tz579 commited on
Commit
f5e27ff
1 Parent(s): d2240f4

End of training

Browse files
README.md CHANGED
@@ -2,6 +2,8 @@
2
  license: apache-2.0
3
  base_model: facebook/wav2vec2-base
4
  tags:
 
 
5
  - generated_from_trainer
6
  datasets:
7
  - timit_asr
@@ -14,15 +16,15 @@ model-index:
14
  name: Automatic Speech Recognition
15
  type: automatic-speech-recognition
16
  dataset:
17
- name: timit_asr
18
  type: timit_asr
19
  config: clean
20
  split: test
21
- args: clean
22
  metrics:
23
  - name: Wer
24
  type: wer
25
- value: 0.4359464627151052
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -30,10 +32,10 @@ should probably proofread and complete it, then remove this comment. -->
30
 
31
  # wav2vec2-base-timit-fine-tuned
32
 
33
- This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the timit_asr dataset.
34
  It achieves the following results on the evaluation set:
35
- - Loss: 0.4279
36
- - Wer: 0.4359
37
 
38
  ## Model description
39
 
 
2
  license: apache-2.0
3
  base_model: facebook/wav2vec2-base
4
  tags:
5
+ - automatic-speech-recognition
6
+ - timit_asr
7
  - generated_from_trainer
8
  datasets:
9
  - timit_asr
 
16
  name: Automatic Speech Recognition
17
  type: automatic-speech-recognition
18
  dataset:
19
+ name: TIMIT_ASR - NA
20
  type: timit_asr
21
  config: clean
22
  split: test
23
+ args: 'Config: na, Training split: train, Eval split: test'
24
  metrics:
25
  - name: Wer
26
  type: wer
27
+ value: 0.4328507693708459
28
  ---
29
 
30
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
32
 
33
  # wav2vec2-base-timit-fine-tuned
34
 
35
+ This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the TIMIT_ASR - NA dataset.
36
  It achieves the following results on the evaluation set:
37
+ - Loss: 0.4233
38
+ - Wer: 0.4329
39
 
40
  ## Model description
41
 
all_results.json CHANGED
@@ -1,15 +1,15 @@
1
  {
2
  "epoch": 20.0,
3
- "eval_loss": 0.42176610231399536,
4
- "eval_runtime": 39.428,
5
  "eval_samples": 1344,
6
- "eval_samples_per_second": 34.087,
7
- "eval_steps_per_second": 34.087,
8
- "eval_wer": 0.4090867704634435,
9
- "total_flos": 2.1476719263248095e+18,
10
- "train_loss": 0.8590125822430027,
11
- "train_runtime": 3151.1477,
12
  "train_samples": 3696,
13
- "train_samples_per_second": 23.458,
14
- "train_steps_per_second": 0.736
15
  }
 
1
  {
2
  "epoch": 20.0,
3
+ "eval_loss": 0.42333292961120605,
4
+ "eval_runtime": 14.6575,
5
  "eval_samples": 1344,
6
+ "eval_samples_per_second": 91.694,
7
+ "eval_steps_per_second": 2.865,
8
+ "eval_wer": 0.4328507693708459,
9
+ "total_flos": 2.2493941126178012e+18,
10
+ "train_loss": 1.5806312146885642,
11
+ "train_runtime": 1021.9437,
12
  "train_samples": 3696,
13
+ "train_samples_per_second": 72.333,
14
+ "train_steps_per_second": 1.135
15
  }
eval_results.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
  "epoch": 20.0,
3
- "eval_loss": 0.42176610231399536,
4
- "eval_runtime": 39.428,
5
  "eval_samples": 1344,
6
- "eval_samples_per_second": 34.087,
7
- "eval_steps_per_second": 34.087,
8
- "eval_wer": 0.4090867704634435
9
  }
 
