DrishtiSharma commited on
Commit
5f68934
1 Parent(s): ebf3100

End of training

Browse files
Files changed (4) hide show
  1. all_results.json +14 -0
  2. eval_results.json +9 -0
  3. train_results.json +8 -0
  4. trainer_state.json +538 -0
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 50.0,
3
+ "eval_loss": 0.7062973976135254,
4
+ "eval_runtime": 74.685,
5
+ "eval_samples": 1700,
6
+ "eval_samples_per_second": 22.762,
7
+ "eval_steps_per_second": 1.433,
8
+ "eval_wer": 0.5557036754903828,
9
+ "train_loss": 0.7598760600297347,
10
+ "train_runtime": 13000.5685,
11
+ "train_samples": 3676,
12
+ "train_samples_per_second": 14.138,
13
+ "train_steps_per_second": 0.442
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 50.0,
3
+ "eval_loss": 0.7062973976135254,
4
+ "eval_runtime": 74.685,
5
+ "eval_samples": 1700,
6
+ "eval_samples_per_second": 22.762,
7
+ "eval_steps_per_second": 1.433,
8
+ "eval_wer": 0.5557036754903828
9
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 50.0,
3
+ "train_loss": 0.7598760600297347,
4
+ "train_runtime": 13000.5685,
5
+ "train_samples": 3676,
6
+ "train_samples_per_second": 14.138,
7
+ "train_steps_per_second": 0.442
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,538 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 50.0,
5
+ "global_step": 5750,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.87,
12
+ "learning_rate": 5.7166666666666664e-06,
13
+ "loss": 13.1691,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 1.74,
18
+ "learning_rate": 1.155e-05,
19
+ "loss": 5.7916,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 2.61,
24
+ "learning_rate": 1.738333333333333e-05,
25
+ "loss": 3.9554,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 2.61,
30
+ "eval_loss": 3.8166096210479736,
31
+ "eval_runtime": 76.9069,
32
+ "eval_samples_per_second": 22.105,
33
+ "eval_steps_per_second": 1.391,
34
+ "eval_wer": 1.0,
35
+ "step": 300
36
+ },
37
+ {
38
+ "epoch": 3.48,
39
+ "learning_rate": 2.3216666666666664e-05,
40
+ "loss": 3.4177,
41
+ "step": 400
42
+ },
43
+ {
44
+ "epoch": 4.35,
45
+ "learning_rate": 2.9049999999999995e-05,
46
+ "loss": 3.1088,
47
+ "step": 500
48
+ },
49
+ {
50
+ "epoch": 5.22,
51
+ "learning_rate": 3.488333333333333e-05,
52
+ "loss": 2.9853,
53
+ "step": 600
54
+ },
55
+ {
56
+ "epoch": 5.22,
57
+ "eval_loss": 2.979321002960205,
58
+ "eval_runtime": 76.5995,
59
+ "eval_samples_per_second": 22.193,
60
+ "eval_steps_per_second": 1.397,
61
+ "eval_wer": 1.0,
62
+ "step": 600
63
+ },
64
+ {
65
+ "epoch": 6.09,
66
+ "learning_rate": 4.071666666666666e-05,
67
+ "loss": 2.6633,
68
+ "step": 700
69
+ },
70
+ {
71
+ "epoch": 6.96,
72
+ "learning_rate": 4.655e-05,
73
