anuragshas commited on
Commit
84fcc84
1 Parent(s): 466bfc4

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 +565 -0
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 60.0,
3
+ "eval_loss": 0.18953098356723785,
4
+ "eval_runtime": 68.8099,
5
+ "eval_samples": 1625,
6
+ "eval_samples_per_second": 23.616,
7
+ "eval_steps_per_second": 1.482,
8
+ "eval_wer": 0.1983627746660922,
9
+ "train_loss": 1.3688410950852585,
10
+ "train_runtime": 13291.4016,
11
+ "train_samples": 3548,
12
+ "train_samples_per_second": 16.016,
13
+ "train_steps_per_second": 0.501
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 60.0,
3
+ "eval_loss": 0.18953098356723785,
4
+ "eval_runtime": 68.8099,
5
+ "eval_samples": 1625,
6
+ "eval_samples_per_second": 23.616,
7
+ "eval_steps_per_second": 1.482,
8
+ "eval_wer": 0.1983627746660922
9
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 60.0,
3
+ "train_loss": 1.3688410950852585,
4
+ "train_runtime": 13291.4016,
5
+ "train_samples": 3548,
6
+ "train_samples_per_second": 16.016,
7
+ "train_steps_per_second": 0.501
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,565 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 60.0,
5
+ "global_step": 6660,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.9,
12
+ "learning_rate": 7.275e-06,
13
+ "loss": 13.5875,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 1.8,
18
+ "learning_rate": 1.4775e-05,
19
+ "loss": 5.5556,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 2.7,
24
+ "learning_rate": 2.2274999999999996e-05,
25
+ "loss": 4.0171,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 3.6,
30
+ "learning_rate": 2.9775e-05,
31
+ "loss": 3.4219,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 3.6,
36
+ "eval_loss": 3.3127081394195557,
37
+ "eval_runtime": 73.7584,
38
+ "eval_samples_per_second": 22.031,
39
+ "eval_steps_per_second": 1.383,
40
+ "eval_wer": 1.0,
41
+ "step": 400
42
+ },
43
+ {
44
+ "epoch": 4.5,
45
+ "learning_rate": 3.7275e-05,
46
+ "loss": 3.2104,
47
+ "step": 500
48
+ },
49
+ {
50
+ "epoch": 5.41,
51
+ "learning_rate": 4.4775e-05,
52
+ "loss": 3.152,
53
+ "step": 600
54
+ },
55
+ {
56
+ "epoch": 6.31,
57
+ "learning_rate": 5.227499999999999e-05,
58
+ "loss": 3.0987,
59
+ "step": 700
60
+ },
61
+ {
62
+ "epoch": 7.21,
63
+ "learning_rate": 5.9774999999999996e-05,
64
+ "loss": 3.0399,
65
+ "step": 800
66
+ },
67
+ {
68
+ "epoch": 7.21,
69
+ "eval_loss": 3.0330417156219482,
70
+ "eval_runtime": 71.6796,
71
+ "eval_samples_per_second": 22.67,
72
+ "eval_steps_per_second": 1.423,
73
+ "eval_wer": 1.0,
74
+ "step": 800
75
+ },
76
+ {
77
+ "epoch": 8.11,
78
+ "learning_rate": 6.7275e-05,
79
+ "loss": 3.0035,
80
+ "step": 900
81
+ },
82
+ {
83
+ "epoch": 9.01,
84
+ "learning_rate": 7.477499999999999e-05,
