DrishtiSharma commited on
Commit
4451829
1 Parent(s): cede41b

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 +790 -0
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 100.0,
3
+ "eval_loss": 0.29942241311073303,
4
+ "eval_runtime": 66.0035,
5
+ "eval_samples": 1625,
6
+ "eval_samples_per_second": 24.62,
7
+ "eval_steps_per_second": 1.545,
8
+ "eval_wer": 0.27806979750107713,
9
+ "train_loss": 1.0701171730660104,
10
+ "train_runtime": 18720.6378,
11
+ "train_samples": 3548,
12
+ "train_samples_per_second": 18.952,
13
+ "train_steps_per_second": 0.593
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 100.0,
3
+ "eval_loss": 0.29942241311073303,
4
+ "eval_runtime": 66.0035,
5
+ "eval_samples": 1625,
6
+ "eval_samples_per_second": 24.62,
7
+ "eval_steps_per_second": 1.545,
8
+ "eval_wer": 0.27806979750107713
9
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 100.0,
3
+ "train_loss": 1.0701171730660104,
4
+ "train_runtime": 18720.6378,
5
+ "train_samples": 3548,
6
+ "train_samples_per_second": 18.952,
7
+ "train_steps_per_second": 0.593
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,790 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 100.0,
5
+ "global_step": 11100,
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": 3.772222222222222e-06,
13
+ "loss": 14.8289,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 1.8,
18
+ "learning_rate": 7.66111111111111e-06,
19
+ "loss": 7.3013,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 2.7,
24
+ "learning_rate": 1.155e-05,
25
+ "loss": 4.522,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 3.6,
30
+ "learning_rate": 1.5438888888888887e-05,
31
+ "loss": 3.9098,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 4.5,
36
+ "learning_rate": 1.9327777777777774e-05,
37
+ "loss": 3.511,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 5.41,
42
+ "learning_rate": 2.3216666666666664e-05,
43
+ "loss": 3.2448,
44
+ "step": 600
45
+ },
46
+ {
47
+ "epoch": 6.31,
48
+ "learning_rate": 2.7105555555555554e-05,
49
+ "loss": 3.1281,
50
+ "step": 700
51
+ },
52
+ {
53
+ "epoch": 7.21,
54
+ "learning_rate": 3.099444444444444e-05,
55
+ "loss": 3.0628,
56
+ "step": 800
57
+ },
58
+ {
59
+ "epoch": 8.11,
60
+ "learning_rate": 3.488333333333333e-05,
61
+ "loss": 3.038,
62
+ "step": 900
63
+ },
64
+ {
65
+ "epoch": 9.01,
66
+ "learning_rate": 3.877222222222222e-05,
67
+ "loss": 3.0174,
68
+ "step": 1000
69
+ },
70
+ {
71
+ "epoch": 9.01,
72
+ "eval_loss": 3.055245876312256,
73
+ "eval_runtime": 67.6307,
74
+ "eval_samples_per_second": 24.028,
75
+ "eval_steps_per_second": 1.508,
76
+ "eval_wer": 1.0,
77
+ "step": 1000
78
+ },
79
+ {
80
+ "epoch": 9.91,
81
+ "learning_rate": 4.2661111111111106e-05,
82
+ "loss": 2.9754,
83
+ "step": 1100
84
+ },
85
+ {
86
+ "epoch": 10.81,
87
+ "learning_rate": 4.655e-05,
88
+ "loss": 2.6398,
89
+ "step": 1200
90
+ },
91
+ {
92
+ "epoch": 11.71,
93
+ "learning_rate": 5.043888888888888e-05,
94
+ "loss": 2.0198,
95
+ "step": 1300
96
+ },
97
+ {
98
+ "epoch": 12.61,
99
