infinitejoy commited on
Commit
10e6825
1 Parent(s): bc3f17b

End of training

Browse files
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 100.0,
3
+ "eval_loss": 0.6002532243728638,
4
+ "eval_runtime": 111.3535,
5
+ "eval_samples": 1880,
6
+ "eval_samples_per_second": 16.883,
7
+ "eval_steps_per_second": 16.883,
8
+ "eval_wer": 0.4472546121421142,
9
+ "train_loss": 1.186835220579117,
10
+ "train_runtime": 23714.6286,
11
+ "train_samples": 4003,
12
+ "train_samples_per_second": 16.88,
13
+ "train_steps_per_second": 0.531
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 100.0,
3
+ "eval_loss": 0.6002532243728638,
4
+ "eval_runtime": 111.3535,
5
+ "eval_samples": 1880,
6
+ "eval_samples_per_second": 16.883,
7
+ "eval_steps_per_second": 16.883,
8
+ "eval_wer": 0.4472546121421142
9
+ }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:77663d208f94582374fed1b63a4070481a1c97acb1f2a5d5eabe0fd665bf0eac
3
  size 1262095857
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:349c2b02411572665815241f425172e5fd8c523637036b3d806e58c074df984d
3
  size 1262095857
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 100.0,
3
+ "train_loss": 1.186835220579117,
4
+ "train_runtime": 23714.6286,
5
+ "train_samples": 4003,
6
+ "train_samples_per_second": 16.88,
7
+ "train_steps_per_second": 0.531
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,835 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 100.0,
5
+ "global_step": 12600,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.79,
12
+ "learning_rate": 1.47e-05,
13
+ "loss": 9.8646,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 1.59,
18
+ "learning_rate": 2.97e-05,
19
+ "loss": 4.237,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 2.38,
24
+ "learning_rate": 4.4699999999999996e-05,
25
+ "loss": 3.3845,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 3.17,
30
+ "learning_rate": 5.97e-05,
31
+ "loss": 3.1402,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 3.97,
36
+ "learning_rate": 7.47e-05,
37
+ "loss": 3.0675,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 4.76,
42
+ "learning_rate": 8.969999999999998e-05,
43
+ "loss": 2.9355,
44
+ "step": 600
45
+ },
46
+ {
47
+ "epoch": 5.56,
48
+ "learning_rate": 0.00010469999999999998,
49
+ "loss": 2.4044,
50
+ "step": 700
51
+ },
52
+ {
53
+ "epoch": 6.35,
54
+ "learning_rate": 0.0001197,
55
+ "loss": 1.8799,
56
+ "step": 800
57
+ },
58
+ {
59
+ "epoch": 7.14,
60
+ "learning_rate": 0.0001347,
61
+ "loss": 1.6588,
62
+ "step": 900
63
+ },
64
+ {
65
+ "epoch": 7.94,
66
+ "learning_rate": 0.00014969999999999998,
67
+ "loss": 1.5757,
68
+ "step": 1000
69
+ },
70
+ {
71
+ "epoch": 8.73,
72
+ "learning_rate": 0.0001647,
73
+ "loss": 1.5135,
74
+ "step": 1100
