126
- "epoch": 4.02,
127
- "learning_rate": 2.485e-05,
128
- "loss": 0.1009,
129
  "step": 500
130
  },
131
  {
132
- "epoch": 5.0,
133
- "learning_rate": 2.4877777777777776e-05,
134
- "loss": 0.1019,
135
  "step": 525
136
  },
137
  {
138
- "epoch": 5.01,
139
- "learning_rate": 2.473888888888889e-05,
140
- "loss": 0.0913,
141
  "step": 550
142
  },
143
  {
144
- "epoch": 5.01,
145
- "learning_rate": 2.46e-05,
146
- "loss": 0.0692,
147
  "step": 575
148
  },
149
  {
150
- "epoch": 5.02,
151
- "learning_rate": 2.446111111111111e-05,
152
- "loss": 0.0553,
153
  "step": 600
154
  },
155
  {
156
- "epoch": 6.0,
157
- "learning_rate": 2.4322222222222224e-05,
158
- "loss": 0.0583,
159
  "step": 625
160
  },
161
  {
162
- "epoch": 6.01,
163
- "learning_rate": 2.4183333333333336e-05,
164
- "loss": 0.0517,
165
  "step": 650
166
  },
167
  {
168
- "epoch": 6.01,
169
- "learning_rate": 2.4044444444444445e-05,
170
- "loss": 0.0395,
171
  "step": 675
172
  },
173
  {
174
- "epoch": 6.02,
175
- "learning_rate": 2.3905555555555555e-05,
176
- "loss": 0.0299,
177
  "step": 700
178
  },
179
  {
180
- "epoch": 7.0,
181
- "learning_rate": 2.3766666666666667e-05,
182
- "loss": 0.0342,
183
  "step": 725
184
  },
185
  {
186
- "epoch": 7.01,
187
- "learning_rate": 2.362777777777778e-05,
188
- "loss": 0.0316,
189
  "step": 750
190
  },
191
  {
192
- "epoch": 7.01,
193
- "learning_rate": 2.3488888888888893e-05,
194
- "loss": 0.0236,
195
  "step": 775
196
  },
197
  {
198
- "epoch": 7.02,
199
- "learning_rate": 2.3350000000000002e-05,
200
- "loss": 0.018,
201
  "step": 800
202
  },
203
  {
204
- "epoch": 8.01,
205
- "learning_rate": 2.321111111111111e-05,
206
- "loss": 0.0203,
207
  "step": 825
208
  },
209
  {
210
- "epoch": 8.01,
211
- "learning_rate": 2.3072222222222224e-05,
212
- "loss": 0.019,
213
  "step": 850
214
  },
215
  {
216
- "epoch": 8.02,
217
- "learning_rate": 2.2933333333333333e-05,
218
- "loss": 0.0161,
219
  "step": 875
220
  },
221
  {
222
- "epoch": 8.02,
223
- "learning_rate": 2.2794444444444445e-05,
224
- "loss": 0.0129,
225
  "step": 900
226
  },
227
  {
228
- "epoch": 9.01,
229
- "learning_rate": 2.2655555555555558e-05,
230
- "loss": 0.0142,
231
  "step": 925
232
  },
233
  {
234
- "epoch": 9.01,
235
- "learning_rate": 2.2516666666666667e-05,
236
- "loss": 0.0128,
237
  "step": 950
238
  },
239
  {
240
- "epoch": 9.02,
241
- "learning_rate": 2.2377777777777777e-05,
242
- "loss": 0.0108,
243
  "step": 975
244
  },
245
  {
246
- "epoch": 9.02,
247
- "learning_rate": 2.223888888888889e-05,
248
- "loss": 0.0088,
249
  "step": 1000
250
  },
251
  {
252
- "epoch": 9.02,
253
- "eval_loss": 0.421886682510376,
254
- "eval_runtime": 714.5524,
255
- "eval_samples_per_second": 9.161,
256
- "eval_steps_per_second": 0.144,
257
- "eval_wer": 18.77759381898455,
258
  "step": 1000
259
  },
260
  {
261
- "epoch": 10.01,
262
- "learning_rate": 2.2100000000000002e-05,
263
- "loss": 0.0099,
264
  "step": 1025
265
  },
266
  {
267
- "epoch": 10.01,
268
- "learning_rate": 2.1961111111111114e-05,
269
- "loss": 0.0085,
270
  "step": 1050
271
  },
272
  {
273
- "epoch": 10.02,
274
- "learning_rate": 2.1822222222222224e-05,
275
- "loss": 0.0075,
276
  "step": 1075
277
  },
278
  {
279
- "epoch": 10.02,
280
- "learning_rate": 2.1683333333333333e-05,
281
- "loss": 0.0065,
282
  "step": 1100
283
  },
284
  {
285
- "epoch": 11.01,
286
- "learning_rate": 2.1544444444444446e-05,
287
- "loss": 0.007,
288
  "step": 1125
289
  },
290
  {
291
- "epoch": 11.01,
292
- "learning_rate": 2.1405555555555555e-05,
293
- "loss": 0.0075,
294
  "step": 1150
295
  },
296
  {
297
- "epoch": 11.02,
298
- "learning_rate": 2.1266666666666667e-05,
299
- "loss": 0.0062,
300
  "step": 1175
301
  },
302
  {
303
- "epoch": 11.02,
304
- "learning_rate": 2.112777777777778e-05,
305
- "loss": 0.0056,
306
  "step": 1200
307
  },
308
  {
309
- "epoch": 12.01,
310
- "learning_rate": 2.0988888888888893e-05,
311
- "loss": 0.0055,
312
  "step": 1225
313
  },
314
  {
315
- "epoch": 12.01,
316
- "learning_rate": 2.085e-05,
317
- "loss": 0.0057,
318
  "step": 1250
319
  },
320
  {
321
- "epoch": 12.02,
322
- "learning_rate": 2.071111111111111e-05,
323
- "loss": 0.0046,
324
  "step": 1275
325
  },
326
  {
327
- "epoch": 12.02,
328
- "learning_rate": 2.0572222222222224e-05,
329
- "loss": 0.0043,
330
  "step": 1300
331
  },
332
  {
333
- "epoch": 13.01,
334
- "learning_rate": 2.0433333333333336e-05,
335
- "loss": 0.0039,
336
  "step": 1325
337
  },
338
  {
339
- "epoch": 13.01,
340
- "learning_rate": 2.0294444444444446e-05,
341
- "loss": 0.004,
342
  "step": 1350
343
  },
344
  {
345
- "epoch": 13.02,
346
- "learning_rate": 2.0155555555555555e-05,
347
- "loss": 0.0035,
348
  "step": 1375
349
  },
350
  {
351
- "epoch": 13.02,
352
- "learning_rate": 2.0016666666666668e-05,
353
- "loss": 0.0032,
354
  "step": 1400
355
  },
356
  {
357
- "epoch": 14.01,
358
- "learning_rate": 1.9877777777777777e-05,
359
- "loss": 0.0031,
360
  "step": 1425
361
  },
362
  {
363
- "epoch": 14.01,
364
- "learning_rate": 1.973888888888889e-05,
365
- "loss": 0.003,
366
  "step": 1450
367
  },
368
  {
369
- "epoch": 14.02,
370
- "learning_rate": 1.9600000000000002e-05,
371
- "loss": 0.0028,
372
  "step": 1475
373
  },
374
  {
375
- "epoch": 14.02,
376
- "learning_rate": 1.9461111111111115e-05,
377
- "loss": 0.0026,
378
  "step": 1500
379
  },
380
  {
381
- "epoch": 15.01,
382
- "learning_rate": 1.932222222222222e-05,
383
- "loss": 0.0026,
384
  "step": 1525
385
  },
386
  {
387
- "epoch": 15.01,
388
- "learning_rate": 1.9183333333333333e-05,
389
- "loss": 0.0026,
390
  "step": 1550
391
  },
392
  {
393
- "epoch": 15.02,
394
- "learning_rate": 1.9044444444444446e-05,
395
- "loss": 0.0025,
396
  "step": 1575
397
  },
398
  {
399
- "epoch": 15.02,
400
- "learning_rate": 1.890555555555556e-05,
401
- "loss": 0.0023,
402
  "step": 1600
403
  },
404
  {
405
- "epoch": 16.0,
406
- "learning_rate": 1.8766666666666668e-05,
407
- "loss": 0.0023,
408
  "step": 1625
409
  },
410
  {
411
- "epoch": 16.01,
412
- "learning_rate": 1.8627777777777777e-05,
413
- "loss": 0.0023,
414
  "step": 1650
415
  },
416
  {
417
- "epoch": 16.02,
418
- "learning_rate": 1.848888888888889e-05,
419
- "loss": 0.0021,
420
  "step": 1675
421
  },
422
  {
423
- "epoch": 16.02,
424
- "learning_rate": 1.8350000000000002e-05,
425
- "loss": 0.002,
426
  "step": 1700
427
  },
428
  {
429
- "epoch": 17.0,
430
- "learning_rate": 1.821111111111111e-05,
431
- "loss": 0.002,
432
  "step": 1725
