freddyw commited on
Commit
0b3cdb9
1 Parent(s): 0ad09a6

End of training

Browse files
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 40.0,
3
+ "eval_loss": 0.35698655247688293,
4
+ "eval_runtime": 52.1875,
5
+ "eval_samples": 825,
6
+ "eval_samples_per_second": 15.808,
7
+ "eval_steps_per_second": 0.996,
8
+ "eval_wer": 0.21414389713529355,
9
+ "train_loss": 0.896209309619406,
10
+ "train_runtime": 25745.3803,
11
+ "train_samples": 7354,
12
+ "train_samples_per_second": 11.426,
13
+ "train_steps_per_second": 0.357
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 40.0,
3
+ "eval_loss": 0.35698655247688293,
4
+ "eval_runtime": 52.1875,
5
+ "eval_samples": 825,
6
+ "eval_samples_per_second": 15.808,
7
+ "eval_steps_per_second": 0.996,
8
+ "eval_wer": 0.21414389713529355
9
+ }
runs/Feb04_15-34-38_dante/events.out.tfevents.1644011658.dante.2164111.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:164e3b812ea42a27ad31e226075ad58a7ca99b0bf6f551fc7b73803719621d0b
3
+ size 358
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 40.0,
3
+ "train_loss": 0.896209309619406,
4
+ "train_runtime": 25745.3803,
5
+ "train_samples": 7354,
6
+ "train_samples_per_second": 11.426,
7
+ "train_steps_per_second": 0.357
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,739 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 40.0,
5
+ "global_step": 9200,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.43,
12
+ "learning_rate": 3.675e-06,
13
+ "loss": 10.3748,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 0.87,
18
+ "learning_rate": 7.35e-06,
19
+ "loss": 9.8141,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 1.3,
24
+ "learning_rate": 1.1099999999999999e-05,
25
+ "loss": 5.6549,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 1.74,
30
+ "learning_rate": 1.485e-05,
31
+ "loss": 3.4055,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 2.17,
36
+ "learning_rate": 1.8599999999999998e-05,
37
+ "loss": 3.086,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 2.17,
42
+ "eval_loss": 3.0773141384124756,
43
+ "eval_runtime": 54.4742,
44
+ "eval_samples_per_second": 15.145,
45
+ "eval_steps_per_second": 0.955,
46
+ "eval_wer": 1.0,
47
+ "step": 500
48
+ },
49
+ {
50
+ "epoch": 2.61,
51
+ "learning_rate": 2.2349999999999998e-05,
52
+ "loss": 2.9844,
53
+ "step": 600
54
+ },
55
+ {
56
+ "epoch": 3.04,
57
+ "learning_rate": 2.6099999999999997e-05,
58
+ "loss": 2.957,
59
+ "step": 700
60
+ },
61
+ {
62
+ "epoch": 3.48,
63
+ "learning_rate": 2.985e-05,
64
+ "loss": 2.912,
65
+ "step": 800
66
+ },
67
+ {
68
+ "epoch": 3.91,
69
+ "learning_rate": 3.36e-05,
70
+ "loss": 2.8929,
71
+ "step": 900
72
+ },
73
+ {
74
+ "epoch": 4.35,
75
+ "learning_rate": 3.735e-05,
76
+ "loss": 2.8532,
77
+ "step": 1000
78
+ },
79
+ {
80
+ "epoch": 4.35,
81
+ "eval_loss": 2.839279890060425,
82
+ "eval_runtime": 54.2914,
83
+ "eval_samples_per_second": 15.196,
84