+ "learning_rate": 5.4327777777777774e-05,
100
+ "loss": 1.6983,
101
+ "step": 1400
102
+ },
103
+ {
104
+ "epoch": 13.51,
105
+ "learning_rate": 5.821666666666666e-05,
106
+ "loss": 1.5587,
107
+ "step": 1500
108
+ },
109
+ {
110
+ "epoch": 14.41,
111
+ "learning_rate": 6.210555555555555e-05,
112
+ "loss": 1.4169,
113
+ "step": 1600
114
+ },
115
+ {
116
+ "epoch": 15.32,
117
+ "learning_rate": 6.599444444444444e-05,
118
+ "loss": 1.2776,
119
+ "step": 1700
120
+ },
121
+ {
122
+ "epoch": 16.22,
123
+ "learning_rate": 6.988333333333333e-05,
124
+ "loss": 1.161,
125
+ "step": 1800
126
+ },
127
+ {
128
+ "epoch": 17.12,
129
+ "learning_rate": 6.926989247311828e-05,
130
+ "loss": 1.0941,
131
+ "step": 1900
132
+ },
133
+ {
134
+ "epoch": 18.02,
135
+ "learning_rate": 6.851720430107527e-05,
136
+ "loss": 1.0446,
137
+ "step": 2000
138
+ },
139
+ {
140
+ "epoch": 18.02,
141
+ "eval_loss": 0.6708454489707947,
142
+ "eval_runtime": 67.0128,
143
+ "eval_samples_per_second": 24.249,
144
+ "eval_steps_per_second": 1.522,
145
+ "eval_wer": 0.7576906505816459,
146
+ "step": 2000
147
+ },
148
+ {
149
+ "epoch": 18.92,
150
+ "learning_rate": 6.776451612903225e-05,
151
+ "loss": 0.9991,
152
+ "step": 2100
153
+ },
154
+ {
155
+ "epoch": 19.82,
156
+ "learning_rate": 6.701182795698924e-05,
157
+ "loss": 0.9554,
158
+ "step": 2200
159
+ },
160
+ {
161
+ "epoch": 20.72,
162
+ "learning_rate": 6.625913978494624e-05,
163
+ "loss": 0.9412,
164
+ "step": 2300
165
+ },
166
+ {
167
+ "epoch": 21.62,
168
+ "learning_rate": 6.550645161290322e-05,
169
+ "loss": 0.8972,
170
+ "step": 2400
171
+ },
172
+ {
173
+ "epoch": 22.52,
174
+ "learning_rate": 6.475376344086021e-05,
175
+ "loss": 0.8861,
176
+ "step": 2500
177
+ },
178
+ {
179
+ "epoch": 23.42,
180
+ "learning_rate": 6.400107526881719e-05,
181
+ "loss": 0.8724,
182
+ "step": 2600
183
+ },
184
+ {
185
+ "epoch": 24.32,
186
+ "learning_rate": 6.324838709677419e-05,
187
+ "loss": 0.8638,
188
+ "step": 2700
189
+ },
190
+ {
191
+ "epoch": 25.23,
192
+ "learning_rate": 6.249569892473118e-05,
193
+ "loss": 0.8306,
194
+ "step": 2800
195
+ },
196
+ {
197
+ "epoch": 26.13,
198
+ "learning_rate": 6.174301075268816e-05,
199
+ "loss": 0.8184,
200
+ "step": 2900
201
+ },
202
+ {
203
+ "epoch": 27.03,
204
+ "learning_rate": 6.0990322580645156e-05,
205
+ "loss": 0.7995,
206
+ "step": 3000
207
+ },
208
+ {
209
+ "epoch": 27.03,
210
+ "eval_loss": 0.42021921277046204,
211
+ "eval_runtime": 67.5681,
212
+ "eval_samples_per_second": 24.05,
213
+ "eval_steps_per_second": 1.51,
214
+ "eval_wer": 0.4770357604480827,
215
+ "step": 3000
216
+ },
217
+ {
218
+ "epoch": 27.93,
219
+ "learning_rate": 6.023763440860215e-05,
220
+ "loss": 0.7849,
221
+ "step": 3100
222
+ },
223
+ {
224
+ "epoch": 28.83,
225
+ "learning_rate": 5.9484946236559134e-05,
226
+ "loss": 0.7679,
227
+ "step": 3200
228
+ },
229
+ {
230
+ "epoch": 29.73,
231
+ "learning_rate": 5.873225806451612e-05,
232
+ "loss": 0.78,
233
+ "step": 3300
234
+ },
235
+ {
236
+ "epoch": 30.63,
237
+ "learning_rate": 5.797956989247311e-05,
238