433
  },
434
  {
435
- "epoch": 17.01,
436
- "learning_rate": 1.8072222222222224e-05,
437
- "loss": 0.002,
438
  "step": 1750
439
  },
440
  {
441
- "epoch": 17.02,
442
- "learning_rate": 1.7933333333333337e-05,
443
- "loss": 0.0019,
444
  "step": 1775
445
  },
446
  {
447
- "epoch": 17.02,
448
- "learning_rate": 1.7794444444444443e-05,
449
- "loss": 0.0018,
450
  "step": 1800
451
  },
452
  {
453
- "epoch": 18.0,
454
- "learning_rate": 1.7655555555555555e-05,
455
- "loss": 0.0018,
456
  "step": 1825
457
  },
458
  {
459
- "epoch": 18.01,
460
- "learning_rate": 1.7516666666666668e-05,
461
- "loss": 0.0018,
462
  "step": 1850
463
  },
464
  {
465
- "epoch": 18.02,
466
- "learning_rate": 1.737777777777778e-05,
467
- "loss": 0.0017,
468
  "step": 1875
469
  },
470
  {
471
- "epoch": 18.02,
472
- "learning_rate": 1.723888888888889e-05,
473
- "loss": 0.0017,
474
  "step": 1900
475
  },
476
  {
477
- "epoch": 19.0,
478
- "learning_rate": 1.7100000000000002e-05,
479
- "loss": 0.0017,
480
  "step": 1925
481
  },
482
  {
483
- "epoch": 19.01,
484
- "learning_rate": 1.696111111111111e-05,
485
- "loss": 0.0017,
486
  "step": 1950
487
  },
488
  {
489
- "epoch": 19.02,
490
- "learning_rate": 1.6822222222222224e-05,
491
- "loss": 0.0016,
492
  "step": 1975
493
  },
494
  {
495
- "epoch": 19.02,
496
- "learning_rate": 1.6683333333333333e-05,
497
- "loss": 0.0015,
498
  "step": 2000
499
  },
500
  {
501
- "epoch": 19.02,
502
- "eval_loss": 0.4754403829574585,
503
- "eval_runtime": 678.1792,
504
- "eval_samples_per_second": 9.652,
505
- "eval_steps_per_second": 0.152,
506
- "eval_wer": 18.687913907284766,
507
  "step": 2000
508
  },
509
  {
510
- "epoch": 20.0,
511
- "learning_rate": 1.6544444444444446e-05,
512
- "loss": 0.0015,
513
  "step": 2025
514
  },
515
  {
516
- "epoch": 20.01,
517
- "learning_rate": 1.640555555555556e-05,
518
- "loss": 0.0015,
519
  "step": 2050
520
  },
521
  {
522
- "epoch": 20.02,
523
- "learning_rate": 1.6266666666666665e-05,
524
- "loss": 0.0014,
525
  "step": 2075
526
  },
527
  {
528
- "epoch": 20.02,
529
- "learning_rate": 1.6127777777777777e-05,
530
- "loss": 0.0014,
531
  "step": 2100
532
  },
533
  {
534
- "epoch": 21.0,
535
- "learning_rate": 1.598888888888889e-05,
536
- "loss": 0.0014,
537
  "step": 2125
538
  },
539
  {
540
- "epoch": 21.01,
541
- "learning_rate": 1.5850000000000002e-05,
542
- "loss": 0.0014,
543
  "step": 2150
544
  },
545
  {
546
- "epoch": 21.02,
547
- "learning_rate": 1.571111111111111e-05,
548
- "loss": 0.0013,
549
  "step": 2175
550
  },
551
  {
552
- "epoch": 21.02,
553
- "learning_rate": 1.5572222222222224e-05,
554
- "loss": 0.0013,
555
  "step": 2200
556
  },
557
  {
558
- "epoch": 22.0,
559
- "learning_rate": 1.5433333333333334e-05,
560
- "loss": 0.0013,
561
  "step": 2225
562
  },
563
  {
564
- "epoch": 22.01,
565
- "learning_rate": 1.5294444444444446e-05,
566
- "loss": 0.0013,
567
  "step": 2250
568
  },
569
  {
570
- "epoch": 22.02,
571
- "learning_rate": 1.5155555555555555e-05,
572
- "loss": 0.0012,
573
  "step": 2275
574
  },
575
  {
576
- "epoch": 22.02,
577
- "learning_rate": 1.5016666666666668e-05,
578
- "loss": 0.0012,
579
  "step": 2300
580
  },
581
  {
582
- "epoch": 23.0,
583
- "learning_rate": 1.4877777777777779e-05,
584
- "loss": 0.0012,
585
  "step": 2325
586
  },
587
  {
588
- "epoch": 23.01,
589
- "learning_rate": 1.4738888888888892e-05,
590
- "loss": 0.0012,
591
  "step": 2350
592
  },
593
  {
594
- "epoch": 23.02,
595
- "learning_rate": 1.4599999999999999e-05,
596
- "loss": 0.0011,
597
  "step": 2375
598
  },
599
  {
600
- "epoch": 23.02,
601
- "learning_rate": 1.4461111111111112e-05,
602
- "loss": 0.0011,
603
  "step": 2400
604
  },
605
  {
606
- "epoch": 24.0,
607
- "learning_rate": 1.4322222222222223e-05,
608
- "loss": 0.0011,
609
  "step": 2425
610
  },
611
  {
612
- "epoch": 24.01,
613
- "learning_rate": 1.4183333333333335e-05,
614
- "loss": 0.0011,
615
  "step": 2450
616
  },
617
  {
618
- "epoch": 24.02,
619
- "learning_rate": 1.4044444444444446e-05,
620
- "loss": 0.0011,
621
  "step": 2475
622
  },
623
  {
624
- "epoch": 24.02,
625
- "learning_rate": 1.3905555555555555e-05,
626
- "loss": 0.001,
627
  "step": 2500
628
  },
629
  {
630
- "epoch": 25.0,
631
- "learning_rate": 1.3766666666666666e-05,
632
- "loss": 0.001,
633
  "step": 2525
634
  },
635
  {
636
- "epoch": 25.01,
637
- "learning_rate": 1.3627777777777777e-05,
638
- "loss": 0.0011,
639
  "step": 2550
640
  },
641
  {
642
- "epoch": 25.02,
643
- "learning_rate": 1.348888888888889e-05,
644
- "loss": 0.001,
645
  "step": 2575
646
  },
647
  {
648
- "epoch": 25.02,
649
- "learning_rate": 1.3350000000000001e-05,
650
- "loss": 0.001,
651
  "step": 2600
652
  },
653
  {
654
- "epoch": 26.0,
655
- "learning_rate": 1.3211111111111114e-05,
656
- "loss": 0.001,
657
  "step": 2625
658
  },
659
  {
660
- "epoch": 26.01,
661
- "learning_rate": 1.3072222222222221e-05,
662
- "loss": 0.001,
663
  "step": 2650
664
  },
665
  {
666
- "epoch": 26.02,
667
- "learning_rate": 1.2933333333333334e-05,
668
- "loss": 0.0009,
669
  "step": 2675
670
  },
671
  {
672
- "epoch": 26.02,
673
- "learning_rate": 1.2794444444444445e-05,
674
- "loss": 0.0009,
675
  "step": 2700
676
  },
677
  {
678
- "epoch": 27.0,
679
- "learning_rate": 1.2655555555555557e-05,
680
- "loss": 0.0009,
681
  "step": 2725
682
  },
683
  {
684
- "epoch": 27.01,
685
- "learning_rate": 1.2516666666666668e-05,
686
- "loss": 0.0009,
687
  "step": 2750
688
  },
689
  {
690
- "epoch": 27.02,
691
- "learning_rate": 1.237777777777778e-05,
692
- "loss": 0.0009,
693
  "step": 2775
694
  },
695
  {
696
- "epoch": 27.02,
697
- "learning_rate": 1.223888888888889e-05,
698
- "loss": 0.0009,
699
  "step": 2800
700
  },
701
  {
702
- "epoch": 28.0,
703
- "learning_rate": 1.2100000000000001e-05,
704
- "loss": 0.0009,
705
  "step": 2825
706
  },
707
  {
708
- "epoch": 28.01,
709
- "learning_rate": 1.1961111111111112e-05,
710
- "loss": 0.0009,
711
  "step": 2850
712
  },
713
  {
714
- "epoch": 28.02,
715
- "learning_rate": 1.1822222222222223e-05,
716
- "loss": 0.0008,
717
  "step": 2875
718
  },
719
  {
720
- "epoch": 28.02,
721
- "learning_rate": 1.1683333333333334e-05,
722
- "loss": 0.0008,
723
  "step": 2900
724
  },
725
  {
726
- "epoch": 29.0,
727
- "learning_rate": 1.1544444444444445e-05,
728
- "loss": 0.0008,
729
  "step": 2925
730
  },
731
  {
732
- "epoch": 29.01,
733
- "learning_rate": 1.1405555555555556e-05,
734
- "loss": 0.0008,
735
  "step": 2950
736
  },
737
  {
738
- "epoch": 29.02,