75
+ },
76
+ {
77
+ "epoch": 9.52,
78
+ "learning_rate": 0.00017969999999999998,
79
+ "loss": 1.4795,
80
+ "step": 1200
81
+ },
82
+ {
83
+ "epoch": 10.32,
84
+ "learning_rate": 0.0001947,
85
+ "loss": 1.4418,
86
+ "step": 1300
87
+ },
88
+ {
89
+ "epoch": 11.11,
90
+ "learning_rate": 0.00020969999999999997,
91
+ "loss": 1.4246,
92
+ "step": 1400
93
+ },
94
+ {
95
+ "epoch": 11.9,
96
+ "learning_rate": 0.0002247,
97
+ "loss": 1.4072,
98
+ "step": 1500
99
+ },
100
+ {
101
+ "epoch": 12.7,
102
+ "learning_rate": 0.0002397,
103
+ "loss": 1.3936,
104
+ "step": 1600
105
+ },
106
+ {
107
+ "epoch": 13.49,
108
+ "learning_rate": 0.00025455,
109
+ "loss": 1.3632,
110
+ "step": 1700
111
+ },
112
+ {
113
+ "epoch": 14.29,
114
+ "learning_rate": 0.00026954999999999997,
115
+ "loss": 1.3721,
116
+ "step": 1800
117
+ },
118
+ {
119
+ "epoch": 15.08,
120
+ "learning_rate": 0.00028455,
121
+ "loss": 1.369,
122
+ "step": 1900
123
+ },
124
+ {
125
+ "epoch": 15.87,
126
+ "learning_rate": 0.00029955,
127
+ "loss": 1.3677,
128
+ "step": 2000
129
+ },
130
+ {
131
+ "epoch": 15.87,
132
+ "eval_loss": 0.6431823968887329,
133
+ "eval_runtime": 111.1097,
134
+ "eval_samples_per_second": 16.92,
135
+ "eval_steps_per_second": 16.92,
136
+ "eval_wer": 0.6198062079159139,
137
+ "step": 2000
138
+ },
139
+ {
140
+ "epoch": 16.67,
141
+ "learning_rate": 0.000297254716981132,
142
+ "loss": 1.3547,
143
+ "step": 2100
144
+ },
145
+ {
146
+ "epoch": 17.46,
147
+ "learning_rate": 0.0002944245283018868,
148
+ "loss": 1.346,
149
+ "step": 2200
150
+ },
151
+ {
152
+ "epoch": 18.25,
153
+ "learning_rate": 0.0002915943396226415,
154
+ "loss": 1.3279,
155
+ "step": 2300
156
+ },
157
+ {
158
+ "epoch": 19.05,
159
+ "learning_rate": 0.0002887641509433962,
160
+ "loss": 1.3309,
161
+ "step": 2400
162
+ },
163
+ {
164
+ "epoch": 19.84,
165
+ "learning_rate": 0.0002859339622641509,
166
+ "loss": 1.3086,
167
+ "step": 2500
168
+ },
169
+ {
170
+ "epoch": 20.63,
171
+ "learning_rate": 0.00028310377358490565,
172
+ "loss": 1.2874,
173
+ "step": 2600
174
+ },
175
+ {
176
+ "epoch": 21.43,
177
+ "learning_rate": 0.00028027358490566036,
178
+ "loss": 1.2788,
179
+ "step": 2700
180
+ },
181
+ {
182
+ "epoch": 22.22,
183
+ "learning_rate": 0.00027744339622641506,
184
+ "loss": 1.2532,
185
+ "step": 2800
186
+ },
187
+ {
188
+ "epoch": 23.02,
189
+ "learning_rate": 0.00027461320754716977,
190
+ "loss": 1.2481,
191
+ "step": 2900
192
+ },
193
+ {
194
+ "epoch": 23.81,
195
+ "learning_rate": 0.00027178301886792447,
196
+ "loss": 1.24,
197
+ "step": 3000
198
+ },
199
+ {
200
+ "epoch": 24.6,
201
+ "learning_rate": 0.00026895283018867923,
202
+ "loss": 1.2139,
203
+ "step": 3100
204
+ },
205
+ {
206
+ "epoch": 25.4,
207
