agnesluhtaru commited on
Commit
63bd7c2
1 Parent(s): b87ec76

End of training

Browse files
all_results.json CHANGED
@@ -1,12 +1,12 @@
1
  {
2
- "epoch": 16.13,
3
- "eval_loss": 0.3861992359161377,
4
- "eval_runtime": 2316.9306,
5
- "eval_samples_per_second": 0.555,
6
- "eval_steps_per_second": 0.555,
7
- "eval_wer": 15.541324622856406,
8
- "train_loss": 0.032722438401542604,
9
- "train_runtime": 36967.2072,
10
- "train_samples_per_second": 1.731,
11
  "train_steps_per_second": 0.054
12
  }
 
1
  {
2
+ "epoch": 32.26,
3
+ "eval_loss": 0.4131947159767151,
4
+ "eval_runtime": 2301.1875,
5
+ "eval_samples_per_second": 0.559,
6
+ "eval_steps_per_second": 0.559,
7
+ "eval_wer": 15.734731594103234,
8
+ "train_loss": 0.017603627420263366,
9
+ "train_runtime": 74040.5343,
10
+ "train_samples_per_second": 1.729,
11
  "train_steps_per_second": 0.054
12
  }
eval_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
- "epoch": 16.13,
3
- "eval_loss": 0.3861992359161377,
4
- "eval_runtime": 2316.9306,
5
- "eval_samples_per_second": 0.555,
6
- "eval_steps_per_second": 0.555,
7
- "eval_wer": 15.541324622856406
8
  }
 
1
  {
2
+ "epoch": 32.26,
3
+ "eval_loss": 0.4131947159767151,
4
+ "eval_runtime": 2301.1875,
5
+ "eval_samples_per_second": 0.559,
6
+ "eval_steps_per_second": 0.559,
7
+ "eval_wer": 15.734731594103234
8
  }
runs/Jan26_14-56-49_nid007241/events.out.tfevents.1674816418.nid007241.27557.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15ba4ccbd1fd46421e5acb084f5941d95259c946b92efbc34c6a6ec0321e5d96
3
+ size 358
train_results.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
- "epoch": 16.13,
3
- "train_loss": 0.032722438401542604,
4
- "train_runtime": 36967.2072,
5
- "train_samples_per_second": 1.731,
6
  "train_steps_per_second": 0.054
7
  }
 
1
  {
2
+ "epoch": 32.26,
3
+ "train_loss": 0.017603627420263366,
4
+ "train_runtime": 74040.5343,
5
+ "train_samples_per_second": 1.729,
6
  "train_steps_per_second": 0.054
7
  }
trainer_state.json CHANGED
@@ -1,541 +1,1057 @@
1
  {
2
- "best_metric": 15.541324622856406,
3
- "best_model_checkpoint": "/scratch/project_465000370/asr_children/models/whisper-large-et-clinic/checkpoint-2000",
4
- "epoch": 16.12806403201601,
5
- "global_step": 2000,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
  "epoch": 0.2,
12
- "learning_rate": 1.1500000000000002e-06,
13
- "loss": 0.495,
14
  "step": 25
15
  },
16
  {
17
  "epoch": 0.4,
18
- "learning_rate": 2.4000000000000003e-06,
19
- "loss": 0.2703,
20
  "step": 50
21
  },
22
  {
23
  "epoch": 0.6,
24
- "learning_rate": 3.65e-06,
25
- "loss": 0.2068,
26
  "step": 75
27
  },
28
  {
29
  "epoch": 0.8,
30
- "learning_rate": 4.9000000000000005e-06,
31
- "loss": 0.2183,
32
  "step": 100
33
  },
34
  {
35
  "epoch": 1.01,
36
- "learning_rate": 6.15e-06,
37
- "loss": 0.1988,
38
  "step": 125
39
  },
40
  {
41
  "epoch": 1.21,
42
  "learning_rate": 7.4e-06,
43
- "loss": 0.1218,
44
  "step": 150
45
  },
46
  {
47
  "epoch": 1.41,
48
  "learning_rate": 8.65e-06,
49
- "loss": 0.1199,
50
  "step": 175
51
  },
52
  {
53
  "epoch": 1.61,
54
  "learning_rate": 9.9e-06,
55
- "loss": 0.1196,
56
  "step": 200
57
  },
58
  {
59
  "epoch": 1.81,
60
- "learning_rate": 9.872222222222223e-06,
61
- "loss": 0.1177,
62
  "step": 225
63
  },
64
  {
65
  "epoch": 2.02,
66
- "learning_rate": 9.733333333333334e-06,
67
- "loss": 0.1277,
68
  "step": 250
69
  },
70
  {
71
  "epoch": 2.22,
72
- "learning_rate": 9.594444444444445e-06,
73
- "loss": 0.0481,
74
  "step": 275
75
  },
76
  {
77
  "epoch": 2.42,
78
- "learning_rate": 9.455555555555557e-06,
79
- "loss": 0.0542,
80
  "step": 300
81
  },
82
  {
83
  "epoch": 2.62,
84
- "learning_rate": 9.316666666666667e-06,
85
- "loss": 0.06,
86
  "step": 325
87
  },
88
  {
89
  "epoch": 2.82,
90
- "learning_rate": 9.17777777777778e-06,
91
- "loss": 0.0543,
92
  "step": 350
93
  },
94
  {
95
  "epoch": 3.02,
96
- "learning_rate": 9.038888888888889e-06,
97
- "loss": 0.0481,
98
  "step": 375
99
  },
100
  {
101
  "epoch": 3.22,
102
- "learning_rate": 8.900000000000001e-06,
103
- "loss": 0.0238,
104
  "step": 400
105
  },
106
  {
107
  "epoch": 3.42,
108
- "learning_rate": 8.761111111111112e-06,
109
- "loss": 0.0236,
110
  "step": 425
111
  },
112
  {
113
  "epoch": 3.62,
114
- "learning_rate": 8.627777777777778e-06,
115
- "loss": 0.0265,
116
  "step": 450
117
  },
118
  {
119
  "epoch": 3.82,
120
- "learning_rate": 8.48888888888889e-06,
121
- "loss": 0.0277,
122
  "step": 475
123
  },
124
  {
125
  "epoch": 4.03,
126
- "learning_rate": 8.35e-06,