+ "loss": 0.7505,
239
+ "step": 3400
240
+ },
241
+ {
242
+ "epoch": 31.53,
243
+ "learning_rate": 5.72268817204301e-05,
244
+ "loss": 0.7414,
245
+ "step": 3500
246
+ },
247
+ {
248
+ "epoch": 32.43,
249
+ "learning_rate": 5.648172043010752e-05,
250
+ "loss": 0.7269,
251
+ "step": 3600
252
+ },
253
+ {
254
+ "epoch": 33.33,
255
+ "learning_rate": 5.5729032258064514e-05,
256
+ "loss": 0.7023,
257
+ "step": 3700
258
+ },
259
+ {
260
+ "epoch": 34.23,
261
+ "learning_rate": 5.49763440860215e-05,
262
+ "loss": 0.7223,
263
+ "step": 3800
264
+ },
265
+ {
266
+ "epoch": 35.14,
267
+ "learning_rate": 5.422365591397849e-05,
268
+ "loss": 0.6899,
269
+ "step": 3900
270
+ },
271
+ {
272
+ "epoch": 36.04,
273
+ "learning_rate": 5.3470967741935484e-05,
274
+ "loss": 0.6978,
275
+ "step": 4000
276
+ },
277
+ {
278
+ "epoch": 36.04,
279
+ "eval_loss": 0.3053520619869232,
280
+ "eval_runtime": 66.0235,
281
+ "eval_samples_per_second": 24.612,
282
+ "eval_steps_per_second": 1.545,
283
+ "eval_wer": 0.3494183541576906,
284
+ "step": 4000
285
+ },
286
+ {
287
+ "epoch": 36.94,
288
+ "learning_rate": 5.271827956989247e-05,
289
+ "loss": 0.683,
290
+ "step": 4100
291
+ },
292
+ {
293
+ "epoch": 37.84,
294
+ "learning_rate": 5.1965591397849456e-05,
295
+ "loss": 0.652,
296
+ "step": 4200
297
+ },
298
+ {
299
+ "epoch": 38.74,
300
+ "learning_rate": 5.121290322580644e-05,
301
+ "loss": 0.6738,
302
+ "step": 4300
303
+ },
304
+ {
305
+ "epoch": 39.64,
306
+ "learning_rate": 5.0460215053763434e-05,
307
+ "loss": 0.661,
308
+ "step": 4400
309
+ },
310
+ {
311
+ "epoch": 40.54,
312
+ "learning_rate": 4.9707526881720426e-05,
313
+ "loss": 0.6532,
314
+ "step": 4500
315
+ },
316
+ {
317
+ "epoch": 41.44,
318
+ "learning_rate": 4.895483870967741e-05,
319
+ "loss": 0.6375,
320
+ "step": 4600
321
+ },
322
+ {
323
+ "epoch": 42.34,
324
+ "learning_rate": 4.8202150537634404e-05,
325
+ "loss": 0.6339,
326
+ "step": 4700
327
+ },
328
+ {
329
+ "epoch": 43.24,
330
+ "learning_rate": 4.74494623655914e-05,
331
+ "loss": 0.624,
332
+ "step": 4800
333
+ },
334
+ {
335
+ "epoch": 44.14,
336
+ "learning_rate": 4.669677419354838e-05,
337
+ "loss": 0.6087,
338
+ "step": 4900
339
+ },
340
+ {
341
+ "epoch": 45.05,
342
+ "learning_rate": 4.5944086021505375e-05,
343
+ "loss": 0.6189,
344
+ "step": 5000
345
+ },
346
+ {
347
+ "epoch": 45.05,
348
+ "eval_loss": 0.2878226935863495,
349
+ "eval_runtime": 65.3432,
350
+ "eval_samples_per_second": 24.869,
351
+ "eval_steps_per_second": 1.561,
352
+ "eval_wer": 0.31538130116329166,
353
+ "step": 5000
354
+ },
355
+ {
356
+ "epoch": 45.95,
357
+ "learning_rate": 4.519139784946237e-05,
358
+ "loss": 0.6067,
359
+ "step": 5100
360
+ },
361
+ {
362
+ "epoch": 46.85,
363
+ "learning_rate": 4.443870967741935e-05,
364
+ "loss": 0.6064,
365
+ "step": 5200
366
+ },
367
+ {
368
+ "epoch": 47.75,
369
+ "learning_rate": 4.3686021505376345e-05,
370
+ "loss": 0.5911,
371
+ "step": 5300
372
+ },
373
+ {
374
+ "epoch": 48.65,
375
+ "learning_rate": 4.2933333333333324e-05,
376
+ "loss": 0.6038,
377
+ "step": 5400