+ "eval_steps_per_second": 0.958,
85
+ "eval_wer": 1.0,
86
+ "step": 1000
87
+ },
88
+ {
89
+ "epoch": 4.78,
90
+ "learning_rate": 4.11e-05,
91
+ "loss": 2.595,
92
+ "step": 1100
93
+ },
94
+ {
95
+ "epoch": 5.22,
96
+ "learning_rate": 4.484999999999999e-05,
97
+ "loss": 1.9229,
98
+ "step": 1200
99
+ },
100
+ {
101
+ "epoch": 5.65,
102
+ "learning_rate": 4.8599999999999995e-05,
103
+ "loss": 1.3521,
104
+ "step": 1300
105
+ },
106
+ {
107
+ "epoch": 6.09,
108
+ "learning_rate": 5.234999999999999e-05,
109
+ "loss": 1.0772,
110
+ "step": 1400
111
+ },
112
+ {
113
+ "epoch": 6.52,
114
+ "learning_rate": 5.6099999999999995e-05,
115
+ "loss": 0.9738,
116
+ "step": 1500
117
+ },
118
+ {
119
+ "epoch": 6.52,
120
+ "eval_loss": 0.728294849395752,
121
+ "eval_runtime": 54.3803,
122
+ "eval_samples_per_second": 15.171,
123
+ "eval_steps_per_second": 0.956,
124
+ "eval_wer": 0.4889855346506571,
125
+ "step": 1500
126
+ },
127
+ {
128
+ "epoch": 6.96,
129
+ "learning_rate": 5.985e-05,
130
+ "loss": 0.891,
131
+ "step": 1600
132
+ },
133
+ {
134
+ "epoch": 7.39,
135
+ "learning_rate": 6.359999999999999e-05,
136
+ "loss": 0.7848,
137
+ "step": 1700
138
+ },
139
+ {
140
+ "epoch": 7.83,
141
+ "learning_rate": 6.735e-05,
142
+ "loss": 0.7527,
143
+ "step": 1800
144
+ },
145
+ {
146
+ "epoch": 8.26,
147
+ "learning_rate": 7.11e-05,
148
+ "loss": 0.7279,
149
+ "step": 1900
150
+ },
151
+ {
152
+ "epoch": 8.7,
153
+ "learning_rate": 7.484999999999999e-05,
154
+ "loss": 0.6763,
155
+ "step": 2000
156
+ },
157
+ {
158
+ "epoch": 8.7,
159
+ "eval_loss": 0.5340386033058167,
160
+ "eval_runtime": 53.7943,
161
+ "eval_samples_per_second": 15.336,
162
+ "eval_steps_per_second": 0.967,
163
+ "eval_wer": 0.36617188238630993,
164
+ "step": 2000
165
+ },
166
+ {
167
+ "epoch": 9.13,
168
+ "learning_rate": 7.4e-05,
169
+ "loss": 0.6165,
170
+ "step": 2100
171
+ },
172
+ {
173
+ "epoch": 9.57,
174
+ "learning_rate": 7.295833333333332e-05,
175
+ "loss": 0.6021,
176
+ "step": 2200
177
+ },
178
+ {
179
+ "epoch": 10.0,
180
+ "learning_rate": 7.191666666666666e-05,
181
+ "loss": 0.5904,
182
+ "step": 2300
183
+ },
184
+ {
185
+ "epoch": 10.43,
186
+ "learning_rate": 7.087499999999999e-05,
187
+ "loss": 0.5469,
188
+ "step": 2400
189
+ },
190
+ {
191
+ "epoch": 10.87,
192
+ "learning_rate": 6.983333333333333e-05,
193
+ "loss": 0.5303,
194
+ "step": 2500
195
+ },
196
+ {
197
+ "epoch": 10.87,
198
+ "eval_loss": 0.45213818550109863,
199
+ "eval_runtime": 53.7095,
200
+ "eval_samples_per_second": 15.36,
201
+ "eval_steps_per_second": 0.968,
202
+ "eval_wer": 0.31398317103148343,
203
+ "step": 2500
204
+ },
205
+ {
206
+ "epoch": 11.3,
207
+ "learning_rate": 6.879166666666667e-05,
208
+ "loss": 0.5061,
209
+ "step": 2600
210
+ },
211
+ {
212
+ "epoch": 11.74,
213
+ "learning_rate": 6.775e-05,
214
+ "loss": 0.4928,
215
+ "step": 2700
216
+ },
217
+ {
218
+ "epoch": 12.17,
219
+ "learning_rate": 6.670833333333333e-05,
