roseman commited on
Commit
d6303c3
1 Parent(s): d865701

End of training

Browse files
Files changed (4) hide show
  1. all_results.json +12 -0
  2. eval_results.json +8 -0
  3. train_results.json +7 -0
  4. trainer_state.json +508 -0
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 6.0,
3
+ "eval_loss": 0.2976975440979004,
4
+ "eval_runtime": 5525.1171,
5
+ "eval_samples_per_second": 0.88,
6
+ "eval_steps_per_second": 0.055,
7
+ "eval_wer": 31.350262297574698,
8
+ "train_loss": 0.2124484316768522,
9
+ "train_runtime": 44455.4113,
10
+ "train_samples_per_second": 1.319,
11
+ "train_steps_per_second": 0.041
12
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 6.0,
3
+ "eval_loss": 0.2976975440979004,
4
+ "eval_runtime": 5525.1171,
5
+ "eval_samples_per_second": 0.88,
6
+ "eval_steps_per_second": 0.055,
7
+ "eval_wer": 31.350262297574698
8
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 6.0,
3
+ "train_loss": 0.2124484316768522,
4
+ "train_runtime": 44455.4113,
5
+ "train_samples_per_second": 1.319,
6
+ "train_steps_per_second": 0.041
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,508 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 31.350262297574698,
3
+ "best_model_checkpoint": "./whisper-medium-ckb/checkpoint-1830",
4
+ "epoch": 6.004909983633388,
5
+ "global_step": 1833,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 4.2000000000000006e-07,
13
+ "loss": 3.1622,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.03,
18
+ "learning_rate": 9.200000000000001e-07,
19
+ "loss": 2.1678,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.04,
24
+ "learning_rate": 1.42e-06,
25
+ "loss": 1.2419,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.05,
30
+ "learning_rate": 1.9200000000000003e-06,
31
+ "loss": 0.7203,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.07,
36
+ "learning_rate": 2.42e-06,
37
+ "loss": 0.5754,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.08,
42
+ "learning_rate": 2.92e-06,
43
+ "loss": 0.4364,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.1,
48
+ "learning_rate": 3.4200000000000007e-06,
49
+ "loss": 0.4198,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.11,
54
+ "learning_rate": 3.920000000000001e-06,
55
+ "loss": 0.404,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.12,
60
+ "learning_rate": 4.42e-06,
61
+ "loss": 0.3441,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.14,
66
+ "learning_rate": 4.92e-06,
67
+ "loss": 0.3334,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.15,
72
+ "learning_rate": 5.420000000000001e-06,
73
+ "loss": 0.2885,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 0.16,
78
+ "learning_rate": 5.92e-06,
79
+ "loss": 0.274,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 1.01,
84
+ "learning_rate": 6.42e-06,
85
+ "loss": 0.249,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 1.03,
90
+ "learning_rate": 6.92e-06,
91
+ "loss": 0.27,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 1.03,
96