+ "loss": 1.3761,
74
+ "step": 800
75
+ },
76
+ {
77
+ "epoch": 7.83,
78
+ "learning_rate": 5.2383333333333324e-05,
79
+ "loss": 0.6939,
80
+ "step": 900
81
+ },
82
+ {
83
+ "epoch": 7.83,
84
+ "eval_loss": 0.972253143787384,
85
+ "eval_runtime": 77.2431,
86
+ "eval_samples_per_second": 22.008,
87
+ "eval_steps_per_second": 1.385,
88
+ "eval_wer": 0.9149368374277915,
89
+ "step": 900
90
+ },
91
+ {
92
+ "epoch": 8.7,
93
+ "learning_rate": 5.821666666666666e-05,
94
+ "loss": 0.5141,
95
+ "step": 1000
96
+ },
97
+ {
98
+ "epoch": 9.57,
99
+ "learning_rate": 6.405e-05,
100
+ "loss": 0.4189,
101
+ "step": 1100
102
+ },
103
+ {
104
+ "epoch": 10.43,
105
+ "learning_rate": 6.988333333333333e-05,
106
+ "loss": 0.352,
107
+ "step": 1200
108
+ },
109
+ {
110
+ "epoch": 10.43,
111
+ "eval_loss": 0.6556435823440552,
112
+ "eval_runtime": 76.3866,
113
+ "eval_samples_per_second": 22.255,
114
+ "eval_steps_per_second": 1.401,
115
+ "eval_wer": 0.7296387989589285,
116
+ "step": 1200
117
+ },
118
+ {
119
+ "epoch": 11.3,
120
+ "learning_rate": 6.849230769230769e-05,
121
+ "loss": 0.307,
122
+ "step": 1300
123
+ },
124
+ {
125
+ "epoch": 12.17,
126
+ "learning_rate": 6.695384615384615e-05,
127
+ "loss": 0.2658,
128
+ "step": 1400
129
+ },
130
+ {
131
+ "epoch": 13.04,
132
+ "learning_rate": 6.541538461538461e-05,
133
+ "loss": 0.2369,
134
+ "step": 1500
135
+ },
136
+ {
137
+ "epoch": 13.04,
138
+ "eval_loss": 0.6314664483070374,
139
+ "eval_runtime": 76.2318,
140
+ "eval_samples_per_second": 22.3,
141
+ "eval_steps_per_second": 1.404,
142
+ "eval_wer": 0.6927569351869485,
143
+ "step": 1500
144
+ },
145
+ {
146
+ "epoch": 13.91,
147
+ "learning_rate": 6.387692307692307e-05,
148
+ "loss": 0.2084,
149
+ "step": 1600
150
+ },
151
+ {
152
+ "epoch": 14.78,
153
+ "learning_rate": 6.233846153846153e-05,
154
+ "loss": 0.1904,
155
+ "step": 1700
156
+ },
157
+ {
158
+ "epoch": 15.65,
159
+ "learning_rate": 6.0799999999999994e-05,
160
+ "loss": 0.1757,
161
+ "step": 1800
162
+ },
163
+ {
164
+ "epoch": 15.65,
165
+ "eval_loss": 0.6387060284614563,
166
+ "eval_runtime": 76.1573,
167
+ "eval_samples_per_second": 22.322,
168
+ "eval_steps_per_second": 1.405,
169
+ "eval_wer": 0.6723798641528598,
170
+ "step": 1800
171
+ },
172
+ {
173
+ "epoch": 16.52,
174
+ "learning_rate": 5.9261538461538453e-05,
175
+ "loss": 0.1693,
176
+ "step": 1900
177
+ },
178
+ {
179
+ "epoch": 17.39,
180
+ "learning_rate": 5.772307692307692e-05,
181
+ "loss": 0.1572,
182
+ "step": 2000
183
+ },
184
+ {
185
+ "epoch": 18.26,
186
+ "learning_rate": 5.618461538461538e-05,
187
+ "loss": 0.1517,
188
+ "step": 2100
189
+ },
190
+ {
191
+ "epoch": 18.26,
192
+ "eval_loss": 0.6607237458229065,
193
