shivam commited on
Commit
dace185
1 Parent(s): a4d3c6b

End of training

Browse files
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 50.0,
3
+ "eval_loss": 1.228236436843872,
4
+ "eval_runtime": 62.4577,
5
+ "eval_samples": 2095,
6
+ "eval_samples_per_second": 33.543,
7
+ "eval_steps_per_second": 4.195,
8
+ "eval_wer": 0.6837796244998461,
9
+ "train_loss": 2.1161629922049388,
10
+ "train_runtime": 13636.3379,
11
+ "train_samples": 4711,
12
+ "train_samples_per_second": 17.274,
13
+ "train_steps_per_second": 0.539
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 50.0,
3
+ "eval_loss": 1.228236436843872,
4
+ "eval_runtime": 62.4577,
5
+ "eval_samples": 2095,
6
+ "eval_samples_per_second": 33.543,
7
+ "eval_steps_per_second": 4.195,
8
+ "eval_wer": 0.6837796244998461
9
+ }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f76b0084d44e3115f1635d30a8c7a275a4b6a654c9796bdd48fec6a4f0746a67
3
  size 1262321393
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aadc2a96ffa1820127aea59ffc22a9de458b2efb93a4f8f92c6b6c935649c389
3
  size 1262321393
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 50.0,
3
+ "train_loss": 2.1161629922049388,
4
+ "train_runtime": 13636.3379,
5
+ "train_samples": 4711,
6
+ "train_samples_per_second": 17.274,
7
+ "train_steps_per_second": 0.539
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,589 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 49.99830220713073,
5
+ "global_step": 7350,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.68,
12
+ "learning_rate": 3.7125e-06,
13
+ "loss": 20.4357,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 1.36,
18
+ "learning_rate": 7.425e-06,
19
+ "loss": 11.928,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 2.04,
24
+ "learning_rate": 1.1174999999999999e-05,
25
+ "loss": 8.0523,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 2.72,
30
+ "learning_rate": 1.4925e-05,
31
+ "loss": 6.5593,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 3.4,
36
+ "learning_rate": 1.8675e-05,
37
+ "loss": 5.3155,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 3.4,
42
+ "eval_loss": 4.558210849761963,
43
+ "eval_runtime": 61.1578,
44
+ "eval_samples_per_second": 34.256,
45
+ "eval_steps_per_second": 4.284,
46
+ "eval_wer": 1.0,
47
+ "step": 500
48
+ },
49
+ {
50
+ "epoch": 4.08,
51
+ "learning_rate": 2.2424999999999996e-05,
52
+ "loss": 4.3699,
53
+ "step": 600
54
+ },
55
+ {
56
+ "epoch": 4.76,
57
+ "learning_rate": 2.6174999999999996e-05,
58
+ "loss": 3.7571,
59
+ "step": 700
60
+ },
61
+ {
62
+ "epoch": 5.44,
63
+ "learning_rate": 2.9925e-05,
64
+ "loss": 3.5182,
65
+ "step": 800
66
+ },
67
+ {
68
+ "epoch": 6.12,
69
+ "learning_rate": 3.3675e-05,
70
+ "loss": 3.4537,
71
+ "step": 900
72
+ },
73
+ {
74
+ "epoch": 6.8,
75
+ "learning_rate": 3.7424999999999995e-05,
76
+ "loss": 3.3369,
77
+ "step": 1000
78
+ },
