sanchit-gandhi HF staff commited on
Commit
09960c9
1 Parent(s): 16b0fc6

End of training

Browse files
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 14.34,
3
+ "eval_loss": 0.2998046875,
4
+ "eval_runtime": 650.2511,
5
+ "eval_samples_per_second": 3.402,
6
+ "eval_steps_per_second": 0.214,
7
+ "eval_wer": 8.957818965817019,
8
+ "train_loss": 0.08977190920710564,
9
+ "train_runtime": 17594.9662,
10
+ "train_samples_per_second": 3.637,
11
+ "train_steps_per_second": 0.227
12
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 14.34,
3
+ "eval_loss": 0.2998046875,
4
+ "eval_runtime": 650.2511,
5
+ "eval_samples_per_second": 3.402,
6
+ "eval_steps_per_second": 0.214,
7
+ "eval_wer": 8.957818965817019
8
+ }
runs/Jun28_13-55-29_vorace/events.out.tfevents.1687979353.vorace.1465295.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68f54a7d088fb42b78a3d834e5cf2f45c7fa95e0fb54cf70dcb8a84f498d9a8e
3
+ size 406
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 14.34,
3
+ "train_loss": 0.08977190920710564,
4
+ "train_runtime": 17594.9662,
5
+ "train_samples_per_second": 3.637,
6
+ "train_steps_per_second": 0.227
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1021 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 14.336917562724015,
5
+ "global_step": 4000,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.09,
12
+ "learning_rate": 4.558957377820063e-06,
13
+ "loss": 4.7495,
14
+ "step": 25
15
+ },
16
+ {
17
+ "epoch": 0.18,
18
+ "learning_rate": 6.014328754265799e-06,
19
+ "loss": 2.6193,
20
+ "step": 50
21
+ },
22
+ {
23
+ "epoch": 0.27,
24
+ "learning_rate": 6.765821034569313e-06,
25
+ "loss": 1.6513,
26
+ "step": 75
27
+ },
28
+ {
29
+ "epoch": 0.36,
30
+ "learning_rate": 7.276063921387198e-06,
31
+ "loss": 0.9913,
32
+ "step": 100
33
+ },
34
+ {
35
+ "epoch": 0.45,
36
+ "learning_rate": 7.662870867121632e-06,
37
+ "loss": 0.4604,
38
+ "step": 125
39
+ },
40
+ {
41
+ "epoch": 0.54,
42
+ "learning_rate": 7.974480416326626e-06,
43
+ "loss": 0.2504,
44
+ "step": 150
45
+ },
46
+ {
47
+ "epoch": 0.63,
48
+ "learning_rate": 8.235424875329062e-06,
49
+ "loss": 0.1962,
50
+ "step": 175
51
+ },
52
+ {
53
+ "epoch": 0.72,
54
+ "learning_rate": 8.459898498446897e-06,
55
+ "loss": 0.1723,
56
+ "step": 200
57
+ },
58
+ {
59
+ "epoch": 0.81,
60
+ "learning_rate": 8.656856986535236e-06,
61
+ "loss": 0.1621,
62
+ "step": 225
63
+ },
64
+ {
65
+ "epoch": 0.9,
66
+ "learning_rate": 8.832315150411909e-06,
67
+ "loss": 0.1584,
68
+ "step": 250
69
+ },
70
+ {
71
+ "epoch": 0.99,
72
+ "learning_rate": 8.990508443836999e-06,
73
+ "loss": 0.1541,
74
+ "step": 275
75
+ },
76
+ {
77
+ "epoch": 1.08,
78
+ "learning_rate": 9.13453223817852e-06,
79
+ "loss": 0.1217,
80
+ "step": 300
81
+ },
82
+ {
83
+ "epoch": 1.16,
84
+ "learning_rate": 9.266717518904195e-06,
85