378
+ },
379
+ {
380
+ "epoch": 49.55,
381
+ "learning_rate": 4.218064516129032e-05,
382
+ "loss": 0.5886,
383
+ "step": 5500
384
+ },
385
+ {
386
+ "epoch": 50.45,
387
+ "learning_rate": 4.142795698924731e-05,
388
+ "loss": 0.5667,
389
+ "step": 5600
390
+ },
391
+ {
392
+ "epoch": 51.35,
393
+ "learning_rate": 4.0675268817204295e-05,
394
+ "loss": 0.5942,
395
+ "step": 5700
396
+ },
397
+ {
398
+ "epoch": 52.25,
399
+ "learning_rate": 3.992258064516129e-05,
400
+ "loss": 0.5785,
401
+ "step": 5800
402
+ },
403
+ {
404
+ "epoch": 53.15,
405
+ "learning_rate": 3.916989247311827e-05,
406
+ "loss": 0.5771,
407
+ "step": 5900
408
+ },
409
+ {
410
+ "epoch": 54.05,
411
+ "learning_rate": 3.8417204301075265e-05,
412
+ "loss": 0.5667,
413
+ "step": 6000
414
+ },
415
+ {
416
+ "epoch": 54.05,
417
+ "eval_loss": 0.31139013171195984,
418
+ "eval_runtime": 65.1359,
419
+ "eval_samples_per_second": 24.948,
420
+ "eval_steps_per_second": 1.566,
421
+ "eval_wer": 0.32856527358897025,
422
+ "step": 6000
423
+ },
424
+ {
425
+ "epoch": 54.95,
426
+ "learning_rate": 3.766451612903226e-05,
427
+ "loss": 0.5553,
428
+ "step": 6100
429
+ },
430
+ {
431
+ "epoch": 55.86,
432
+ "learning_rate": 3.6911827956989244e-05,
433
+ "loss": 0.5524,
434
+ "step": 6200
435
+ },
436
+ {
437
+ "epoch": 56.76,
438
+ "learning_rate": 3.616666666666667e-05,
439
+ "loss": 0.5389,
440
+ "step": 6300
441
+ },
442
+ {
443
+ "epoch": 57.66,
444
+ "learning_rate": 3.5413978494623646e-05,
445
+ "loss": 0.5376,
446
+ "step": 6400
447
+ },
448
+ {
449
+ "epoch": 58.56,
450
+ "learning_rate": 3.4661290322580645e-05,
451
+ "loss": 0.5379,
452
+ "step": 6500
453
+ },
454
+ {
455
+ "epoch": 59.46,
456
+ "learning_rate": 3.390860215053763e-05,
457
+ "loss": 0.5398,
458
+ "step": 6600
459
+ },
460
+ {
461
+ "epoch": 60.36,
462
+ "learning_rate": 3.3155913978494617e-05,
463
+ "loss": 0.5388,
464
+ "step": 6700
465
+ },
466
+ {
467
+ "epoch": 61.26,
468
+ "learning_rate": 3.240322580645161e-05,
469
+ "loss": 0.5181,
470
+ "step": 6800
471
+ },
472
+ {
473
+ "epoch": 62.16,
474
+ "learning_rate": 3.16505376344086e-05,
475
+ "loss": 0.516,
476
+ "step": 6900
477
+ },
478
+ {
479
+ "epoch": 63.06,
480
+ "learning_rate": 3.089784946236559e-05,
481
+ "loss": 0.5173,
482
+ "step": 7000
483
+ },
484
+ {
485
+ "epoch": 63.06,
486
+ "eval_loss": 0.3085043430328369,
487
+ "eval_runtime": 65.8426,
488
+ "eval_samples_per_second": 24.68,
489
+ "eval_steps_per_second": 1.549,
490
+ "eval_wer": 0.3021111589831969,
491
+ "step": 7000
492
+ },
493
+ {
494
+ "epoch": 63.96,
495
+ "learning_rate": 3.014516129032258e-05,
496
+ "loss": 0.515,
497
+ "step": 7100
498
+ },
499
+ {
500
+ "epoch": 64.86,
501
+ "learning_rate": 2.939247311827957e-05,
502
+ "loss": 0.5092,
503
+ "step": 7200
504
+ },
505
+ {
506
+ "epoch": 65.77,
507
+ "learning_rate": 2.8639784946236554e-05,
508
+ "loss": 0.5202,
509
+ "step": 7300
510
+ },
511
+ {
512
+ "epoch": 66.67,
513
+ "learning_rate": 2.7887096774193547e-05,
514
+ "loss": 0.5008,
515
+ "step": 7400
516
+ },
517
+ {
518
