anuragshas commited on
Commit
fb99d58
1 Parent(s): 594a567

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 +817 -0
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 50.0,
3
+ "eval_loss": 0.6779686808586121,
4
+ "eval_runtime": 162.8436,
5
+ "eval_samples": 2693,
6
+ "eval_samples_per_second": 16.537,
7
+ "eval_steps_per_second": 1.038,
8
+ "eval_wer": 0.36695595974517586,
9
+ "train_loss": 1.1252575924730053,
10
+ "train_runtime": 34837.5253,
11
+ "train_samples": 6176,
12
+ "train_samples_per_second": 8.864,
13
+ "train_steps_per_second": 0.277
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 50.0,
3
+ "eval_loss": 0.6779686808586121,
4
+ "eval_runtime": 162.8436,
5
+ "eval_samples": 2693,
6
+ "eval_samples_per_second": 16.537,
7
+ "eval_steps_per_second": 1.038,
8
+ "eval_wer": 0.36695595974517586
9
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 50.0,
3
+ "train_loss": 1.1252575924730053,
4
+ "train_runtime": 34837.5253,
5
+ "train_samples": 6176,
6
+ "train_samples_per_second": 8.864,
7
+ "train_steps_per_second": 0.277
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,817 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 50.0,
5
+ "global_step": 9650,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.52,
12
+ "learning_rate": 3.2333333333333334e-06,
13
+ "loss": 12.3601,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 1.04,
18
+ "learning_rate": 6.566666666666667e-06,
19
+ "loss": 3.4654,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 1.55,
24
+ "learning_rate": 9.900000000000002e-06,
25
+ "loss": 3.3012,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 2.07,
30
+ "learning_rate": 1.3233333333333334e-05,
31
+ "loss": 2.514,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 2.07,
36
+ "eval_loss": 1.4589147567749023,
37
+ "eval_runtime": 164.3831,
38
+ "eval_samples_per_second": 16.382,
39
+ "eval_steps_per_second": 1.028,
40
+ "eval_wer": 0.8530606592189087,
41
+ "step": 400
42
+ },
43
+ {
44
+ "epoch": 2.59,
45
+ "learning_rate": 1.6566666666666665e-05,
46
+ "loss": 1.8333,
47
+ "step": 500
48
+ },
49
+ {
50
+ "epoch": 3.11,
51
+ "learning_rate": 1.9900000000000003e-05,
52
+ "loss": 1.6083,
53
+ "step": 600
54
+ },
55
+ {
56
+ "epoch": 3.63,
57
+ "learning_rate": 2.3233333333333333e-05,
58
+ "loss": 1.4999,
59
+ "step": 700
60
+ },
61
+ {
62
+ "epoch": 4.15,
63
+ "learning_rate": 2.6566666666666668e-05,
64
+ "loss": 1.4289,
65
+ "step": 800
66
+ },
67
+ {
68
+ "epoch": 4.15,
69
+ "eval_loss": 0.8939768075942993,
70
+ "eval_runtime": 163.9044,
71
+ "eval_samples_per_second": 16.43,
72
+ "eval_steps_per_second": 1.031,
73
+ "eval_wer": 0.6475394700397008,
74
+ "step": 800
75
+ },
76
+ {
77
+ "epoch": 4.66,
78
+ "learning_rate": 2.9900000000000002e-05,
79
+ "loss": 1.3809,
80
+ "step": 900