127
- "loss": 0.0278,
128
  "step": 500
129
  },
130
  {
131
  "epoch": 4.03,
132
- "eval_loss": 0.2985570430755615,
133
- "eval_runtime": 2329.8998,
134
- "eval_samples_per_second": 0.552,
135
- "eval_steps_per_second": 0.552,
136
- "eval_wer": 16.99402587355482,
137
  "step": 500
138
  },
139
  {
140
  "epoch": 4.23,
141
- "learning_rate": 8.211111111111112e-06,
142
- "loss": 0.013,
143
  "step": 525
144
  },
145
  {
146
  "epoch": 4.43,
147
- "learning_rate": 8.072222222222223e-06,
148
- "loss": 0.0138,
149
  "step": 550
150
  },
151
  {
152
  "epoch": 4.63,
153
- "learning_rate": 7.933333333333334e-06,
154
  "loss": 0.0165,
155
  "step": 575
156
  },
157
  {
158
  "epoch": 4.83,
159
- "learning_rate": 7.794444444444445e-06,
160
- "loss": 0.0145,
161
  "step": 600
162
  },
163
  {
164
  "epoch": 5.04,
165
- "learning_rate": 7.655555555555556e-06,
166
- "loss": 0.0143,
167
  "step": 625
168
  },
169
  {
170
  "epoch": 5.24,
171
- "learning_rate": 7.516666666666668e-06,
172
- "loss": 0.008,
173
  "step": 650
174
  },
175
  {
176
  "epoch": 5.44,
177
- "learning_rate": 7.377777777777778e-06,
178
- "loss": 0.0103,
179
  "step": 675
180
  },
181
  {
182
  "epoch": 5.64,
183
- "learning_rate": 7.23888888888889e-06,
184
- "loss": 0.0081,
185
  "step": 700
186
  },
187
  {
188
  "epoch": 5.84,
189
- "learning_rate": 7.100000000000001e-06,
190
- "loss": 0.0083,
191
  "step": 725
192
  },
193
  {
194
  "epoch": 6.05,
195
- "learning_rate": 6.9611111111111116e-06,
196
- "loss": 0.0087,
197
  "step": 750
198
  },
199
  {
200
  "epoch": 6.25,
201
- "learning_rate": 6.8222222222222225e-06,
202
- "loss": 0.007,
203
  "step": 775
204
  },
205
  {
206
  "epoch": 6.45,
207
- "learning_rate": 6.683333333333334e-06,
208
- "loss": 0.0054,
209
  "step": 800
210
  },
211
  {
212
  "epoch": 6.65,
213
- "learning_rate": 6.544444444444445e-06,
214
- "loss": 0.0072,
215
  "step": 825
216
  },
217
  {
218
  "epoch": 6.85,
219
- "learning_rate": 6.405555555555556e-06,
220
- "loss": 0.0046,
221
  "step": 850
222
  },
223
  {
224
  "epoch": 7.06,
225
- "learning_rate": 6.266666666666668e-06,
226
- "loss": 0.0053,
227
  "step": 875
228
  },
229
  {
230
  "epoch": 7.26,
231
- "learning_rate": 6.127777777777778e-06,
232
- "loss": 0.0036,
233
  "step": 900
234
  },
235
  {
236
  "epoch": 7.46,
237
- "learning_rate": 5.98888888888889e-06,
238
- "loss": 0.0035,
239
  "step": 925
240
  },
241
  {
242
  "epoch": 7.66,
243
- "learning_rate": 5.85e-06,
244
- "loss": 0.0039,
245
  "step": 950
246
  },
247
  {
248
  "epoch": 7.86,
249
- "learning_rate": 5.711111111111112e-06,
250
- "loss": 0.004,
251
  "step": 975
252
  },
253
  {
254
  "epoch": 8.06,
255
- "learning_rate": 5.572222222222223e-06,
256
- "loss": 0.0037,
257
  "step": 1000
258
  },
259
  {
260
  "epoch": 8.06,
261
- "eval_loss": 0.3382585346698761,
262
- "eval_runtime": 2322.4054,
263
- "eval_samples_per_second": 0.554,
264
- "eval_steps_per_second": 0.554,
265
- "eval_wer": 16.17741866162376,
266
  "step": 1000
267
  },
268
  {
269
  "epoch": 8.26,
270
- "learning_rate": 5.4333333333333335e-06,
271
- "loss": 0.002,
272
  "step": 1025
273
  },
274
  {
275
  "epoch": 8.46,
276
- "learning_rate": 5.294444444444445e-06,
277
- "loss": 0.0036,
278
  "step": 1050
279
  },
280
  {
281
  "epoch": 8.66,
282
- "learning_rate": 5.155555555555556e-06,
283
- "loss": 0.0029,
284
  "step": 1075
285
  },
286
  {
287
  "epoch": 8.86,
288
- "learning_rate": 5.016666666666667e-06,
289
- "loss": 0.0026,
290
  "step": 1100
291
  },
292
  {
293
  "epoch": 9.07,
294
- "learning_rate": 4.877777777777778e-06,
295
- "loss": 0.0023,
296
  "step": 1125
297
  },
298
  {
299
  "epoch": 9.27,
300
- "learning_rate": 4.73888888888889e-06,
301
- "loss": 0.0023,
302
  "step": 1150
303
  },
304
  {
305
  "epoch": 9.47,
306
- "learning_rate": 4.600000000000001e-06,
307
- "loss": 0.0023,
308
  "step": 1175
309
  },
310
  {
311
  "epoch": 9.67,
312
- "learning_rate": 4.461111111111112e-06,
313
- "loss": 0.0019,
314
  "step": 1200
315
  },
316
  {
317
  "epoch": 9.87,
318
- "learning_rate": 4.322222222222223e-06,
319
- "loss": 0.002,
320
  "step": 1225
321
  },
322
  {
323
  "epoch": 10.08,
324
- "learning_rate": 4.183333333333334e-06,
325
- "loss": 0.0017,
326
  "step": 1250
327
  },
328
  {
329
  "epoch": 10.28,
330
- "learning_rate": 4.044444444444445e-06,
331
- "loss": 0.0014,
332
  "step": 1275
333
  },
334
  {
335
  "epoch": 10.48,
336
- "learning_rate": 3.9055555555555555e-06,
337
- "loss": 0.0021,
338
  "step": 1300
339
  },
340