+ "eval_runtime": 77.027,
194
+ "eval_samples_per_second": 22.07,
195
+ "eval_steps_per_second": 1.389,
196
+ "eval_wer": 0.6265473243191773,
197
+ "step": 2100
198
+ },
199
+ {
200
+ "epoch": 19.13,
201
+ "learning_rate": 5.464615384615384e-05,
202
+ "loss": 0.1403,
203
+ "step": 2200
204
+ },
205
+ {
206
+ "epoch": 20.0,
207
+ "learning_rate": 5.3107692307692305e-05,
208
+ "loss": 0.1382,
209
+ "step": 2300
210
+ },
211
+ {
212
+ "epoch": 20.87,
213
+ "learning_rate": 5.1569230769230765e-05,
214
+ "loss": 0.1229,
215
+ "step": 2400
216
+ },
217
+ {
218
+ "epoch": 20.87,
219
+ "eval_loss": 0.6519985198974609,
220
+ "eval_runtime": 76.1295,
221
+ "eval_samples_per_second": 22.33,
222
+ "eval_steps_per_second": 1.405,
223
+ "eval_wer": 0.6182949279502317,
224
+ "step": 2400
225
+ },
226
+ {
227
+ "epoch": 21.74,
228
+ "learning_rate": 5.0030769230769225e-05,
229
+ "loss": 0.129,
230
+ "step": 2500
231
+ },
232
+ {
233
+ "epoch": 22.61,
234
+ "learning_rate": 4.849230769230769e-05,
235
+ "loss": 0.1172,
236
+ "step": 2600
237
+ },
238
+ {
239
+ "epoch": 23.48,
240
+ "learning_rate": 4.695384615384615e-05,
241
+ "loss": 0.1201,
242
+ "step": 2700
243
+ },
244
+ {
245
+ "epoch": 23.48,
246
+ "eval_loss": 0.67499840259552,
247
+ "eval_runtime": 76.4187,
248
+ "eval_samples_per_second": 22.246,
249
+ "eval_steps_per_second": 1.4,
250
+ "eval_wer": 0.6115660509109376,
251
+ "step": 2700
252
+ },
253
+ {
254
+ "epoch": 24.35,
255
+ "learning_rate": 4.541538461538461e-05,
256
+ "loss": 0.1112,
257
+ "step": 2800
258
+ },
259
+ {
260
+ "epoch": 25.22,
261
+ "learning_rate": 4.387692307692307e-05,
262
+ "loss": 0.1131,
263
+ "step": 2900
264
+ },
265
+ {
266
+ "epoch": 26.09,
267
+ "learning_rate": 4.2338461538461536e-05,
268
+ "loss": 0.1076,
269
+ "step": 3000
270
+ },
271
+ {
272
+ "epoch": 26.09,
273
+ "eval_loss": 0.6698060035705566,
274
+ "eval_runtime": 75.169,
275
+ "eval_samples_per_second": 22.616,
276
+ "eval_steps_per_second": 1.423,
277
+ "eval_wer": 0.6006474957151019,
278
+ "step": 3000
279
+ },
280
+ {
281
+ "epoch": 26.96,
282
+ "learning_rate": 4.0799999999999996e-05,
283
+ "loss": 0.104,
284
+ "step": 3100
285
+ },
286
+ {
287
+ "epoch": 27.83,
288
+ "learning_rate": 3.9261538461538455e-05,
289
+ "loss": 0.1014,
290
+ "step": 3200
291
+ },
292
+ {
293
+ "epoch": 28.7,
294
+ "learning_rate": 3.772307692307692e-05,
295
+ "loss": 0.1006,
296
+ "step": 3300
297
+ },
298
+ {
299
+ "epoch": 28.7,
300
+ "eval_loss": 0.680385947227478,
301
+ "eval_runtime": 75.8528,
302
+ "eval_samples_per_second": 22.412,
303
+ "eval_steps_per_second": 1.411,
304
+ "eval_wer": 0.5871897416365137,
305
+ "step": 3300
306
+ },
307
+ {
308
+ "epoch": 29.57,
309