79
+ {
80
+ "epoch": 6.8,
81
+ "eval_loss": 3.426917791366577,
82
+ "eval_runtime": 62.6973,
83
+ "eval_samples_per_second": 33.415,
84
+ "eval_steps_per_second": 4.179,
85
+ "eval_wer": 1.0,
86
+ "step": 1000
87
+ },
88
+ {
89
+ "epoch": 7.48,
90
+ "learning_rate": 4.1175e-05,
91
+ "loss": 3.2812,
92
+ "step": 1100
93
+ },
94
+ {
95
+ "epoch": 8.16,
96
+ "learning_rate": 4.4924999999999994e-05,
97
+ "loss": 3.2389,
98
+ "step": 1200
99
+ },
100
+ {
101
+ "epoch": 8.84,
102
+ "learning_rate": 4.8675e-05,
103
+ "loss": 3.1102,
104
+ "step": 1300
105
+ },
106
+ {
107
+ "epoch": 9.52,
108
+ "learning_rate": 5.2424999999999994e-05,
109
+ "loss": 2.7161,
110
+ "step": 1400
111
+ },
112
+ {
113
+ "epoch": 10.2,
114
+ "learning_rate": 5.6175e-05,
115
+ "loss": 2.1785,
116
+ "step": 1500
117
+ },
118
+ {
119
+ "epoch": 10.2,
120
+ "eval_loss": 1.7190676927566528,
121
+ "eval_runtime": 62.8007,
122
+ "eval_samples_per_second": 33.36,
123
+ "eval_steps_per_second": 4.172,
124
+ "eval_wer": 0.8831024930747923,
125
+ "step": 1500
126
+ },
127
+ {
128
+ "epoch": 10.88,
129
+ "learning_rate": 5.9925e-05,
130
+ "loss": 1.9322,
131
+ "step": 1600
132
+ },
133
+ {
134
+ "epoch": 11.56,
135
+ "learning_rate": 6.367499999999999e-05,
136
+ "loss": 1.7738,
137
+ "step": 1700
138
+ },
139
+ {
140
+ "epoch": 12.24,
141
+ "learning_rate": 6.7425e-05,
142
+ "loss": 1.6853,
143
+ "step": 1800
144
+ },
145
+ {
146
+ "epoch": 12.92,
147
+ "learning_rate": 7.1175e-05,
148
+ "loss": 1.6283,
149
+ "step": 1900
150
+ },
151
+ {
152
+ "epoch": 13.6,
153
+ "learning_rate": 7.492499999999999e-05,
154
+ "loss": 1.579,
155
+ "step": 2000
156
+ },
157
+ {
158
+ "epoch": 13.6,
159
+ "eval_loss": 1.3603801727294922,
160
+ "eval_runtime": 61.857,
161
+ "eval_samples_per_second": 33.868,
162
+ "eval_steps_per_second": 4.236,
163
+ "eval_wer": 0.7647276084949215,
164
+ "step": 2000
165
+ },
166
+ {
167
+ "epoch": 14.29,
168
+ "learning_rate": 7.362616822429906e-05,
169
+ "loss": 1.5102,
170
+ "step": 2100
171
+ },
172
+ {
173
+ "epoch": 14.96,
174
+ "learning_rate": 7.222429906542056e-05,
175
+ "loss": 1.478,
176
+ "step": 2200
177
+ },
178
+ {
179
+ "epoch": 15.65,
180
+ "learning_rate": 7.082242990654205e-05,
181
+ "loss": 1.472,
182
+ "step": 2300
183
+ },
184
+ {
185
+ "epoch": 16.33,
186
+ "learning_rate": 6.942056074766355e-05,
187
+ "loss": 1.4333,
188
+ "step": 2400
189
+ },
190
+ {
191
+ "epoch": 17.01,
192
+ "learning_rate": 6.801869158878504e-05,
193
+ "loss": 1.3773,
194
+ "step": 2500
195
+ },
196
+ {
197
+ "epoch": 17.01,
198
+ "eval_loss": 1.2737088203430176,
199
+ "eval_runtime": 62.7361,
200
+ "eval_samples_per_second": 33.394,
201
+ "eval_steps_per_second": 4.176,
202
+ "eval_wer": 0.7519236688211758,
203
+ "step": 2500
204
+ },
205
+ {
206
+ "epoch": 17.69,
207
+ "learning_rate": 6.661682242990654e-05,
208
+ "loss": 1.3823,
209