  {
341
  "epoch": 10.68,
342
- "learning_rate": 3.766666666666667e-06,
343
- "loss": 0.0017,
344
  "step": 1325
345
  },
346
  {
347
  "epoch": 10.88,
348
- "learning_rate": 3.627777777777778e-06,
349
- "loss": 0.0024,
350
  "step": 1350
351
  },
352
  {
353
  "epoch": 11.09,
354
- "learning_rate": 3.4888888888888896e-06,
355
- "loss": 0.0015,
356
  "step": 1375
357
  },
358
  {
359
  "epoch": 11.29,
360
- "learning_rate": 3.3500000000000005e-06,
361
- "loss": 0.0012,
362
  "step": 1400
363
  },
364
  {
365
  "epoch": 11.49,
366
- "learning_rate": 3.2111111111111115e-06,
367
- "loss": 0.0012,
368
  "step": 1425
369
  },
370
  {
371
  "epoch": 11.69,
372
- "learning_rate": 3.0722222222222224e-06,
373
- "loss": 0.0017,
374
  "step": 1450
375
  },
376
  {
377
  "epoch": 11.89,
378
- "learning_rate": 2.9333333333333338e-06,
379
- "loss": 0.0022,
380
  "step": 1475
381
  },
382
  {
383
  "epoch": 12.1,
384
- "learning_rate": 2.7944444444444447e-06,
385
- "loss": 0.0017,
386
  "step": 1500
387
  },
388
  {
389
  "epoch": 12.1,
390
- "eval_loss": 0.3618781566619873,
391
- "eval_runtime": 2327.1444,
392
- "eval_samples_per_second": 0.553,
393
- "eval_steps_per_second": 0.553,
394
- "eval_wer": 15.803498517213221,
395
  "step": 1500
396
  },
397
  {
398
  "epoch": 12.3,
399
- "learning_rate": 2.6555555555555556e-06,
400
- "loss": 0.0012,
401
  "step": 1525
402
  },
403
  {
404
  "epoch": 12.5,
405
- "learning_rate": 2.5166666666666666e-06,
406
- "loss": 0.0016,
407
  "step": 1550
408
  },
409
  {
410
  "epoch": 12.7,
411
- "learning_rate": 2.377777777777778e-06,
412
- "loss": 0.0016,
413
  "step": 1575
414
  },
415
  {
416
  "epoch": 12.9,
417
- "learning_rate": 2.238888888888889e-06,
418
- "loss": 0.0012,
419
  "step": 1600
420
  },
421
  {
422
  "epoch": 13.1,
423
- "learning_rate": 2.1000000000000002e-06,
424
- "loss": 0.0012,
425
  "step": 1625
426
  },
427
  {
428
  "epoch": 13.3,
429
- "learning_rate": 1.9611111111111116e-06,
430
- "loss": 0.0011,
431
  "step": 1650
432
  },
433
  {
434
  "epoch": 13.5,
435
- "learning_rate": 1.8222222222222225e-06,
436
- "loss": 0.0014,
437
  "step": 1675
438
  },
439
  {
440
  "epoch": 13.7,
441
- "learning_rate": 1.6833333333333335e-06,
442
- "loss": 0.0012,
443
  "step": 1700
444
  },
445
  {
446
  "epoch": 13.9,
447
- "learning_rate": 1.5444444444444446e-06,
448
- "loss": 0.0013,
449
  "step": 1725
450
  },
451
  {
452
  "epoch": 14.11,
453
- "learning_rate": 1.4055555555555555e-06,
454
- "loss": 0.0013,
455
  "step": 1750
456
  },
457
  {
458
  "epoch": 14.31,
459
- "learning_rate": 1.2666666666666669e-06,
460
- "loss": 0.0012,
461
  "step": 1775
462
  },
463
  {
464
  "epoch": 14.51,
465
- "learning_rate": 1.1277777777777778e-06,
466
- "loss": 0.001,
467
  "step": 1800
468
  },
469
  {
470
  "epoch": 14.71,
471
- "learning_rate": 9.88888888888889e-07,
472
- "loss": 0.001,
473
  "step": 1825
474
  },
475
  {
476
  "epoch": 14.91,
477
- "learning_rate": 8.500000000000001e-07,
478
- "loss": 0.0011,
479
  "step": 1850
480
  },
481
  {
482
  "epoch": 15.12,
483
- "learning_rate": 7.111111111111112e-07,
484
- "loss": 0.0011,
485
  "step": 1875
486
  },
487
  {
488
  "epoch": 15.32,
489
- "learning_rate": 5.722222222222223e-07,
490
- "loss": 0.001,
491
  "step": 1900
492
  },
493
  {
494
  "epoch": 15.52,
495
- "learning_rate": 4.333333333333334e-07,
496
- "loss": 0.0009,
497
  "step": 1925
498
  },
499
  {
500
  "epoch": 15.72,
501
- "learning_rate": 2.9444444444444444e-07,
502
- "loss": 0.0012,
503
  "step": 1950
504
  },
505
  {
506
  "epoch": 15.92,
507
- "learning_rate": 1.5555555555555556e-07,
508
- "loss": 0.0012,
509
  "step": 1975
510
  },
511
  {
512
  "epoch": 16.13,
513
- "learning_rate": 1.666666666666667e-08,
514
- "loss": 0.0008,
515
  "step": 2000
516
  },
517
  {
518
  "epoch": 16.13,
519
- "eval_loss": 0.3861992359161377,
520
- "eval_runtime": 2318.6816,
521
- "eval_samples_per_second": 0.555,
522
- "eval_steps_per_second": 0.555,
523
- "eval_wer": 15.541324622856406,
524
  "step": 2000
525
  },
526
  {
527
- "epoch": 16.13,
528
- "step": 2000,
529
- "total_flos": 1.36902067421184e+20,
530
- "train_loss": 0.032722438401542604,
531
- "train_runtime": 36967.2072,
532
- "train_samples_per_second": 1.731,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
533
  "train_steps_per_second": 0.054
534
  }
535
  ],
536
- "max_steps": 2000,
537
- "num_train_epochs": 17,
538
- "total_flos": 1.36902067421184e+20,
539
  "trial_name": null,
540
  "trial_params": null
541
  }
 
1
  {
2