+ "learning_rate": 3.618461538461538e-05,
310
+ "loss": 0.1013,
311
+ "step": 3400
312
+ },
313
+ {
314
+ "epoch": 30.43,
315
+ "learning_rate": 3.464615384615384e-05,
316
+ "loss": 0.0972,
317
+ "step": 3500
318
+ },
319
+ {
320
+ "epoch": 31.3,
321
+ "learning_rate": 3.310769230769231e-05,
322
+ "loss": 0.0952,
323
+ "step": 3600
324
+ },
325
+ {
326
+ "epoch": 31.3,
327
+ "eval_loss": 0.7009897828102112,
328
+ "eval_runtime": 74.9836,
329
+ "eval_samples_per_second": 22.672,
330
+ "eval_steps_per_second": 1.427,
331
+ "eval_wer": 0.600266615882689,
332
+ "step": 3600
333
+ },
334
+ {
335
+ "epoch": 32.17,
336
+ "learning_rate": 3.156923076923077e-05,
337
+ "loss": 0.091,
338
+ "step": 3700
339
+ },
340
+ {
341
+ "epoch": 33.04,
342
+ "learning_rate": 3.0030769230769226e-05,
343
+ "loss": 0.0924,
344
+ "step": 3800
345
+ },
346
+ {
347
+ "epoch": 33.91,
348
+ "learning_rate": 2.849230769230769e-05,
349
+ "loss": 0.0894,
350
+ "step": 3900
351
+ },
352
+ {
353
+ "epoch": 33.91,
354
+ "eval_loss": 0.7068695425987244,
355
+ "eval_runtime": 74.6323,
356
+ "eval_samples_per_second": 22.778,
357
+ "eval_steps_per_second": 1.434,
358
+ "eval_wer": 0.584015743033073,
359
+ "step": 3900
360
+ },
361
+ {
362
+ "epoch": 34.78,
363
+ "learning_rate": 2.695384615384615e-05,
364
+ "loss": 0.0869,
365
+ "step": 4000
366
+ },
367
+ {
368
+ "epoch": 35.65,
369
+ "learning_rate": 2.5415384615384612e-05,
370
+ "loss": 0.0848,
371
+ "step": 4100
372
+ },
373
+ {
374
+ "epoch": 36.52,
375
+ "learning_rate": 2.3876923076923075e-05,
376
+ "loss": 0.0873,
377
+ "step": 4200
378
+ },
379
+ {
380
+ "epoch": 36.52,
381
+ "eval_loss": 0.6764819622039795,
382
+ "eval_runtime": 75.5212,
383
+ "eval_samples_per_second": 22.51,
384
+ "eval_steps_per_second": 1.417,
385
+ "eval_wer": 0.5781121056306735,
386
+ "step": 4200
387
+ },
388
+ {
389
+ "epoch": 37.39,
390
+ "learning_rate": 2.2338461538461534e-05,
391
+ "loss": 0.0845,
392
+ "step": 4300
393
+ },
394
+ {
395
+ "epoch": 38.26,
396
+ "learning_rate": 2.0799999999999997e-05,
397
+ "loss": 0.0813,
398
+ "step": 4400
399
+ },
400
+ {
401
+ "epoch": 39.13,
402
+ "learning_rate": 1.9261538461538457e-05,
403
+ "loss": 0.0798,
404
+ "step": 4500
405
+ },
406
+ {
407
+ "epoch": 39.13,
408
+ "eval_loss": 0.6821776032447815,
409
+ "eval_runtime": 76.5806,
410
+ "eval_samples_per_second": 22.199,
411
+ "eval_steps_per_second": 1.397,
412
+ "eval_wer": 0.5617342728369199,
413
+ "step": 4500
414
+ },
415
+ {
416
+ "epoch": 40.0,
417
+ "learning_rate": 1.772307692307692e-05,
418
+ "loss": 0.0783,
419
+ "step": 4600
420
+ },
421
+ {
422
+ "epoch": 40.87,
423
+ "learning_rate": 1.6199999999999997e-05,
424
+ "loss": 0.0753,
425