+ "learning_rate": 0.00026612264150943394,
208
+ "loss": 1.2218,
209
+ "step": 3200
210
+ },
211
+ {
212
+ "epoch": 26.19,
213
+ "learning_rate": 0.00026329245283018864,
214
+ "loss": 1.2139,
215
+ "step": 3300
216
+ },
217
+ {
218
+ "epoch": 26.98,
219
+ "learning_rate": 0.00026046226415094335,
220
+ "loss": 1.2061,
221
+ "step": 3400
222
+ },
223
+ {
224
+ "epoch": 27.78,
225
+ "learning_rate": 0.0002576320754716981,
226
+ "loss": 1.1801,
227
+ "step": 3500
228
+ },
229
+ {
230
+ "epoch": 28.57,
231
+ "learning_rate": 0.0002548018867924528,
232
+ "loss": 1.1654,
233
+ "step": 3600
234
+ },
235
+ {
236
+ "epoch": 29.37,
237
+ "learning_rate": 0.0002519716981132075,
238
+ "loss": 1.1647,
239
+ "step": 3700
240
+ },
241
+ {
242
+ "epoch": 30.16,
243
+ "learning_rate": 0.0002491415094339622,
244
+ "loss": 1.1677,
245
+ "step": 3800
246
+ },
247
+ {
248
+ "epoch": 30.95,
249
+ "learning_rate": 0.00024631132075471693,
250
+ "loss": 1.1487,
251
+ "step": 3900
252
+ },
253
+ {
254
+ "epoch": 31.75,
255
+ "learning_rate": 0.00024348113207547166,
256
+ "loss": 1.1379,
257
+ "step": 4000
258
+ },
259
+ {
260
+ "epoch": 31.75,
261
+ "eval_loss": 0.6196324229240417,
262
+ "eval_runtime": 111.1854,
263
+ "eval_samples_per_second": 16.909,
264
+ "eval_steps_per_second": 16.909,
265
+ "eval_wer": 0.5592051239940877,
266
+ "step": 4000
267
+ },
268
+ {
269
+ "epoch": 32.54,
270
+ "learning_rate": 0.0002406509433962264,
271
+ "loss": 1.1574,
272
+ "step": 4100
273
+ },
274
+ {
275
+ "epoch": 33.33,
276
+ "learning_rate": 0.0002378207547169811,
277
+ "loss": 1.1431,
278
+ "step": 4200
279
+ },
280
+ {
281
+ "epoch": 34.13,
282
+ "learning_rate": 0.00023499056603773583,
283
+ "loss": 1.1192,
284
+ "step": 4300
285
+ },
286
+ {
287
+ "epoch": 34.92,
288
+ "learning_rate": 0.00023216037735849054,
289
+ "loss": 1.1371,
290
+ "step": 4400
291
+ },
292
+ {
293
+ "epoch": 35.71,
294
+ "learning_rate": 0.00022933018867924527,
295
+ "loss": 1.1085,
296
+ "step": 4500
297
+ },
298
+ {
299
+ "epoch": 36.51,
300
+ "learning_rate": 0.00022649999999999998,
301
+ "loss": 1.109,
302
+ "step": 4600
303
+ },
304
+ {
305
+ "epoch": 37.3,
306
+ "learning_rate": 0.0002236698113207547,
307
+ "loss": 1.1021,
308
+ "step": 4700
309
+ },
310
+ {
311
+ "epoch": 38.1,
312
+ "learning_rate": 0.00022083962264150941,
313
+ "loss": 1.1072,
314
+ "step": 4800
315
+ },
316
+ {
317
+ "epoch": 38.89,
318
+ "learning_rate": 0.00021800943396226415,
319
+ "loss": 1.1001,
320
+ "step": 4900
321
+ },
322
+ {
323
+ "epoch": 39.68,
324
+ "learning_rate": 0.00021517924528301885,
325
+ "loss": 1.0972,
326
+ "step": 5000
327
+ },
328
+ {
329
+ "epoch": 40.48,
330
+ "learning_rate": 0.00021234905660377356,
331
+ "loss": 1.0675,
332
+ "step": 5100
333