739
- "learning_rate": 1.1266666666666667e-05,
740
- "loss": 0.0008,
741
  "step": 2975
742
  },
743
  {
744
- "epoch": 29.02,
745
- "learning_rate": 1.112777777777778e-05,
746
- "loss": 0.0008,
747
  "step": 3000
748
  },
749
  {
750
- "epoch": 29.02,
751
- "eval_loss": 0.5036433935165405,
752
- "eval_runtime": 722.6283,
753
- "eval_samples_per_second": 9.059,
754
- "eval_steps_per_second": 0.143,
755
- "eval_wer": 18.90004139072848,
756
  "step": 3000
757
  },
758
  {
759
- "epoch": 30.0,
760
- "learning_rate": 1.0988888888888889e-05,
761
- "loss": 0.0008,
762
  "step": 3025
763
  },
764
  {
765
- "epoch": 30.01,
766
- "learning_rate": 1.0850000000000001e-05,
767
- "loss": 0.0008,
768
  "step": 3050
769
  },
770
  {
771
- "epoch": 30.02,
772
- "learning_rate": 1.0711111111111112e-05,
773
- "loss": 0.0008,
774
  "step": 3075
775
  },
776
  {
777
- "epoch": 30.02,
778
- "learning_rate": 1.0572222222222223e-05,
779
- "loss": 0.0007,
780
  "step": 3100
781
  },
782
  {
783
- "epoch": 31.0,
784
- "learning_rate": 1.0433333333333334e-05,
785
- "loss": 0.0007,
786
  "step": 3125
787
  },
788
  {
789
- "epoch": 31.01,
790
- "learning_rate": 1.0294444444444445e-05,
791
- "loss": 0.0008,
792
  "step": 3150
793
  },
794
  {
795
- "epoch": 31.02,
796
- "learning_rate": 1.0155555555555556e-05,
797
- "loss": 0.0007,
798
  "step": 3175
799
  },
800
  {
801
- "epoch": 31.02,
802
- "learning_rate": 1.0016666666666667e-05,
803
- "loss": 0.0007,
804
  "step": 3200
805
  },
806
  {
807
- "epoch": 32.01,
808
- "learning_rate": 9.87777777777778e-06,
809
- "loss": 0.0007,
810
  "step": 3225
811
  },
812
  {
813
- "epoch": 32.01,
814
- "learning_rate": 9.738888888888889e-06,
815
- "loss": 0.0007,
816
  "step": 3250
817
  },
818
  {
819
- "epoch": 32.02,
820
- "learning_rate": 9.600000000000001e-06,
821
- "loss": 0.0007,
822
  "step": 3275
823
  },
824
  {
825
- "epoch": 32.02,
826
- "learning_rate": 9.461111111111112e-06,
827
- "loss": 0.0007,
828
  "step": 3300
829
  },
830
  {
831
- "epoch": 33.01,
832
- "learning_rate": 9.322222222222223e-06,
833
- "loss": 0.0007,
834
  "step": 3325
835
  },
836
  {
837
- "epoch": 33.01,
838
- "learning_rate": 9.183333333333334e-06,
839
- "loss": 0.0007,
840
  "step": 3350
841
  },
842
  {
843
- "epoch": 33.02,
844
- "learning_rate": 9.044444444444445e-06,
845
- "loss": 0.0007,
846
  "step": 3375
847
  },
848
  {
849
- "epoch": 33.02,
850
- "learning_rate": 8.905555555555556e-06,
851
- "loss": 0.0007,
852
  "step": 3400
853
  },
854
  {
855
- "epoch": 34.01,
856
- "learning_rate": 8.766666666666667e-06,
857
- "loss": 0.0007,
858
  "step": 3425
859
  },
860
  {
861
- "epoch": 34.01,
862
- "learning_rate": 8.627777777777778e-06,
863
- "loss": 0.0007,
864
  "step": 3450
865
  },
866
  {
867
- "epoch": 34.02,
868
- "learning_rate": 8.488888888888889e-06,
869
- "loss": 0.0006,
870
  "step": 3475
871
  },
872
  {
873
- "epoch": 34.02,
874
- "learning_rate": 8.350000000000001e-06,
875
- "loss": 0.0006,
876
  "step": 3500
877
  },
878
  {
879
- "epoch": 35.01,
880
- "learning_rate": 8.21111111111111e-06,
881
- "loss": 0.0006,
882
  "step": 3525
883
  },
884
  {
885
- "epoch": 35.01,
886
- "learning_rate": 8.072222222222223e-06,
887
- "loss": 0.0006,
888
  "step": 3550
889
  },
890
  {
891
- "epoch": 35.02,
892
- "learning_rate": 7.933333333333334e-06,
893
- "loss": 0.0006,
894
  "step": 3575
895
  },
896
  {
897
- "epoch": 35.02,
898
- "learning_rate": 7.794444444444445e-06,
899
- "loss": 0.0006,
900
  "step": 3600
901
  },
902
  {
903
- "epoch": 36.01,
904
- "learning_rate": 7.655555555555556e-06,
905
- "loss": 0.0006,
906
  "step": 3625
907
  },
908
  {
909
- "epoch": 36.01,
910
- "learning_rate": 7.516666666666668e-06,
911
- "loss": 0.0006,
912
  "step": 3650
913
  },
914
  {
915
- "epoch": 36.02,
916
- "learning_rate": 7.377777777777778e-06,
917
- "loss": 0.0006,
918
  "step": 3675
919
  },
920
  {
921
- "epoch": 36.02,
922
- "learning_rate": 7.23888888888889e-06,
923
- "loss": 0.0006,
924
  "step": 3700
925
  },
926
  {
927
- "epoch": 37.01,
928
- "learning_rate": 7.1e-06,
929
- "loss": 0.0006,
930
  "step": 3725
931
  },
932
  {
933
- "epoch": 37.01,
934
- "learning_rate": 6.9611111111111116e-06,
935
- "loss": 0.0006,
936
  "step": 3750
937
  },
938
  {
939
- "epoch": 37.02,
940
- "learning_rate": 6.8222222222222225e-06,
941
- "loss": 0.0006,
942
  "step": 3775
943
  },
944
  {
945
- "epoch": 37.02,
946
- "learning_rate": 6.6833333333333334e-06,
947
- "loss": 0.0006,
948
  "step": 3800
949
  },
950
  {
951
- "epoch": 38.01,
952
- "learning_rate": 6.544444444444444e-06,
953
- "loss": 0.0006,
954
  "step": 3825
955
  },
956
  {
957
- "epoch": 38.01,
958
- "learning_rate": 6.405555555555556e-06,
959
- "loss": 0.0006,
960
  "step": 3850
961
  },
962
  {
963
- "epoch": 38.02,
964
- "learning_rate": 6.266666666666666e-06,
965
- "loss": 0.0006,
966
  "step": 3875
967
  },
968
  {
969
- "epoch": 38.02,
970
- "learning_rate": 6.127777777777778e-06,
971
- "loss": 0.0005,
972
  "step": 3900
973
  },
974
  {
975
- "epoch": 39.01,
976
- "learning_rate": 5.988888888888889e-06,
977
- "loss": 0.0005,
978
  "step": 3925
979
  },
980
  {
981
- "epoch": 39.01,
982
- "learning_rate": 5.850000000000001e-06,
983
- "loss": 0.0006,
984
  "step": 3950
985
  },
986
  {
987
- "epoch": 39.02,
988
- "learning_rate": 5.711111111111112e-06,
989
- "loss": 0.0005,
990
  "step": 3975
991
  },
992
  {
993
- "epoch": 39.02,
994
- "learning_rate": 5.572222222222223e-06,
995
- "loss": 0.0005,
996
  "step": 4000
997
  },
998
  {
999
- "epoch": 39.02,
1000
- "eval_loss": 0.5224926471710205,
1001
- "eval_runtime": 756.7286,
1002
- "eval_samples_per_second": 8.65,
1003
- "eval_steps_per_second": 0.136,
1004
- "eval_wer": 19.055256622516556,
1005
  "step": 4000
1006
  },
1007
  {
1008
- "epoch": 40.01,
1009
- "learning_rate": 5.4333333333333335e-06,
1010
- "loss": 0.0005,
1011
  "step": 4025
1012
  },
1013
  {
1014
- "epoch": 40.01,
1015
- "learning_rate": 5.294444444444445e-06,
1016
- "loss": 0.0005,
1017
  "step": 4050
1018
  },
1019
  {
1020
- "epoch": 40.02,
1021
- "learning_rate": 5.155555555555555e-06,
1022
- "loss": 0.0005,
1023
  "step": 4075
1024
  },
1025
  {
1026
- "epoch": 40.02,
1027
- "learning_rate": 5.016666666666666e-06,
1028
- "loss": 0.0005,
1029
  "step": 4100
1030
  },
1031
  {
1032
- "epoch": 41.01,
1033
- "learning_rate": 4.877777777777778e-06,
1034
- "loss": 0.0005,
1035
  "step": 4125
1036
  },
1037
  {
1038
- "epoch": 41.01,
1039