+ "loss": 0.1027,
86
+ "step": 325
87
+ },
88
+ {
89
+ "epoch": 1.25,
90
+ "learning_rate": 9.388863536775534e-06,
91
+ "loss": 0.1145,
92
+ "step": 350
93
+ },
94
+ {
95
+ "epoch": 1.34,
96
+ "learning_rate": 9.502388170790473e-06,
97
+ "loss": 0.0976,
98
+ "step": 375
99
+ },
100
+ {
101
+ "epoch": 1.43,
102
+ "learning_rate": 9.608428633346146e-06,
103
+ "loss": 0.1059,
104
+ "step": 400
105
+ },
106
+ {
107
+ "epoch": 1.52,
108
+ "learning_rate": 9.707910983687811e-06,
109
+ "loss": 0.1032,
110
+ "step": 425
111
+ },
112
+ {
113
+ "epoch": 1.61,
114
+ "learning_rate": 9.801599369756241e-06,
115
+ "loss": 0.1071,
116
+ "step": 450
117
+ },
118
+ {
119
+ "epoch": 1.7,
120
+ "learning_rate": 9.890131703122795e-06,
121
+ "loss": 0.0938,
122
+ "step": 475
123
+ },
124
+ {
125
+ "epoch": 1.79,
126
+ "learning_rate": 9.974046019194712e-06,
127
+ "loss": 0.1066,
128
+ "step": 500
129
+ },
130
+ {
131
+ "epoch": 1.88,
132
+ "learning_rate": 9.954285714285715e-06,
133
+ "loss": 0.0994,
134
+ "step": 525
135
+ },
136
+ {
137
+ "epoch": 1.97,
138
+ "learning_rate": 9.882857142857144e-06,
139
+ "loss": 0.0884,
140
+ "step": 550
141
+ },
142
+ {
143
+ "epoch": 2.06,
144
+ "learning_rate": 9.811428571428571e-06,
145
+ "loss": 0.0736,
146
+ "step": 575
147
+ },
148
+ {
149
+ "epoch": 2.15,
150
+ "learning_rate": 9.74e-06,
151
+ "loss": 0.0493,
152
+ "step": 600
153
+ },
154
+ {
155
+ "epoch": 2.24,
156
+ "learning_rate": 9.668571428571429e-06,
157
+ "loss": 0.0636,
158
+ "step": 625
159
+ },
160
+ {
161
+ "epoch": 2.33,
162
+ "learning_rate": 9.597142857142858e-06,
163
+ "loss": 0.0534,
164
+ "step": 650
165
+ },
166
+ {
167
+ "epoch": 2.42,
168
+ "learning_rate": 9.525714285714287e-06,
169
+ "loss": 0.0651,
170
+ "step": 675
171
+ },
172
+ {
173
+ "epoch": 2.51,
174
+ "learning_rate": 9.454285714285714e-06,
175
+ "loss": 0.052,
176
+ "step": 700
177
+ },
178
+ {
179
+ "epoch": 2.6,
180
+ "learning_rate": 9.382857142857143e-06,
181
+ "loss": 0.0632,
182
+ "step": 725
183
+ },
184
+ {
185
+ "epoch": 2.69,
186
+ "learning_rate": 9.311428571428572e-06,
187
+ "loss": 0.058,
188
+ "step": 750
189
+ },
190
+ {
191
+ "epoch": 2.78,
192
+ "learning_rate": 9.240000000000001e-06,
193
+ "loss": 0.0638,
194
+ "step": 775
195
+ },
196
+ {
197
+ "epoch": 2.87,
198
+ "learning_rate": 9.168571428571429e-06,
199
+ "loss": 0.0553,
200
+ "step": 800
201
+ },
202
+ {
203
+ "epoch": 2.96,
204
+ "learning_rate": 9.097142857142858e-06,
205
+ "loss": 0.0561,
206
+ "step": 825
207
+ },
208
+ {
209
+ "epoch": 3.05,
210
+ "learning_rate": 9.025714285714287e-06,
211
+ "loss": 0.0483,
212
+ "step": 850
213
+ },
214
+ {
215
+ "epoch": 3.14,
216
+ "learning_rate": 8.954285714285716e-06,
217
+ "loss": 0.0293,
218
+ "step": 875
219
+ },
220
+ {
221
+ "epoch": 3.23,
222
+ "learning_rate": 8.882857142857143e-06,