1
  {
2
  "epoch": 20.0,
3
+ "eval_loss": 0.42333292961120605,
4
+ "eval_runtime": 14.6575,
5
  "eval_samples": 1344,
6
+ "eval_samples_per_second": 91.694,
7
+ "eval_steps_per_second": 2.865,
8
+ "eval_wer": 0.4328507693708459
9
  }
runs/May26_14-23-40_tz579-raptorlake/events.out.tfevents.1716752577.tz579-raptorlake.17539.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74d784d8e4b12d527e75df42e434482184417ea064d6cc1d258a33b6a1331d8e
3
+ size 406
train_results.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
  "epoch": 20.0,
3
- "total_flos": 2.1476719263248095e+18,
4
- "train_loss": 0.8590125822430027,
5
- "train_runtime": 3151.1477,
6
  "train_samples": 3696,
7
- "train_samples_per_second": 23.458,
8
- "train_steps_per_second": 0.736
9
  }
 
1
  {
2
  "epoch": 20.0,
3
+ "total_flos": 2.2493941126178012e+18,
4
+ "train_loss": 1.5806312146885642,
5
+ "train_runtime": 1021.9437,
6
  "train_samples": 3696,
7
+ "train_samples_per_second": 72.333,
8
+ "train_steps_per_second": 1.135
9
  }
trainer_state.json CHANGED
@@ -3,1854 +3,934 @@
3
  "best_model_checkpoint": null,
4
  "epoch": 20.0,
5
  "eval_steps": 100,
6
- "global_step": 2320,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.08620689655172414,
13
- "grad_norm": 9.590513229370117,
14
  "learning_rate": 9e-07,
15
- "loss": 9.1142,
16
  "step": 10
17
  },
18
  {
19
- "epoch": 0.1724137931034483,
20
- "grad_norm": 9.729034423828125,
21
  "learning_rate": 1.9e-06,
22
- "loss": 8.3444,
23
  "step": 20
24
  },
25
  {
26
- "epoch": 0.25862068965517243,
27
- "grad_norm": 14.263296127319336,
28
  "learning_rate": 2.8000000000000003e-06,
29
- "loss": 8.6571,
30
  "step": 30
31
  },
32
  {
33
- "epoch": 0.3448275862068966,
34
- "grad_norm": 15.165566444396973,
35
  "learning_rate": 3.8e-06,
36
- "loss": 7.6933,
37
  "step": 40
38
  },
39
  {
40
- "epoch": 0.43103448275862066,
41
- "grad_norm": 16.664026260375977,
42
  "learning_rate": 4.800000000000001e-06,
43
- "loss": 6.9526,
44
  "step": 50
45
  },
46
  {
47
- "epoch": 0.5172413793103449,
48
- "grad_norm": 17.301790237426758,
49
  "learning_rate": 5.8e-06,
50
- "loss": 6.2005,
51
  "step": 60
52
  },
53
  {
54
- "epoch": 0.603448275862069,
55
- "grad_norm": 11.270517349243164,
56
  "learning_rate": 6.800000000000001e-06,
57
- "loss": 4.7052,
58
  "step": 70
59
  },
60
  {
61
- "epoch": 0.6896551724137931,
62
- "grad_norm": 4.181836128234863,
63
  "learning_rate": 7.8e-06,
64
- "loss": 3.7839,
65
  "step": 80
66
  },
67
  {
68
- "epoch": 0.7758620689655172,
69
- "grad_norm": 1.8710567951202393,
70
  "learning_rate": 8.8e-06,
71
- "loss": 3.3947,
72
  "step": 90
73
  },
74
  {
75
- "epoch": 0.8620689655172413,
76
- "grad_norm": 1.407578706741333,
77
  "learning_rate": 9.800000000000001e-06,
78
- "loss": 3.1612,
79
  "step": 100
80
  },
81
  {
82
- "epoch": 0.8620689655172413,
83
- "eval_loss": 3.118124485015869,
84
- "eval_runtime": 39.9765,
85
- "eval_samples_per_second": 33.62,
86
- "eval_steps_per_second": 33.62,
87
  "eval_wer": 1.0,
88
  "step": 100
89
  },
90
  {
91
- "epoch": 0.9482758620689655,