+ "step": 2600
210
+ },
211
+ {
212
+ "epoch": 18.37,
213
+ "learning_rate": 6.52429906542056e-05,
214
+ "loss": 1.3638,
215
+ "step": 2700
216
+ },
217
+ {
218
+ "epoch": 19.05,
219
+ "learning_rate": 6.384112149532709e-05,
220
+ "loss": 1.3431,
221
+ "step": 2800
222
+ },
223
+ {
224
+ "epoch": 19.73,
225
+ "learning_rate": 6.243925233644859e-05,
226
+ "loss": 1.3223,
227
+ "step": 2900
228
+ },
229
+ {
230
+ "epoch": 20.41,
231
+ "learning_rate": 6.103738317757008e-05,
232
+ "loss": 1.3165,
233
+ "step": 3000
234
+ },
235
+ {
236
+ "epoch": 20.41,
237
+ "eval_loss": 1.245711326599121,
238
+ "eval_runtime": 62.7019,
239
+ "eval_samples_per_second": 33.412,
240
+ "eval_steps_per_second": 4.179,
241
+ "eval_wer": 0.740104647583872,
242
+ "step": 3000
243
+ },
244
+ {
245
+ "epoch": 21.09,
246
+ "learning_rate": 5.9635514018691585e-05,
247
+ "loss": 1.292,
248
+ "step": 3100
249
+ },
250
+ {
251
+ "epoch": 21.77,
252
+ "learning_rate": 5.823364485981308e-05,
253
+ "loss": 1.2607,
254
+ "step": 3200
255
+ },
256
+ {
257
+ "epoch": 22.45,
258
+ "learning_rate": 5.6831775700934575e-05,
259
+ "loss": 1.2767,
260
+ "step": 3300
261
+ },
262
+ {
263
+ "epoch": 23.13,
264
+ "learning_rate": 5.542990654205607e-05,
265
+ "loss": 1.2518,
266
+ "step": 3400
267
+ },
268
+ {
269
+ "epoch": 23.81,
270
+ "learning_rate": 5.4028037383177566e-05,
271
+ "loss": 1.2274,
272
+ "step": 3500
273
+ },
274
+ {
275
+ "epoch": 23.81,
276
+ "eval_loss": 1.3616658449172974,
277
+ "eval_runtime": 63.5478,
278
+ "eval_samples_per_second": 32.967,
279
+ "eval_steps_per_second": 4.123,
280
+ "eval_wer": 0.7300707910126193,
281
+ "step": 3500
282
+ },
283
+ {
284
+ "epoch": 24.49,
285
+ "learning_rate": 5.262616822429906e-05,
286
+ "loss": 1.1919,
287
+ "step": 3600
288
+ },
289
+ {
290
+ "epoch": 25.17,
291
+ "learning_rate": 5.1224299065420557e-05,
292
+ "loss": 1.2453,
293
+ "step": 3700
294
+ },
295
+ {
296
+ "epoch": 25.85,
297
+ "learning_rate": 4.982242990654205e-05,
298
+ "loss": 1.2109,
299
+ "step": 3800
300
+ },
301
+ {
302
+ "epoch": 26.53,
303
+ "learning_rate": 4.842056074766355e-05,
304
+ "loss": 1.1759,
305
+ "step": 3900
306
+ },
307
+ {
308
+ "epoch": 27.21,
309
+ "learning_rate": 4.701869158878504e-05,
310
+ "loss": 1.1787,
311
+ "step": 4000
312
+ },
313
+ {
314
+ "epoch": 27.21,
315
+ "eval_loss": 1.2067540884017944,
316
+ "eval_runtime": 62.6504,
317
+ "eval_samples_per_second": 33.44,
318
+ "eval_steps_per_second": 4.182,
319
+ "eval_wer": 0.7009541397353032,
320
+ "step": 4000
321
+ },
322
+ {
323
+ "epoch": 27.89,
324
+ "learning_rate": 4.561682242990654e-05,
325
+ "loss": 1.1783,
326
+ "step": 4100
327
+ },
328
+ {
329
+ "epoch": 28.57,
330
+ "learning_rate": 4.421495327102803e-05,
331
+ "loss": 1.1709,
332
+ "step": 4200
333
+ },
334
+ {
335
+ "epoch": 29.25,
336
+ "learning_rate": 4.281308411214953e-05,
337