223
+ "loss": 0.0385,
224
+ "step": 900
225
+ },
226
+ {
227
+ "epoch": 3.32,
228
+ "learning_rate": 8.811428571428572e-06,
229
+ "loss": 0.0348,
230
+ "step": 925
231
+ },
232
+ {
233
+ "epoch": 3.41,
234
+ "learning_rate": 8.740000000000001e-06,
235
+ "loss": 0.0406,
236
+ "step": 950
237
+ },
238
+ {
239
+ "epoch": 3.49,
240
+ "learning_rate": 8.66857142857143e-06,
241
+ "loss": 0.0329,
242
+ "step": 975
243
+ },
244
+ {
245
+ "epoch": 3.58,
246
+ "learning_rate": 8.597142857142859e-06,
247
+ "loss": 0.0349,
248
+ "step": 1000
249
+ },
250
+ {
251
+ "epoch": 3.58,
252
+ "eval_loss": 0.1622314453125,
253
+ "eval_runtime": 624.6921,
254
+ "eval_samples_per_second": 3.541,
255
+ "eval_steps_per_second": 0.223,
256
+ "eval_wer": 9.943665890044164,
257
+ "step": 1000
258
+ },
259
+ {
260
+ "epoch": 3.67,
261
+ "learning_rate": 8.525714285714286e-06,
262
+ "loss": 0.036,
263
+ "step": 1025
264
+ },
265
+ {
266
+ "epoch": 3.76,
267
+ "learning_rate": 8.454285714285715e-06,
268
+ "loss": 0.0446,
269
+ "step": 1050
270
+ },
271
+ {
272
+ "epoch": 3.85,
273
+ "learning_rate": 8.382857142857144e-06,
274
+ "loss": 0.0332,
275
+ "step": 1075
276
+ },
277
+ {
278
+ "epoch": 3.94,
279
+ "learning_rate": 8.311428571428572e-06,
280
+ "loss": 0.0336,
281
+ "step": 1100
282
+ },
283
+ {
284
+ "epoch": 4.03,
285
+ "learning_rate": 8.24e-06,
286
+ "loss": 0.0308,
287
+ "step": 1125
288
+ },
289
+ {
290
+ "epoch": 4.12,
291
+ "learning_rate": 8.16857142857143e-06,
292
+ "loss": 0.0185,
293
+ "step": 1150
294
+ },
295
+ {
296
+ "epoch": 4.21,
297
+ "learning_rate": 8.097142857142857e-06,
298
+ "loss": 0.0243,
299
+ "step": 1175
300
+ },
301
+ {
302
+ "epoch": 4.3,
303
+ "learning_rate": 8.025714285714286e-06,
304
+ "loss": 0.0195,
305
+ "step": 1200
306
+ },
307
+ {
308
+ "epoch": 4.39,
309
+ "learning_rate": 7.954285714285715e-06,
310
+ "loss": 0.0253,
311
+ "step": 1225
312
+ },
313
+ {
314
+ "epoch": 4.48,
315
+ "learning_rate": 7.882857142857142e-06,
316
+ "loss": 0.019,
317
+ "step": 1250
318
+ },
319
+ {
320
+ "epoch": 4.57,
321
+ "learning_rate": 7.811428571428571e-06,
322
+ "loss": 0.0248,
323
+ "step": 1275
324
+ },
325
+ {
326
+ "epoch": 4.66,
327
+ "learning_rate": 7.74e-06,
328
+ "loss": 0.0216,
329
+ "step": 1300
330
+ },
331
+ {
332
+ "epoch": 4.75,
333
+ "learning_rate": 7.66857142857143e-06,
334
+ "loss": 0.0236,
335
+ "step": 1325
336
+ },
337
+ {
338
+ "epoch": 4.84,
339
+ "learning_rate": 7.597142857142858e-06,
340
+ "loss": 0.0211,
341
+ "step": 1350
342
+ },
343
+ {
344
+ "epoch": 4.93,
345
+ "learning_rate": 7.525714285714286e-06,
346
+ "loss": 0.0212,
347
+ "step": 1375
348
+ },
349
+ {
350
+ "epoch": 5.02,
351
+ "learning_rate": 7.454285714285715e-06,
352
+ "loss": 0.0188,
353
+ "step": 1400
354
+ },
355
+ {
356
+ "epoch": 5.11,
357