+ "step": 4700
426
+ },
427
+ {
428
+ "epoch": 41.74,
429
+ "learning_rate": 1.466153846153846e-05,
430
+ "loss": 0.0767,
431
+ "step": 4800
432
+ },
433
+ {
434
+ "epoch": 41.74,
435
+ "eval_loss": 0.7003459334373474,
436
+ "eval_runtime": 74.7077,
437
+ "eval_samples_per_second": 22.755,
438
+ "eval_steps_per_second": 1.432,
439
+ "eval_wer": 0.5637656319431219,
440
+ "step": 4800
441
+ },
442
+ {
443
+ "epoch": 42.61,
444
+ "learning_rate": 1.3123076923076922e-05,
445
+ "loss": 0.0739,
446
+ "step": 4900
447
+ },
448
+ {
449
+ "epoch": 43.48,
450
+ "learning_rate": 1.1584615384615385e-05,
451
+ "loss": 0.0766,
452
+ "step": 5000
453
+ },
454
+ {
455
+ "epoch": 44.35,
456
+ "learning_rate": 1.0046153846153846e-05,
457
+ "loss": 0.0717,
458
+ "step": 5100
459
+ },
460
+ {
461
+ "epoch": 44.35,
462
+ "eval_loss": 0.6969868540763855,
463
+ "eval_runtime": 74.6644,
464
+ "eval_samples_per_second": 22.769,
465
+ "eval_steps_per_second": 1.433,
466
+ "eval_wer": 0.5590681140100299,
467
+ "step": 5100
468
+ },
469
+ {
470
+ "epoch": 45.22,
471
+ "learning_rate": 8.507692307692307e-06,
472
+ "loss": 0.0732,
473
+ "step": 5200
474
+ },
475
+ {
476
+ "epoch": 46.09,
477
+ "learning_rate": 6.9692307692307684e-06,
478
+ "loss": 0.0713,
479
+ "step": 5300
480
+ },
481
+ {
482
+ "epoch": 46.96,
483
+ "learning_rate": 5.4307692307692306e-06,
484
+ "loss": 0.0687,
485
+ "step": 5400
486
+ },
487
+ {
488
+ "epoch": 46.96,
489
+ "eval_loss": 0.7072671055793762,
490
+ "eval_runtime": 74.5796,
491
+ "eval_samples_per_second": 22.794,
492
+ "eval_steps_per_second": 1.435,
493
+ "eval_wer": 0.560337713451406,
494
+ "step": 5400
495
+ },
496
+ {
497
+ "epoch": 47.83,
498
+ "learning_rate": 3.892307692307692e-06,
499
+ "loss": 0.0674,
500
+ "step": 5500
501
+ },
502
+ {
503
+ "epoch": 48.7,
504
+ "learning_rate": 2.3538461538461536e-06,
505
+ "loss": 0.0702,
506
+ "step": 5600
507
+ },
508
+ {
509
+ "epoch": 49.57,
510
+ "learning_rate": 8.153846153846153e-07,
511
+ "loss": 0.0669,
512
+ "step": 5700
513
+ },
514
+ {
515
+ "epoch": 49.57,
516
+ "eval_loss": 0.7059715390205383,
517
+ "eval_runtime": 74.624,
518
+ "eval_samples_per_second": 22.781,
519
+ "eval_steps_per_second": 1.434,
520
+ "eval_wer": 0.5558941154065892,
521
+ "step": 5700
522
+ },
523
+ {
524
+ "epoch": 50.0,
525
+ "step": 5750,
526
+ "total_flos": 2.9609940258263142e+19,
527
+ "train_loss": 0.7598760600297347,
528
+ "train_runtime": 13000.5685,
529
+ "train_samples_per_second": 14.138,
530
+ "train_steps_per_second": 0.442
531
+ }
532
+ ],
533
+ "max_steps": 5750,
534
+ "num_train_epochs": 50,
535
+ "total_flos": 2.9609940258263142e+19,
536
+ "trial_name": null,
537
+ "trial_params": null
538
+ }