+ "loss": 1.1346,
338
+ "step": 4300
339
+ },
340
+ {
341
+ "epoch": 29.93,
342
+ "learning_rate": 4.1411214953271024e-05,
343
+ "loss": 1.1589,
344
+ "step": 4400
345
+ },
346
+ {
347
+ "epoch": 30.61,
348
+ "learning_rate": 4.000934579439252e-05,
349
+ "loss": 1.1467,
350
+ "step": 4500
351
+ },
352
+ {
353
+ "epoch": 30.61,
354
+ "eval_loss": 1.241578459739685,
355
+ "eval_runtime": 62.8423,
356
+ "eval_samples_per_second": 33.337,
357
+ "eval_steps_per_second": 4.169,
358
+ "eval_wer": 0.6946137273007079,
359
+ "step": 4500
360
+ },
361
+ {
362
+ "epoch": 31.29,
363
+ "learning_rate": 3.8621495327102795e-05,
364
+ "loss": 1.127,
365
+ "step": 4600
366
+ },
367
+ {
368
+ "epoch": 31.97,
369
+ "learning_rate": 3.72196261682243e-05,
370
+ "loss": 1.1027,
371
+ "step": 4700
372
+ },
373
+ {
374
+ "epoch": 32.65,
375
+ "learning_rate": 3.581775700934579e-05,
376
+ "loss": 1.125,
377
+ "step": 4800
378
+ },
379
+ {
380
+ "epoch": 33.33,
381
+ "learning_rate": 3.4429906542056075e-05,
382
+ "loss": 1.1193,
383
+ "step": 4900
384
+ },
385
+ {
386
+ "epoch": 34.01,
387
+ "learning_rate": 3.302803738317757e-05,
388
+ "loss": 1.0801,
389
+ "step": 5000
390
+ },
391
+ {
392
+ "epoch": 34.01,
393
+ "eval_loss": 1.2311749458312988,
394
+ "eval_runtime": 61.8112,
395
+ "eval_samples_per_second": 33.894,
396
+ "eval_steps_per_second": 4.239,
397
+ "eval_wer": 0.6990458602646968,
398
+ "step": 5000
399
+ },
400
+ {
401
+ "epoch": 34.69,
402
+ "learning_rate": 3.1626168224299066e-05,
403
+ "loss": 1.0891,
404
+ "step": 5100
405
+ },
406
+ {
407
+ "epoch": 35.37,
408
+ "learning_rate": 3.0224299065420558e-05,
409
+ "loss": 1.0841,
410
+ "step": 5200
411
+ },
412
+ {
413
+ "epoch": 36.05,
414
+ "learning_rate": 2.8822429906542053e-05,
415
+ "loss": 1.0679,
416
+ "step": 5300
417
+ },
418
+ {
419
+ "epoch": 36.73,
420
+ "learning_rate": 2.742056074766355e-05,
421
+ "loss": 1.0555,
422
+ "step": 5400
423
+ },
424
+ {
425
+ "epoch": 37.41,
426
+ "learning_rate": 2.6018691588785044e-05,
427
+ "loss": 1.0709,
428
+ "step": 5500
429
+ },
430
+ {
431
+ "epoch": 37.41,
432
+ "eval_loss": 1.2984226942062378,
433
+ "eval_runtime": 62.056,
434
+ "eval_samples_per_second": 33.76,
435
+ "eval_steps_per_second": 4.222,
436
+ "eval_wer": 0.7137580794090489,
437
+ "step": 5500
438
+ },
439
+ {
440
+ "epoch": 38.1,
441
+ "learning_rate": 2.461682242990654e-05,
442
+ "loss": 1.051,
443
+ "step": 5600
444
+ },
445
+ {
446
+ "epoch": 38.77,
447
+ "learning_rate": 2.3214953271028034e-05,
448
+ "loss": 1.0309,
449
+ "step": 5700
450
+ },
451
+ {
452
+ "epoch": 39.46,
453
+ "learning_rate": 2.181308411214953e-05,
454
+ "loss": 1.0343,
455
+ "step": 5800
456
+ },
457
+ {
458
+ "epoch": 40.14,
459
+ "learning_rate": 2.0411214953271025e-05,
460
+ "loss": 1.0456,
461
+ "step": 5900
462
+ },
463
+ {
464
+ "epoch": 40.81,
465
+ "learning_rate": 1.900934579439252e-05,