+ "learning_rate": 7.382857142857144e-06,
358
+ "loss": 0.0116,
359
+ "step": 1425
360
+ },
361
+ {
362
+ "epoch": 5.2,
363
+ "learning_rate": 7.311428571428571e-06,
364
+ "loss": 0.0143,
365
+ "step": 1450
366
+ },
367
+ {
368
+ "epoch": 5.29,
369
+ "learning_rate": 7.24e-06,
370
+ "loss": 0.0137,
371
+ "step": 1475
372
+ },
373
+ {
374
+ "epoch": 5.38,
375
+ "learning_rate": 7.168571428571429e-06,
376
+ "loss": 0.013,
377
+ "step": 1500
378
+ },
379
+ {
380
+ "epoch": 5.47,
381
+ "learning_rate": 7.097142857142858e-06,
382
+ "loss": 0.0121,
383
+ "step": 1525
384
+ },
385
+ {
386
+ "epoch": 5.56,
387
+ "learning_rate": 7.025714285714287e-06,
388
+ "loss": 0.0123,
389
+ "step": 1550
390
+ },
391
+ {
392
+ "epoch": 5.65,
393
+ "learning_rate": 6.9542857142857145e-06,
394
+ "loss": 0.0122,
395
+ "step": 1575
396
+ },
397
+ {
398
+ "epoch": 5.73,
399
+ "learning_rate": 6.8828571428571435e-06,
400
+ "loss": 0.0123,
401
+ "step": 1600
402
+ },
403
+ {
404
+ "epoch": 5.82,
405
+ "learning_rate": 6.8114285714285725e-06,
406
+ "loss": 0.0125,
407
+ "step": 1625
408
+ },
409
+ {
410
+ "epoch": 5.91,
411
+ "learning_rate": 6.740000000000001e-06,
412
+ "loss": 0.0146,
413
+ "step": 1650
414
+ },
415
+ {
416
+ "epoch": 6.0,
417
+ "learning_rate": 6.668571428571429e-06,
418
+ "loss": 0.0121,
419
+ "step": 1675
420
+ },
421
+ {
422
+ "epoch": 6.09,
423
+ "learning_rate": 6.597142857142858e-06,
424
+ "loss": 0.0075,
425
+ "step": 1700
426
+ },
427
+ {
428
+ "epoch": 6.18,
429
+ "learning_rate": 6.525714285714286e-06,
430
+ "loss": 0.0086,
431
+ "step": 1725
432
+ },
433
+ {
434
+ "epoch": 6.27,
435
+ "learning_rate": 6.454285714285715e-06,
436
+ "loss": 0.0099,
437
+ "step": 1750
438
+ },
439
+ {
440
+ "epoch": 6.36,
441
+ "learning_rate": 6.382857142857143e-06,
442
+ "loss": 0.0058,
443
+ "step": 1775
444
+ },
445
+ {
446
+ "epoch": 6.45,
447
+ "learning_rate": 6.311428571428571e-06,
448
+ "loss": 0.0093,
449
+ "step": 1800
450
+ },
451
+ {
452
+ "epoch": 6.54,
453
+ "learning_rate": 6.24e-06,
454
+ "loss": 0.0063,
455
+ "step": 1825
456
+ },
457
+ {
458
+ "epoch": 6.63,
459
+ "learning_rate": 6.168571428571429e-06,
460
+ "loss": 0.009,
461
+ "step": 1850
462
+ },
463
+ {
464
+ "epoch": 6.72,
465
+ "learning_rate": 6.097142857142858e-06,
466
+ "loss": 0.0051,
467
+ "step": 1875
468
+ },
469
+ {
470
+ "epoch": 6.81,
471
+ "learning_rate": 6.025714285714286e-06,
472
+ "loss": 0.0079,
473
+ "step": 1900
474
+ },
475
+ {
476
+ "epoch": 6.9,
477
+ "learning_rate": 5.954285714285715e-06,
478
+ "loss": 0.0061,
479
+ "step": 1925
480
+ },
481
+ {
482
+ "epoch": 6.99,
483
+ "learning_rate": 5.882857142857144e-06,
484
+ "loss": 0.0087,
485
+ "step": 1950
486
+ },
487
+ {
488
+ "epoch": 7.08,
489
+ "learning_rate": 5.811428571428572e-06,
490
+ "loss": 0.0076,
491
+ "step": 1975
492