+ "epoch": 67.57,
519
+ "learning_rate": 2.7134408602150536e-05,
520
+ "loss": 0.4936,
521
+ "step": 7500
522
+ },
523
+ {
524
+ "epoch": 68.47,
525
+ "learning_rate": 2.6381720430107525e-05,
526
+ "loss": 0.4843,
527
+ "step": 7600
528
+ },
529
+ {
530
+ "epoch": 69.37,
531
+ "learning_rate": 2.5629032258064514e-05,
532
+ "loss": 0.4771,
533
+ "step": 7700
534
+ },
535
+ {
536
+ "epoch": 70.27,
537
+ "learning_rate": 2.4876344086021503e-05,
538
+ "loss": 0.4865,
539
+ "step": 7800
540
+ },
541
+ {
542
+ "epoch": 71.17,
543
+ "learning_rate": 2.4123655913978492e-05,
544
+ "loss": 0.4812,
545
+ "step": 7900
546
+ },
547
+ {
548
+ "epoch": 72.07,
549
+ "learning_rate": 2.337096774193548e-05,
550
+ "loss": 0.4682,
551
+ "step": 8000
552
+ },
553
+ {
554
+ "epoch": 72.07,
555
+ "eval_loss": 0.305768221616745,
556
+ "eval_runtime": 65.7283,
557
+ "eval_samples_per_second": 24.723,
558
+ "eval_steps_per_second": 1.552,
559
+ "eval_wer": 0.2969409737182249,
560
+ "step": 8000
561
+ },
562
+ {
563
+ "epoch": 72.97,
564
+ "learning_rate": 2.261827956989247e-05,
565
+ "loss": 0.4835,
566
+ "step": 8100
567
+ },
568
+ {
569
+ "epoch": 73.87,
570
+ "learning_rate": 2.1865591397849463e-05,
571
+ "loss": 0.4734,
572
+ "step": 8200
573
+ },
574
+ {
575
+ "epoch": 74.77,
576
+ "learning_rate": 2.111290322580645e-05,
577
+ "loss": 0.4685,
578
+ "step": 8300
579
+ },
580
+ {
581
+ "epoch": 75.68,
582
+ "learning_rate": 2.0360215053763437e-05,
583
+ "loss": 0.4556,
584
+ "step": 8400
585
+ },
586
+ {
587
+ "epoch": 76.58,
588
+ "learning_rate": 1.9607526881720426e-05,
589
+ "loss": 0.4655,
590
+ "step": 8500
591
+ },
592
+ {
593
+ "epoch": 77.48,
594
+ "learning_rate": 1.885483870967742e-05,
595
+ "loss": 0.4623,
596
+ "step": 8600
597
+ },
598
+ {
599
+ "epoch": 78.38,
600
+ "learning_rate": 1.8102150537634408e-05,
601
+ "loss": 0.4661,
602
+ "step": 8700
603
+ },
604
+ {
605
+ "epoch": 79.28,
606
+ "learning_rate": 1.7356989247311825e-05,
607
+ "loss": 0.451,
608
+ "step": 8800
609
+ },
610
+ {
611
+ "epoch": 80.18,
612
+ "learning_rate": 1.6604301075268817e-05,
613
+ "loss": 0.4482,
614
+ "step": 8900
615
+ },
616
+ {
617
+ "epoch": 81.08,
618
+ "learning_rate": 1.5851612903225803e-05,
619
+ "loss": 0.451,
620
+ "step": 9000
621
+ },
622
+ {
623
+ "epoch": 81.08,
624
+ "eval_loss": 0.31458327174186707,
625
+ "eval_runtime": 66.1124,
626
+ "eval_samples_per_second": 24.579,
627
+ "eval_steps_per_second": 1.543,
628
+ "eval_wer": 0.2907367514002585,
629
+ "step": 9000
630
+ },
631
+ {
632
+ "epoch": 81.98,
633
+ "learning_rate": 1.5098924731182794e-05,
634
+ "loss": 0.4504,
635
+ "step": 9100
636
+ },
637
+ {
638
+ "epoch": 82.88,
639
+ "learning_rate": 1.4346236559139784e-05,
640
+ "loss": 0.4373,
641
+ "step": 9200
642
+ },
643
+ {
644
+ "epoch": 83.78,
645
+ "learning_rate": 1.3593548387096772e-05,
646
+ "loss": 0.4257,
647
+ "step": 9300
648
+ },
649
+ {
650
+ "epoch": 84.68,
651
+ "learning_rate": 1.2840860215053762e-05,
652
+ "loss": 0.4454,
653
+ "step": 9400
654
+ },
655
+ {
656
+ "epoch": 85.59,