92
- "grad_norm": 1.6558985710144043,
93
  "learning_rate": 1.08e-05,
94
- "loss": 3.0858,
95
  "step": 110
96
  },
97
  {
98
- "epoch": 1.0344827586206897,
99
- "grad_norm": 1.8191890716552734,
100
  "learning_rate": 1.18e-05,
101
- "loss": 3.0791,
102
  "step": 120
103
  },
104
  {
105
- "epoch": 1.1206896551724137,
106
- "grad_norm": 0.9610480070114136,
107
  "learning_rate": 1.2800000000000001e-05,
108
- "loss": 3.034,
109
  "step": 130
110
  },
111
  {
112
- "epoch": 1.206896551724138,
113
- "grad_norm": 0.7719367742538452,
114
  "learning_rate": 1.3800000000000002e-05,
115
- "loss": 3.0098,
116
  "step": 140
117
  },
118
  {
119
- "epoch": 1.293103448275862,
120
- "grad_norm": 0.7965385913848877,
121
  "learning_rate": 1.48e-05,
122
- "loss": 3.0138,
123
  "step": 150
124
  },
125
  {
126
- "epoch": 1.3793103448275863,
127
- "grad_norm": 0.9387674927711487,
128
  "learning_rate": 1.58e-05,
129
- "loss": 2.9973,
130
  "step": 160
131
  },
132
  {
133
- "epoch": 1.4655172413793103,
134
- "grad_norm": 0.8531718254089355,
135
  "learning_rate": 1.6800000000000002e-05,
136
- "loss": 2.9875,
137
  "step": 170
138
  },
139
  {
140
- "epoch": 1.5517241379310345,
141
- "grad_norm": 0.5282021760940552,
142
  "learning_rate": 1.78e-05,
143
- "loss": 2.9905,
144
  "step": 180
145
  },
146
  {
147
- "epoch": 1.6379310344827587,
148
- "grad_norm": 1.1362160444259644,
149
  "learning_rate": 1.88e-05,
150
- "loss": 2.9881,
151
  "step": 190
152
  },
153
  {
154
- "epoch": 1.7241379310344827,
155
- "grad_norm": 1.0381989479064941,
156
  "learning_rate": 1.9800000000000004e-05,
157
- "loss": 2.978,
158
  "step": 200
159
  },
160
  {
161
- "epoch": 1.7241379310344827,
162
- "eval_loss": 2.9721522331237793,
163
- "eval_runtime": 39.4986,
164
- "eval_samples_per_second": 34.026,
165
- "eval_steps_per_second": 34.026,
166
  "eval_wer": 1.0,
167
  "step": 200
168
  },
169
  {
170
- "epoch": 1.8103448275862069,
171
- "grad_norm": 1.0819815397262573,
172
  "learning_rate": 2.08e-05,
173
- "loss": 2.9728,
174
  "step": 210
175
  },
176
  {
177
- "epoch": 1.896551724137931,
178
- "grad_norm": 0.5016360878944397,
179
  "learning_rate": 2.18e-05,
180
- "loss": 2.9769,
181
  "step": 220
182
  },
183
  {
184
- "epoch": 1.9827586206896552,
185
- "grad_norm": 1.476746678352356,
186
  "learning_rate": 2.2800000000000002e-05,
187
- "loss": 2.9638,
188
  "step": 230
189
  },
190
  {
191
- "epoch": 2.0689655172413794,
192
- "grad_norm": 0.416255384683609,
193
  "learning_rate": 2.38e-05,
194
- "loss": 2.9566,
195
  "step": 240
196
  },
197
  {
198
- "epoch": 2.1551724137931036,
199
- "grad_norm": 0.5918602347373962,
200
  "learning_rate": 2.48e-05,
201
- "loss": 2.9507,
202
  "step": 250
203
  },
204
  {
205
- "epoch": 2.2413793103448274,
206
- "grad_norm": 0.8085893988609314,
207
  "learning_rate": 2.58e-05,
208
- "loss": 2.9469,
209
  "step": 260
210
  },
211
  {
212
- "epoch": 2.3275862068965516,
213
- "grad_norm": 0.8484460115432739,
214
  "learning_rate": 2.6800000000000004e-05,
215
- "loss": 2.9438,
216