+ },
334
+ {
335
+ "epoch": 41.27,
336
+ "learning_rate": 0.0002095188679245283,
337
+ "loss": 1.0626,
338
+ "step": 5200
339
+ },
340
+ {
341
+ "epoch": 42.06,
342
+ "learning_rate": 0.000206688679245283,
343
+ "loss": 1.0505,
344
+ "step": 5300
345
+ },
346
+ {
347
+ "epoch": 42.86,
348
+ "learning_rate": 0.00020385849056603773,
349
+ "loss": 1.0665,
350
+ "step": 5400
351
+ },
352
+ {
353
+ "epoch": 43.65,
354
+ "learning_rate": 0.00020102830188679243,
355
+ "loss": 1.0456,
356
+ "step": 5500
357
+ },
358
+ {
359
+ "epoch": 44.44,
360
+ "learning_rate": 0.00019819811320754717,
361
+ "loss": 1.0474,
362
+ "step": 5600
363
+ },
364
+ {
365
+ "epoch": 45.24,
366
+ "learning_rate": 0.00019536792452830187,
367
+ "loss": 1.0439,
368
+ "step": 5700
369
+ },
370
+ {
371
+ "epoch": 46.03,
372
+ "learning_rate": 0.0001925377358490566,
373
+ "loss": 1.0231,
374
+ "step": 5800
375
+ },
376
+ {
377
+ "epoch": 46.83,
378
+ "learning_rate": 0.0001897075471698113,
379
+ "loss": 1.0248,
380
+ "step": 5900
381
+ },
382
+ {
383
+ "epoch": 47.62,
384
+ "learning_rate": 0.00018687735849056604,
385
+ "loss": 1.0093,
386
+ "step": 6000
387
+ },
388
+ {
389
+ "epoch": 47.62,
390
+ "eval_loss": 0.5827558040618896,
391
+ "eval_runtime": 112.3252,
392
+ "eval_samples_per_second": 16.737,
393
+ "eval_steps_per_second": 16.737,
394
+ "eval_wer": 0.5117424864509772,
395
+ "step": 6000
396
+ },
397
+ {
398
+ "epoch": 48.41,
399
+ "learning_rate": 0.00018404716981132075,
400
+ "loss": 1.0113,
401
+ "step": 6100
402
+ },
403
+ {
404
+ "epoch": 49.21,
405
+ "learning_rate": 0.00018121698113207548,
406
+ "loss": 1.0139,
407
+ "step": 6200
408
+ },
409
+ {
410
+ "epoch": 50.0,
411
+ "learning_rate": 0.00017838679245283019,
412
+ "loss": 1.0,
413
+ "step": 6300
414
+ },
415
+ {
416
+ "epoch": 50.79,
417
+ "learning_rate": 0.0001755566037735849,
418
+ "loss": 1.0081,
419
+ "step": 6400
420
+ },
421
+ {
422
+ "epoch": 51.59,
423
+ "learning_rate": 0.00017272641509433962,
424
+ "loss": 0.9896,
425
+ "step": 6500
426
+ },
427
+ {
428
+ "epoch": 52.38,
429
+ "learning_rate": 0.00016989622641509433,
430
+ "loss": 0.997,
431
+ "step": 6600
432
+ },
433
+ {
434
+ "epoch": 53.17,
435
+ "learning_rate": 0.00016709433962264152,
436
+ "loss": 0.9755,
437
+ "step": 6700
438
+ },
439
+ {
440
+ "epoch": 53.97,
441
+ "learning_rate": 0.00016426415094339622,
442
+ "loss": 0.9592,
443
+ "step": 6800
444
+ },
445
+ {
446
+ "epoch": 54.76,
447
+ "learning_rate": 0.00016143396226415093,
448
+ "loss": 0.9542,
449
+ "step": 6900
450
+ },
451
+ {
452
+ "epoch": 55.56,
453
+ "learning_rate": 0.00015860377358490566,
454
+ "loss": 0.9563,
455
+ "step": 7000
456
+ },
457
+ {
458
+ "epoch": 56.35,
459
+ "learning_rate": 0.00015577358490566036,