81
+ },
82
+ {
83
+ "epoch": 5.18,
84
+ "learning_rate": 3.323333333333333e-05,
85
+ "loss": 1.3202,
86
+ "step": 1000
87
+ },
88
+ {
89
+ "epoch": 5.7,
90
+ "learning_rate": 3.656666666666666e-05,
91
+ "loss": 1.2969,
92
+ "step": 1100
93
+ },
94
+ {
95
+ "epoch": 6.22,
96
+ "learning_rate": 3.99e-05,
97
+ "loss": 1.276,
98
+ "step": 1200
99
+ },
100
+ {
101
+ "epoch": 6.22,
102
+ "eval_loss": 0.7743439078330994,
103
+ "eval_runtime": 165.0198,
104
+ "eval_samples_per_second": 16.319,
105
+ "eval_steps_per_second": 1.024,
106
+ "eval_wer": 0.6089003785430708,
107
+ "step": 1200
108
+ },
109
+ {
110
+ "epoch": 6.74,
111
+ "learning_rate": 4.323333333333334e-05,
112
+ "loss": 1.2398,
113
+ "step": 1300
114
+ },
115
+ {
116
+ "epoch": 7.25,
117
+ "learning_rate": 4.656666666666667e-05,
118
+ "loss": 1.2339,
119
+ "step": 1400
120
+ },
121
+ {
122
+ "epoch": 7.77,
123
+ "learning_rate": 4.99e-05,
124
+ "loss": 1.2393,
125
+ "step": 1500
126
+ },
127
+ {
128
+ "epoch": 8.29,
129
+ "learning_rate": 4.940490797546012e-05,
130
+ "loss": 1.2213,
131
+ "step": 1600
132
+ },
133
+ {
134
+ "epoch": 8.29,
135
+ "eval_loss": 0.6918994784355164,
136
+ "eval_runtime": 163.5897,
137
+ "eval_samples_per_second": 16.462,
138
+ "eval_steps_per_second": 1.033,
139
+ "eval_wer": 0.4973225002308189,
140
+ "step": 1600
141
+ },
142
+ {
143
+ "epoch": 8.81,
144
+ "learning_rate": 4.879141104294479e-05,
145
+ "loss": 1.2088,
146
+ "step": 1700
147
+ },
148
+ {
149
+ "epoch": 9.33,
150
+ "learning_rate": 4.817791411042945e-05,
151
+ "loss": 1.189,
152
+ "step": 1800
153
+ },
154
+ {
155
+ "epoch": 9.84,
156
+ "learning_rate": 4.756441717791411e-05,
157
+ "loss": 1.1805,
158
+ "step": 1900
159
+ },
160
+ {
161
+ "epoch": 10.36,
162
+ "learning_rate": 4.695092024539878e-05,
163
+ "loss": 1.1522,
164
+ "step": 2000
165
+ },
166
+ {
167
+ "epoch": 10.36,
168
+ "eval_loss": 0.6634941697120667,
169
+ "eval_runtime": 162.8319,
170
+ "eval_samples_per_second": 16.539,
171
+ "eval_steps_per_second": 1.038,
172
+ "eval_wer": 0.4588219001015603,
173
+ "step": 2000
174
+ },
175
+ {
176
+ "epoch": 10.88,
177
+ "learning_rate": 4.633742331288344e-05,
178
+ "loss": 1.1468,
179
+ "step": 2100
180
+ },
181
+ {
182
+ "epoch": 11.4,
183
+ "learning_rate": 4.5723926380368096e-05,
184
+ "loss": 1.129,
185
+ "step": 2200
186
+ },
187
+ {
188
+ "epoch": 11.92,
189
+ "learning_rate": 4.511042944785277e-05,
190
+ "loss": 1.1133,
191
+ "step": 2300
192
+ },
193
+ {
194
+ "epoch": 12.44,
195
+ "learning_rate": 4.4496932515337425e-05,
196
+ "loss": 1.0914,
197
+ "step": 2400
198
+ },
199
+ {
200
+ "epoch": 12.44,
201
+ "eval_loss": 0.6839072704315186,
202
+ "eval_runtime": 163.9109,
203
+ "eval_samples_per_second": 16.43,
204
+ "eval_steps_per_second": 1.031,
205
+ "eval_wer": 0.45859108115594127,