  "step": 270
217
  },
218
  {
219
- "epoch": 2.413793103448276,
220
- "grad_norm": 0.4717480540275574,
221
  "learning_rate": 2.7800000000000005e-05,
222
- "loss": 2.9421,
223
  "step": 280
224
  },
225
  {
226
- "epoch": 2.5,
227
- "grad_norm": 2.0097577571868896,
228
  "learning_rate": 2.88e-05,
229
- "loss": 2.9234,
230
  "step": 290
231
  },
232
  {
233
- "epoch": 2.586206896551724,
234
- "grad_norm": 1.3988488912582397,
235
  "learning_rate": 2.98e-05,
236
- "loss": 2.9185,
237
  "step": 300
238
  },
239
  {
240
- "epoch": 2.586206896551724,
241
- "eval_loss": 2.909769058227539,
242
- "eval_runtime": 39.5502,
243
- "eval_samples_per_second": 33.982,
244
- "eval_steps_per_second": 33.982,
245
  "eval_wer": 1.0,
246
  "step": 300
247
  },
248
  {
249
- "epoch": 2.6724137931034484,
250
- "grad_norm": 2.138706922531128,
251
  "learning_rate": 3.08e-05,
252
- "loss": 2.9028,
253
  "step": 310
254
  },
255
  {
256
- "epoch": 2.7586206896551726,
257
- "grad_norm": 0.854528546333313,
258
  "learning_rate": 3.18e-05,
259
- "loss": 2.8889,
260
  "step": 320
261
  },
262
  {
263
- "epoch": 2.844827586206897,
264
- "grad_norm": 0.8741695284843445,
265
  "learning_rate": 3.2800000000000004e-05,
266
  "loss": 2.8698,
267
  "step": 330
268
  },
269
  {
270
- "epoch": 2.9310344827586206,
271
- "grad_norm": 2.8528709411621094,
272
  "learning_rate": 3.38e-05,
273
- "loss": 2.7851,
274
  "step": 340
275
  },
276
  {
277
- "epoch": 3.0172413793103448,
278
- "grad_norm": 2.0051960945129395,
279
  "learning_rate": 3.48e-05,
280
- "loss": 2.7537,
281
  "step": 350
282
  },
283
  {
284
- "epoch": 3.103448275862069,
285
- "grad_norm": 1.6813557147979736,
286
  "learning_rate": 3.58e-05,
287
- "loss": 2.6125,
288
  "step": 360
289
  },
290
  {
291
- "epoch": 3.189655172413793,
292
- "grad_norm": 2.679445266723633,
293
  "learning_rate": 3.68e-05,
294
- "loss": 2.4925,
295
  "step": 370
296
  },
297
  {
298
- "epoch": 3.2758620689655173,
299
- "grad_norm": 1.8131614923477173,
300
  "learning_rate": 3.7800000000000004e-05,
301
- "loss": 2.4007,
302
  "step": 380
303
  },
304
  {
305
- "epoch": 3.3620689655172415,
306
- "grad_norm": 2.168797016143799,
307
  "learning_rate": 3.88e-05,
308
- "loss": 2.2533,
309
  "step": 390
310
  },
311
  {
312
- "epoch": 3.4482758620689653,
313
- "grad_norm": 2.1998753547668457,
314
  "learning_rate": 3.9800000000000005e-05,
315
- "loss": 2.1282,
316
  "step": 400
317
  },
318
  {
319
- "epoch": 3.4482758620689653,
320
- "eval_loss": 2.0066494941711426,
321
- "eval_runtime": 39.5789,
322
- "eval_samples_per_second": 33.958,
323
- "eval_steps_per_second": 33.958,
324
- "eval_wer": 1.0246744969498316,
325
  "step": 400
326
  },
327
  {
328
- "epoch": 3.5344827586206895,
329
- "grad_norm": 4.27025842666626,
330
  "learning_rate": 4.08e-05,
331
- "loss": 2.0277,
332
  "step": 410
333
  },
334
  {
335
- "epoch": 3.6206896551724137,
336
- "grad_norm": 2.3157191276550293,
337
  "learning_rate": 4.18e-05,
338
- "loss": 1.8844,
339
  "step": 420
340
  },
341
  {
342
- "epoch": 3.706896551724138,