466
+ "loss": 1.0307,
467
+ "step": 6000
468
+ },
469
+ {
470
+ "epoch": 40.81,
471
+ "eval_loss": 1.2048739194869995,
472
+ "eval_runtime": 63.482,
473
+ "eval_samples_per_second": 33.001,
474
+ "eval_steps_per_second": 4.127,
475
+ "eval_wer": 0.6871037242228378,
476
+ "step": 6000
477
+ },
478
+ {
479
+ "epoch": 41.5,
480
+ "learning_rate": 1.760747663551402e-05,
481
+ "loss": 1.0092,
482
+ "step": 6100
483
+ },
484
+ {
485
+ "epoch": 42.18,
486
+ "learning_rate": 1.6205607476635514e-05,
487
+ "loss": 1.0258,
488
+ "step": 6200
489
+ },
490
+ {
491
+ "epoch": 42.86,
492
+ "learning_rate": 1.4803738317757008e-05,
493
+ "loss": 1.0129,
494
+ "step": 6300
495
+ },
496
+ {
497
+ "epoch": 43.54,
498
+ "learning_rate": 1.3401869158878503e-05,
499
+ "loss": 1.0044,
500
+ "step": 6400
501
+ },
502
+ {
503
+ "epoch": 44.22,
504
+ "learning_rate": 1.1999999999999999e-05,
505
+ "loss": 1.0003,
506
+ "step": 6500
507
+ },
508
+ {
509
+ "epoch": 44.22,
510
+ "eval_loss": 1.1955562829971313,
511
+ "eval_runtime": 63.8,
512
+ "eval_samples_per_second": 32.837,
513
+ "eval_steps_per_second": 4.107,
514
+ "eval_wer": 0.6841489689135118,
515
+ "step": 6500
516
+ },
517
+ {
518
+ "epoch": 44.9,
519
+ "learning_rate": 1.0598130841121494e-05,
520
+ "loss": 1.0068,
521
+ "step": 6600
522
+ },
523
+ {
524
+ "epoch": 45.58,
525
+ "learning_rate": 9.19626168224299e-06,
526
+ "loss": 1.0071,
527
+ "step": 6700
528
+ },
529
+ {
530
+ "epoch": 46.26,
531
+ "learning_rate": 7.794392523364485e-06,
532
+ "loss": 0.9813,
533
+ "step": 6800
534
+ },
535
+ {
536
+ "epoch": 46.94,
537
+ "learning_rate": 6.392523364485981e-06,
538
+ "loss": 0.9888,
539
+ "step": 6900
540
+ },
541
+ {
542
+ "epoch": 47.62,
543
+ "learning_rate": 4.990654205607477e-06,
544
+ "loss": 1.004,
545
+ "step": 7000
546
+ },
547
+ {
548
+ "epoch": 47.62,
549
+ "eval_loss": 1.2101075649261475,
550
+ "eval_runtime": 62.342,
551
+ "eval_samples_per_second": 33.605,
552
+ "eval_steps_per_second": 4.203,
553
+ "eval_wer": 0.6792859341335795,
554
+ "step": 7000
555
+ },
556
+ {
557
+ "epoch": 48.3,
558
+ "learning_rate": 3.5887850467289714e-06,
559
+ "loss": 0.9865,
560
+ "step": 7100
561
+ },
562
+ {
563
+ "epoch": 48.98,
564
+ "learning_rate": 2.186915887850467e-06,
565
+ "loss": 0.976,
566
+ "step": 7200
567
+ },
568
+ {
569
+ "epoch": 49.66,
570
+ "learning_rate": 7.850467289719626e-07,
571
+ "loss": 1.0041,
572
+ "step": 7300
573
+ },
574
+ {
575
+ "epoch": 50.0,
576
+ "step": 7350,
577
+ "total_flos": 2.9231564377949348e+19,
578
+ "train_loss": 2.1161629922049388,
579
+ "train_runtime": 13636.3379,
580
+ "train_samples_per_second": 17.274,
581
+ "train_steps_per_second": 0.539
582
+ }
583
+ ],
584
+ "max_steps": 7350,
585
+ "num_train_epochs": 50,
586
+ "total_flos": 2.9231564377949348e+19,
587
+ "trial_name": null,
588
+ "trial_params": null
589
+ }