460
+ "loss": 0.9531,
461
+ "step": 7100
462
+ },
463
+ {
464
+ "epoch": 57.14,
465
+ "learning_rate": 0.0001529433962264151,
466
+ "loss": 0.9386,
467
+ "step": 7200
468
+ },
469
+ {
470
+ "epoch": 57.94,
471
+ "learning_rate": 0.0001501132075471698,
472
+ "loss": 0.9424,
473
+ "step": 7300
474
+ },
475
+ {
476
+ "epoch": 58.73,
477
+ "learning_rate": 0.0001472830188679245,
478
+ "loss": 0.9233,
479
+ "step": 7400
480
+ },
481
+ {
482
+ "epoch": 59.52,
483
+ "learning_rate": 0.00014445283018867924,
484
+ "loss": 0.912,
485
+ "step": 7500
486
+ },
487
+ {
488
+ "epoch": 60.32,
489
+ "learning_rate": 0.00014162264150943395,
490
+ "loss": 0.9165,
491
+ "step": 7600
492
+ },
493
+ {
494
+ "epoch": 61.11,
495
+ "learning_rate": 0.00013879245283018868,
496
+ "loss": 0.9143,
497
+ "step": 7700
498
+ },
499
+ {
500
+ "epoch": 61.9,
501
+ "learning_rate": 0.00013596226415094338,
502
+ "loss": 0.9175,
503
+ "step": 7800
504
+ },
505
+ {
506
+ "epoch": 62.7,
507
+ "learning_rate": 0.0001331320754716981,
508
+ "loss": 0.8966,
509
+ "step": 7900
510
+ },
511
+ {
512
+ "epoch": 63.49,
513
+ "learning_rate": 0.00013030188679245282,
514
+ "loss": 0.8888,
515
+ "step": 8000
516
+ },
517
+ {
518
+ "epoch": 63.49,
519
+ "eval_loss": 0.5754138827323914,
520
+ "eval_runtime": 112.1089,
521
+ "eval_samples_per_second": 16.769,
522
+ "eval_steps_per_second": 16.769,
523
+ "eval_wer": 0.4822357256254448,
524
+ "step": 8000
525
+ },
526
+ {
527
+ "epoch": 64.29,
528
+ "learning_rate": 0.00012747169811320753,
529
+ "loss": 0.8999,
530
+ "step": 8100
531
+ },
532
+ {
533
+ "epoch": 65.08,
534
+ "learning_rate": 0.00012464150943396226,
535
+ "loss": 0.8865,
536
+ "step": 8200
537
+ },
538
+ {
539
+ "epoch": 65.87,
540
+ "learning_rate": 0.00012181132075471698,
541
+ "loss": 0.8739,
542
+ "step": 8300
543
+ },
544
+ {
545
+ "epoch": 66.67,
546
+ "learning_rate": 0.00011898113207547169,
547
+ "loss": 0.8812,
548
+ "step": 8400
549
+ },
550
+ {
551
+ "epoch": 67.46,
552
+ "learning_rate": 0.0001161509433962264,
553
+ "loss": 0.8808,
554
+ "step": 8500
555
+ },
556
+ {
557
+ "epoch": 68.25,
558
+ "learning_rate": 0.00011332075471698112,
559
+ "loss": 0.8658,
560
+ "step": 8600
561
+ },
562
+ {
563
+ "epoch": 69.05,
564
+ "learning_rate": 0.00011049056603773584,
565
+ "loss": 0.8582,
566
+ "step": 8700
567
+ },
568
+ {
569
+ "epoch": 69.84,
570
+ "learning_rate": 0.00010766037735849056,
571
+ "loss": 0.8524,
572
+ "step": 8800
573
+ },
574
+ {
575
+ "epoch": 70.63,
576
+ "learning_rate": 0.00010483018867924528,
577
+ "loss": 0.8402,
578
+ "step": 8900
579
+ },
580
+ {
581
+ "epoch": 71.43,
582
+ "learning_rate": 0.000102,
583
+ "loss": 0.8417,
584
+ "step": 9000
585
+ },
586
+ {
587
+ "epoch": 72.22,
588