206
+ "step": 2400
207
+ },
208
+ {
209
+ "epoch": 12.95,
210
+ "learning_rate": 4.388343558282208e-05,
211
+ "loss": 1.1104,
212
+ "step": 2500
213
+ },
214
+ {
215
+ "epoch": 13.47,
216
+ "learning_rate": 4.3269938650306755e-05,
217
+ "loss": 1.0732,
218
+ "step": 2600
219
+ },
220
+ {
221
+ "epoch": 13.99,
222
+ "learning_rate": 4.265644171779141e-05,
223
+ "loss": 1.0746,
224
+ "step": 2700
225
+ },
226
+ {
227
+ "epoch": 14.51,
228
+ "learning_rate": 4.204294478527608e-05,
229
+ "loss": 1.0499,
230
+ "step": 2800
231
+ },
232
+ {
233
+ "epoch": 14.51,
234
+ "eval_loss": 0.7151051163673401,
235
+ "eval_runtime": 165.6452,
236
+ "eval_samples_per_second": 16.258,
237
+ "eval_steps_per_second": 1.02,
238
+ "eval_wer": 0.44672698735112176,
239
+ "step": 2800
240
+ },
241
+ {
242
+ "epoch": 15.03,
243
+ "learning_rate": 4.142944785276074e-05,
244
+ "loss": 1.0641,
245
+ "step": 2900
246
+ },
247
+ {
248
+ "epoch": 15.54,
249
+ "learning_rate": 4.08159509202454e-05,
250
+ "loss": 1.0386,
251
+ "step": 3000
252
+ },
253
+ {
254
+ "epoch": 16.06,
255
+ "learning_rate": 4.0202453987730065e-05,
256
+ "loss": 1.0328,
257
+ "step": 3100
258
+ },
259
+ {
260
+ "epoch": 16.58,
261
+ "learning_rate": 3.958895705521473e-05,
262
+ "loss": 1.0238,
263
+ "step": 3200
264
+ },
265
+ {
266
+ "epoch": 16.58,
267
+ "eval_loss": 0.6823991537094116,
268
+ "eval_runtime": 163.1559,
269
+ "eval_samples_per_second": 16.506,
270
+ "eval_steps_per_second": 1.036,
271
+ "eval_wer": 0.44358784969070264,
272
+ "step": 3200
273
+ },
274
+ {
275
+ "epoch": 17.1,
276
+ "learning_rate": 3.897546012269939e-05,
277
+ "loss": 1.0055,
278
+ "step": 3300
279
+ },
280
+ {
281
+ "epoch": 17.62,
282
+ "learning_rate": 3.836196319018405e-05,
283
+ "loss": 1.006,
284
+ "step": 3400
285
+ },
286
+ {
287
+ "epoch": 18.13,
288
+ "learning_rate": 3.774846625766871e-05,
289
+ "loss": 1.0008,
290
+ "step": 3500
291
+ },
292
+ {
293
+ "epoch": 18.65,
294
+ "learning_rate": 3.7134969325153375e-05,
295
+ "loss": 0.9963,
296
+ "step": 3600
297
+ },
298
+ {
299
+ "epoch": 18.65,
300
+ "eval_loss": 0.6872182488441467,
301
+ "eval_runtime": 163.684,
302
+ "eval_samples_per_second": 16.452,
303
+ "eval_steps_per_second": 1.032,
304
+ "eval_wer": 0.44368017726895026,
305
+ "step": 3600
306
+ },
307
+ {
308
+ "epoch": 19.17,
309
+ "learning_rate": 3.652147239263804e-05,
310
+ "loss": 0.9852,
311
+ "step": 3700
312
+ },
313
+ {
314
+ "epoch": 19.69,
315
+ "learning_rate": 3.59079754601227e-05,
316
+ "loss": 0.9905,
317
+ "step": 3800
318
+ },
319
+ {
320
+ "epoch": 20.21,
321
+ "learning_rate": 3.529447852760737e-05,
322
+ "loss": 0.9621,
323
+ "step": 3900
324
+ },
325
+ {
326
+ "epoch": 20.73,
327
+ "learning_rate": 3.468711656441718e-05,
328
+ "loss": 0.9728,
329
+ "step": 4000
330
+ },
331