+ },
493
+ {
494
+ "epoch": 7.17,
495
+ "learning_rate": 5.74e-06,
496
+ "loss": 0.0046,
497
+ "step": 2000
498
+ },
499
+ {
500
+ "epoch": 7.17,
501
+ "eval_loss": 0.228759765625,
502
+ "eval_runtime": 624.7011,
503
+ "eval_samples_per_second": 3.541,
504
+ "eval_steps_per_second": 0.223,
505
+ "eval_wer": 9.50898911569357,
506
+ "step": 2000
507
+ },
508
+ {
509
+ "epoch": 7.26,
510
+ "learning_rate": 5.668571428571429e-06,
511
+ "loss": 0.0061,
512
+ "step": 2025
513
+ },
514
+ {
515
+ "epoch": 7.35,
516
+ "learning_rate": 5.597142857142857e-06,
517
+ "loss": 0.0018,
518
+ "step": 2050
519
+ },
520
+ {
521
+ "epoch": 7.44,
522
+ "learning_rate": 5.525714285714286e-06,
523
+ "loss": 0.0049,
524
+ "step": 2075
525
+ },
526
+ {
527
+ "epoch": 7.53,
528
+ "learning_rate": 5.4542857142857145e-06,
529
+ "loss": 0.0032,
530
+ "step": 2100
531
+ },
532
+ {
533
+ "epoch": 7.62,
534
+ "learning_rate": 5.382857142857143e-06,
535
+ "loss": 0.0064,
536
+ "step": 2125
537
+ },
538
+ {
539
+ "epoch": 7.71,
540
+ "learning_rate": 5.311428571428572e-06,
541
+ "loss": 0.0041,
542
+ "step": 2150
543
+ },
544
+ {
545
+ "epoch": 7.8,
546
+ "learning_rate": 5.240000000000001e-06,
547
+ "loss": 0.005,
548
+ "step": 2175
549
+ },
550
+ {
551
+ "epoch": 7.89,
552
+ "learning_rate": 5.16857142857143e-06,
553
+ "loss": 0.0027,
554
+ "step": 2200
555
+ },
556
+ {
557
+ "epoch": 7.97,
558
+ "learning_rate": 5.097142857142857e-06,
559
+ "loss": 0.0043,
560
+ "step": 2225
561
+ },
562
+ {
563
+ "epoch": 8.06,
564
+ "learning_rate": 5.025714285714286e-06,
565
+ "loss": 0.0026,
566
+ "step": 2250
567
+ },
568
+ {
569
+ "epoch": 8.15,
570
+ "learning_rate": 4.954285714285715e-06,
571
+ "loss": 0.0019,
572
+ "step": 2275
573
+ },
574
+ {
575
+ "epoch": 8.24,
576
+ "learning_rate": 4.882857142857143e-06,
577
+ "loss": 0.0035,
578
+ "step": 2300
579
+ },
580
+ {
581
+ "epoch": 8.33,
582
+ "learning_rate": 4.811428571428572e-06,
583
+ "loss": 0.0017,
584
+ "step": 2325
585
+ },
586
+ {
587
+ "epoch": 8.42,
588
+ "learning_rate": 4.74e-06,
589
+ "loss": 0.003,
590
+ "step": 2350
591
+ },
592
+ {
593
+ "epoch": 8.51,
594
+ "learning_rate": 4.668571428571429e-06,
595
+ "loss": 0.0015,
596
+ "step": 2375
597
+ },
598
+ {
599
+ "epoch": 8.6,
600
+ "learning_rate": 4.5971428571428575e-06,
601
+ "loss": 0.0032,
602
+ "step": 2400
603
+ },
604
+ {
605
+ "epoch": 8.69,
606
+ "learning_rate": 4.525714285714286e-06,
607
+ "loss": 0.0017,
608
+ "step": 2425
609
+ },
610
+ {
611
+ "epoch": 8.78,
612
+ "learning_rate": 4.454285714285715e-06,
613
+ "loss": 0.0021,
614
+ "step": 2450
615
+ },
616
+ {
617
+ "epoch": 8.87,
618
+ "learning_rate": 4.382857142857143e-06,
619
+ "loss": 0.0025,
620
+ "step": 2475
621
+ },
622
+ {
623
+ "epoch": 8.96,
624
+ "learning_rate": 4.311428571428572e-06,
625
+ "loss": 0.0035,