220
+ "loss": 0.4693,
221
+ "step": 2800
222
+ },
223
+ {
224
+ "epoch": 12.61,
225
+ "learning_rate": 6.566666666666666e-05,
226
+ "loss": 0.4733,
227
+ "step": 2900
228
+ },
229
+ {
230
+ "epoch": 13.04,
231
+ "learning_rate": 6.4625e-05,
232
+ "loss": 0.4765,
233
+ "step": 3000
234
+ },
235
+ {
236
+ "epoch": 13.04,
237
+ "eval_loss": 0.4181167483329773,
238
+ "eval_runtime": 53.0777,
239
+ "eval_samples_per_second": 15.543,
240
+ "eval_steps_per_second": 0.98,
241
+ "eval_wer": 0.2853361066464971,
242
+ "step": 3000
243
+ },
244
+ {
245
+ "epoch": 13.48,
246
+ "learning_rate": 6.358333333333332e-05,
247
+ "loss": 0.4396,
248
+ "step": 3100
249
+ },
250
+ {
251
+ "epoch": 13.91,
252
+ "learning_rate": 6.254166666666666e-05,
253
+ "loss": 0.4359,
254
+ "step": 3200
255
+ },
256
+ {
257
+ "epoch": 14.35,
258
+ "learning_rate": 6.149999999999999e-05,
259
+ "loss": 0.4173,
260
+ "step": 3300
261
+ },
262
+ {
263
+ "epoch": 14.78,
264
+ "learning_rate": 6.045833333333333e-05,
265
+ "loss": 0.408,
266
+ "step": 3400
267
+ },
268
+ {
269
+ "epoch": 15.22,
270
+ "learning_rate": 5.941666666666666e-05,
271
+ "loss": 0.4219,
272
+ "step": 3500
273
+ },
274
+ {
275
+ "epoch": 15.22,
276
+ "eval_loss": 0.41555172204971313,
277
+ "eval_runtime": 52.8273,
278
+ "eval_samples_per_second": 15.617,
279
+ "eval_steps_per_second": 0.984,
280
+ "eval_wer": 0.2933724118370048,
281
+ "step": 3500
282
+ },
283
+ {
284
+ "epoch": 15.65,
285
+ "learning_rate": 5.8374999999999996e-05,
286
+ "loss": 0.4022,
287
+ "step": 3600
288
+ },
289
+ {
290
+ "epoch": 16.09,
291
+ "learning_rate": 5.733333333333333e-05,
292
+ "loss": 0.3932,
293
+ "step": 3700
294
+ },
295
+ {
296
+ "epoch": 16.52,
297
+ "learning_rate": 5.629166666666666e-05,
298
+ "loss": 0.374,
299
+ "step": 3800
300
+ },
301
+ {
302
+ "epoch": 16.96,
303
+ "learning_rate": 5.5249999999999994e-05,
304
+ "loss": 0.3697,
305
+ "step": 3900
306
+ },
307
+ {
308
+ "epoch": 17.39,
309
+ "learning_rate": 5.420833333333333e-05,
310
+ "loss": 0.3564,
311
+ "step": 4000
312
+ },
313
+ {
314
+ "epoch": 17.39,
315
+ "eval_loss": 0.3925444483757019,
316
+ "eval_runtime": 54.1337,
317
+ "eval_samples_per_second": 15.24,
318
+ "eval_steps_per_second": 0.961,
319
+ "eval_wer": 0.2509218114777347,
320
+ "step": 4000
321
+ },
322
+ {
323
+ "epoch": 17.83,
324
+ "learning_rate": 5.316666666666666e-05,
325
+ "loss": 0.3784,
326
+ "step": 4100
327
+ },
328
+ {
329
+ "epoch": 18.26,
330
+ "learning_rate": 5.212499999999999e-05,
331
+ "loss": 0.3805,
332
+ "step": 4200
333
+ },
334
+ {
335
+ "epoch": 18.7,
336
+ "learning_rate": 5.1083333333333326e-05,
337
+ "loss": 0.3419,
338
+ "step": 4300
339
+ },
340
+ {
341
+ "epoch": 19.13,
342
+ "learning_rate": 5.0041666666666666e-05,
343
+ "loss": 0.3344,
344
+ "step": 4400
345
+ },
346
+ {
347
+ "epoch": 19.57,
348
+ "learning_rate": 4.899999999999999e-05,
349
+ "loss": 0.3282,
350
+ "step": 4500
351