343
- "grad_norm": 3.1167590618133545,
344
  "learning_rate": 4.2800000000000004e-05,
345
- "loss": 1.7777,
346
  "step": 430
347
  },
348
  {
349
- "epoch": 3.793103448275862,
350
- "grad_norm": 4.118265151977539,
351
  "learning_rate": 4.38e-05,
352
- "loss": 1.7015,
353
  "step": 440
354
  },
355
  {
356
- "epoch": 3.8793103448275863,
357
- "grad_norm": 2.6515376567840576,
358
  "learning_rate": 4.4800000000000005e-05,
359
- "loss": 1.5855,
360
  "step": 450
361
  },
362
  {
363
- "epoch": 3.9655172413793105,
364
- "grad_norm": 3.6305439472198486,
365
  "learning_rate": 4.58e-05,
366
- "loss": 1.5015,
367
  "step": 460
368
  },
369
  {
370
- "epoch": 4.051724137931035,
371
- "grad_norm": 1.8111392259597778,
372
  "learning_rate": 4.6800000000000006e-05,
373
- "loss": 1.3926,
374
  "step": 470
375
  },
376
  {
377
- "epoch": 4.137931034482759,
378
- "grad_norm": 2.117809295654297,
379
  "learning_rate": 4.78e-05,
380
- "loss": 1.2363,
381
  "step": 480
382
  },
383
  {
384
- "epoch": 4.224137931034483,
385
- "grad_norm": 2.7456188201904297,
386
  "learning_rate": 4.88e-05,
387
- "loss": 1.156,
388
  "step": 490
389
  },
390
  {
391
- "epoch": 4.310344827586207,
392
- "grad_norm": 2.3859684467315674,
393
  "learning_rate": 4.9800000000000004e-05,
394
- "loss": 1.1234,
395
  "step": 500
396
  },
397
  {
398
- "epoch": 4.310344827586207,
399
- "eval_loss": 1.0196667909622192,
400
- "eval_runtime": 39.5574,
401
- "eval_samples_per_second": 33.976,
402
- "eval_steps_per_second": 33.976,
403
- "eval_wer": 0.8392970955112447,
404
  "step": 500
405
  },
406
  {
407
- "epoch": 4.396551724137931,
408
- "grad_norm": 2.3013603687286377,
409
  "learning_rate": 5.08e-05,
410
- "loss": 1.0335,
411
  "step": 510
412
  },
413
  {
414
- "epoch": 4.482758620689655,
415
- "grad_norm": 5.266668796539307,
416
  "learning_rate": 5.1800000000000005e-05,
417
- "loss": 1.002,
418
  "step": 520
419
  },
420
  {
421
- "epoch": 4.568965517241379,
422
- "grad_norm": 2.516477346420288,
423
  "learning_rate": 5.28e-05,
424
- "loss": 0.915,
425
  "step": 530
426
  },
427
  {
428
- "epoch": 4.655172413793103,
429
- "grad_norm": 2.389324903488159,
430
  "learning_rate": 5.380000000000001e-05,
431
- "loss": 0.8828,
432
  "step": 540
433
  },
434
  {
435
- "epoch": 4.741379310344827,
436
- "grad_norm": 2.378732204437256,
437
  "learning_rate": 5.4800000000000004e-05,
438
- "loss": 0.8822,
439
  "step": 550
440
  },
441
  {
442
- "epoch": 4.827586206896552,
443
- "grad_norm": 2.126372814178467,
444
  "learning_rate": 5.580000000000001e-05,
445
- "loss": 0.7924,
446
  "step": 560
447
  },
448
  {
449
- "epoch": 4.913793103448276,
450
- "grad_norm": 2.4130356311798096,
451
  "learning_rate": 5.68e-05,
452
- "loss": 0.7577,
453
  "step": 570
454
  },
455
  {
456
- "epoch": 5.0,
457
- "grad_norm": 4.412507057189941,
458
  "learning_rate": 5.7799999999999995e-05,
459
- "loss": 0.8124,
460
  "step": 580
461
  },
462
  {
463
- "epoch": 5.086206896551724,
464
- "grad_norm": 2.4421167373657227,
465
  "learning_rate": 5.88e-05,
466
- "loss": 0.6599,