- "learning_rate": 4.738888888888889e-06,
1040
- "loss": 0.0005,
1041
  "step": 4150
1042
  },
1043
  {
1044
- "epoch": 41.02,
1045
- "learning_rate": 4.6e-06,
1046
- "loss": 0.0005,
1047
  "step": 4175
1048
  },
1049
  {
1050
- "epoch": 41.02,
1051
- "learning_rate": 4.461111111111111e-06,
1052
- "loss": 0.0005,
1053
  "step": 4200
1054
  },
1055
  {
1056
- "epoch": 42.01,
1057
- "learning_rate": 4.322222222222223e-06,
1058
- "loss": 0.0005,
1059
  "step": 4225
1060
  },
1061
  {
1062
- "epoch": 42.01,
1063
- "learning_rate": 4.183333333333334e-06,
1064
- "loss": 0.0005,
1065
  "step": 4250
1066
  },
1067
  {
1068
- "epoch": 42.02,
1069
- "learning_rate": 4.044444444444445e-06,
1070
- "loss": 0.0005,
1071
  "step": 4275
1072
  },
1073
  {
1074
- "epoch": 42.02,
1075
- "learning_rate": 3.905555555555556e-06,
1076
- "loss": 0.0005,
1077
  "step": 4300
1078
  },
1079
  {
1080
- "epoch": 43.01,
1081
- "learning_rate": 3.766666666666667e-06,
1082
- "loss": 0.0005,
1083
  "step": 4325
1084
  },
1085
  {
1086
- "epoch": 43.01,
1087
- "learning_rate": 3.627777777777778e-06,
1088
- "loss": 0.0005,
1089
  "step": 4350
1090
  },
1091
  {
1092
- "epoch": 43.02,
1093
- "learning_rate": 3.4888888888888888e-06,
1094
- "loss": 0.0005,
1095
  "step": 4375
1096
  },
1097
  {
1098
- "epoch": 43.02,
1099
- "learning_rate": 3.3500000000000005e-06,
1100
- "loss": 0.0005,
1101
  "step": 4400
1102
  },
1103
  {
1104
- "epoch": 44.01,
1105
- "learning_rate": 3.2111111111111115e-06,
1106
- "loss": 0.0005,
1107
  "step": 4425
1108
  },
1109
  {
1110
- "epoch": 44.01,
1111
- "learning_rate": 3.0722222222222224e-06,
1112
- "loss": 0.0005,
1113
  "step": 4450
1114
  },
1115
  {
1116
- "epoch": 44.02,
1117
- "learning_rate": 2.9333333333333333e-06,
1118
- "loss": 0.0005,
1119
  "step": 4475
1120
  },
1121
  {
1122
- "epoch": 44.02,
1123
- "learning_rate": 2.7944444444444447e-06,
1124
- "loss": 0.0005,
1125
  "step": 4500
1126
  },
1127
  {
1128
- "epoch": 45.01,
1129
- "learning_rate": 2.6555555555555556e-06,
1130
- "loss": 0.0005,
1131
  "step": 4525
1132
  },
1133
  {
1134
- "epoch": 45.01,
1135
- "learning_rate": 2.516666666666667e-06,
1136
- "loss": 0.0005,
1137
  "step": 4550
1138
  },
1139
  {
1140
- "epoch": 45.02,
1141
- "learning_rate": 2.377777777777778e-06,
1142
- "loss": 0.0005,
1143
  "step": 4575
1144
  },
1145
  {
1146
- "epoch": 45.02,
1147
- "learning_rate": 2.238888888888889e-06,
1148
- "loss": 0.0005,
1149
  "step": 4600
1150
  },
1151
  {
1152
- "epoch": 46.01,
1153
- "learning_rate": 2.1000000000000002e-06,
1154
- "loss": 0.0005,
1155
  "step": 4625
1156
  },
1157
  {
1158
- "epoch": 46.01,
1159
- "learning_rate": 1.961111111111111e-06,
1160
- "loss": 0.0005,
1161
  "step": 4650
1162
  },
1163
  {
1164
- "epoch": 46.02,
1165
- "learning_rate": 1.8222222222222223e-06,
1166
- "loss": 0.0005,
1167
  "step": 4675
1168
  },
1169
  {
1170
- "epoch": 46.02,
1171
- "learning_rate": 1.6833333333333332e-06,
1172
- "loss": 0.0005,
1173
  "step": 4700
1174
  },
1175
  {
1176
- "epoch": 47.01,
1177
- "learning_rate": 1.5444444444444446e-06,
1178
- "loss": 0.0005,
1179
  "step": 4725
1180
  },
1181
  {
1182
- "epoch": 47.01,
1183
- "learning_rate": 1.4055555555555555e-06,
1184
- "loss": 0.0005,
1185
  "step": 4750
1186
  },
1187
  {
1188
- "epoch": 47.02,
1189
- "learning_rate": 1.2666666666666667e-06,
1190
- "loss": 0.0005,
1191
  "step": 4775
1192
  },
1193
  {
1194
- "epoch": 47.02,
1195
- "learning_rate": 1.1277777777777778e-06,
1196
- "loss": 0.0004,
1197
  "step": 4800
1198
  },
1199
  {
1200
- "epoch": 48.01,
1201
- "learning_rate": 9.888888888888888e-07,
1202
- "loss": 0.0004,
1203
  "step": 4825
1204
  },
1205
  {
1206
- "epoch": 48.01,
1207
- "learning_rate": 8.500000000000001e-07,
1208
- "loss": 0.0005,
1209
  "step": 4850
1210
  },
1211
  {
1212
- "epoch": 48.02,
1213
- "learning_rate": 7.111111111111112e-07,
1214
- "loss": 0.0004,
1215
  "step": 4875
1216
  },
1217
  {
1218
- "epoch": 48.02,
1219
- "learning_rate": 5.722222222222223e-07,
1220
- "loss": 0.0004,
1221
  "step": 4900
1222
  },
1223
  {
1224
- "epoch": 49.01,
1225
- "learning_rate": 4.3333333333333335e-07,
1226
- "loss": 0.0004,
1227
  "step": 4925
1228
  },
1229
  {
1230
- "epoch": 49.01,
1231
- "learning_rate": 2.9444444444444444e-07,
1232
- "loss": 0.0005,
1233
  "step": 4950
1234
  },
1235
  {
1236
- "epoch": 49.02,
1237
- "learning_rate": 1.5555555555555556e-07,
1238
- "loss": 0.0004,
1239
  "step": 4975
1240
  },
1241
  {
1242
- "epoch": 49.02,
1243
- "learning_rate": 1.6666666666666667e-08,
1244
- "loss": 0.0004,
1245
  "step": 5000
1246
  },
1247
  {
1248
- "epoch": 49.02,
1249
- "eval_loss": 0.5307408571243286,
1250
- "eval_runtime": 729.6417,
1251
- "eval_samples_per_second": 8.972,
1252
- "eval_steps_per_second": 0.141,
1253
- "eval_wer": 19.112168874172188,
1254
  "step": 5000
1255
  },
1256
  {
1257
- "epoch": 49.02,
1258
  "step": 5000,
1259
  "total_flos": 4.1490939543552e+19,
1260
- "train_loss": 0.054723825027979914,
1261
- "train_runtime": 73614.7434,
1262
- "train_samples_per_second": 8.694,
1263
- "train_steps_per_second": 0.068
1264
  }
1265
  ],
1266
  "logging_steps": 25,
1267
  "max_steps": 5000,
1268
- "num_train_epochs": 9223372036854775807,
 
1269
  "save_steps": 1000,
1270
  "total_flos": 4.1490939543552e+19,
 
1271
  "trial_name": null,
1272
  "trial_params": null
1273
  }
 
1
  {
2
+ "best_metric": 17.290976821192054,
3
+ "best_model_checkpoint": "./checkpoint-5000",
4
+ "epoch": 50.0,
5
  "eval_steps": 1000,
6
  "global_step": 5000,
7
  "is_hyper_param_search": false,
 
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.25,
13
+ "learning_rate": 1.0500000000000001e-06,
14
+ "loss": 10.5596,
15
  "step": 25
16
  },
17
  {
18
+ "epoch": 0.5,
19
+ "learning_rate": 2.3e-06,
20
+ "loss": 8.9561,
21
  "step": 50
22
  },
23
  {
24
+ "epoch": 0.75,
25
+ "learning_rate": 3.55e-06,
26
+ "loss": 7.3512,
27
  "step": 75
28
  },
29
  {
30
+ "epoch": 1.0,
31
+ "learning_rate": 4.800000000000001e-06,
32
+ "loss": 6.3362,
33
  "step": 100
34
  },
35
  {
36
+ "epoch": 1.25,
37
+ "learning_rate": 6.0500000000000005e-06,
38
+ "loss": 5.6865,
39
  "step": 125
40
  },
41
  {
42
+ "epoch": 1.5,
43
+ "learning_rate": 7.2999999999999996e-06,
44
+ "loss": 5.3089,
45
  "step": 150
46
  },
47
  {
48
+ "epoch": 1.75,
49
+ "learning_rate": 8.550000000000001e-06,
50
+ "loss": 5.0576,
51
  "step": 175
52
  },
53
  {
54
+ "epoch": 2.0,
55
+ "learning_rate": 9.800000000000001e-06,
56
+ "loss": 4.7969,
57
  "step": 200