+ "best_metric": 15.734731594103234,
3
+ "best_model_checkpoint": "/scratch/project_465000370/asr_children/models/whisper-large-et-clinic/checkpoint-4000",
4
+ "epoch": 32.25612806403202,
5
+ "global_step": 4000,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
  {
11
  "epoch": 0.2,
12
+ "learning_rate": 1.2000000000000002e-06,
13
+ "loss": 0.4973,
14
  "step": 25
15
  },
16
  {
17
  "epoch": 0.4,
18
+ "learning_rate": 2.4500000000000003e-06,
19
+ "loss": 0.2667,
20
  "step": 50
21
  },
22
  {
23
  "epoch": 0.6,
24
+ "learning_rate": 3.7e-06,
25
+ "loss": 0.2056,
26
  "step": 75
27
  },
28
  {
29
  "epoch": 0.8,
30
+ "learning_rate": 4.95e-06,
31
+ "loss": 0.2174,
32
  "step": 100
33
  },
34
  {
35
  "epoch": 1.01,
36
+ "learning_rate": 6.200000000000001e-06,
37
+ "loss": 0.198,
38
  "step": 125
39
  },
40
  {
41
  "epoch": 1.21,
42
  "learning_rate": 7.4e-06,
43
+ "loss": 0.1203,
44
  "step": 150
45
  },
46
  {
47
  "epoch": 1.41,
48
  "learning_rate": 8.65e-06,
49
+ "loss": 0.1191,
50
  "step": 175
51
  },
52
  {
53
  "epoch": 1.61,
54
  "learning_rate": 9.9e-06,
55
+ "loss": 0.1183,
56
  "step": 200
57
  },
58
  {
59
  "epoch": 1.81,
60
+ "learning_rate": 9.939473684210526e-06,
61
+ "loss": 0.1172,
62
  "step": 225
63
  },
64
  {
65
  "epoch": 2.02,
66
+ "learning_rate": 9.873684210526317e-06,
67
+ "loss": 0.1264,
68
  "step": 250
69
  },
70
  {
71
  "epoch": 2.22,
72
+ "learning_rate": 9.807894736842106e-06,
73
+ "loss": 0.0478,
74
  "step": 275
75
  },
76
  {
77
  "epoch": 2.42,
78
+ "learning_rate": 9.742105263157897e-06,
79
+ "loss": 0.0544,
80
  "step": 300
81
  },
82
  {
83
  "epoch": 2.62,
84
+ "learning_rate": 9.678947368421053e-06,
85
+ "loss": 0.0605,
86
  "step": 325
87
  },
88
  {
89
  "epoch": 2.82,
90
+ "learning_rate": 9.613157894736844e-06,
91
+ "loss": 0.0575,
92
  "step": 350
93
  },
94
  {
95
  "epoch": 3.02,
96
+ "learning_rate": 9.547368421052631e-06,
97
+ "loss": 0.0494,
98
  "step": 375
99
  },
100
  {
101
  "epoch": 3.22,
102
+ "learning_rate": 9.481578947368422e-06,
103
+ "loss": 0.0262,
104
  "step": 400
105
  },
106
  {
107
  "epoch": 3.42,
108
+ "learning_rate": 9.415789473684211e-06,
109
+ "loss": 0.0237,
110
  "step": 425
111
  },
112
  {
113
  "epoch": 3.62,
114
+ "learning_rate": 9.350000000000002e-06,
115
+ "loss": 0.0293,
116
  "step": 450
117
  },
118
  {
119
  "epoch": 3.82,
120
+ "learning_rate": 9.28421052631579e-06,
121
+ "loss": 0.0284,
122
  "step": 475
123
  },
124
  {
125
  "epoch": 4.03,
126
+ "learning_rate": 9.21842105263158e-06,
127
+ "loss": 0.0297,
128
  "step": 500
129
  },
130
  {
131
  "epoch": 4.03,
132
+ "eval_loss": 0.3037151098251343,
133
+ "eval_runtime": 2297.7169,
134
+ "eval_samples_per_second": 0.56,
135
+ "eval_steps_per_second": 0.56,
136
+ "eval_wer": 16.71466024842051,
137
  "step": 500
138
  },
139
  {
140
  "epoch": 4.23,
141
+ "learning_rate": 9.15263157894737e-06,
142
+ "loss": 0.0151,
143
  "step": 525
144
  },
145
  {
146
  "epoch": 4.43,
147
+ "learning_rate": 9.086842105263158e-06,
148
+ "loss": 0.0161,
149
  "step": 550
150
  },
151
  {
152
  "epoch": 4.63,
153
+ "learning_rate": 9.021052631578948e-06,
154
  "loss": 0.0165,
155
  "step": 575
156
  },
157
  {
158
  "epoch": 4.83,
159
+ "learning_rate": 8.955263157894738e-06,
160
+ "loss": 0.017,
161
  "step": 600
162
  },
163
  {
164
  "epoch": 5.04,
165
+ "learning_rate": 8.889473684210528e-06,
166
+ "loss": 0.0154,
167
  "step": 625
168
  },
169
  {
170
  "epoch": 5.24,
171
+ "learning_rate": 8.823684210526317e-06,
172
+ "loss": 0.0093,
173
  "step": 650
174
  },
175
  {
176
  "epoch": 5.44,
177
+ "learning_rate": 8.757894736842106e-06,
178
+ "loss": 0.0109,
179
  "step": 675
180
  },
181
  {
182
  "epoch": 5.64,
183
+ "learning_rate": 8.692105263157895e-06,
184
+ "loss": 0.009,
185
  "step": 700
186
  },
187
  {
188
  "epoch": 5.84,
189
+ "learning_rate": 8.626315789473686e-06,
190
+ "loss": 0.0109,
191
  "step": 725
192
  },
193
  {
194
  "epoch": 6.05,
195
+ "learning_rate": 8.560526315789475e-06,
196
+ "loss": 0.0105,
197
  "step": 750
198
  },
199
  {
200
  "epoch": 6.25,
201
+ "learning_rate": 8.494736842105264e-06,
202
+ "loss": 0.0077,
203
  "step": 775
204
  },
205
  {
206
  "epoch": 6.45,
207
+ "learning_rate": 8.428947368421053e-06,
208
+ "loss": 0.0063,
209
  "step": 800
210
  },
211
  {
212
  "epoch": 6.65,
213
+ "learning_rate": 8.363157894736842e-06,
214
+ "loss": 0.0093,
215
  "step": 825
216
  },
217
  {
218
  "epoch": 6.85,
219
+ "learning_rate": 8.297368421052631e-06,
220
+ "loss": 0.0072,