+ {
332
+ "epoch": 20.73,
333
+ "eval_loss": 0.704698383808136,
334
+ "eval_runtime": 163.0374,
335
+ "eval_samples_per_second": 16.518,
336
+ "eval_steps_per_second": 1.037,
337
+ "eval_wer": 0.4243837134151971,
338
+ "step": 4000
339
+ },
340
+ {
341
+ "epoch": 21.24,
342
+ "learning_rate": 3.407361963190184e-05,
343
+ "loss": 0.9534,
344
+ "step": 4100
345
+ },
346
+ {
347
+ "epoch": 21.76,
348
+ "learning_rate": 3.34601226993865e-05,
349
+ "loss": 0.9599,
350
+ "step": 4200
351
+ },
352
+ {
353
+ "epoch": 22.28,
354
+ "learning_rate": 3.284662576687117e-05,
355
+ "loss": 0.9427,
356
+ "step": 4300
357
+ },
358
+ {
359
+ "epoch": 22.8,
360
+ "learning_rate": 3.223312883435583e-05,
361
+ "loss": 0.9373,
362
+ "step": 4400
363
+ },
364
+ {
365
+ "epoch": 22.8,
366
+ "eval_loss": 0.656895637512207,
367
+ "eval_runtime": 162.8754,
368
+ "eval_samples_per_second": 16.534,
369
+ "eval_steps_per_second": 1.038,
370
+ "eval_wer": 0.41889022250946356,
371
+ "step": 4400
372
+ },
373
+ {
374
+ "epoch": 23.32,
375
+ "learning_rate": 3.161963190184049e-05,
376
+ "loss": 0.9348,
377
+ "step": 4500
378
+ },
379
+ {
380
+ "epoch": 23.83,
381
+ "learning_rate": 3.100613496932516e-05,
382
+ "loss": 0.9295,
383
+ "step": 4600
384
+ },
385
+ {
386
+ "epoch": 24.35,
387
+ "learning_rate": 3.0392638036809817e-05,
388
+ "loss": 0.9162,
389
+ "step": 4700
390
+ },
391
+ {
392
+ "epoch": 24.87,
393
+ "learning_rate": 2.977914110429448e-05,
394
+ "loss": 0.9028,
395
+ "step": 4800
396
+ },
397
+ {
398
+ "epoch": 24.87,
399
+ "eval_loss": 0.6622523665428162,
400
+ "eval_runtime": 163.2025,
401
+ "eval_samples_per_second": 16.501,
402
+ "eval_steps_per_second": 1.036,
403
+ "eval_wer": 0.40938048194995846,
404
+ "step": 4800
405
+ },
406
+ {
407
+ "epoch": 25.39,
408
+ "learning_rate": 2.9165644171779143e-05,
409
+ "loss": 0.9002,
410
+ "step": 4900
411
+ },
412
+ {
413
+ "epoch": 25.91,
414
+ "learning_rate": 2.8552147239263805e-05,
415
+ "loss": 0.8869,
416
+ "step": 5000
417
+ },
418
+ {
419
+ "epoch": 26.42,
420
+ "learning_rate": 2.7938650306748466e-05,
421
+ "loss": 0.8741,
422
+ "step": 5100
423
+ },
424
+ {
425
+ "epoch": 26.94,
426
+ "learning_rate": 2.732515337423313e-05,
427
+ "loss": 0.8759,
428
+ "step": 5200
429
+ },
430
+ {
431
+ "epoch": 26.94,
432
+ "eval_loss": 0.6723098158836365,
433
+ "eval_runtime": 162.4177,
434
+ "eval_samples_per_second": 16.581,
435
+ "eval_steps_per_second": 1.041,
436
+ "eval_wer": 0.41519711937955867,
437
+ "step": 5200
438
+ },
439
+ {
440
+ "epoch": 27.46,
441
+ "learning_rate": 2.6711656441717792e-05,
442
+ "loss": 0.8702,
443
+ "step": 5300
444
+ },
445
+ {
446
+ "epoch": 27.98,
447
+ "learning_rate": 2.6098159509202457e-05,
448
+ "loss": 0.8609,
449
+ "step": 5400
450
+ },
451
+ {
452
+ "epoch": 28.5,
453