58
  },
59
  {
60
+ "epoch": 2.25,
61
+ "learning_rate": 1.1050000000000001e-05,
62
+ "loss": 4.0424,
63
  "step": 225
64
  },
65
  {
66
+ "epoch": 2.5,
67
+ "learning_rate": 1.23e-05,
68
+ "loss": 3.0227,
69
  "step": 250
70
  },
71
  {
72
+ "epoch": 2.75,
73
+ "learning_rate": 1.3550000000000002e-05,
74
+ "loss": 2.3498,
75
  "step": 275
76
  },
77
  {
78
+ "epoch": 3.0,
79
+ "learning_rate": 1.48e-05,
80
+ "loss": 1.9522,
81
  "step": 300
82
  },
83
  {
84
+ "epoch": 3.25,
85
+ "learning_rate": 1.605e-05,
86
+ "loss": 1.653,
87
  "step": 325
88
  },
89
  {
90
+ "epoch": 3.5,
91
+ "learning_rate": 1.73e-05,
92
+ "loss": 1.4606,
93
  "step": 350
94
  },
95
  {
96
+ "epoch": 3.75,
97
+ "learning_rate": 1.855e-05,
98
+ "loss": 1.3339,
99
  "step": 375
100
  },
101
  {
102
+ "epoch": 4.0,
103
+ "learning_rate": 1.9800000000000004e-05,
104
+ "loss": 1.227,
105
  "step": 400
106
  },
107
  {
108
+ "epoch": 4.25,
109
+ "learning_rate": 2.105e-05,
110
+ "loss": 1.0619,
111
  "step": 425
112
  },
113
  {
114
+ "epoch": 4.5,
115
+ "learning_rate": 2.23e-05,
116
+ "loss": 1.0081,
117
  "step": 450
118
  },
119
  {
120
+ "epoch": 4.75,
121
+ "learning_rate": 2.355e-05,
122
+ "loss": 0.9446,
123
  "step": 475
124
  },
125
  {
126
+ "epoch": 5.0,
127
+ "learning_rate": 2.48e-05,
128
+ "loss": 0.8899,
129
  "step": 500
130
  },
131
  {
132
+ "epoch": 5.25,
133
+ "learning_rate": 2.4883333333333333e-05,
134
+ "loss": 0.7849,
135
  "step": 525
136
  },
137
  {
138
+ "epoch": 5.5,
139
+ "learning_rate": 2.4744444444444445e-05,
140
+ "loss": 0.7657,
141
  "step": 550
142
  },
143
  {
144
+ "epoch": 5.75,
145
+ "learning_rate": 2.4605555555555558e-05,
146
+ "loss": 0.724,
147
  "step": 575
148
  },
149
  {
150
+ "epoch": 6.0,
151
+ "learning_rate": 2.4466666666666667e-05,
152
+ "loss": 0.6948,
153
  "step": 600
154
  },
155
  {
156
+ "epoch": 6.25,
157
+ "learning_rate": 2.432777777777778e-05,
158
+ "loss": 0.6251,
159
  "step": 625
160
  },
161
  {
162
+ "epoch": 6.5,
163
+ "learning_rate": 2.418888888888889e-05,
164
+ "loss": 0.6195,
165
  "step": 650
166
  },
167
  {
168
+ "epoch": 6.75,
169
+ "learning_rate": 2.4050000000000002e-05,
170
+ "loss": 0.5928,
171
  "step": 675
172
  },
173
  {
174
+ "epoch": 7.0,
175
+ "learning_rate": 2.391111111111111e-05,
176
+ "loss": 0.5598,
177
  "step": 700
178
  },
179
  {
180
+ "epoch": 7.25,
181
+ "learning_rate": 2.3772222222222224e-05,
182
+ "loss": 0.5138,
183
  "step": 725
184
  },
185
  {
186
+ "epoch": 7.5,
187
+ "learning_rate": 2.3633333333333336e-05,
188
+ "loss": 0.5117,
189
  "step": 750
190
  },
191
  {
192
+ "epoch": 7.75,
193
+ "learning_rate": 2.3494444444444446e-05,
194
+ "loss": 0.5036,
195
  "step": 775
196
  },
197
  {
198
+ "epoch": 8.0,
199
+ "learning_rate": 2.3355555555555555e-05,
200
+ "loss": 0.4974,
201
  "step": 800
202
  },
203
  {
204
+ "epoch": 8.25,
205
+ "learning_rate": 2.3216666666666667e-05,
206
+ "loss": 0.4315,
207
  "step": 825
208
  },
209
  {
210
+ "epoch": 8.5,
211
+ "learning_rate": 2.307777777777778e-05,
212
+ "loss": 0.4377,
213
  "step": 850
214
  },
215
  {
216
+ "epoch": 8.75,
217
+ "learning_rate": 2.293888888888889e-05,
218
+ "loss": 0.4326,
219
  "step": 875
220
  },
221
  {
222
+ "epoch": 9.0,
223
+ "learning_rate": 2.2800000000000002e-05,
224
+ "loss": 0.4239,
225
  "step": 900
226
  },
227
  {
228
+ "epoch": 9.25,
229
+ "learning_rate": 2.2661111111111115e-05,
230
+ "loss": 0.3777,
231
  "step": 925
232
  },
233
  {
234
+ "epoch": 9.5,
235
+ "learning_rate": 2.2522222222222224e-05,
236
+ "loss": 0.3767,
237
  "step": 950
238
  },
239
  {
240
+ "epoch": 9.75,
241
+ "learning_rate": 2.2383333333333333e-05,
242
+ "loss": 0.3797,
243
  "step": 975
244
  },
245
  {
246
+ "epoch": 10.0,
247
+ "learning_rate": 2.2244444444444446e-05,
248
+ "loss": 0.372,
249
  "step": 1000
250
  },
251
  {
252
+ "epoch": 10.0,
253
+ "eval_loss": 0.41730424761772156,
254
+ "eval_runtime": 79.3095,
255
+ "eval_samples_per_second": 82.537,
256
+ "eval_steps_per_second": 1.299,
257
+ "eval_wer": 21.002345474613687,
258
  "step": 1000
259
  },
260
  {
261
+ "epoch": 10.25,
262
+ "learning_rate": 2.2105555555555558e-05,
263
+ "loss": 0.3307,
264
  "step": 1025
265
  },
266
  {
267
+ "epoch": 10.5,
268
+ "learning_rate": 2.1966666666666668e-05,
269
+ "loss": 0.3349,
270
  "step": 1050
271
  },
272
  {
273
+ "epoch": 10.75,
274
+ "learning_rate": 2.1827777777777777e-05,
275
+ "loss": 0.3286,
276
  "step": 1075
277
  },
278
  {
279
+ "epoch": 11.0,
280
+ "learning_rate": 2.168888888888889e-05,
281
+ "loss": 0.3373,
282
  "step": 1100
283
  },
284
  {
285
+ "epoch": 11.25,
286
+ "learning_rate": 2.1550000000000002e-05,
287
+ "loss": 0.2945,
288
  "step": 1125
289
  },
290
  {
291
+ "epoch": 11.5,
292
+ "learning_rate": 2.141111111111111e-05,
293
+ "loss": 0.291,
294
  "step": 1150
295
  },
296
  {
297
+ "epoch": 11.75,
298
+ "learning_rate": 2.1272222222222224e-05,
299
+ "loss": 0.2983,
300
  "step": 1175
301
  },
302
  {
303
+ "epoch": 12.0,
304
+ "learning_rate": 2.1133333333333337e-05,
305
+ "loss": 0.2892,
306
  "step": 1200
307
  },
308
  {
309
+ "epoch": 12.25,
310
+ "learning_rate": 2.0994444444444446e-05,
311
+ "loss": 0.2618,
312
  "step": 1225
313
  },
314
  {
315
+ "epoch": 12.5,
316
+ "learning_rate": 2.0855555555555555e-05,
317
+ "loss": 0.2616,
318
  "step": 1250
319
  },
320
  {
321
+ "epoch": 12.75,
322
+ "learning_rate": 2.0716666666666668e-05,
323
+ "loss": 0.2663,
324
  "step": 1275
325
  },
326
  {
327
+ "epoch": 13.0,
328
+ "learning_rate": 2.057777777777778e-05,
329
+ "loss": 0.271,
330
  "step": 1300
331
  },
332
  {
333
+ "epoch": 13.25,
334
+ "learning_rate": 2.043888888888889e-05,
335
+ "loss": 0.2376,
336
  "step": 1325
337
  },
338
  {
339
+ "epoch": 13.5,
340
+ "learning_rate": 2.0300000000000002e-05,
341
+ "loss": 0.24,
342
  "step": 1350
343
  },
344
  {
345
+ "epoch": 13.75,
346
+ "learning_rate": 2.016111111111111e-05,
347
+ "loss": 0.232,
348
  "step": 1375
349
  },
350
  {
351
+ "epoch": 14.0,
352
+ "learning_rate": 2.0022222222222224e-05,
353
+ "loss": 0.2366,
354
  "step": 1400
355
  },
356
  {
357
+ "epoch": 14.25,
358
+ "learning_rate": 1.9883333333333333e-05,
359
+ "loss": 0.21,
360
  "step": 1425
361
  },
362
  {
363
+ "epoch": 14.5,
364
+ "learning_rate": 1.9744444444444446e-05,
365
+ "loss": 0.2101,
366
  "step": 1450
367
  },
368
  {
369
+ "epoch": 14.75,
370
+ "learning_rate": 1.960555555555556e-05,
371
+ "loss": 0.2151,
372
  "step": 1475
373
  },
374
  {
375