85
+ "loss": 2.7063,
86
+ "step": 1000
87
+ },
88
+ {
89
+ "epoch": 9.91,
90
+ "learning_rate": 7.371466431095406e-05,
91
+ "loss": 2.0137,
92
+ "step": 1100
93
+ },
94
+ {
95
+ "epoch": 10.81,
96
+ "learning_rate": 7.238957597173144e-05,
97
+ "loss": 1.5756,
98
+ "step": 1200
99
+ },
100
+ {
101
+ "epoch": 10.81,
102
+ "eval_loss": 0.6108289361000061,
103
+ "eval_runtime": 70.7969,
104
+ "eval_samples_per_second": 22.953,
105
+ "eval_steps_per_second": 1.441,
106
+ "eval_wer": 0.572425678586816,
107
+ "step": 1200
108
+ },
109
+ {
110
+ "epoch": 11.71,
111
+ "learning_rate": 7.106448763250883e-05,
112
+ "loss": 1.3527,
113
+ "step": 1300
114
+ },
115
+ {
116
+ "epoch": 12.61,
117
+ "learning_rate": 6.973939929328621e-05,
118
+ "loss": 1.2067,
119
+ "step": 1400
120
+ },
121
+ {
122
+ "epoch": 13.51,
123
+ "learning_rate": 6.841431095406359e-05,
124
+ "loss": 1.1508,
125
+ "step": 1500
126
+ },
127
+ {
128
+ "epoch": 14.41,
129
+ "learning_rate": 6.708922261484098e-05,
130
+ "loss": 1.0995,
131
+ "step": 1600
132
+ },
133
+ {
134
+ "epoch": 14.41,
135
+ "eval_loss": 0.30910709500312805,
136
+ "eval_runtime": 70.2841,
137
+ "eval_samples_per_second": 23.12,
138
+ "eval_steps_per_second": 1.451,
139
+ "eval_wer": 0.31538130116329166,
140
+ "step": 1600
141
+ },
142
+ {
143
+ "epoch": 15.32,
144
+ "learning_rate": 6.576413427561838e-05,
145
+ "loss": 1.0495,
146
+ "step": 1700
147
+ },
148
+ {
149
+ "epoch": 16.22,
150
+ "learning_rate": 6.443904593639576e-05,
151
+ "loss": 1.0183,
152
+ "step": 1800
153
+ },
154
+ {
155
+ "epoch": 17.12,
156
+ "learning_rate": 6.311395759717314e-05,
157
+ "loss": 0.9867,
158
+ "step": 1900
159
+ },
160
+ {
161
+ "epoch": 18.02,
162
+ "learning_rate": 6.178886925795053e-05,
163
+ "loss": 0.9639,
164
+ "step": 2000
165
+ },
166
+ {
167
+ "epoch": 18.02,
168
+ "eval_loss": 0.2596471905708313,
169
+ "eval_runtime": 70.5333,
170
+ "eval_samples_per_second": 23.039,
171
+ "eval_steps_per_second": 1.446,
172
+ "eval_wer": 0.2841016803102111,
173
+ "step": 2000
174
+ },
175
+ {
176
+ "epoch": 18.92,
177
+ "learning_rate": 6.046378091872791e-05,
178
+ "loss": 0.9383,
179
+ "step": 2100
180
+ },
181
+ {
182
+ "epoch": 19.82,
183
+ "learning_rate": 5.913869257950529e-05,
184
+ "loss": 0.9041,
185
+ "step": 2200
186
+ },
187
+ {
188
+ "epoch": 20.72,
189
+ "learning_rate": 5.781360424028268e-05,
190
+ "loss": 0.8936,
191
+ "step": 2300
192
+ },
193
+ {
194
+ "epoch": 21.62,
195
+ "learning_rate": 5.648851590106007e-05,
196
+ "loss": 0.9032,
197
+ "step": 2400
198
+ },
199
+ {
200
+ "epoch": 21.62,
201
+ "eval_loss": 0.22702418267726898,
202
+ "eval_runtime": 71.1235,
203
+ "eval_samples_per_second": 22.848,
204
+ "eval_steps_per_second": 1.434,
205