+ "learning_rate": 2.5484662576687118e-05,
454
+ "loss": 0.8593,
455
+ "step": 5500
456
+ },
457
+ {
458
+ "epoch": 29.02,
459
+ "learning_rate": 2.4871165644171783e-05,
460
+ "loss": 0.8824,
461
+ "step": 5600
462
+ },
463
+ {
464
+ "epoch": 29.02,
465
+ "eval_loss": 0.6467483043670654,
466
+ "eval_runtime": 163.0935,
467
+ "eval_samples_per_second": 16.512,
468
+ "eval_steps_per_second": 1.036,
469
+ "eval_wer": 0.40167112916628195,
470
+ "step": 5600
471
+ },
472
+ {
473
+ "epoch": 29.53,
474
+ "learning_rate": 2.425766871165644e-05,
475
+ "loss": 0.8717,
476
+ "step": 5700
477
+ },
478
+ {
479
+ "epoch": 30.05,
480
+ "learning_rate": 2.3644171779141105e-05,
481
+ "loss": 0.8514,
482
+ "step": 5800
483
+ },
484
+ {
485
+ "epoch": 30.57,
486
+ "learning_rate": 2.303067484662577e-05,
487
+ "loss": 0.851,
488
+ "step": 5900
489
+ },
490
+ {
491
+ "epoch": 31.09,
492
+ "learning_rate": 2.2423312883435586e-05,
493
+ "loss": 0.8371,
494
+ "step": 6000
495
+ },
496
+ {
497
+ "epoch": 31.09,
498
+ "eval_loss": 0.6910686492919922,
499
+ "eval_runtime": 162.9708,
500
+ "eval_samples_per_second": 16.524,
501
+ "eval_steps_per_second": 1.037,
502
+ "eval_wer": 0.4079955682762441,
503
+ "step": 6000
504
+ },
505
+ {
506
+ "epoch": 31.61,
507
+ "learning_rate": 2.1809815950920244e-05,
508
+ "loss": 0.8276,
509
+ "step": 6100
510
+ },
511
+ {
512
+ "epoch": 32.12,
513
+ "learning_rate": 2.119631901840491e-05,
514
+ "loss": 0.8282,
515
+ "step": 6200
516
+ },
517
+ {
518
+ "epoch": 32.64,
519
+ "learning_rate": 2.0582822085889574e-05,
520
+ "loss": 0.8195,
521
+ "step": 6300
522
+ },
523
+ {
524
+ "epoch": 33.16,
525
+ "learning_rate": 1.9969325153374235e-05,
526
+ "loss": 0.8205,
527
+ "step": 6400
528
+ },
529
+ {
530
+ "epoch": 33.16,
531
+ "eval_loss": 0.7144636511802673,
532
+ "eval_runtime": 163.1126,
533
+ "eval_samples_per_second": 16.51,
534
+ "eval_steps_per_second": 1.036,
535
+ "eval_wer": 0.4062875080786631,
536
+ "step": 6400
537
+ },
538
+ {
539
+ "epoch": 33.68,
540
+ "learning_rate": 1.9355828220858896e-05,
541
+ "loss": 0.8073,
542
+ "step": 6500
543
+ },
544
+ {
545
+ "epoch": 34.2,
546
+ "learning_rate": 1.874233128834356e-05,
547
+ "loss": 0.8088,
548
+ "step": 6600
549
+ },
550
+ {
551
+ "epoch": 34.72,
552
+ "learning_rate": 1.8128834355828222e-05,
553
+ "loss": 0.7828,
554
+ "step": 6700
555
+ },
556
+ {
557
+ "epoch": 35.23,
558
+ "learning_rate": 1.7515337423312884e-05,
559
+ "loss": 0.7837,
560
+ "step": 6800
561
+ },
562
+ {
563
+ "epoch": 35.23,
564
+ "eval_loss": 0.7037481069564819,
565
+ "eval_runtime": 163.3942,
566
+ "eval_samples_per_second": 16.482,
567
+ "eval_steps_per_second": 1.034,
568
+ "eval_wer": 0.3929923368110054,
569
+ "step": 6800
570
+ },
571
+ {
572
+ "epoch": 35.75,
573
+ "learning_rate": 1.6901840490797545e-05,
574