+ "epoch": 15.0,
376
+ "learning_rate": 1.9466666666666668e-05,
377
+ "loss": 0.2157,
378
  "step": 1500
379
  },
380
  {
381
+ "epoch": 15.25,
382
+ "learning_rate": 1.9327777777777777e-05,
383
+ "loss": 0.192,
384
  "step": 1525
385
  },
386
  {
387
+ "epoch": 15.5,
388
+ "learning_rate": 1.918888888888889e-05,
389
+ "loss": 0.1885,
390
  "step": 1550
391
  },
392
  {
393
+ "epoch": 15.75,
394
+ "learning_rate": 1.9050000000000002e-05,
395
+ "loss": 0.1953,
396
  "step": 1575
397
  },
398
  {
399
+ "epoch": 16.0,
400
+ "learning_rate": 1.891111111111111e-05,
401
+ "loss": 0.1973,
402
  "step": 1600
403
  },
404
  {
405
+ "epoch": 16.25,
406
+ "learning_rate": 1.8772222222222224e-05,
407
+ "loss": 0.1735,
408
  "step": 1625
409
  },
410
  {
411
+ "epoch": 16.5,
412
+ "learning_rate": 1.8633333333333333e-05,
413
+ "loss": 0.1723,
414
  "step": 1650
415
  },
416
  {
417
+ "epoch": 16.75,
418
+ "learning_rate": 1.8494444444444446e-05,
419
+ "loss": 0.1776,
420
  "step": 1675
421
  },
422
  {
423
+ "epoch": 17.0,
424
+ "learning_rate": 1.8355555555555555e-05,
425
+ "loss": 0.1755,
426
  "step": 1700
427
  },
428
  {
429
+ "epoch": 17.25,
430
+ "learning_rate": 1.8216666666666668e-05,
431
+ "loss": 0.1581,
432
  "step": 1725
433
  },
434
  {
435
+ "epoch": 17.5,
436
+ "learning_rate": 1.807777777777778e-05,
437
+ "loss": 0.1598,
438
  "step": 1750
439
  },
440
  {
441
+ "epoch": 17.75,
442
+ "learning_rate": 1.793888888888889e-05,
443
+ "loss": 0.1606,
444
  "step": 1775
445
  },
446
  {
447
+ "epoch": 18.0,
448
+ "learning_rate": 1.78e-05,
449
+ "loss": 0.1635,
450
  "step": 1800
451
  },
452
  {
453
+ "epoch": 18.25,
454
+ "learning_rate": 1.766111111111111e-05,
455
+ "loss": 0.1423,
456
  "step": 1825
457
  },
458
  {
459
+ "epoch": 18.5,
460
+ "learning_rate": 1.7522222222222224e-05,
461
+ "loss": 0.1469,
462
  "step": 1850
463
  },
464
  {
465
+ "epoch": 18.75,
466
+ "learning_rate": 1.7383333333333333e-05,
467
+ "loss": 0.1436,
468
  "step": 1875
469
  },
470
  {
471
+ "epoch": 19.0,
472
+ "learning_rate": 1.7244444444444446e-05,
473
+ "loss": 0.1506,
474
  "step": 1900
475
  },
476
  {
477
+ "epoch": 19.25,
478
+ "learning_rate": 1.7105555555555555e-05,
479
+ "loss": 0.1308,
480
  "step": 1925
481
  },
482
  {
483
+ "epoch": 19.5,
484
+ "learning_rate": 1.6966666666666668e-05,
485
+ "loss": 0.1331,
486
  "step": 1950
487
  },
488
  {
489
+ "epoch": 19.75,
490
+ "learning_rate": 1.6827777777777777e-05,
491
+ "loss": 0.1369,
492
  "step": 1975
493
  },
494
  {
495
+ "epoch": 20.0,
496
+ "learning_rate": 1.668888888888889e-05,
497
+ "loss": 0.1352,
498
  "step": 2000
499
  },
500
  {
501
+ "epoch": 20.0,
502
+ "eval_loss": 0.3981981873512268,
503
+ "eval_runtime": 75.1441,
504
+ "eval_samples_per_second": 87.113,
505
+ "eval_steps_per_second": 1.371,
506
+ "eval_wer": 18.3619619205298,
507
  "step": 2000
508
  },
509
  {
510
+ "epoch": 20.25,
511
+ "learning_rate": 1.6550000000000002e-05,
512
+ "loss": 0.1218,
513
  "step": 2025
514
  },
515
  {
516
+ "epoch": 20.5,
517
+ "learning_rate": 1.6411111111111112e-05,
518
+ "loss": 0.1216,
519
  "step": 2050
520
  },
521
  {
522
+ "epoch": 20.75,
523
+ "learning_rate": 1.627222222222222e-05,
524
+ "loss": 0.1231,
525
  "step": 2075
526
  },
527
  {
528
+ "epoch": 21.0,
529
+ "learning_rate": 1.6133333333333334e-05,
530
+ "loss": 0.1266,
531
  "step": 2100
532
  },
533
  {
534
+ "epoch": 21.25,
535
+ "learning_rate": 1.5994444444444446e-05,
536
+ "loss": 0.1149,
537
  "step": 2125
538
  },
539
  {
540
+ "epoch": 21.5,
541
+ "learning_rate": 1.5855555555555555e-05,
542
+ "loss": 0.1103,
543
  "step": 2150
544
  },
545
  {
546
+ "epoch": 21.75,
547
+ "learning_rate": 1.5716666666666668e-05,
548
+ "loss": 0.1155,
549
  "step": 2175
550
  },
551
  {
552
+ "epoch": 22.0,
553
+ "learning_rate": 1.5577777777777777e-05,
554
+ "loss": 0.1156,
555
  "step": 2200
556
  },
557
  {
558
+ "epoch": 22.25,
559
+ "learning_rate": 1.543888888888889e-05,
560
+ "loss": 0.1008,
561
  "step": 2225
562
  },
563
  {
564
+ "epoch": 22.5,
565
+ "learning_rate": 1.53e-05,
566
+ "loss": 0.1038,
567
  "step": 2250
568
  },
569
  {
570
+ "epoch": 22.75,
571
+ "learning_rate": 1.5161111111111112e-05,
572
+ "loss": 0.105,
573
  "step": 2275
574
  },
575
  {
576
+ "epoch": 23.0,
577
+ "learning_rate": 1.5022222222222224e-05,
578
+ "loss": 0.105,
579
  "step": 2300
580
  },
581
  {
582
+ "epoch": 23.25,
583
+ "learning_rate": 1.4883333333333335e-05,
584
+ "loss": 0.0928,
585
  "step": 2325
586
  },
587
  {
588
+ "epoch": 23.5,
589
+ "learning_rate": 1.4744444444444445e-05,
590
+ "loss": 0.0967,
591
  "step": 2350
592
  },
593
  {
594
+ "epoch": 23.75,
595
+ "learning_rate": 1.4605555555555556e-05,
596
+ "loss": 0.0973,
597
  "step": 2375
598
  },
599
  {
600
+ "epoch": 24.0,
601
+ "learning_rate": 1.4466666666666667e-05,
602
+ "loss": 0.0977,
603
  "step": 2400
604
  },
605
  {
606
+ "epoch": 24.25,
607
+ "learning_rate": 1.4327777777777779e-05,
608
+ "loss": 0.0884,
609
  "step": 2425
610
  },
611
  {
612
+ "epoch": 24.5,
613
+ "learning_rate": 1.418888888888889e-05,
614
+ "loss": 0.0859,
615
  "step": 2450
616
  },
617
  {
618
+ "epoch": 24.75,
619
+ "learning_rate": 1.4050000000000003e-05,
620
+ "loss": 0.089,
621
  "step": 2475
622
  },
623
  {
624
+ "epoch": 25.0,
625
+ "learning_rate": 1.391111111111111e-05,
626
+ "loss": 0.0906,
627
  "step": 2500
628
  },
629
  {
630
+ "epoch": 25.25,
631
+ "learning_rate": 1.3772222222222223e-05,
632
+ "loss": 0.0818,
633
  "step": 2525
634
  },
635
  {
636
+ "epoch": 25.5,
637
+ "learning_rate": 1.3633333333333334e-05,
638
+ "loss": 0.0827,
639
  "step": 2550
640
  },
641
  {
642
+ "epoch": 25.75,
643
+ "learning_rate": 1.3494444444444446e-05,
644
+ "loss": 0.0834,
645
  "step": 2575
646
  },
647
  {
648
+ "epoch": 26.0,
649
+ "learning_rate": 1.3355555555555557e-05,
650
+ "loss": 0.0809,
651
  "step": 2600
652
  },
653
  {
654
+ "epoch": 26.25,
655
+ "learning_rate": 1.3216666666666667e-05,
656
+ "loss": 0.0749,
657
  "step": 2625
658
  },
659
  {
660
+ "epoch": 26.5,
661
+ "learning_rate": 1.3077777777777778e-05,
662
+ "loss": 0.0775,
663
  "step": 2650
664
  },
665
  {
666
+ "epoch": 26.75,
667
+ "learning_rate": 1.2938888888888888e-05,
668
+ "loss": 0.079,
669
  "step": 2675
670
  },
671
  {
672
+ "epoch": 27.0,
673
+ "learning_rate": 1.2800000000000001e-05,
674
+ "loss": 0.0777,
675
  "step": 2700
676
  },
677
  {
678
+ "epoch": 27.25,
679
+ "learning_rate": 1.2661111111111112e-05,
680