221
  "step": 850
222
  },
223
  {
224
  "epoch": 7.06,
225
+ "learning_rate": 8.231578947368422e-06,
226
+ "loss": 0.0081,
227
  "step": 875
228
  },
229
  {
230
  "epoch": 7.26,
231
+ "learning_rate": 8.165789473684211e-06,
232
+ "loss": 0.0047,
233
  "step": 900
234
  },
235
  {
236
  "epoch": 7.46,
237
+ "learning_rate": 8.1e-06,
238
+ "loss": 0.0058,
239
  "step": 925
240
  },
241
  {
242
  "epoch": 7.66,
243
+ "learning_rate": 8.03421052631579e-06,
244
+ "loss": 0.0051,
245
  "step": 950
246
  },
247
  {
248
  "epoch": 7.86,
249
+ "learning_rate": 7.96842105263158e-06,
250
+ "loss": 0.0056,
251
  "step": 975
252
  },
253
  {
254
  "epoch": 8.06,
255
+ "learning_rate": 7.90263157894737e-06,
256
+ "loss": 0.0057,
257
  "step": 1000
258
  },
259
  {
260
  "epoch": 8.06,
261
+ "eval_loss": 0.3434995114803314,
262
+ "eval_runtime": 2295.9725,
263
+ "eval_samples_per_second": 0.56,
264
+ "eval_steps_per_second": 0.56,
265
+ "eval_wer": 15.949628228821936,
266
  "step": 1000
267
  },
268
  {
269
  "epoch": 8.26,
270
+ "learning_rate": 7.836842105263159e-06,
271
+ "loss": 0.0031,
272
  "step": 1025
273
  },
274
  {
275
  "epoch": 8.46,
276
+ "learning_rate": 7.771052631578948e-06,
277
+ "loss": 0.0055,
278
  "step": 1050
279
  },
280
  {
281
  "epoch": 8.66,
282
+ "learning_rate": 7.705263157894738e-06,
283
+ "loss": 0.0056,
284
  "step": 1075
285
  },
286
  {
287
  "epoch": 8.86,
288
+ "learning_rate": 7.639473684210526e-06,
289
+ "loss": 0.0049,
290
  "step": 1100
291
  },
292
  {
293
  "epoch": 9.07,
294
+ "learning_rate": 7.573684210526317e-06,
295
+ "loss": 0.004,
296
  "step": 1125
297
  },
298
  {
299
  "epoch": 9.27,
300
+ "learning_rate": 7.507894736842106e-06,
301
+ "loss": 0.0036,
302
  "step": 1150
303
  },
304
  {
305
  "epoch": 9.47,
306
+ "learning_rate": 7.442105263157895e-06,
307
+ "loss": 0.0028,
308
  "step": 1175
309
  },
310
  {
311
  "epoch": 9.67,
312
+ "learning_rate": 7.376315789473685e-06,
313
+ "loss": 0.0034,
314
  "step": 1200
315
  },
316
  {
317
  "epoch": 9.87,
318
+ "learning_rate": 7.310526315789475e-06,
319
+ "loss": 0.0027,
320
  "step": 1225
321
  },
322
  {
323
  "epoch": 10.08,
324
+ "learning_rate": 7.244736842105263e-06,
325
+ "loss": 0.0027,
326
  "step": 1250
327
  },
328
  {
329
  "epoch": 10.28,
330
+ "learning_rate": 7.178947368421053e-06,
331
+ "loss": 0.0022,
332
  "step": 1275
333
  },
334
  {
335
  "epoch": 10.48,
336
+ "learning_rate": 7.113157894736843e-06,
337
+ "loss": 0.0033,
338
  "step": 1300
339
  },
340
  {
341
  "epoch": 10.68,
342
+ "learning_rate": 7.047368421052631e-06,
343
+ "loss": 0.0033,
344
  "step": 1325
345
  },
346
  {
347
  "epoch": 10.88,
348
+ "learning_rate": 6.981578947368421e-06,
349
+ "loss": 0.005,
350
  "step": 1350
351
  },
352
  {
353
  "epoch": 11.09,
354
+ "learning_rate": 6.915789473684211e-06,
355
+ "loss": 0.0034,
356
  "step": 1375
357
  },
358
  {
359
  "epoch": 11.29,
360
+ "learning_rate": 6.850000000000001e-06,
361
+ "loss": 0.0035,
362
  "step": 1400
363
  },
364
  {
365
  "epoch": 11.49,
366
+ "learning_rate": 6.78421052631579e-06,
367
+ "loss": 0.0033,
368
  "step": 1425
369
  },
370
  {
371
  "epoch": 11.69,
372
+ "learning_rate": 6.71842105263158e-06,
373
+ "loss": 0.0041,
374
  "step": 1450
375
  },
376
  {
377
  "epoch": 11.89,
378
+ "learning_rate": 6.6526315789473695e-06,
379
+ "loss": 0.0047,
380
  "step": 1475
381
  },
382
  {
383
  "epoch": 12.1,
384
+ "learning_rate": 6.586842105263158e-06,
385
+ "loss": 0.0049,
386
  "step": 1500
387
  },
388
  {
389
  "epoch": 12.1,
390
+ "eval_loss": 0.3610152006149292,
391
+ "eval_runtime": 2310.3916,
392
+ "eval_samples_per_second": 0.557,
393
+ "eval_steps_per_second": 0.557,
394
+ "eval_wer": 16.87368375811235,
395
  "step": 1500
396
  },
397
  {
398
  "epoch": 12.3,
399
+ "learning_rate": 6.521052631578948e-06,
400
+ "loss": 0.0035,
401
  "step": 1525
402
  },
403
  {
404
  "epoch": 12.5,
405
+ "learning_rate": 6.455263157894738e-06,
406
+ "loss": 0.0036,
407
  "step": 1550
408
  },
409
  {
410
  "epoch": 12.7,
411
+ "learning_rate": 6.389473684210527e-06,
412
+ "loss": 0.0038,
413
  "step": 1575
414
  },
415
  {
416
  "epoch": 12.9,
417
+ "learning_rate": 6.323684210526316e-06,
418
+ "loss": 0.0033,
419
  "step": 1600
420
  },
421
  {
422
  "epoch": 13.1,
423
+ "learning_rate": 6.257894736842106e-06,
424
+ "loss": 0.0028,
425
  "step": 1625
426
  },
427
  {
428
  "epoch": 13.3,
429
+ "learning_rate": 6.192105263157895e-06,
430
+ "loss": 0.0029,
431
  "step": 1650
432
  },
433
  {
434