657
+ "learning_rate": 1.2088172043010753e-05,
658
+ "loss": 0.4303,
659
+ "step": 9500
660
+ },
661
+ {
662
+ "epoch": 86.49,
663
+ "learning_rate": 1.133548387096774e-05,
664
+ "loss": 0.4348,
665
+ "step": 9600
666
+ },
667
+ {
668
+ "epoch": 87.39,
669
+ "learning_rate": 1.058279569892473e-05,
670
+ "loss": 0.4411,
671
+ "step": 9700
672
+ },
673
+ {
674
+ "epoch": 88.29,
675
+ "learning_rate": 9.83010752688172e-06,
676
+ "loss": 0.4285,
677
+ "step": 9800
678
+ },
679
+ {
680
+ "epoch": 89.19,
681
+ "learning_rate": 9.077419354838708e-06,
682
+ "loss": 0.4235,
683
+ "step": 9900
684
+ },
685
+ {
686
+ "epoch": 90.09,
687
+ "learning_rate": 8.324731182795698e-06,
688
+ "loss": 0.4213,
689
+ "step": 10000
690
+ },
691
+ {
692
+ "epoch": 90.09,
693
+ "eval_loss": 0.3030133545398712,
694
+ "eval_runtime": 66.0396,
695
+ "eval_samples_per_second": 24.606,
696
+ "eval_steps_per_second": 1.545,
697
+ "eval_wer": 0.2880654890133563,
698
+ "step": 10000
699
+ },
700
+ {
701
+ "epoch": 90.99,
702
+ "learning_rate": 7.5720430107526875e-06,
703
+ "loss": 0.4148,
704
+ "step": 10100
705
+ },
706
+ {
707
+ "epoch": 91.89,
708
+ "learning_rate": 6.8193548387096766e-06,
709
+ "loss": 0.4157,
710
+ "step": 10200
711
+ },
712
+ {
713
+ "epoch": 92.79,
714
+ "learning_rate": 6.0666666666666665e-06,
715
+ "loss": 0.4146,
716
+ "step": 10300
717
+ },
718
+ {
719
+ "epoch": 93.69,
720
+ "learning_rate": 5.3139784946236555e-06,
721
+ "loss": 0.4101,
722
+ "step": 10400
723
+ },
724
+ {
725
+ "epoch": 94.59,
726
+ "learning_rate": 4.561290322580645e-06,
727
+ "loss": 0.428,
728
+ "step": 10500
729
+ },
730
+ {
731
+ "epoch": 95.5,
732
+ "learning_rate": 3.808602150537634e-06,
733
+ "loss": 0.4138,
734
+ "step": 10600
735
+ },
736
+ {
737
+ "epoch": 96.4,
738
+ "learning_rate": 3.0559139784946235e-06,
739
+ "loss": 0.4012,
740
+ "step": 10700
741
+ },
742
+ {
743
+ "epoch": 97.3,
744
+ "learning_rate": 2.303225806451613e-06,
745
+ "loss": 0.3957,
746
+ "step": 10800
747
+ },
748
+ {
749
+ "epoch": 98.2,
750
+ "learning_rate": 1.550537634408602e-06,
751
+ "loss": 0.4195,
752
+ "step": 10900
753
+ },
754
+ {
755
+ "epoch": 99.1,
756
+ "learning_rate": 8.053763440860214e-07,
757
+ "loss": 0.4005,
758
+ "step": 11000
759
+ },
760
+ {
761
+ "epoch": 99.1,
762
+ "eval_loss": 0.300111323595047,
763
+ "eval_runtime": 66.7539,
764
+ "eval_samples_per_second": 24.343,
765
+ "eval_steps_per_second": 1.528,
766
+ "eval_wer": 0.2789314950452391,
767
+ "step": 11000
768
+ },
769
+ {
770
+ "epoch": 100.0,
771
+ "learning_rate": 5.268817204301075e-08,
772
+ "loss": 0.4141,
773
+ "step": 11100
774
+ },
775
+ {
776
+ "epoch": 100.0,
777
+ "step": 11100,
778
+ "total_flos": 5.044434281896765e+19,
779
+ "train_loss": 1.0701171730660104,
780
+ "train_runtime": 18720.6378,
781
+ "train_samples_per_second": 18.952,
782
+ "train_steps_per_second": 0.593
783
+ }
784
+ ],
785
+ "max_steps": 11100,
786
+ "num_train_epochs": 100,
787
+ "total_flos": 5.044434281896765e+19,
788
+ "trial_name": null,
789
+ "trial_params": null
790
+ }