+ },
352
+ {
353
+ "epoch": 19.57,
354
+ "eval_loss": 0.3824474811553955,
355
+ "eval_runtime": 54.4571,
356
+ "eval_samples_per_second": 15.15,
357
+ "eval_steps_per_second": 0.955,
358
+ "eval_wer": 0.24203460338470265,
359
+ "step": 4500
360
+ },
361
+ {
362
+ "epoch": 20.0,
363
+ "learning_rate": 4.795833333333333e-05,
364
+ "loss": 0.3447,
365
+ "step": 4600
366
+ },
367
+ {
368
+ "epoch": 20.43,
369
+ "learning_rate": 4.691666666666666e-05,
370
+ "loss": 0.3308,
371
+ "step": 4700
372
+ },
373
+ {
374
+ "epoch": 20.87,
375
+ "learning_rate": 4.5875e-05,
376
+ "loss": 0.3056,
377
+ "step": 4800
378
+ },
379
+ {
380
+ "epoch": 21.3,
381
+ "learning_rate": 4.483333333333333e-05,
382
+ "loss": 0.3069,
383
+ "step": 4900
384
+ },
385
+ {
386
+ "epoch": 21.74,
387
+ "learning_rate": 4.379166666666666e-05,
388
+ "loss": 0.3118,
389
+ "step": 5000
390
+ },
391
+ {
392
+ "epoch": 21.74,
393
+ "eval_loss": 0.36364665627479553,
394
+ "eval_runtime": 52.4627,
395
+ "eval_samples_per_second": 15.725,
396
+ "eval_steps_per_second": 0.991,
397
+ "eval_wer": 0.2354164696984022,
398
+ "step": 5000
399
+ },
400
+ {
401
+ "epoch": 22.17,
402
+ "learning_rate": 4.2749999999999996e-05,
403
+ "loss": 0.3108,
404
+ "step": 5100
405
+ },
406
+ {
407
+ "epoch": 22.61,
408
+ "learning_rate": 4.170833333333333e-05,
409
+ "loss": 0.3034,
410
+ "step": 5200
411
+ },
412
+ {
413
+ "epoch": 23.04,
414
+ "learning_rate": 4.066666666666667e-05,
415
+ "loss": 0.3132,
416
+ "step": 5300
417
+ },
418
+ {
419
+ "epoch": 23.48,
420
+ "learning_rate": 3.9624999999999994e-05,
421
+ "loss": 0.3024,
422
+ "step": 5400
423
+ },
424
+ {
425
+ "epoch": 23.91,
426
+ "learning_rate": 3.8583333333333334e-05,
427
+ "loss": 0.2919,
428
+ "step": 5500
429
+ },
430
+ {
431
+ "epoch": 23.91,
432
+ "eval_loss": 0.3614845275878906,
433
+ "eval_runtime": 52.502,
434
+ "eval_samples_per_second": 15.714,
435
+ "eval_steps_per_second": 0.99,
436
+ "eval_wer": 0.22813652264347167,
437
+ "step": 5500
438
+ },
439
+ {
440
+ "epoch": 24.35,
441
+ "learning_rate": 3.754166666666666e-05,
442
+ "loss": 0.2868,
443
+ "step": 5600
444
+ },
445
+ {
446
+ "epoch": 24.78,
447
+ "learning_rate": 3.65e-05,
448
+ "loss": 0.2929,
449
+ "step": 5700
450
+ },
451
+ {
452
+ "epoch": 25.22,
453
+ "learning_rate": 3.545833333333333e-05,
454
+ "loss": 0.2956,
455
+ "step": 5800
456
+ },
457
+ {
458
+ "epoch": 25.65,
459
+ "learning_rate": 3.4416666666666665e-05,
460
+ "loss": 0.2818,
461
+ "step": 5900
462
+ },
463
+ {
464
+ "epoch": 26.09,
465
+ "learning_rate": 3.3375e-05,
466
+ "loss": 0.2961,
467
+ "step": 6000
468
+ },
469
+ {
470
+ "epoch": 26.09,
471
+ "eval_loss": 0.35476088523864746,
472
+ "eval_runtime": 52.7192,
473
+ "eval_samples_per_second": 15.649,
474
+ "eval_steps_per_second": 0.986,
475
+ "eval_wer": 0.2254892691689515,
476
+ "step": 6000
477
+ },
478
+ {
479
+ "epoch": 26.52,
480
+ "learning_rate": 3.233333333333333e-05,