+ "eval_wer": 0.2513571736320552,
206
+ "step": 2400
207
+ },
208
+ {
209
+ "epoch": 22.52,
210
+ "learning_rate": 5.516342756183745e-05,
211
+ "loss": 0.8739,
212
+ "step": 2500
213
+ },
214
+ {
215
+ "epoch": 23.42,
216
+ "learning_rate": 5.383833922261484e-05,
217
+ "loss": 0.8602,
218
+ "step": 2600
219
+ },
220
+ {
221
+ "epoch": 24.32,
222
+ "learning_rate": 5.2513250883392223e-05,
223
+ "loss": 0.8483,
224
+ "step": 2700
225
+ },
226
+ {
227
+ "epoch": 25.23,
228
+ "learning_rate": 5.11881625441696e-05,
229
+ "loss": 0.8145,
230
+ "step": 2800
231
+ },
232
+ {
233
+ "epoch": 25.23,
234
+ "eval_loss": 0.21715673804283142,
235
+ "eval_runtime": 70.4062,
236
+ "eval_samples_per_second": 23.08,
237
+ "eval_steps_per_second": 1.449,
238
+ "eval_wer": 0.24834123222748816,
239
+ "step": 2800
240
+ },
241
+ {
242
+ "epoch": 26.13,
243
+ "learning_rate": 4.986307420494699e-05,
244
+ "loss": 0.8245,
245
+ "step": 2900
246
+ },
247
+ {
248
+ "epoch": 27.03,
249
+ "learning_rate": 4.853798586572438e-05,
250
+ "loss": 0.8085,
251
+ "step": 3000
252
+ },
253
+ {
254
+ "epoch": 27.93,
255
+ "learning_rate": 4.721289752650177e-05,
256
+ "loss": 0.8101,
257
+ "step": 3100
258
+ },
259
+ {
260
+ "epoch": 28.83,
261
+ "learning_rate": 4.588780918727915e-05,
262
+ "loss": 0.7845,
263
+ "step": 3200
264
+ },
265
+ {
266
+ "epoch": 28.83,
267
+ "eval_loss": 0.2083793580532074,
268
+ "eval_runtime": 70.1439,
269
+ "eval_samples_per_second": 23.167,
270
+ "eval_steps_per_second": 1.454,
271
+ "eval_wer": 0.23326152520465315,
272
+ "step": 3200
273
+ },
274
+ {
275
+ "epoch": 29.73,
276
+ "learning_rate": 4.4562720848056537e-05,
277
+ "loss": 0.7816,
278
+ "step": 3300
279
+ },
280
+ {
281
+ "epoch": 30.63,
282
+ "learning_rate": 4.3237632508833916e-05,
283
+ "loss": 0.7655,
284
+ "step": 3400
285
+ },
286
+ {
287
+ "epoch": 31.53,
288
+ "learning_rate": 4.19125441696113e-05,
289
+ "loss": 0.762,
290
+ "step": 3500
291
+ },
292
+ {
293
+ "epoch": 32.43,
294
+ "learning_rate": 4.058745583038869e-05,
295
+ "loss": 0.7694,
296
+ "step": 3600
297
+ },
298
+ {
299
+ "epoch": 32.43,
300
+ "eval_loss": 0.19744575023651123,
301
+ "eval_runtime": 70.5594,
302
+ "eval_samples_per_second": 23.03,
303
+ "eval_steps_per_second": 1.446,
304
+ "eval_wer": 0.22343817320120637,
305
+ "step": 3600
306
+ },
307
+ {
308
+ "epoch": 33.33,
309
+ "learning_rate": 3.926236749116607e-05,
310
+ "loss": 0.7517,
311
+ "step": 3700
312
+ },
313
+ {
314
+ "epoch": 34.23,
315
+ "learning_rate": 3.7937279151943456e-05,
316
+ "loss": 0.7401,
317
+ "step": 3800
318
+ },
319
+ {
320
+ "epoch": 35.14,
321
+ "learning_rate": 3.661219081272084e-05,
322
+ "loss": 0.7196,
323
+ "step": 3900
324
+ },
325
+ {
326
+ "epoch": 36.04,
327
+ "learning_rate": 3.528710247349823e-05,