467
  "step": 590
468
  },
469
  {
470
- "epoch": 5.172413793103448,
471
- "grad_norm": 2.1765213012695312,
472
  "learning_rate": 5.9800000000000003e-05,
473
- "loss": 0.602,
474
  "step": 600
475
  },
476
  {
477
- "epoch": 5.172413793103448,
478
- "eval_loss": 0.6714467406272888,
479
- "eval_runtime": 39.6477,
480
- "eval_samples_per_second": 33.899,
481
- "eval_steps_per_second": 33.899,
482
- "eval_wer": 0.6600200309569334,
483
  "step": 600
484
  },
485
  {
486
- "epoch": 5.258620689655173,
487
- "grad_norm": 2.5929956436157227,
488
  "learning_rate": 6.08e-05,
489
- "loss": 0.6666,
490
  "step": 610
491
  },
492
  {
493
- "epoch": 5.344827586206897,
494
- "grad_norm": 2.023226261138916,
495
  "learning_rate": 6.18e-05,
496
- "loss": 0.557,
497
  "step": 620
498
  },
499
  {
500
- "epoch": 5.431034482758621,
501
- "grad_norm": 2.348583936691284,
502
  "learning_rate": 6.280000000000001e-05,
503
- "loss": 0.536,
504
  "step": 630
505
  },
506
  {
507
- "epoch": 5.517241379310345,
508
- "grad_norm": 1.9607300758361816,
509
  "learning_rate": 6.38e-05,
510
- "loss": 0.618,
511
  "step": 640
512
  },
513
  {
514
- "epoch": 5.603448275862069,
515
- "grad_norm": 2.136683940887451,
516
  "learning_rate": 6.48e-05,
517
- "loss": 0.5107,
518
  "step": 650
519
  },
520
  {
521
- "epoch": 5.689655172413794,
522
- "grad_norm": 2.5568346977233887,
523
  "learning_rate": 6.58e-05,
524
- "loss": 0.512,
525
  "step": 660
526
  },
527
  {
528
- "epoch": 5.775862068965517,
529
- "grad_norm": 1.6805388927459717,
530
  "learning_rate": 6.680000000000001e-05,
531
- "loss": 0.5663,
532
  "step": 670
533
  },
534
  {
535
- "epoch": 5.862068965517241,
536
- "grad_norm": 3.271713972091675,
537
  "learning_rate": 6.780000000000001e-05,
538
- "loss": 0.4748,
539
  "step": 680
540
  },
541
  {
542
- "epoch": 5.948275862068965,
543
- "grad_norm": 12.830371856689453,
544
  "learning_rate": 6.879999999999999e-05,
545
- "loss": 0.4823,
546
  "step": 690
547
  },
548
  {
549
- "epoch": 6.0344827586206895,
550
- "grad_norm": 1.5290584564208984,
551
  "learning_rate": 6.98e-05,
552
- "loss": 0.5032,
553
  "step": 700
554
  },
555
  {
556
- "epoch": 6.0344827586206895,
557
- "eval_loss": 0.5284600257873535,
558
- "eval_runtime": 39.6104,
559
- "eval_samples_per_second": 33.931,
560
- "eval_steps_per_second": 33.931,
561
- "eval_wer": 0.5658745333697532,
562
  "step": 700
563
  },
564
  {
565
- "epoch": 6.120689655172414,
566
- "grad_norm": 1.902786374092102,
567
  "learning_rate": 7.08e-05,
568
- "loss": 0.387,
569
  "step": 710
570
  },
571
  {
572
- "epoch": 6.206896551724138,
573
- "grad_norm": 2.5520503520965576,
574
  "learning_rate": 7.18e-05,
575
- "loss": 0.3752,
576
  "step": 720
577
  },
578
  {
579
- "epoch": 6.293103448275862,
580
- "grad_norm": 1.8766177892684937,
581
  "learning_rate": 7.280000000000001e-05,
582
- "loss": 0.422,
583
  "step": 730
584
  },
585
  {
586
- "epoch": 6.379310344827586,
587
- "grad_norm": 1.9960404634475708,
588
  "learning_rate": 7.38e-05,
589
- "loss": 0.3703,
590
  "step": 740