+ "loss": 0.7893,
575
+ "step": 6900
576
+ },
577
+ {
578
+ "epoch": 36.27,
579
+ "learning_rate": 1.628834355828221e-05,
580
+ "loss": 0.7777,
581
+ "step": 7000
582
+ },
583
+ {
584
+ "epoch": 36.79,
585
+ "learning_rate": 1.5674846625766874e-05,
586
+ "loss": 0.7659,
587
+ "step": 7100
588
+ },
589
+ {
590
+ "epoch": 37.31,
591
+ "learning_rate": 1.5067484662576689e-05,
592
+ "loss": 0.7708,
593
+ "step": 7200
594
+ },
595
+ {
596
+ "epoch": 37.31,
597
+ "eval_loss": 0.6925445199012756,
598
+ "eval_runtime": 163.9386,
599
+ "eval_samples_per_second": 16.427,
600
+ "eval_steps_per_second": 1.031,
601
+ "eval_wer": 0.38399039793186224,
602
+ "step": 7200
603
+ },
604
+ {
605
+ "epoch": 37.82,
606
+ "learning_rate": 1.445398773006135e-05,
607
+ "loss": 0.7568,
608
+ "step": 7300
609
+ },
610
+ {
611
+ "epoch": 38.34,
612
+ "learning_rate": 1.3840490797546013e-05,
613
+ "loss": 0.7467,
614
+ "step": 7400
615
+ },
616
+ {
617
+ "epoch": 38.86,
618
+ "learning_rate": 1.3226993865030676e-05,
619
+ "loss": 0.7456,
620
+ "step": 7500
621
+ },
622
+ {
623
+ "epoch": 39.38,
624
+ "learning_rate": 1.2613496932515337e-05,
625
+ "loss": 0.7359,
626
+ "step": 7600
627
+ },
628
+ {
629
+ "epoch": 39.38,
630
+ "eval_loss": 0.7034336924552917,
631
+ "eval_runtime": 163.3955,
632
+ "eval_samples_per_second": 16.481,
633
+ "eval_steps_per_second": 1.034,
634
+ "eval_wer": 0.3829286307820146,
635
+ "step": 7600
636
+ },
637
+ {
638
+ "epoch": 39.9,
639
+ "learning_rate": 1.2e-05,
640
+ "loss": 0.7464,
641
+ "step": 7700
642
+ },
643
+ {
644
+ "epoch": 40.41,
645
+ "learning_rate": 1.1386503067484663e-05,
646
+ "loss": 0.7381,
647
+ "step": 7800
648
+ },
649
+ {
650
+ "epoch": 40.93,
651
+ "learning_rate": 1.0773006134969325e-05,
652
+ "loss": 0.726,
653
+ "step": 7900
654
+ },
655
+ {
656
+ "epoch": 41.45,
657
+ "learning_rate": 1.015950920245399e-05,
658
+ "loss": 0.7153,
659
+ "step": 8000
660
+ },
661
+ {
662
+ "epoch": 41.45,
663
+ "eval_loss": 0.7030363082885742,
664
+ "eval_runtime": 163.6509,
665
+ "eval_samples_per_second": 16.456,
666
+ "eval_steps_per_second": 1.033,
667
+ "eval_wer": 0.37937401901948115,
668
+ "step": 8000
669
+ },
670
+ {
671
+ "epoch": 41.97,
672
+ "learning_rate": 9.54601226993865e-06,
673
+ "loss": 0.708,
674
+ "step": 8100
675
+ },
676
+ {
677
+ "epoch": 42.49,
678
+ "learning_rate": 8.932515337423314e-06,
679
+ "loss": 0.7127,
680
+ "step": 8200
681
+ },
682
+ {
683
+ "epoch": 43.01,
684
+ "learning_rate": 8.319018404907975e-06,
685
+ "loss": 0.7117,
686
+ "step": 8300
687
+ },
688
+ {
689
+ "epoch": 43.52,
690
+ "learning_rate": 7.705521472392638e-06,
691
+ "loss": 0.7127,
692
+ "step": 8400
693
+ },
694
+ {
695
+ "epoch": 43.52,
696
+ "eval_loss": 0.6823467016220093,
697
+ "eval_runtime": 163.2294,
698
+ "eval_samples_per_second": 16.498,