328
+ "loss": 0.7333,
329
+ "step": 4000
330
+ },
331
+ {
332
+ "epoch": 36.04,
333
+ "eval_loss": 0.20204411447048187,
334
+ "eval_runtime": 69.6934,
335
+ "eval_samples_per_second": 23.316,
336
+ "eval_steps_per_second": 1.464,
337
+ "eval_wer": 0.21852649719948297,
338
+ "step": 4000
339
+ },
340
+ {
341
+ "epoch": 36.94,
342
+ "learning_rate": 3.3962014134275616e-05,
343
+ "loss": 0.7251,
344
+ "step": 4100
345
+ },
346
+ {
347
+ "epoch": 37.84,
348
+ "learning_rate": 3.2636925795053e-05,
349
+ "loss": 0.7116,
350
+ "step": 4200
351
+ },
352
+ {
353
+ "epoch": 38.74,
354
+ "learning_rate": 3.131183745583038e-05,
355
+ "loss": 0.7236,
356
+ "step": 4300
357
+ },
358
+ {
359
+ "epoch": 39.64,
360
+ "learning_rate": 2.998674911660777e-05,
361
+ "loss": 0.693,
362
+ "step": 4400
363
+ },
364
+ {
365
+ "epoch": 39.64,
366
+ "eval_loss": 0.194662407040596,
367
+ "eval_runtime": 69.8883,
368
+ "eval_samples_per_second": 23.251,
369
+ "eval_steps_per_second": 1.459,
370
+ "eval_wer": 0.21482119775958639,
371
+ "step": 4400
372
+ },
373
+ {
374
+ "epoch": 40.54,
375
+ "learning_rate": 2.866166077738516e-05,
376
+ "loss": 0.6943,
377
+ "step": 4500
378
+ },
379
+ {
380
+ "epoch": 41.44,
381
+ "learning_rate": 2.7336572438162543e-05,
382
+ "loss": 0.6938,
383
+ "step": 4600
384
+ },
385
+ {
386
+ "epoch": 42.34,
387
+ "learning_rate": 2.6011484098939926e-05,
388
+ "loss": 0.6831,
389
+ "step": 4700
390
+ },
391
+ {
392
+ "epoch": 43.24,
393
+ "learning_rate": 2.469964664310954e-05,
394
+ "loss": 0.6802,
395
+ "step": 4800
396
+ },
397
+ {
398
+ "epoch": 43.24,
399
+ "eval_loss": 0.19601202011108398,
400
+ "eval_runtime": 69.3985,
401
+ "eval_samples_per_second": 23.416,
402
+ "eval_steps_per_second": 1.47,
403
+ "eval_wer": 0.2101680310211116,
404
+ "step": 4800
405
+ },
406
+ {
407
+ "epoch": 44.14,
408
+ "learning_rate": 2.3374558303886924e-05,
409
+ "loss": 0.6756,
410
+ "step": 4900
411
+ },
412
+ {
413
+ "epoch": 45.05,
414
+ "learning_rate": 2.2049469964664307e-05,
415
+ "loss": 0.67,
416
+ "step": 5000
417
+ },
418
+ {
419
+ "epoch": 45.95,
420
+ "learning_rate": 2.0724381625441694e-05,
421
+ "loss": 0.6753,
422
+ "step": 5100
423
+ },
424
+ {
425
+ "epoch": 46.85,
426
+ "learning_rate": 1.939929328621908e-05,
427
+ "loss": 0.667,
428
+ "step": 5200
429
+ },
430
+ {
431
+ "epoch": 46.85,
432
+ "eval_loss": 0.19041335582733154,
433
+ "eval_runtime": 70.1706,
434
+ "eval_samples_per_second": 23.158,
435
+ "eval_steps_per_second": 1.454,
436
+ "eval_wer": 0.2072382593709608,
437
+ "step": 5200
438
+ },
439
+ {
440
+ "epoch": 47.75,
441
+ "learning_rate": 1.8074204946996464e-05,
442
+ "loss": 0.6562,
443
+ "step": 5300
444
+ },
445
+ {
446
+ "epoch": 48.65,
447
+ "learning_rate": 1.674911660777385e-05,