591
  },
592
  {
593
- "epoch": 6.4655172413793105,
594
- "grad_norm": 2.9510915279388428,
595
  "learning_rate": 7.48e-05,
596
- "loss": 0.3777,
597
  "step": 750
598
  },
599
  {
600
- "epoch": 6.551724137931035,
601
- "grad_norm": 1.5135278701782227,
602
  "learning_rate": 7.58e-05,
603
- "loss": 0.3978,
604
  "step": 760
605
  },
606
  {
607
- "epoch": 6.637931034482759,
608
- "grad_norm": 2.198090076446533,
609
  "learning_rate": 7.680000000000001e-05,
610
- "loss": 0.306,
611
  "step": 770
612
  },
613
  {
614
- "epoch": 6.724137931034483,
615
- "grad_norm": 4.511343955993652,
616
  "learning_rate": 7.780000000000001e-05,
617
- "loss": 0.4127,
618
  "step": 780
619
  },
620
  {
621
- "epoch": 6.810344827586206,
622
- "grad_norm": 1.5059940814971924,
623
  "learning_rate": 7.88e-05,
624
- "loss": 0.3259,
625
  "step": 790
626
  },
627
  {
628
- "epoch": 6.896551724137931,
629
- "grad_norm": 2.252959728240967,
630
  "learning_rate": 7.98e-05,
631
- "loss": 0.3101,
632
  "step": 800
633
  },
634
  {
635
- "epoch": 6.896551724137931,
636
- "eval_loss": 0.48188281059265137,
637
- "eval_runtime": 39.7499,
638
- "eval_samples_per_second": 33.811,
639
- "eval_steps_per_second": 33.811,
640
- "eval_wer": 0.528179914413184,
641
  "step": 800
642
  },
643
  {
644
- "epoch": 6.982758620689655,
645
- "grad_norm": 2.095763683319092,
646
  "learning_rate": 8.080000000000001e-05,
647
- "loss": 0.3926,
648
  "step": 810
649
  },
650
  {
651
- "epoch": 7.068965517241379,
652
- "grad_norm": 1.679442048072815,
653
  "learning_rate": 8.18e-05,
654
- "loss": 0.3035,
655
  "step": 820
656
  },
657
  {
658
- "epoch": 7.155172413793103,
659
- "grad_norm": 1.725831151008606,
660
  "learning_rate": 8.28e-05,
661
- "loss": 0.2546,
662
  "step": 830
663
  },
664
  {
665
- "epoch": 7.241379310344827,
666
- "grad_norm": 4.802426338195801,
667
  "learning_rate": 8.38e-05,
668
- "loss": 0.3201,
669
  "step": 840
670
  },
671
  {
672
- "epoch": 7.327586206896552,
673
- "grad_norm": 1.871408462524414,
674
  "learning_rate": 8.48e-05,
675
- "loss": 0.2891,
676
  "step": 850
677
  },
678
  {
679
- "epoch": 7.413793103448276,
680
- "grad_norm": 2.0499789714813232,
681
  "learning_rate": 8.58e-05,
682
- "loss": 0.2513,
683
  "step": 860
684
  },
685
  {
686
- "epoch": 7.5,
687
- "grad_norm": 1.3894842863082886,
688
  "learning_rate": 8.680000000000001e-05,
689
- "loss": 0.3401,
690
  "step": 870
691
  },
692
  {
693
- "epoch": 7.586206896551724,
694
- "grad_norm": 1.56522798538208,
695
  "learning_rate": 8.78e-05,
696
- "loss": 0.2398,
697
  "step": 880
698
  },
699
  {
700
- "epoch": 7.672413793103448,
701
- "grad_norm": 1.908964991569519,
702
  "learning_rate": 8.88e-05,
703
- "loss": 0.2656,
704
  "step": 890
705
  },
706
  {
707
- "epoch": 7.758620689655173,
708
- "grad_norm": 1.5925772190093994,
709
  "learning_rate": 8.98e-05,
710
- "loss": 0.3432,
711
  "step": 900
712
  },
713
  {
714
- "epoch": 7.758620689655173,
715
- "eval_loss": 0.46526795625686646,
716
- "eval_runtime": 39.52,
717
- "eval_samples_per_second": 34.008,