699
+ "eval_steps_per_second": 1.035,
700
+ "eval_wer": 0.37609638999169054,
701
+ "step": 8400
702
+ },
703
+ {
704
+ "epoch": 44.04,
705
+ "learning_rate": 7.092024539877301e-06,
706
+ "loss": 0.7081,
707
+ "step": 8500
708
+ },
709
+ {
710
+ "epoch": 44.56,
711
+ "learning_rate": 6.478527607361963e-06,
712
+ "loss": 0.6977,
713
+ "step": 8600
714
+ },
715
+ {
716
+ "epoch": 45.08,
717
+ "learning_rate": 5.865030674846626e-06,
718
+ "loss": 0.6943,
719
+ "step": 8700
720
+ },
721
+ {
722
+ "epoch": 45.6,
723
+ "learning_rate": 5.251533742331288e-06,
724
+ "loss": 0.6884,
725
+ "step": 8800
726
+ },
727
+ {
728
+ "epoch": 45.6,
729
+ "eval_loss": 0.6854104399681091,
730
+ "eval_runtime": 163.0739,
731
+ "eval_samples_per_second": 16.514,
732
+ "eval_steps_per_second": 1.036,
733
+ "eval_wer": 0.3710645369771951,
734
+ "step": 8800
735
+ },
736
+ {
737
+ "epoch": 46.11,
738
+ "learning_rate": 4.6380368098159506e-06,
739
+ "loss": 0.6803,
740
+ "step": 8900
741
+ },
742
+ {
743
+ "epoch": 46.63,
744
+ "learning_rate": 4.0245398773006136e-06,
745
+ "loss": 0.6902,
746
+ "step": 9000
747
+ },
748
+ {
749
+ "epoch": 47.15,
750
+ "learning_rate": 3.411042944785276e-06,
751
+ "loss": 0.6827,
752
+ "step": 9100
753
+ },
754
+ {
755
+ "epoch": 47.67,
756
+ "learning_rate": 2.7975460122699388e-06,
757
+ "loss": 0.6835,
758
+ "step": 9200
759
+ },
760
+ {
761
+ "epoch": 47.67,
762
+ "eval_loss": 0.6723486185073853,
763
+ "eval_runtime": 162.6892,
764
+ "eval_samples_per_second": 16.553,
765
+ "eval_steps_per_second": 1.039,
766
+ "eval_wer": 0.36649432185393777,
767
+ "step": 9200
768
+ },
769
+ {
770
+ "epoch": 48.19,
771
+ "learning_rate": 2.1840490797546013e-06,
772
+ "loss": 0.6825,
773
+ "step": 9300
774
+ },
775
+ {
776
+ "epoch": 48.7,
777
+ "learning_rate": 1.570552147239264e-06,
778
+ "loss": 0.6611,
779
+ "step": 9400
780
+ },
781
+ {
782
+ "epoch": 49.22,
783
+ "learning_rate": 9.570552147239263e-07,
784
+ "loss": 0.6666,
785
+ "step": 9500
786
+ },
787
+ {
788
+ "epoch": 49.74,
789
+ "learning_rate": 3.43558282208589e-07,
790
+ "loss": 0.6703,
791
+ "step": 9600
792
+ },
793
+ {
794
+ "epoch": 49.74,
795
+ "eval_loss": 0.6773473024368286,
796
+ "eval_runtime": 162.1677,
797
+ "eval_samples_per_second": 16.606,
798
+ "eval_steps_per_second": 1.042,
799
+ "eval_wer": 0.3667713045886806,
800
+ "step": 9600
801
+ },
802
+ {
803
+ "epoch": 50.0,
804
+ "step": 9650,
805
+ "total_flos": 1.2240973116802213e+20,
806
+ "train_loss": 1.1252575924730053,
807
+ "train_runtime": 34837.5253,
808
+ "train_samples_per_second": 8.864,
809
+ "train_steps_per_second": 0.277
810
+ }
811
+ ],
812
+ "max_steps": 9650,
813
+ "num_train_epochs": 50,
814
+ "total_flos": 1.2240973116802213e+20,
815
+ "trial_name": null,
816
+ "trial_params": null
817
+ }