626
+ "step": 2500
627
+ },
628
+ {
629
+ "epoch": 9.05,
630
+ "learning_rate": 4.24e-06,
631
+ "loss": 0.0036,
632
+ "step": 2525
633
+ },
634
+ {
635
+ "epoch": 9.14,
636
+ "learning_rate": 4.168571428571429e-06,
637
+ "loss": 0.0009,
638
+ "step": 2550
639
+ },
640
+ {
641
+ "epoch": 9.23,
642
+ "learning_rate": 4.097142857142857e-06,
643
+ "loss": 0.0009,
644
+ "step": 2575
645
+ },
646
+ {
647
+ "epoch": 9.32,
648
+ "learning_rate": 4.025714285714286e-06,
649
+ "loss": 0.0014,
650
+ "step": 2600
651
+ },
652
+ {
653
+ "epoch": 9.41,
654
+ "learning_rate": 3.954285714285714e-06,
655
+ "loss": 0.0017,
656
+ "step": 2625
657
+ },
658
+ {
659
+ "epoch": 9.5,
660
+ "learning_rate": 3.8828571428571434e-06,
661
+ "loss": 0.0006,
662
+ "step": 2650
663
+ },
664
+ {
665
+ "epoch": 9.59,
666
+ "learning_rate": 3.8114285714285716e-06,
667
+ "loss": 0.0011,
668
+ "step": 2675
669
+ },
670
+ {
671
+ "epoch": 9.68,
672
+ "learning_rate": 3.74e-06,
673
+ "loss": 0.0012,
674
+ "step": 2700
675
+ },
676
+ {
677
+ "epoch": 9.77,
678
+ "learning_rate": 3.668571428571429e-06,
679
+ "loss": 0.0026,
680
+ "step": 2725
681
+ },
682
+ {
683
+ "epoch": 9.86,
684
+ "learning_rate": 3.5971428571428574e-06,
685
+ "loss": 0.0011,
686
+ "step": 2750
687
+ },
688
+ {
689
+ "epoch": 9.95,
690
+ "learning_rate": 3.5257142857142864e-06,
691
+ "loss": 0.0006,
692
+ "step": 2775
693
+ },
694
+ {
695
+ "epoch": 10.04,
696
+ "learning_rate": 3.4542857142857145e-06,
697
+ "loss": 0.0012,
698
+ "step": 2800
699
+ },
700
+ {
701
+ "epoch": 10.13,
702
+ "learning_rate": 3.382857142857143e-06,
703
+ "loss": 0.0004,
704
+ "step": 2825
705
+ },
706
+ {
707
+ "epoch": 10.22,
708
+ "learning_rate": 3.3114285714285717e-06,
709
+ "loss": 0.0003,
710
+ "step": 2850
711
+ },
712
+ {
713
+ "epoch": 10.3,
714
+ "learning_rate": 3.2400000000000003e-06,
715
+ "loss": 0.0003,
716
+ "step": 2875
717
+ },
718
+ {
719
+ "epoch": 10.39,
720
+ "learning_rate": 3.1685714285714285e-06,
721
+ "loss": 0.0004,
722
+ "step": 2900
723
+ },
724
+ {
725
+ "epoch": 10.48,
726
+ "learning_rate": 3.0971428571428575e-06,
727
+ "loss": 0.0004,
728
+ "step": 2925
729
+ },
730
+ {
731
+ "epoch": 10.57,
732
+ "learning_rate": 3.0257142857142857e-06,
733
+ "loss": 0.0008,
734
+ "step": 2950
735
+ },
736
+ {
737
+ "epoch": 10.66,
738
+ "learning_rate": 2.9542857142857147e-06,
739
+ "loss": 0.0003,
740
+ "step": 2975
741
+ },
742
+ {
743
+ "epoch": 10.75,
744
+ "learning_rate": 2.8828571428571433e-06,
745
+ "loss": 0.0007,
746
+ "step": 3000
747
+ },
748
+ {
749
+ "epoch": 10.75,
750
+ "eval_loss": 0.281982421875,
751
+ "eval_runtime": 624.9016,
752
+ "eval_samples_per_second": 3.54,
753
+ "eval_steps_per_second": 0.222,
754
+ "eval_wer": 9.095176826511805,
755
+ "step": 3000
756
+ },
757
+ {
758
+ "epoch": 10.84,
759