448
+ "loss": 0.666,
449
+ "step": 5400
450
+ },
451
+ {
452
+ "epoch": 49.55,
453
+ "learning_rate": 1.5424028268551237e-05,
454
+ "loss": 0.6572,
455
+ "step": 5500
456
+ },
457
+ {
458
+ "epoch": 50.45,
459
+ "learning_rate": 1.409893992932862e-05,
460
+ "loss": 0.6486,
461
+ "step": 5600
462
+ },
463
+ {
464
+ "epoch": 50.45,
465
+ "eval_loss": 0.18806982040405273,
466
+ "eval_runtime": 70.1099,
467
+ "eval_samples_per_second": 23.178,
468
+ "eval_steps_per_second": 1.455,
469
+ "eval_wer": 0.2009478672985782,
470
+ "step": 5600
471
+ },
472
+ {
473
+ "epoch": 51.35,
474
+ "learning_rate": 1.2773851590106007e-05,
475
+ "loss": 0.6484,
476
+ "step": 5700
477
+ },
478
+ {
479
+ "epoch": 52.25,
480
+ "learning_rate": 1.1448763250883392e-05,
481
+ "loss": 0.6549,
482
+ "step": 5800
483
+ },
484
+ {
485
+ "epoch": 53.15,
486
+ "learning_rate": 1.0123674911660777e-05,
487
+ "loss": 0.6322,
488
+ "step": 5900
489
+ },
490
+ {
491
+ "epoch": 54.05,
492
+ "learning_rate": 8.811837455830388e-06,
493
+ "loss": 0.6339,
494
+ "step": 6000
495
+ },
496
+ {
497
+ "epoch": 54.05,
498
+ "eval_loss": 0.1877446174621582,
499
+ "eval_runtime": 70.1417,
500
+ "eval_samples_per_second": 23.167,
501
+ "eval_steps_per_second": 1.454,
502
+ "eval_wer": 0.1988797931925894,
503
+ "step": 6000
504
+ },
505
+ {
506
+ "epoch": 54.95,
507
+ "learning_rate": 7.486749116607773e-06,
508
+ "loss": 0.6413,
509
+ "step": 6100
510
+ },
511
+ {
512
+ "epoch": 55.86,
513
+ "learning_rate": 6.161660777385159e-06,
514
+ "loss": 0.6392,
515
+ "step": 6200
516
+ },
517
+ {
518
+ "epoch": 56.76,
519
+ "learning_rate": 4.836572438162544e-06,
520
+ "loss": 0.6363,
521
+ "step": 6300
522
+ },
523
+ {
524
+ "epoch": 57.66,
525
+ "learning_rate": 3.511484098939929e-06,
526
+ "loss": 0.6254,
527
+ "step": 6400
528
+ },
529
+ {
530
+ "epoch": 57.66,
531
+ "eval_loss": 0.18933533132076263,
532
+ "eval_runtime": 69.6111,
533
+ "eval_samples_per_second": 23.344,
534
+ "eval_steps_per_second": 1.465,
535
+ "eval_wer": 0.2002585092632486,
536
+ "step": 6400
537
+ },
538
+ {
539
+ "epoch": 58.56,
540
+ "learning_rate": 2.1863957597173144e-06,
541
+ "loss": 0.625,
542
+ "step": 6500
543
+ },
544
+ {
545
+ "epoch": 59.46,
546
+ "learning_rate": 8.613074204946996e-07,
547
+ "loss": 0.6379,
548
+ "step": 6600
549
+ },
550
+ {
551
+ "epoch": 60.0,
552
+ "step": 6660,
553
+ "total_flos": 3.0269424793140363e+19,
554
+ "train_loss": 1.3688410950852585,
555
+ "train_runtime": 13291.4016,
556
+ "train_samples_per_second": 16.016,
557
+ "train_steps_per_second": 0.501
558
+ }
559
+ ],
560
+ "max_steps": 6660,
561
+ "num_train_epochs": 60,
562
+ "total_flos": 3.0269424793140363e+19,
563
+ "trial_name": null,
564
+ "trial_params": null
565
+ }