+ "learning_rate": 9.919811320754716e-05,
589
+ "loss": 0.8386,
590
+ "step": 9100
591
+ },
592
+ {
593
+ "epoch": 73.02,
594
+ "learning_rate": 9.636792452830188e-05,
595
+ "loss": 0.8319,
596
+ "step": 9200
597
+ },
598
+ {
599
+ "epoch": 73.81,
600
+ "learning_rate": 9.35377358490566e-05,
601
+ "loss": 0.8271,
602
+ "step": 9300
603
+ },
604
+ {
605
+ "epoch": 74.6,
606
+ "learning_rate": 9.070754716981132e-05,
607
+ "loss": 0.8182,
608
+ "step": 9400
609
+ },
610
+ {
611
+ "epoch": 75.4,
612
+ "learning_rate": 8.787735849056603e-05,
613
+ "loss": 0.8223,
614
+ "step": 9500
615
+ },
616
+ {
617
+ "epoch": 76.19,
618
+ "learning_rate": 8.504716981132075e-05,
619
+ "loss": 0.8074,
620
+ "step": 9600
621
+ },
622
+ {
623
+ "epoch": 76.98,
624
+ "learning_rate": 8.221698113207547e-05,
625
+ "loss": 0.7932,
626
+ "step": 9700
627
+ },
628
+ {
629
+ "epoch": 77.78,
630
+ "learning_rate": 7.938679245283019e-05,
631
+ "loss": 0.8135,
632
+ "step": 9800
633
+ },
634
+ {
635
+ "epoch": 78.57,
636
+ "learning_rate": 7.655660377358491e-05,
637
+ "loss": 0.7849,
638
+ "step": 9900
639
+ },
640
+ {
641
+ "epoch": 79.37,
642
+ "learning_rate": 7.372641509433962e-05,
643
+ "loss": 0.7985,
644
+ "step": 10000
645
+ },
646
+ {
647
+ "epoch": 79.37,
648
+ "eval_loss": 0.5987282395362854,
649
+ "eval_runtime": 112.2997,
650
+ "eval_samples_per_second": 16.741,
651
+ "eval_steps_per_second": 16.741,
652
+ "eval_wer": 0.4690425357201511,
653
+ "step": 10000
654
+ },
655
+ {
656
+ "epoch": 80.16,
657
+ "learning_rate": 7.089622641509434e-05,
658
+ "loss": 0.7807,
659
+ "step": 10100
660
+ },
661
+ {
662
+ "epoch": 80.95,
663
+ "learning_rate": 6.806603773584905e-05,
664
+ "loss": 0.7713,
665
+ "step": 10200
666
+ },
667
+ {
668
+ "epoch": 81.75,
669
+ "learning_rate": 6.523584905660377e-05,
670
+ "loss": 0.7628,
671
+ "step": 10300
672
+ },
673
+ {
674
+ "epoch": 82.54,
675
+ "learning_rate": 6.243396226415093e-05,
676
+ "loss": 0.7789,
677
+ "step": 10400
678
+ },
679
+ {
680
+ "epoch": 83.33,
681
+ "learning_rate": 5.960377358490565e-05,
682
+ "loss": 0.7727,
683
+ "step": 10500
684
+ },
685
+ {
686
+ "epoch": 84.13,
687
+ "learning_rate": 5.677358490566037e-05,
688
+ "loss": 0.7555,
689
+ "step": 10600
690
+ },
691
+ {
692
+ "epoch": 84.92,
693
+ "learning_rate": 5.394339622641509e-05,
694
+ "loss": 0.7549,
695
+ "step": 10700
696
+ },
697
+ {
698
+ "epoch": 85.71,
699
+ "learning_rate": 5.111320754716981e-05,
700
+ "loss": 0.7517,
701
+ "step": 10800
702
+ },
703
+ {
704
+ "epoch": 86.51,
705
+ "learning_rate": 4.828301886792453e-05,
706
+ "loss": 0.7329,
707
+ "step": 10900
708
+ },
709
+ {
710
+ "epoch": 87.3,
711
+ "learning_rate": 4.545283018867924e-05,
712
+ "loss": 0.7509,
713
+ "step": 11000