+ "learning_rate": 2.8114285714285714e-06,
760
+ "loss": 0.0003,
761
+ "step": 3025
762
+ },
763
+ {
764
+ "epoch": 10.93,
765
+ "learning_rate": 2.7400000000000004e-06,
766
+ "loss": 0.0005,
767
+ "step": 3050
768
+ },
769
+ {
770
+ "epoch": 11.02,
771
+ "learning_rate": 2.6685714285714286e-06,
772
+ "loss": 0.0001,
773
+ "step": 3075
774
+ },
775
+ {
776
+ "epoch": 11.11,
777
+ "learning_rate": 2.5971428571428576e-06,
778
+ "loss": 0.0002,
779
+ "step": 3100
780
+ },
781
+ {
782
+ "epoch": 11.2,
783
+ "learning_rate": 2.525714285714286e-06,
784
+ "loss": 0.0003,
785
+ "step": 3125
786
+ },
787
+ {
788
+ "epoch": 11.29,
789
+ "learning_rate": 2.454285714285715e-06,
790
+ "loss": 0.0001,
791
+ "step": 3150
792
+ },
793
+ {
794
+ "epoch": 11.38,
795
+ "learning_rate": 2.382857142857143e-06,
796
+ "loss": 0.0003,
797
+ "step": 3175
798
+ },
799
+ {
800
+ "epoch": 11.47,
801
+ "learning_rate": 2.3114285714285716e-06,
802
+ "loss": 0.0001,
803
+ "step": 3200
804
+ },
805
+ {
806
+ "epoch": 11.56,
807
+ "learning_rate": 2.24e-06,
808
+ "loss": 0.0007,
809
+ "step": 3225
810
+ },
811
+ {
812
+ "epoch": 11.65,
813
+ "learning_rate": 2.1685714285714287e-06,
814
+ "loss": 0.0002,
815
+ "step": 3250
816
+ },
817
+ {
818
+ "epoch": 11.74,
819
+ "learning_rate": 2.0971428571428573e-06,
820
+ "loss": 0.0001,
821
+ "step": 3275
822
+ },
823
+ {
824
+ "epoch": 11.83,
825
+ "learning_rate": 2.025714285714286e-06,
826
+ "loss": 0.0001,
827
+ "step": 3300
828
+ },
829
+ {
830
+ "epoch": 11.92,
831
+ "learning_rate": 1.9542857142857145e-06,
832
+ "loss": 0.0001,
833
+ "step": 3325
834
+ },
835
+ {
836
+ "epoch": 12.01,
837
+ "learning_rate": 1.8828571428571429e-06,
838
+ "loss": 0.0006,
839
+ "step": 3350
840
+ },
841
+ {
842
+ "epoch": 12.1,
843
+ "learning_rate": 1.8114285714285715e-06,
844
+ "loss": 0.0002,
845
+ "step": 3375
846
+ },
847
+ {
848
+ "epoch": 12.19,
849
+ "learning_rate": 1.74e-06,
850
+ "loss": 0.0,
851
+ "step": 3400
852
+ },
853
+ {
854
+ "epoch": 12.28,
855
+ "learning_rate": 1.6685714285714289e-06,
856
+ "loss": 0.0001,
857
+ "step": 3425
858
+ },
859
+ {
860
+ "epoch": 12.37,
861
+ "learning_rate": 1.5971428571428575e-06,
862
+ "loss": 0.0001,
863
+ "step": 3450
864
+ },
865
+ {
866
+ "epoch": 12.46,
867
+ "learning_rate": 1.5257142857142858e-06,
868
+ "loss": 0.0001,
869
+ "step": 3475
870
+ },
871
+ {
872
+ "epoch": 12.54,
873
+ "learning_rate": 1.4542857142857144e-06,
874
+ "loss": 0.0,
875
+ "step": 3500
876
+ },
877
+ {
878
+ "epoch": 12.63,
879
+ "learning_rate": 1.382857142857143e-06,
880
+ "loss": 0.0001,
881
+ "step": 3525
882
+ },
883
+ {
884
+ "epoch": 12.72,
885
+ "learning_rate": 1.3114285714285716e-06,
886
+ "loss": 0.0,
887
+ "step": 3550
888
+ },
889
+ {
890
+ "epoch": 12.81,
891
+ "learning_rate": 1.2400000000000002e-06,
892
+ "loss": 0.0001,