+ "eval_loss": 0.27222105860710144,
97
+ "eval_runtime": 5782.8476,
98
+ "eval_samples_per_second": 0.841,
99
+ "eval_steps_per_second": 0.053,
100
+ "eval_wer": 43.91773739831217,
101
+ "step": 366
102
+ },
103
+ {
104
+ "epoch": 1.04,
105
+ "learning_rate": 7.420000000000001e-06,
106
+ "loss": 0.2623,
107
+ "step": 375
108
+ },
109
+ {
110
+ "epoch": 1.05,
111
+ "learning_rate": 7.92e-06,
112
+ "loss": 0.2521,
113
+ "step": 400
114
+ },
115
+ {
116
+ "epoch": 1.07,
117
+ "learning_rate": 8.42e-06,
118
+ "loss": 0.258,
119
+ "step": 425
120
+ },
121
+ {
122
+ "epoch": 1.08,
123
+ "learning_rate": 8.920000000000001e-06,
124
+ "loss": 0.2201,
125
+ "step": 450
126
+ },
127
+ {
128
+ "epoch": 1.09,
129
+ "learning_rate": 9.42e-06,
130
+ "loss": 0.2285,
131
+ "step": 475
132
+ },
133
+ {
134
+ "epoch": 1.11,
135
+ "learning_rate": 9.920000000000002e-06,
136
+ "loss": 0.2239,
137
+ "step": 500
138
+ },
139
+ {
140
+ "epoch": 1.12,
141
+ "learning_rate": 9.84246061515379e-06,
142
+ "loss": 0.1928,
143
+ "step": 525
144
+ },
145
+ {
146
+ "epoch": 1.13,
147
+ "learning_rate": 9.654913728432109e-06,
148
+ "loss": 0.197,
149
+ "step": 550
150
+ },
151
+ {
152
+ "epoch": 1.15,
153
+ "learning_rate": 9.467366841710428e-06,
154
+ "loss": 0.1732,
155
+ "step": 575
156
+ },
157
+ {
158
+ "epoch": 1.16,
159
+ "learning_rate": 9.279819954988747e-06,
160
+ "loss": 0.1454,
161
+ "step": 600
162
+ },
163
+ {
164
+ "epoch": 2.01,
165
+ "learning_rate": 9.092273068267068e-06,
166
+ "loss": 0.142,
167
+ "step": 625
168
+ },
169
+ {
170
+ "epoch": 2.02,
171
+ "learning_rate": 8.904726181545387e-06,
172
+ "loss": 0.1486,
173
+ "step": 650
174
+ },
175
+ {
176
+ "epoch": 2.04,
177
+ "learning_rate": 8.717179294823706e-06,
178
+ "loss": 0.1446,
179
+ "step": 675
180
+ },
181
+ {
182
+ "epoch": 2.05,
183
+ "learning_rate": 8.529632408102027e-06,
184
+ "loss": 0.1329,
185
+ "step": 700
186
+ },
187
+ {
188
+ "epoch": 2.06,
189
+ "learning_rate": 8.342085521380345e-06,
190
+ "loss": 0.1498,
191
+ "step": 725
192
+ },
193
+ {
194
+ "epoch": 2.07,
195
+ "eval_loss": 0.23542030155658722,
196
+ "eval_runtime": 5576.1832,
197
+ "eval_samples_per_second": 0.872,
198
+ "eval_steps_per_second": 0.055,
199
+ "eval_wer": 36.284497833193946,
200
+ "step": 732
201
+ },
202
+ {
203
+ "epoch": 2.08,
204
+ "learning_rate": 8.154538634658665e-06,
205
+ "loss": 0.1265,
206
+ "step": 750
207
+ },
208
+ {
209
+ "epoch": 2.09,
210
+ "learning_rate": 7.966991747936984e-06,
211
+ "loss": 0.1154,
212
+ "step": 775
213
+ },
214
+ {
215
+ "epoch": 2.1,
216
+ "learning_rate": 7.779444861215303e-06,
217
+ "loss": 0.1095,
218
+ "step": 800
219
+ },
220
+ {
221
+ "epoch": 2.12,
222
+ "learning_rate": 7.591897974493624e-06,
223
+ "loss": 0.1,
224
+ "step": 825
225
+ },
226
+ {
227
+ "epoch": 2.13,
228
+ "learning_rate": 7.404351087771943e-06,
229
+ "loss": 0.1,
230
+ "step": 850
231
+ },
232
+ {
233
+ "epoch": 2.15,
234
+ "learning_rate": 7.216804201050263e-06,
235
+ "loss": 0.0924,