481
+ "loss": 0.2816,
482
+ "step": 6100
483
+ },
484
+ {
485
+ "epoch": 26.96,
486
+ "learning_rate": 3.1291666666666664e-05,
487
+ "loss": 0.2788,
488
+ "step": 6200
489
+ },
490
+ {
491
+ "epoch": 27.39,
492
+ "learning_rate": 3.0249999999999997e-05,
493
+ "loss": 0.272,
494
+ "step": 6300
495
+ },
496
+ {
497
+ "epoch": 27.83,
498
+ "learning_rate": 2.920833333333333e-05,
499
+ "loss": 0.2756,
500
+ "step": 6400
501
+ },
502
+ {
503
+ "epoch": 28.26,
504
+ "learning_rate": 2.8166666666666662e-05,
505
+ "loss": 0.284,
506
+ "step": 6500
507
+ },
508
+ {
509
+ "epoch": 28.26,
510
+ "eval_loss": 0.3526020646095276,
511
+ "eval_runtime": 52.8188,
512
+ "eval_samples_per_second": 15.619,
513
+ "eval_steps_per_second": 0.984,
514
+ "eval_wer": 0.2208565755885412,
515
+ "step": 6500
516
+ },
517
+ {
518
+ "epoch": 28.7,
519
+ "learning_rate": 2.7125e-05,
520
+ "loss": 0.2721,
521
+ "step": 6600
522
+ },
523
+ {
524
+ "epoch": 29.13,
525
+ "learning_rate": 2.608333333333333e-05,
526
+ "loss": 0.2601,
527
+ "step": 6700
528
+ },
529
+ {
530
+ "epoch": 29.57,
531
+ "learning_rate": 2.5041666666666664e-05,
532
+ "loss": 0.2462,
533
+ "step": 6800
534
+ },
535
+ {
536
+ "epoch": 30.0,
537
+ "learning_rate": 2.3999999999999997e-05,
538
+ "loss": 0.271,
539
+ "step": 6900
540
+ },
541
+ {
542
+ "epoch": 30.43,
543
+ "learning_rate": 2.295833333333333e-05,
544
+ "loss": 0.2566,
545
+ "step": 7000
546
+ },
547
+ {
548
+ "epoch": 30.43,
549
+ "eval_loss": 0.35256850719451904,
550
+ "eval_runtime": 52.6434,
551
+ "eval_samples_per_second": 15.671,
552
+ "eval_steps_per_second": 0.988,
553
+ "eval_wer": 0.22047839652075257,
554
+ "step": 7000
555
+ },
556
+ {
557
+ "epoch": 30.87,
558
+ "learning_rate": 2.1916666666666663e-05,
559
+ "loss": 0.2612,
560
+ "step": 7100
561
+ },
562
+ {
563
+ "epoch": 31.3,
564
+ "learning_rate": 2.0874999999999996e-05,
565
+ "loss": 0.2524,
566
+ "step": 7200
567
+ },
568
+ {
569
+ "epoch": 31.74,
570
+ "learning_rate": 1.983333333333333e-05,
571
+ "loss": 0.254,
572
+ "step": 7300
573
+ },
574
+ {
575
+ "epoch": 32.17,
576
+ "learning_rate": 1.879166666666666e-05,
577
+ "loss": 0.2446,
578
+ "step": 7400
579
+ },
580
+ {
581
+ "epoch": 32.61,
582
+ "learning_rate": 1.7749999999999998e-05,
583
+ "loss": 0.2422,
584
+ "step": 7500
585
+ },
586
+ {
587
+ "epoch": 32.61,
588
+ "eval_loss": 0.35690072178840637,
589
+ "eval_runtime": 53.8708,
590
+ "eval_samples_per_second": 15.314,
591
+ "eval_steps_per_second": 0.965,
592
+ "eval_wer": 0.2172638744445495,
593
+ "step": 7500
594
+ },
595
+ {
596
+ "epoch": 33.04,
597
+ "learning_rate": 1.670833333333333e-05,
598
+ "loss": 0.2549,
599
+ "step": 7600
600
+ },
601
+ {
602
+ "epoch": 33.48,
603
+ "learning_rate": 1.5666666666666667e-05,
604
+ "loss": 0.2547,
605
+ "step": 7700
606
+ },
607
+ {
608
+ "epoch": 33.91,
609
+ "learning_rate": 1.4625e-05,
610
+ "loss": 0.2553,
611
+ "step": 7800
612
+ },