  "epoch": 13.5,
435
+ "learning_rate": 6.126315789473685e-06,
436
+ "loss": 0.0033,
437
  "step": 1675
438
  },
439
  {
440
  "epoch": 13.7,
441
+ "learning_rate": 6.060526315789474e-06,
442
+ "loss": 0.003,
443
  "step": 1700
444
  },
445
  {
446
  "epoch": 13.9,
447
+ "learning_rate": 5.994736842105263e-06,
448
+ "loss": 0.0036,
449
  "step": 1725
450
  },
451
  {
452
  "epoch": 14.11,
453
+ "learning_rate": 5.928947368421053e-06,
454
+ "loss": 0.0029,
455
  "step": 1750
456
  },
457
  {
458
  "epoch": 14.31,
459
+ "learning_rate": 5.863157894736842e-06,
460
+ "loss": 0.0027,
461
  "step": 1775
462
  },
463
  {
464
  "epoch": 14.51,
465
+ "learning_rate": 5.7973684210526315e-06,
466
+ "loss": 0.0023,
467
  "step": 1800
468
  },
469
  {
470
  "epoch": 14.71,
471
+ "learning_rate": 5.7315789473684215e-06,
472
+ "loss": 0.0023,
473
  "step": 1825
474
  },
475
  {
476
  "epoch": 14.91,
477
+ "learning_rate": 5.6657894736842115e-06,
478
+ "loss": 0.0033,
479
  "step": 1850
480
  },
481
  {
482
  "epoch": 15.12,
483
+ "learning_rate": 5.600000000000001e-06,
484
+ "loss": 0.0033,
485
  "step": 1875
486
  },
487
  {
488
  "epoch": 15.32,
489
+ "learning_rate": 5.53421052631579e-06,
490
+ "loss": 0.0037,
491
  "step": 1900
492
  },
493
  {
494
  "epoch": 15.52,
495
+ "learning_rate": 5.46842105263158e-06,
496
+ "loss": 0.0035,
497
  "step": 1925
498
  },
499
  {
500
  "epoch": 15.72,
501
+ "learning_rate": 5.40263157894737e-06,
502
+ "loss": 0.0037,
503
  "step": 1950
504
  },
505
  {
506
  "epoch": 15.92,
507
+ "learning_rate": 5.336842105263158e-06,
508
+ "loss": 0.0024,
509
  "step": 1975
510
  },
511
  {
512
  "epoch": 16.13,
513
+ "learning_rate": 5.271052631578948e-06,
514
+ "loss": 0.0015,
515
  "step": 2000
516
  },
517
  {
518
  "epoch": 16.13,
519
+ "eval_loss": 0.36096739768981934,
520
+ "eval_runtime": 2288.7771,
521
+ "eval_samples_per_second": 0.562,
522
+ "eval_steps_per_second": 0.562,
523
+ "eval_wer": 16.125843469291272,
524
  "step": 2000
525
  },
526
  {
527
+ "epoch": 16.33,
528
+ "learning_rate": 5.205263157894738e-06,
529
+ "loss": 0.0017,
530
+ "step": 2025
531
+ },
532
+ {
533
+ "epoch": 16.53,
534
+ "learning_rate": 5.139473684210526e-06,
535
+ "loss": 0.0017,
536
+ "step": 2050
537
+ },
538
+ {
539
+ "epoch": 16.73,
540
+ "learning_rate": 5.073684210526316e-06,
541
+ "loss": 0.0017,
542
+ "step": 2075
543
+ },
544
+ {
545
+ "epoch": 16.93,
546
+ "learning_rate": 5.007894736842106e-06,
547
+ "loss": 0.003,
548
+ "step": 2100
549
+ },
550
+ {
551
+ "epoch": 17.14,
552
+ "learning_rate": 4.942105263157895e-06,
553
+ "loss": 0.0017,
554
+ "step": 2125
555
+ },
556
+ {
557
+ "epoch": 17.34,
558
+ "learning_rate": 4.876315789473684e-06,
559
+ "loss": 0.0014,
560
+ "step": 2150
561
+ },
562
+ {
563
+ "epoch": 17.54,
564
+ "learning_rate": 4.8105263157894735e-06,
565
+ "loss": 0.0022,
566
+ "step": 2175
567
+ },
568
+ {
569
+ "epoch": 17.74,
570
+ "learning_rate": 4.7447368421052634e-06,
571
+ "loss": 0.0023,
572
+ "step": 2200
573
+ },
574
+ {
575
+ "epoch": 17.94,
576
+ "learning_rate": 4.6789473684210525e-06,
577
+ "loss": 0.0026,
578
+ "step": 2225
579
+ },
580
+ {
581
+ "epoch": 18.14,
582
+ "learning_rate": 4.6131578947368425e-06,
583
+ "loss": 0.0013,
584
+ "step": 2250
585
+ },
586
+ {
587
+ "epoch": 18.34,
588
+ "learning_rate": 4.547368421052632e-06,
589
+ "loss": 0.0012,
590
+ "step": 2275
591
+ },
592
+ {
593
+ "epoch": 18.54,
594
+ "learning_rate": 4.481578947368422e-06,
595
+ "loss": 0.0014,
596
+ "step": 2300
597
+ },
598
+ {
599
+ "epoch": 18.74,
600
+ "learning_rate": 4.415789473684211e-06,
601
+ "loss": 0.0017,
602
+ "step": 2325
603
+ },
604
+ {
605
+ "epoch": 18.94,
606
+ "learning_rate": 4.350000000000001e-06,
607
+ "loss": 0.002,
608
+ "step": 2350
609
+ },
610
+ {
611
+ "epoch": 19.15,
612
+ "learning_rate": 4.28421052631579e-06,
613
+ "loss": 0.0015,
614
+ "step": 2375
615
+ },
616
+ {
617
+ "epoch": 19.35,
618
+ "learning_rate": 4.21842105263158e-06,
619
+ "loss": 0.0016,
620
+ "step": 2400
621
+ },
622
+ {
623
+ "epoch": 19.55,
624
+ "learning_rate": 4.152631578947369e-06,
625
+ "loss": 0.0013,
626
+ "step": 2425
627
+ },
628
+ {
629
+ "epoch": 19.75,
630
+ "learning_rate": 4.086842105263159e-06,
631
+ "loss": 0.0014,
632
+ "step": 2450
633
+ },
634
+ {
635
+ "epoch": 19.95,
636
+ "learning_rate": 4.021052631578948e-06,
637
+ "loss": 0.0011,
638
+ "step": 2475
639
+ },
640
+ {
641
+ "epoch": 20.16,
642
+ "learning_rate": 3.955263157894737e-06,
643
+ "loss": 0.001,
644
+ "step": 2500
645
+ },