236
+ "step": 875
237
+ },
238
+ {
239
+ "epoch": 2.16,
240
+ "learning_rate": 7.0292573143285824e-06,
241
+ "loss": 0.0773,
242
+ "step": 900
243
+ },
244
+ {
245
+ "epoch": 3.01,
246
+ "learning_rate": 6.841710427606902e-06,
247
+ "loss": 0.0687,
248
+ "step": 925
249
+ },
250
+ {
251
+ "epoch": 3.02,
252
+ "learning_rate": 6.654163540885222e-06,
253
+ "loss": 0.0778,
254
+ "step": 950
255
+ },
256
+ {
257
+ "epoch": 3.03,
258
+ "learning_rate": 6.4666166541635414e-06,
259
+ "loss": 0.0677,
260
+ "step": 975
261
+ },
262
+ {
263
+ "epoch": 3.05,
264
+ "learning_rate": 6.279069767441861e-06,
265
+ "loss": 0.0723,
266
+ "step": 1000
267
+ },
268
+ {
269
+ "epoch": 3.06,
270
+ "learning_rate": 6.091522880720181e-06,
271
+ "loss": 0.0732,
272
+ "step": 1025
273
+ },
274
+ {
275
+ "epoch": 3.08,
276
+ "learning_rate": 5.9039759939984996e-06,
277
+ "loss": 0.062,
278
+ "step": 1050
279
+ },
280
+ {
281
+ "epoch": 3.09,
282
+ "learning_rate": 5.7164291072768195e-06,
283
+ "loss": 0.0538,
284
+ "step": 1075
285
+ },
286
+ {
287
+ "epoch": 3.1,
288
+ "eval_loss": 0.24223530292510986,
289
+ "eval_runtime": 5591.471,
290
+ "eval_samples_per_second": 0.87,
291
+ "eval_steps_per_second": 0.054,
292
+ "eval_wer": 33.65391925796396,
293
+ "step": 1098
294
+ },
295
+ {
296
+ "epoch": 3.1,
297
+ "learning_rate": 5.5288822205551395e-06,
298
+ "loss": 0.0549,
299
+ "step": 1100
300
+ },
301
+ {
302
+ "epoch": 3.12,
303
+ "learning_rate": 5.3413353338334586e-06,
304
+ "loss": 0.0473,
305
+ "step": 1125
306
+ },
307
+ {
308
+ "epoch": 3.13,
309
+ "learning_rate": 5.1537884471117785e-06,
310
+ "loss": 0.0491,
311
+ "step": 1150
312
+ },
313
+ {
314
+ "epoch": 3.14,
315
+ "learning_rate": 4.9662415603900985e-06,
316
+ "loss": 0.0421,
317
+ "step": 1175
318
+ },
319
+ {
320
+ "epoch": 3.16,
321
+ "learning_rate": 4.7786946736684176e-06,
322
+ "loss": 0.0343,
323
+ "step": 1200
324
+ },
325
+ {
326
+ "epoch": 4.0,
327
+ "learning_rate": 4.591147786946737e-06,
328
+ "loss": 0.0336,
329
+ "step": 1225
330
+ },
331
+ {
332
+ "epoch": 4.02,
333
+ "learning_rate": 4.403600900225057e-06,
334
+ "loss": 0.0316,
335
+ "step": 1250
336
+ },
337
+ {
338
+ "epoch": 4.03,
339
+ "learning_rate": 4.2160540135033765e-06,
340
+ "loss": 0.0324,
341
+ "step": 1275
342
+ },
343
+ {
344
+ "epoch": 4.05,
345
+ "learning_rate": 4.028507126781696e-06,
346
+ "loss": 0.031,
347
+ "step": 1300
348
+ },
349
+ {
350
+ "epoch": 4.06,
351
+ "learning_rate": 3.840960240060016e-06,
352
+ "loss": 0.031,
353
+ "step": 1325
354
+ },
355
+ {
356
+ "epoch": 4.07,
357
+ "learning_rate": 3.6534133533383347e-06,
358
+ "loss": 0.028,
359
+ "step": 1350
360
+ },
361
+ {
362
+ "epoch": 4.09,
363
+ "learning_rate": 3.4658664666166546e-06,
364
+ "loss": 0.0218,
365
+ "step": 1375
366
+ },
367
+ {
368
+ "epoch": 4.1,
369
+ "learning_rate": 3.278319579894974e-06,
370
+ "loss": 0.0221,
371
+ "step": 1400
372
+ },
373
+ {
374
+ "epoch": 4.11,
375
+ "learning_rate": 3.0907726931732933e-06,
376