714
+ },
715
+ {
716
+ "epoch": 88.1,
717
+ "learning_rate": 4.262264150943396e-05,
718
+ "loss": 0.7255,
719
+ "step": 11100
720
+ },
721
+ {
722
+ "epoch": 88.89,
723
+ "learning_rate": 3.9820754716981125e-05,
724
+ "loss": 0.7431,
725
+ "step": 11200
726
+ },
727
+ {
728
+ "epoch": 89.68,
729
+ "learning_rate": 3.6990566037735845e-05,
730
+ "loss": 0.7341,
731
+ "step": 11300
732
+ },
733
+ {
734
+ "epoch": 90.48,
735
+ "learning_rate": 3.4160377358490564e-05,
736
+ "loss": 0.7026,
737
+ "step": 11400
738
+ },
739
+ {
740
+ "epoch": 91.27,
741
+ "learning_rate": 3.133018867924528e-05,
742
+ "loss": 0.7257,
743
+ "step": 11500
744
+ },
745
+ {
746
+ "epoch": 92.06,
747
+ "learning_rate": 2.8499999999999998e-05,
748
+ "loss": 0.7195,
749
+ "step": 11600
750
+ },
751
+ {
752
+ "epoch": 92.86,
753
+ "learning_rate": 2.5669811320754717e-05,
754
+ "loss": 0.7091,
755
+ "step": 11700
756
+ },
757
+ {
758
+ "epoch": 93.65,
759
+ "learning_rate": 2.2839622641509433e-05,
760
+ "loss": 0.7017,
761
+ "step": 11800
762
+ },
763
+ {
764
+ "epoch": 94.44,
765
+ "learning_rate": 2.000943396226415e-05,
766
+ "loss": 0.7071,
767
+ "step": 11900
768
+ },
769
+ {
770
+ "epoch": 95.24,
771
+ "learning_rate": 1.7179245283018867e-05,
772
+ "loss": 0.697,
773
+ "step": 12000
774
+ },
775
+ {
776
+ "epoch": 95.24,
777
+ "eval_loss": 0.6013949513435364,
778
+ "eval_runtime": 112.0467,
779
+ "eval_samples_per_second": 16.779,
780
+ "eval_steps_per_second": 16.779,
781
+ "eval_wer": 0.44709038156238023,
782
+ "step": 12000
783
+ },
784
+ {
785
+ "epoch": 96.03,
786
+ "learning_rate": 1.4349056603773583e-05,
787
+ "loss": 0.7027,
788
+ "step": 12100
789
+ },
790
+ {
791
+ "epoch": 96.83,
792
+ "learning_rate": 1.15188679245283e-05,
793
+ "loss": 0.7046,
794
+ "step": 12200
795
+ },
796
+ {
797
+ "epoch": 97.62,
798
+ "learning_rate": 8.688679245283018e-06,
799
+ "loss": 0.6978,
800
+ "step": 12300
801
+ },
802
+ {
803
+ "epoch": 98.41,
804
+ "learning_rate": 5.858490566037735e-06,
805
+ "loss": 0.7058,
806
+ "step": 12400
807
+ },
808
+ {
809
+ "epoch": 99.21,
810
+ "learning_rate": 3.0283018867924524e-06,
811
+ "loss": 0.6868,
812
+ "step": 12500
813
+ },
814
+ {
815
+ "epoch": 100.0,
816
+ "learning_rate": 1.9811320754716982e-07,
817
+ "loss": 0.6835,
818
+ "step": 12600
819
+ },
820
+ {
821
+ "epoch": 100.0,
822
+ "step": 12600,
823
+ "total_flos": 6.584079374045402e+19,
824
+ "train_loss": 1.186835220579117,
825
+ "train_runtime": 23714.6286,
826
+ "train_samples_per_second": 16.88,
827
+ "train_steps_per_second": 0.531
828
+ }
829
+ ],
830
+ "max_steps": 12600,
831
+ "num_train_epochs": 100,
832
+ "total_flos": 6.584079374045402e+19,
833
+ "trial_name": null,
834
+ "trial_params": null
835
+ }