646
+ {
647
+ "epoch": 20.16,
648
+ "eval_loss": 0.38679206371307373,
649
+ "eval_runtime": 2297.8177,
650
+ "eval_samples_per_second": 0.56,
651
+ "eval_steps_per_second": 0.56,
652
+ "eval_wer": 16.035586882709417,
653
+ "step": 2500
654
+ },
655
+ {
656
+ "epoch": 20.36,
657
+ "learning_rate": 3.889473684210527e-06,
658
+ "loss": 0.0013,
659
+ "step": 2525
660
+ },
661
+ {
662
+ "epoch": 20.56,
663
+ "learning_rate": 3.823684210526316e-06,
664
+ "loss": 0.0012,
665
+ "step": 2550
666
+ },
667
+ {
668
+ "epoch": 20.76,
669
+ "learning_rate": 3.7578947368421053e-06,
670
+ "loss": 0.001,
671
+ "step": 2575
672
+ },
673
+ {
674
+ "epoch": 20.96,
675
+ "learning_rate": 3.6921052631578953e-06,
676
+ "loss": 0.0012,
677
+ "step": 2600
678
+ },
679
+ {
680
+ "epoch": 21.17,
681
+ "learning_rate": 3.6263157894736844e-06,
682
+ "loss": 0.0013,
683
+ "step": 2625
684
+ },
685
+ {
686
+ "epoch": 21.37,
687
+ "learning_rate": 3.5605263157894736e-06,
688
+ "loss": 0.0008,
689
+ "step": 2650
690
+ },
691
+ {
692
+ "epoch": 21.57,
693
+ "learning_rate": 3.4947368421052635e-06,
694
+ "loss": 0.0013,
695
+ "step": 2675
696
+ },
697
+ {
698
+ "epoch": 21.77,
699
+ "learning_rate": 3.4289473684210527e-06,
700
+ "loss": 0.001,
701
+ "step": 2700
702
+ },
703
+ {
704
+ "epoch": 21.97,
705
+ "learning_rate": 3.3631578947368426e-06,
706
+ "loss": 0.0014,
707
+ "step": 2725
708
+ },
709
+ {
710
+ "epoch": 22.18,
711
+ "learning_rate": 3.2973684210526318e-06,
712
+ "loss": 0.0012,
713
+ "step": 2750
714
+ },
715
+ {
716
+ "epoch": 22.38,
717
+ "learning_rate": 3.2315789473684213e-06,
718
+ "loss": 0.0008,
719
+ "step": 2775
720
+ },
721
+ {
722
+ "epoch": 22.58,
723
+ "learning_rate": 3.165789473684211e-06,
724
+ "loss": 0.001,
725
+ "step": 2800
726
+ },
727
+ {
728
+ "epoch": 22.78,
729
+ "learning_rate": 3.1000000000000004e-06,
730
+ "loss": 0.0012,
731
+ "step": 2825
732
+ },
733
+ {
734
+ "epoch": 22.98,
735
+ "learning_rate": 3.0342105263157895e-06,
736
+ "loss": 0.001,
737
+ "step": 2850
738
+ },
739
+ {
740
+ "epoch": 23.18,
741
+ "learning_rate": 2.9684210526315795e-06,
742
+ "loss": 0.0008,
743
+ "step": 2875
744
+ },
745
+ {
746
+ "epoch": 23.38,
747
+ "learning_rate": 2.9026315789473686e-06,
748
+ "loss": 0.0007,
749
+ "step": 2900
750
+ },
751
+ {
752
+ "epoch": 23.58,
753
+ "learning_rate": 2.8368421052631586e-06,
754
+ "loss": 0.0011,
755
+ "step": 2925
756
+ },
757
+ {
758
+ "epoch": 23.78,
759
+ "learning_rate": 2.7710526315789477e-06,
760
+ "loss": 0.0015,
761
+ "step": 2950
762
+ },
763
+ {
764
+ "epoch": 23.98,
765
+ "learning_rate": 2.705263157894737e-06,
766
+ "loss": 0.0011,
767
+ "step": 2975
768
+ },
769
+ {
770
+ "epoch": 24.19,
771
+ "learning_rate": 2.639473684210527e-06,
772
+ "loss": 0.0011,
773
+ "step": 3000
774
+ },
775
+ {
776
+ "epoch": 24.19,
777
+ "eval_loss": 0.394901841878891,
778
+ "eval_runtime": 2297.2378,
779
+ "eval_samples_per_second": 0.56,
780
+ "eval_steps_per_second": 0.56,
781
+ "eval_wer": 15.829286113379466,
782
+ "step": 3000
783
+ },
784
+ {
785
+ "epoch": 24.39,
786
+ "learning_rate": 2.573684210526316e-06,
787
+ "loss": 0.0007,
788
+ "step": 3025
789
+ },
790
+ {
791
+ "epoch": 24.59,
792
+ "learning_rate": 2.5078947368421055e-06,
793
+ "loss": 0.0012,
794
+ "step": 3050
795
+ },
796
+ {
797
+ "epoch": 24.79,
798
+ "learning_rate": 2.442105263157895e-06,
799
+ "loss": 0.0011,
800
+ "step": 3075
801
+ },
802
+ {
803
+ "epoch": 24.99,
804
+ "learning_rate": 2.3763157894736846e-06,
805
+ "loss": 0.0009,
806
+ "step": 3100
807
+ },
808
+ {
809
+ "epoch": 25.2,
810
+ "learning_rate": 2.310526315789474e-06,
811
+ "loss": 0.0009,
812
+ "step": 3125
813
+ },
814
+ {
815
+ "epoch": 25.4,
816
+ "learning_rate": 2.2447368421052636e-06,
817
+ "loss": 0.0008,
818
+ "step": 3150
819
+ },
820
+ {
821
+ "epoch": 25.6,
822
+ "learning_rate": 2.1789473684210528e-06,
823
+ "loss": 0.0009,
824
+ "step": 3175
825
+ },
826
+ {
827
+ "epoch": 25.8,
828
+ "learning_rate": 2.1131578947368423e-06,
829
+ "loss": 0.0009,
830
+ "step": 3200
831
+ },
832
+ {
833
+ "epoch": 26.01,
834
+ "learning_rate": 2.047368421052632e-06,
835
+ "loss": 0.0011,
836
+ "step": 3225
837
+ },
838
+ {
839
+ "epoch": 26.21,
840
+ "learning_rate": 1.9815789473684214e-06,
841
+ "loss": 0.0007,
842
+ "step": 3250
843
+ },
844
+ {
845
+ "epoch": 26.41,
846
+ "learning_rate": 1.9157894736842105e-06,
847
+ "loss": 0.001,
848
+ "step": 3275
849
+ },
850
+ {
851
+ "epoch": 26.61,
852
+ "learning_rate": 1.85e-06,
853
+ "loss": 0.0008,
854
+ "step": 3300
855