893
+ "step": 3575
894
+ },
895
+ {
896
+ "epoch": 12.9,
897
+ "learning_rate": 1.1685714285714286e-06,
898
+ "loss": 0.0,
899
+ "step": 3600
900
+ },
901
+ {
902
+ "epoch": 12.99,
903
+ "learning_rate": 1.0971428571428572e-06,
904
+ "loss": 0.0,
905
+ "step": 3625
906
+ },
907
+ {
908
+ "epoch": 13.08,
909
+ "learning_rate": 1.0257142857142858e-06,
910
+ "loss": 0.0001,
911
+ "step": 3650
912
+ },
913
+ {
914
+ "epoch": 13.17,
915
+ "learning_rate": 9.542857142857143e-07,
916
+ "loss": 0.0,
917
+ "step": 3675
918
+ },
919
+ {
920
+ "epoch": 13.26,
921
+ "learning_rate": 8.828571428571429e-07,
922
+ "loss": 0.0001,
923
+ "step": 3700
924
+ },
925
+ {
926
+ "epoch": 13.35,
927
+ "learning_rate": 8.114285714285715e-07,
928
+ "loss": 0.0,
929
+ "step": 3725
930
+ },
931
+ {
932
+ "epoch": 13.44,
933
+ "learning_rate": 7.4e-07,
934
+ "loss": 0.0,
935
+ "step": 3750
936
+ },
937
+ {
938
+ "epoch": 13.53,
939
+ "learning_rate": 6.685714285714286e-07,
940
+ "loss": 0.0,
941
+ "step": 3775
942
+ },
943
+ {
944
+ "epoch": 13.62,
945
+ "learning_rate": 5.971428571428572e-07,
946
+ "loss": 0.0007,
947
+ "step": 3800
948
+ },
949
+ {
950
+ "epoch": 13.71,
951
+ "learning_rate": 5.257142857142858e-07,
952
+ "loss": 0.0,
953
+ "step": 3825
954
+ },
955
+ {
956
+ "epoch": 13.8,
957
+ "learning_rate": 4.542857142857143e-07,
958
+ "loss": 0.0001,
959
+ "step": 3850
960
+ },
961
+ {
962
+ "epoch": 13.89,
963
+ "learning_rate": 3.8285714285714285e-07,
964
+ "loss": 0.0,
965
+ "step": 3875
966
+ },
967
+ {
968
+ "epoch": 13.98,
969
+ "learning_rate": 3.1142857142857144e-07,
970
+ "loss": 0.0,
971
+ "step": 3900
972
+ },
973
+ {
974
+ "epoch": 14.07,
975
+ "learning_rate": 2.4000000000000003e-07,
976
+ "loss": 0.0,
977
+ "step": 3925
978
+ },
979
+ {
980
+ "epoch": 14.16,
981
+ "learning_rate": 1.685714285714286e-07,
982
+ "loss": 0.0,
983
+ "step": 3950
984
+ },
985
+ {
986
+ "epoch": 14.25,
987
+ "learning_rate": 9.714285714285714e-08,
988
+ "loss": 0.0,
989
+ "step": 3975
990
+ },
991
+ {
992
+ "epoch": 14.34,
993
+ "learning_rate": 2.5714285714285715e-08,
994
+ "loss": 0.0,
995
+ "step": 4000
996
+ },
997
+ {
998
+ "epoch": 14.34,
999
+ "eval_loss": 0.2998046875,
1000
+ "eval_runtime": 641.2744,
1001
+ "eval_samples_per_second": 3.449,
1002
+ "eval_steps_per_second": 0.217,
1003
+ "eval_wer": 8.957818965817019,
1004
+ "step": 4000
1005
+ },
1006
+ {
1007
+ "epoch": 14.34,
1008
+ "step": 4000,
1009
+ "total_flos": 6.517583013067358e+19,
1010
+ "train_loss": 0.08977190920710564,
1011
+ "train_runtime": 17594.9662,
1012
+ "train_samples_per_second": 3.637,
1013
+ "train_steps_per_second": 0.227
1014
+ }
1015
+ ],
1016
+ "max_steps": 4000,
1017
+ "num_train_epochs": 15,
1018
+ "total_flos": 6.517583013067358e+19,
1019
+ "trial_name": null,
1020
+ "trial_params": null
1021
+ }