+ "loss": 0.0704,
681
  "step": 2725
682
  },
683
  {
684
+ "epoch": 27.5,
685
+ "learning_rate": 1.2522222222222225e-05,
686
+ "loss": 0.0706,
687
  "step": 2750
688
  },
689
  {
690
+ "epoch": 27.75,
691
+ "learning_rate": 1.2383333333333334e-05,
692
+ "loss": 0.072,
693
  "step": 2775
694
  },
695
  {
696
+ "epoch": 28.0,
697
+ "learning_rate": 1.2244444444444445e-05,
698
+ "loss": 0.0732,
699
  "step": 2800
700
  },
701
  {
702
+ "epoch": 28.25,
703
+ "learning_rate": 1.2105555555555556e-05,
704
+ "loss": 0.0645,
705
  "step": 2825
706
  },
707
  {
708
+ "epoch": 28.5,
709
+ "learning_rate": 1.1966666666666668e-05,
710
+ "loss": 0.0659,
711
  "step": 2850
712
  },
713
  {
714
+ "epoch": 28.75,
715
+ "learning_rate": 1.1827777777777778e-05,
716
+ "loss": 0.0663,
717
  "step": 2875
718
  },
719
  {
720
+ "epoch": 29.0,
721
+ "learning_rate": 1.168888888888889e-05,
722
+ "loss": 0.0704,
723
  "step": 2900
724
  },
725
  {
726
+ "epoch": 29.25,
727
+ "learning_rate": 1.1550000000000001e-05,
728
+ "loss": 0.0635,
729
  "step": 2925
730
  },
731
  {
732
+ "epoch": 29.5,
733
+ "learning_rate": 1.141111111111111e-05,
734
+ "loss": 0.0615,
735
  "step": 2950
736
  },
737
  {
738
+ "epoch": 29.75,
739
+ "learning_rate": 1.1272222222222223e-05,
740
+ "loss": 0.0611,
741
  "step": 2975
742
  },
743
  {
744
+ "epoch": 30.0,
745
+ "learning_rate": 1.1133333333333334e-05,
746
+ "loss": 0.0638,
747
  "step": 3000
748
  },
749
  {
750
+ "epoch": 30.0,
751
+ "eval_loss": 0.4174603819847107,
752
+ "eval_runtime": 75.4718,
753
+ "eval_samples_per_second": 86.734,
754
+ "eval_steps_per_second": 1.365,
755
+ "eval_wer": 17.884243929359823,
756
  "step": 3000
757
  },
758
  {
759
+ "epoch": 30.25,
760
+ "learning_rate": 1.0994444444444445e-05,
761
+ "loss": 0.06,
762
  "step": 3025
763
  },
764
  {
765
+ "epoch": 30.5,
766
+ "learning_rate": 1.0855555555555556e-05,
767
+ "loss": 0.0583,
768
  "step": 3050
769
  },
770
  {
771
+ "epoch": 30.75,
772
+ "learning_rate": 1.0716666666666667e-05,
773
+ "loss": 0.0587,
774
  "step": 3075
775
  },
776
  {
777
+ "epoch": 31.0,
778
+ "learning_rate": 1.0577777777777778e-05,
779
+ "loss": 0.0615,
780
  "step": 3100
781
  },
782
  {
783
+ "epoch": 31.25,
784
+ "learning_rate": 1.043888888888889e-05,
785
+ "loss": 0.054,
786
  "step": 3125
787
  },
788
  {
789
+ "epoch": 31.5,
790
+ "learning_rate": 1.03e-05,
791
+ "loss": 0.0542,
792
  "step": 3150
793
  },
794
  {
795
+ "epoch": 31.75,
796
+ "learning_rate": 1.0161111111111112e-05,
797
+ "loss": 0.0559,
798
  "step": 3175
799
  },
800
  {
801
+ "epoch": 32.0,
802
+ "learning_rate": 1.0022222222222223e-05,
803
+ "loss": 0.055,
804
  "step": 3200
805
  },
806
  {
807
+ "epoch": 32.25,
808
+ "learning_rate": 9.883333333333334e-06,
809
+ "loss": 0.0518,
810
  "step": 3225
811
  },
812
  {
813
+ "epoch": 32.5,
814
+ "learning_rate": 9.744444444444445e-06,
815
+ "loss": 0.0524,
816
  "step": 3250
817
  },
818
  {
819
+ "epoch": 32.75,
820
+ "learning_rate": 9.605555555555556e-06,
821
+ "loss": 0.0512,
822
  "step": 3275
823
  },
824
  {
825
+ "epoch": 33.0,
826
+ "learning_rate": 9.466666666666667e-06,
827
+ "loss": 0.0538,
828
  "step": 3300
829
  },
830
  {
831
+ "epoch": 33.25,
832
+ "learning_rate": 9.327777777777778e-06,
833
+ "loss": 0.0483,
834
  "step": 3325
835
  },
836
  {
837
+ "epoch": 33.5,
838
+ "learning_rate": 9.18888888888889e-06,
839
+ "loss": 0.0489,
840
  "step": 3350
841
  },
842
  {
843
+ "epoch": 33.75,
844
+ "learning_rate": 9.05e-06,
845
+ "loss": 0.0506,
846
  "step": 3375
847
  },
848
  {
849
+ "epoch": 34.0,
850
+ "learning_rate": 8.911111111111112e-06,
851
+ "loss": 0.05,
852
  "step": 3400
853
  },
854
  {
855
+ "epoch": 34.25,
856
+ "learning_rate": 8.772222222222222e-06,
857
+ "loss": 0.0456,
858
  "step": 3425
859
  },
860
  {
861
+ "epoch": 34.5,
862
+ "learning_rate": 8.633333333333334e-06,
863
+ "loss": 0.0463,
864
  "step": 3450
865
  },
866
  {
867
+ "epoch": 34.75,
868
+ "learning_rate": 8.494444444444445e-06,
869
+ "loss": 0.0485,
870
  "step": 3475
871
  },
872
  {
873
+ "epoch": 35.0,
874
+ "learning_rate": 8.355555555555556e-06,
875
+ "loss": 0.0476,
876
  "step": 3500
877
  },
878
  {
879
+ "epoch": 35.25,
880
+ "learning_rate": 8.216666666666667e-06,
881
+ "loss": 0.0449,
882
  "step": 3525
883
  },
884
  {
885
+ "epoch": 35.5,
886
+ "learning_rate": 8.077777777777778e-06,
887
+ "loss": 0.0438,
888
  "step": 3550
889
  },
890
  {
891
+ "epoch": 35.75,
892
+ "learning_rate": 7.938888888888889e-06,
893
+ "loss": 0.0447,
894
  "step": 3575
895
  },
896
  {
897
+ "epoch": 36.0,
898
+ "learning_rate": 7.8e-06,
899
+ "loss": 0.0461,
900
  "step": 3600
901
  },
902
  {
903
+ "epoch": 36.25,
904
+ "learning_rate": 7.661111111111112e-06,
905
+ "loss": 0.0413,
906
  "step": 3625
907
  },
908
  {
909
+ "epoch": 36.5,
910
+ "learning_rate": 7.5222222222222226e-06,
911
+ "loss": 0.0419,
912
  "step": 3650
913
  },
914
  {
915
+ "epoch": 36.75,
916
+ "learning_rate": 7.3833333333333335e-06,
917
+ "loss": 0.0426,
918
  "step": 3675
919
  },
920
  {
921
+ "epoch": 37.0,
922
+ "learning_rate": 7.244444444444445e-06,
923
+ "loss": 0.0429,
924
  "step": 3700
925
  },
926
  {
927
+ "epoch": 37.25,
928
+ "learning_rate": 7.105555555555555e-06,
929
+ "loss": 0.0402,
930
  "step": 3725
931
  },
932
  {
933
+ "epoch": 37.5,
934
+ "learning_rate": 6.966666666666667e-06,
935
+ "loss": 0.0395,
936
  "step": 3750
937
  },
938
  {
939
+ "epoch": 37.75,
940
+ "learning_rate": 6.827777777777779e-06,
941
+ "loss": 0.0402,
942
  "step": 3775
943
  },
944
  {
945
+ "epoch": 38.0,
946
+ "learning_rate": 6.688888888888889e-06,
947
+ "loss": 0.041,
948
  "step": 3800
949
  },
950
  {
951
+ "epoch": 38.25,
952
+ "learning_rate": 6.550000000000001e-06,
953
+ "loss": 0.0388,
954
  "step": 3825
955
  },
956
  {
957
+ "epoch": 38.5,
958
+ "learning_rate": 6.411111111111111e-06,
959
+ "loss": 0.0381,
960
  "step": 3850
961
  },
962
  {
963
+ "epoch": 38.75,
964
+ "learning_rate": 6.272222222222223e-06,
965
+ "loss": 0.0394,
966
  "step": 3875
967
  },
968
  {
969
+ "epoch": 39.0,
970
+ "learning_rate": 6.133333333333334e-06,
971
+ "loss": 0.0394,
972
  "step": 3900
973
  },
974
  {
975
+ "epoch": 39.25,
976
+ "learning_rate": 5.9944444444444446e-06,
977
+ "loss": 0.0363,
978
  "step": 3925
979
  },
980
  {
981
+ "epoch": 39.5,
982
+ "learning_rate": 5.8555555555555555e-06,
983
+ "loss": 0.0363,
984
  "step": 3950
985
  },
986
  {
987
+ "epoch": 39.75,
988
+ "learning_rate": 5.7166666666666664e-06,