613
+ {
614
+ "epoch": 34.35,
615
+ "learning_rate": 1.3583333333333333e-05,
616
+ "loss": 0.2445,
617
+ "step": 7900
618
+ },
619
+ {
620
+ "epoch": 34.78,
621
+ "learning_rate": 1.2541666666666665e-05,
622
+ "loss": 0.2472,
623
+ "step": 8000
624
+ },
625
+ {
626
+ "epoch": 34.78,
627
+ "eval_loss": 0.35918259620666504,
628
+ "eval_runtime": 52.3221,
629
+ "eval_samples_per_second": 15.768,
630
+ "eval_steps_per_second": 0.994,
631
+ "eval_wer": 0.21660206107591945,
632
+ "step": 8000
633
+ },
634
+ {
635
+ "epoch": 35.22,
636
+ "learning_rate": 1.1499999999999998e-05,
637
+ "loss": 0.2608,
638
+ "step": 8100
639
+ },
640
+ {
641
+ "epoch": 35.65,
642
+ "learning_rate": 1.0458333333333333e-05,
643
+ "loss": 0.2482,
644
+ "step": 8200
645
+ },
646
+ {
647
+ "epoch": 36.09,
648
+ "learning_rate": 9.416666666666666e-06,
649
+ "loss": 0.2464,
650
+ "step": 8300
651
+ },
652
+ {
653
+ "epoch": 36.52,
654
+ "learning_rate": 8.374999999999999e-06,
655
+ "loss": 0.2347,
656
+ "step": 8400
657
+ },
658
+ {
659
+ "epoch": 36.96,
660
+ "learning_rate": 7.333333333333333e-06,
661
+ "loss": 0.2337,
662
+ "step": 8500
663
+ },
664
+ {
665
+ "epoch": 36.96,
666
+ "eval_loss": 0.36245495080947876,
667
+ "eval_runtime": 54.1316,
668
+ "eval_samples_per_second": 15.241,
669
+ "eval_steps_per_second": 0.961,
670
+ "eval_wer": 0.21716932967760236,
671
+ "step": 8500
672
+ },
673
+ {
674
+ "epoch": 37.39,
675
+ "learning_rate": 6.291666666666666e-06,
676
+ "loss": 0.2392,
677
+ "step": 8600
678
+ },
679
+ {
680
+ "epoch": 37.83,
681
+ "learning_rate": 5.25e-06,
682
+ "loss": 0.2347,
683
+ "step": 8700
684
+ },
685
+ {
686
+ "epoch": 38.26,
687
+ "learning_rate": 4.208333333333333e-06,
688
+ "loss": 0.2455,
689
+ "step": 8800
690
+ },
691
+ {
692
+ "epoch": 38.7,
693
+ "learning_rate": 3.1666666666666663e-06,
694
+ "loss": 0.243,
695
+ "step": 8900
696
+ },
697
+ {
698
+ "epoch": 39.13,
699
+ "learning_rate": 2.1249999999999996e-06,
700
+ "loss": 0.2315,
701
+ "step": 9000
702
+ },
703
+ {
704
+ "epoch": 39.13,
705
+ "eval_loss": 0.3579612374305725,
706
+ "eval_runtime": 52.3246,
707
+ "eval_samples_per_second": 15.767,
708
+ "eval_steps_per_second": 0.994,
709
+ "eval_wer": 0.21546752387255366,
710
+ "step": 9000
711
+ },
712
+ {
713
+ "epoch": 39.57,
714
+ "learning_rate": 1.0833333333333333e-06,
715
+ "loss": 0.2489,
716
+ "step": 9100
717
+ },
718
+ {
719
+ "epoch": 40.0,
720
+ "learning_rate": 4.166666666666666e-08,
721
+ "loss": 0.227,
722
+ "step": 9200
723
+ },
724
+ {
725
+ "epoch": 40.0,
726
+ "step": 9200,
727
+ "total_flos": 6.4078116974997864e+19,
728
+ "train_loss": 0.896209309619406,
729
+ "train_runtime": 25745.3803,
730
+ "train_samples_per_second": 11.426,
731
+ "train_steps_per_second": 0.357
732
+ }
733
+ ],
734
+ "max_steps": 9200,
735
+ "num_train_epochs": 40,
736
+ "total_flos": 6.4078116974997864e+19,
737
+ "trial_name": null,
738
+ "trial_params": null
739
+ }