+ },
856
+ {
857
+ "epoch": 26.81,
858
+ "learning_rate": 1.7842105263157896e-06,
859
+ "loss": 0.0008,
860
+ "step": 3325
861
+ },
862
+ {
863
+ "epoch": 27.02,
864
+ "learning_rate": 1.7184210526315792e-06,
865
+ "loss": 0.0012,
866
+ "step": 3350
867
+ },
868
+ {
869
+ "epoch": 27.22,
870
+ "learning_rate": 1.6526315789473685e-06,
871
+ "loss": 0.0009,
872
+ "step": 3375
873
+ },
874
+ {
875
+ "epoch": 27.42,
876
+ "learning_rate": 1.586842105263158e-06,
877
+ "loss": 0.0009,
878
+ "step": 3400
879
+ },
880
+ {
881
+ "epoch": 27.62,
882
+ "learning_rate": 1.5210526315789476e-06,
883
+ "loss": 0.0009,
884
+ "step": 3425
885
+ },
886
+ {
887
+ "epoch": 27.82,
888
+ "learning_rate": 1.4552631578947371e-06,
889
+ "loss": 0.0009,
890
+ "step": 3450
891
+ },
892
+ {
893
+ "epoch": 28.02,
894
+ "learning_rate": 1.3894736842105263e-06,
895
+ "loss": 0.001,
896
+ "step": 3475
897
+ },
898
+ {
899
+ "epoch": 28.22,
900
+ "learning_rate": 1.3236842105263158e-06,
901
+ "loss": 0.0006,
902
+ "step": 3500
903
+ },
904
+ {
905
+ "epoch": 28.22,
906
+ "eval_loss": 0.4028264284133911,
907
+ "eval_runtime": 2295.0076,
908
+ "eval_samples_per_second": 0.56,
909
+ "eval_steps_per_second": 0.56,
910
+ "eval_wer": 15.74762539218636,
911
+ "step": 3500
912
+ },
913
+ {
914
+ "epoch": 28.42,
915
+ "learning_rate": 1.2578947368421054e-06,
916
+ "loss": 0.0005,
917
+ "step": 3525
918
+ },
919
+ {
920
+ "epoch": 28.62,
921
+ "learning_rate": 1.192105263157895e-06,
922
+ "loss": 0.0006,
923
+ "step": 3550
924
+ },
925
+ {
926
+ "epoch": 28.82,
927
+ "learning_rate": 1.1263157894736842e-06,
928
+ "loss": 0.0011,
929
+ "step": 3575
930
+ },
931
+ {
932
+ "epoch": 29.03,
933
+ "learning_rate": 1.0605263157894738e-06,
934
+ "loss": 0.0012,
935
+ "step": 3600
936
+ },
937
+ {
938
+ "epoch": 29.23,
939
+ "learning_rate": 9.947368421052631e-07,
940
+ "loss": 0.0006,
941
+ "step": 3625
942
+ },
943
+ {
944
+ "epoch": 29.43,
945
+ "learning_rate": 9.289473684210528e-07,
946
+ "loss": 0.0008,
947
+ "step": 3650
948
+ },
949
+ {
950
+ "epoch": 29.63,
951
+ "learning_rate": 8.631578947368421e-07,
952
+ "loss": 0.0009,
953
+ "step": 3675
954
+ },
955
+ {
956
+ "epoch": 29.83,
957
+ "learning_rate": 7.973684210526317e-07,
958
+ "loss": 0.0008,
959
+ "step": 3700
960
+ },
961
+ {
962
+ "epoch": 30.04,
963
+ "learning_rate": 7.315789473684211e-07,
964
+ "loss": 0.0008,
965
+ "step": 3725
966
+ },
967
+ {
968
+ "epoch": 30.24,
969
+ "learning_rate": 6.657894736842106e-07,
970
+ "loss": 0.0007,
971
+ "step": 3750
972
+ },
973
+ {
974
+ "epoch": 30.44,
975
+ "learning_rate": 6.000000000000001e-07,
976
+ "loss": 0.0008,
977
+ "step": 3775
978
+ },
979
+ {
980
+ "epoch": 30.64,
981
+ "learning_rate": 5.342105263157895e-07,
982
+ "loss": 0.0009,
983
+ "step": 3800
984
+ },
985
+ {
986
+ "epoch": 30.84,
987
+ "learning_rate": 4.6842105263157896e-07,
988
+ "loss": 0.0008,
989
+ "step": 3825
990
+ },
991
+ {
992
+ "epoch": 31.05,
993
+ "learning_rate": 4.0263157894736845e-07,
994
+ "loss": 0.0007,
995
+ "step": 3850
996
+ },
997
+ {
998
+ "epoch": 31.25,
999
+ "learning_rate": 3.368421052631579e-07,
1000
+ "loss": 0.0006,
1001
+ "step": 3875
1002
+ },
1003
+ {
1004
+ "epoch": 31.45,
1005
+ "learning_rate": 2.710526315789474e-07,
1006
+ "loss": 0.0008,
1007
+ "step": 3900
1008
+ },
1009
+ {
1010
+ "epoch": 31.65,
1011
+ "learning_rate": 2.0526315789473685e-07,
1012
+ "loss": 0.0007,
1013
+ "step": 3925
1014
+ },
1015
+ {
1016
+ "epoch": 31.85,
1017
+ "learning_rate": 1.3947368421052632e-07,
1018
+ "loss": 0.001,
1019
+ "step": 3950
1020
+ },
1021
+ {
1022
+ "epoch": 32.06,
1023
+ "learning_rate": 7.368421052631579e-08,
1024
+ "loss": 0.0006,
1025
+ "step": 3975
1026
+ },
1027
+ {
1028
+ "epoch": 32.26,
1029
+ "learning_rate": 7.894736842105264e-09,
1030
+ "loss": 0.0008,
1031
+ "step": 4000
1032
+ },
1033
+ {
1034
+ "epoch": 32.26,
1035
+ "eval_loss": 0.4131947159767151,
1036
+ "eval_runtime": 2384.3215,
1037
+ "eval_samples_per_second": 0.539,
1038
+ "eval_steps_per_second": 0.539,
1039
+ "eval_wer": 15.734731594103234,
1040
+ "step": 4000
1041
+ },
1042
+ {
1043
+ "epoch": 32.26,
1044
+ "step": 4000,
1045
+ "total_flos": 2.73804134842368e+20,
1046
+ "train_loss": 0.017603627420263366,
1047
+ "train_runtime": 74040.5343,
1048
+ "train_samples_per_second": 1.729,
1049
  "train_steps_per_second": 0.054
1050
  }
1051
  ],
1052
+ "max_steps": 4000,
1053
+ "num_train_epochs": 33,
1054
+ "total_flos": 2.73804134842368e+20,
1055
  "trial_name": null,
1056
  "trial_params": null
1057
  }