989
+ "loss": 0.0367,
990
  "step": 3975
991
  },
992
  {
993
+ "epoch": 40.0,
994
+ "learning_rate": 5.577777777777778e-06,
995
+ "loss": 0.0371,
996
  "step": 4000
997
  },
998
  {
999
+ "epoch": 40.0,
1000
+ "eval_loss": 0.4310153126716614,
1001
+ "eval_runtime": 75.5592,
1002
+ "eval_samples_per_second": 86.634,
1003
+ "eval_steps_per_second": 1.363,
1004
+ "eval_wer": 17.472061258278146,
1005
  "step": 4000
1006
  },
1007
  {
1008
+ "epoch": 40.25,
1009
+ "learning_rate": 5.438888888888889e-06,
1010
+ "loss": 0.0371,
1011
  "step": 4025
1012
  },
1013
  {
1014
+ "epoch": 40.5,
1015
+ "learning_rate": 5.3e-06,
1016
+ "loss": 0.0359,
1017
  "step": 4050
1018
  },
1019
  {
1020
+ "epoch": 40.75,
1021
+ "learning_rate": 5.161111111111112e-06,
1022
+ "loss": 0.036,
1023
  "step": 4075
1024
  },
1025
  {
1026
+ "epoch": 41.0,
1027
+ "learning_rate": 5.022222222222223e-06,
1028
+ "loss": 0.0358,
1029
  "step": 4100
1030
  },
1031
  {
1032
+ "epoch": 41.25,
1033
+ "learning_rate": 4.883333333333334e-06,
1034
+ "loss": 0.0347,
1035
  "step": 4125
1036
  },
1037
  {
1038
+ "epoch": 41.5,
1039
+ "learning_rate": 4.744444444444445e-06,
1040
+ "loss": 0.0344,
1041
  "step": 4150
1042
  },
1043
  {
1044
+ "epoch": 41.75,
1045
+ "learning_rate": 4.605555555555556e-06,
1046
+ "loss": 0.0353,
1047
  "step": 4175
1048
  },
1049
  {
1050
+ "epoch": 42.0,
1051
+ "learning_rate": 4.4666666666666665e-06,
1052
+ "loss": 0.0347,
1053
  "step": 4200
1054
  },
1055
  {
1056
+ "epoch": 42.25,
1057
+ "learning_rate": 4.3277777777777775e-06,
1058
+ "loss": 0.0324,
1059
  "step": 4225
1060
  },
1061
  {
1062
+ "epoch": 42.5,
1063
+ "learning_rate": 4.188888888888889e-06,
1064
+ "loss": 0.0349,
1065
  "step": 4250
1066
  },
1067
  {
1068
+ "epoch": 42.75,
1069
+ "learning_rate": 4.05e-06,
1070
+ "loss": 0.0342,
1071
  "step": 4275
1072
  },
1073
  {
1074
+ "epoch": 43.0,
1075
+ "learning_rate": 3.911111111111111e-06,
1076
+ "loss": 0.0343,
1077
  "step": 4300
1078
  },
1079
  {
1080
+ "epoch": 43.25,
1081
+ "learning_rate": 3.772222222222222e-06,
1082
+ "loss": 0.0323,
1083
  "step": 4325
1084
  },
1085
  {
1086
+ "epoch": 43.5,
1087
+ "learning_rate": 3.633333333333334e-06,
1088
+ "loss": 0.034,
1089
  "step": 4350
1090
  },
1091
  {
1092
+ "epoch": 43.75,
1093
+ "learning_rate": 3.4944444444444448e-06,
1094
+ "loss": 0.0334,
1095
  "step": 4375
1096
  },
1097
  {
1098
+ "epoch": 44.0,
1099
+ "learning_rate": 3.3555555555555557e-06,
1100
+ "loss": 0.0322,
1101
  "step": 4400
1102
  },
1103
  {
1104
+ "epoch": 44.25,
1105
+ "learning_rate": 3.216666666666667e-06,
1106
+ "loss": 0.0317,
1107
  "step": 4425
1108
  },
1109
  {
1110
+ "epoch": 44.5,
1111
+ "learning_rate": 3.077777777777778e-06,
1112
+ "loss": 0.0323,
1113
  "step": 4450
1114
  },
1115
  {
1116
+ "epoch": 44.75,
1117
+ "learning_rate": 2.938888888888889e-06,
1118
+ "loss": 0.0312,
1119
  "step": 4475
1120
  },
1121
  {
1122
+ "epoch": 45.0,
1123
+ "learning_rate": 2.8000000000000003e-06,
1124
+ "loss": 0.0308,
1125
  "step": 4500
1126
  },
1127
  {
1128
+ "epoch": 45.25,
1129
+ "learning_rate": 2.6611111111111112e-06,
1130
+ "loss": 0.0312,
1131
  "step": 4525
1132
  },
1133
  {
1134
+ "epoch": 45.5,
1135
+ "learning_rate": 2.522222222222222e-06,
1136
+ "loss": 0.03,
1137
  "step": 4550
1138
  },
1139
  {
1140
+ "epoch": 45.75,
1141
+ "learning_rate": 2.3833333333333335e-06,
1142
+ "loss": 0.0313,
1143
  "step": 4575
1144
  },
1145
  {
1146
+ "epoch": 46.0,
1147
+ "learning_rate": 2.2444444444444445e-06,
1148
+ "loss": 0.031,
1149
  "step": 4600
1150
  },
1151
  {
1152
+ "epoch": 46.25,
1153
+ "learning_rate": 2.105555555555556e-06,
1154
+ "loss": 0.0297,
1155
  "step": 4625
1156
  },
1157
  {
1158
+ "epoch": 46.5,
1159
+ "learning_rate": 1.9666666666666668e-06,
1160
+ "loss": 0.0305,
1161
  "step": 4650
1162
  },
1163
  {
1164
+ "epoch": 46.75,
1165
+ "learning_rate": 1.827777777777778e-06,
1166
+ "loss": 0.031,
1167
  "step": 4675
1168
  },
1169
  {
1170
+ "epoch": 47.0,
1171
+ "learning_rate": 1.6888888888888888e-06,
1172
+ "loss": 0.0308,
1173
  "step": 4700
1174
  },
1175
  {
1176
+ "epoch": 47.25,
1177
+ "learning_rate": 1.55e-06,
1178
+ "loss": 0.0297,
1179
  "step": 4725
1180
  },
1181
  {
1182
+ "epoch": 47.5,
1183
+ "learning_rate": 1.4111111111111111e-06,
1184
+ "loss": 0.0301,
1185
  "step": 4750
1186
  },
1187
  {
1188
+ "epoch": 47.75,
1189
+ "learning_rate": 1.2722222222222223e-06,
1190
+ "loss": 0.0298,
1191
  "step": 4775
1192
  },
1193
  {
1194
+ "epoch": 48.0,
1195
+ "learning_rate": 1.1333333333333334e-06,
1196
+ "loss": 0.0299,
1197
  "step": 4800
1198
  },
1199
  {
1200
+ "epoch": 48.25,
1201
+ "learning_rate": 9.944444444444446e-07,
1202
+ "loss": 0.029,
1203
  "step": 4825
1204
  },
1205
  {
1206
+ "epoch": 48.5,
1207
+ "learning_rate": 8.555555555555556e-07,
1208
+ "loss": 0.03,
1209
  "step": 4850
1210
  },
1211
  {
1212
+ "epoch": 48.75,
1213
+ "learning_rate": 7.166666666666667e-07,
1214
+ "loss": 0.0289,
1215
  "step": 4875
1216
  },
1217
  {
1218
+ "epoch": 49.0,
1219
+ "learning_rate": 5.777777777777778e-07,
1220
+ "loss": 0.0302,
1221
  "step": 4900
1222
  },
1223
  {
1224
+ "epoch": 49.25,
1225
+ "learning_rate": 4.3888888888888895e-07,
1226
+ "loss": 0.0298,
1227
  "step": 4925
1228
  },
1229
  {
1230
+ "epoch": 49.5,
1231
+ "learning_rate": 3.0000000000000004e-07,
1232
+ "loss": 0.0293,
1233
  "step": 4950
1234
  },
1235
  {
1236
+ "epoch": 49.75,
1237
+ "learning_rate": 1.611111111111111e-07,
1238
+ "loss": 0.0292,
1239
  "step": 4975
1240
  },
1241
  {
1242
+ "epoch": 50.0,
1243
+ "learning_rate": 2.2222222222222224e-08,
1244
+ "loss": 0.0279,
1245
  "step": 5000
1246
  },
1247
  {
1248
+ "epoch": 50.0,
1249
+ "eval_loss": 0.43604204058647156,
1250
+ "eval_runtime": 75.2936,
1251
+ "eval_samples_per_second": 86.94,
1252
+ "eval_steps_per_second": 1.368,
1253
+ "eval_wer": 17.290976821192054,
1254
  "step": 5000
1255
  },
1256
  {
1257
+ "epoch": 50.0,
1258
  "step": 5000,
1259
  "total_flos": 4.1490939543552e+19,
1260
+ "train_loss": 0.5045194145917893,
1261
+ "train_runtime": 4484.1133,
1262
+ "train_samples_per_second": 142.726,
1263
+ "train_steps_per_second": 1.115
1264
  }
1265
  ],
1266
  "logging_steps": 25,
1267
  "max_steps": 5000,
1268
+ "num_input_tokens_seen": 0,
1269
+ "num_train_epochs": 50,
1270
  "save_steps": 1000,
1271
  "total_flos": 4.1490939543552e+19,
1272
+ "train_batch_size": 128,
1273
  "trial_name": null,
1274
  "trial_params": null
1275
  }