+ "loss": 0.0183,
377
+ "step": 1425
378
+ },
379
+ {
380
+ "epoch": 4.13,
381
+ "learning_rate": 2.903225806451613e-06,
382
+ "loss": 0.0168,
383
+ "step": 1450
384
+ },
385
+ {
386
+ "epoch": 4.14,
387
+ "eval_loss": 0.27173367142677307,
388
+ "eval_runtime": 5547.7291,
389
+ "eval_samples_per_second": 0.876,
390
+ "eval_steps_per_second": 0.055,
391
+ "eval_wer": 32.384246939861626,
392
+ "step": 1464
393
+ },
394
+ {
395
+ "epoch": 4.14,
396
+ "learning_rate": 2.7156789197299327e-06,
397
+ "loss": 0.0175,
398
+ "step": 1475
399
+ },
400
+ {
401
+ "epoch": 4.15,
402
+ "learning_rate": 2.5281320330082523e-06,
403
+ "loss": 0.0121,
404
+ "step": 1500
405
+ },
406
+ {
407
+ "epoch": 5.0,
408
+ "learning_rate": 2.3405851462865718e-06,
409
+ "loss": 0.0103,
410
+ "step": 1525
411
+ },
412
+ {
413
+ "epoch": 5.02,
414
+ "learning_rate": 2.1530382595648913e-06,
415
+ "loss": 0.0098,
416
+ "step": 1550
417
+ },
418
+ {
419
+ "epoch": 5.03,
420
+ "learning_rate": 1.965491372843211e-06,
421
+ "loss": 0.0124,
422
+ "step": 1575
423
+ },
424
+ {
425
+ "epoch": 5.04,
426
+ "learning_rate": 1.7779444861215306e-06,
427
+ "loss": 0.0102,
428
+ "step": 1600
429
+ },
430
+ {
431
+ "epoch": 5.06,
432
+ "learning_rate": 1.5903975993998503e-06,
433
+ "loss": 0.0101,
434
+ "step": 1625
435
+ },
436
+ {
437
+ "epoch": 5.07,
438
+ "learning_rate": 1.4028507126781696e-06,
439
+ "loss": 0.0081,
440
+ "step": 1650
441
+ },
442
+ {
443
+ "epoch": 5.08,
444
+ "learning_rate": 1.2153038259564891e-06,
445
+ "loss": 0.0076,
446
+ "step": 1675
447
+ },
448
+ {
449
+ "epoch": 5.1,
450
+ "learning_rate": 1.0277569392348089e-06,
451
+ "loss": 0.0076,
452
+ "step": 1700
453
+ },
454
+ {
455
+ "epoch": 5.11,
456
+ "learning_rate": 8.402100525131284e-07,
457
+ "loss": 0.0055,
458
+ "step": 1725
459
+ },
460
+ {
461
+ "epoch": 5.13,
462
+ "learning_rate": 6.52663165791448e-07,
463
+ "loss": 0.0046,
464
+ "step": 1750
465
+ },
466
+ {
467
+ "epoch": 5.14,
468
+ "learning_rate": 4.651162790697675e-07,
469
+ "loss": 0.005,
470
+ "step": 1775
471
+ },
472
+ {
473
+ "epoch": 5.15,
474
+ "learning_rate": 2.7756939234808706e-07,
475
+ "loss": 0.0031,
476
+ "step": 1800
477
+ },
478
+ {
479
+ "epoch": 6.0,
480
+ "learning_rate": 9.002250562640661e-08,
481
+ "loss": 0.0025,
482
+ "step": 1825
483
+ },
484
+ {
485
+ "epoch": 6.0,
486
+ "eval_loss": 0.2976975440979004,
487
+ "eval_runtime": 5536.8386,
488
+ "eval_samples_per_second": 0.878,
489
+ "eval_steps_per_second": 0.055,
490
+ "eval_wer": 31.350262297574698,
491
+ "step": 1830
492
+ },
493
+ {
494
+ "epoch": 6.0,
495
+ "step": 1833,
496
+ "total_flos": 5.977887064326144e+19,
497
+ "train_loss": 0.2124484316768522,
498
+ "train_runtime": 44455.4113,
499
+ "train_samples_per_second": 1.319,
500
+ "train_steps_per_second": 0.041
501
+ }
502
+ ],
503
+ "max_steps": 1833,
504
+ "num_train_epochs": 9223372036854775807,
505
+ "total_flos": 5.977887064326144e+19,
506
+ "trial_name": null,
507
+ "trial_params": null
508
+ }