718
- "eval_steps_per_second": 34.008,
719
- "eval_wer": 0.5271783665665118,
720
  "step": 900
721
  },
722
  {
723
- "epoch": 7.844827586206897,
724
- "grad_norm": 1.7011960744857788,
725
  "learning_rate": 9.080000000000001e-05,
726
- "loss": 0.2539,
727
  "step": 910
728
  },
729
  {
730
- "epoch": 7.931034482758621,
731
- "grad_norm": 1.718232274055481,
732
  "learning_rate": 9.180000000000001e-05,
733
- "loss": 0.2653,
734
  "step": 920
735
  },
736
  {
737
- "epoch": 8.017241379310345,
738
- "grad_norm": 1.6879725456237793,
739
  "learning_rate": 9.28e-05,
740
- "loss": 0.3313,
741
  "step": 930
742
  },
743
  {
744
- "epoch": 8.10344827586207,
745
- "grad_norm": 2.626966714859009,
746
  "learning_rate": 9.38e-05,
747
- "loss": 0.2116,
748
  "step": 940
749
  },
750
  {
751
- "epoch": 8.189655172413794,
752
- "grad_norm": 2.0977375507354736,
753
  "learning_rate": 9.48e-05,
754
- "loss": 0.2445,
755
  "step": 950
756
  },
757
  {
758
- "epoch": 8.275862068965518,
759
- "grad_norm": 1.6694329977035522,
760
  "learning_rate": 9.58e-05,
761
- "loss": 0.2928,
762
  "step": 960
763
  },
764
  {
765
- "epoch": 8.362068965517242,
766
- "grad_norm": 1.8011162281036377,
767
  "learning_rate": 9.680000000000001e-05,
768
- "loss": 0.2008,
769
  "step": 970
770
  },
771
  {
772
- "epoch": 8.448275862068966,
773
- "grad_norm": 2.6468820571899414,
774
  "learning_rate": 9.78e-05,
775
- "loss": 0.2089,
776
  "step": 980
777
  },
778
  {
779
- "epoch": 8.53448275862069,
780
- "grad_norm": 1.5480060577392578,
781
  "learning_rate": 9.88e-05,
782
- "loss": 0.2593,
783
  "step": 990
784
  },
785
  {
786
- "epoch": 8.620689655172415,
787
- "grad_norm": 2.0152957439422607,
788
  "learning_rate": 9.98e-05,
789
- "loss": 0.1922,
790
  "step": 1000
791
  },
792
  {
793
- "epoch": 8.620689655172415,
794
- "eval_loss": 0.46724727749824524,
795
- "eval_runtime": 39.7331,
796
- "eval_samples_per_second": 33.826,
797
- "eval_steps_per_second": 33.826,
798
- "eval_wer": 0.49175999271601567,
799
  "step": 1000
800
  },
801
  {
802
- "epoch": 8.706896551724139,
803
- "grad_norm": 2.7271480560302734,
804
- "learning_rate": 9.939393939393939e-05,
805
- "loss": 0.2466,
806
  "step": 1010
807
  },
808
  {
809
- "epoch": 8.793103448275861,
810
- "grad_norm": 1.4664121866226196,
811
- "learning_rate": 9.863636363636364e-05,
812
- "loss": 0.259,
813
  "step": 1020
814
  },
815
  {
816
- "epoch": 8.879310344827585,
817
- "grad_norm": 2.0848429203033447,
818
- "learning_rate": 9.787878787878789e-05,
819
- "loss": 0.1964,
820
  "step": 1030
821
  },
822
  {
823
- "epoch": 8.96551724137931,
824
- "grad_norm": 2.68035888671875,
825
- "learning_rate": 9.712121212121212e-05,
826
- "loss": 0.2528,
827
  "step": 1040
828
  },
829
  {
830
- "epoch": 9.051724137931034,
831
- "grad_norm": 1.2563536167144775,
832
- "learning_rate": 9.636363636363637e-05,
833
- "loss": 0.2322,
834
  "step": 1050
835
  },
836
  {
837
- "epoch": 9.137931034482758,
838
- "grad_norm": 1.8890479803085327,
839
- "learning_rate": 9.560606060606061e-05,
840
- "loss": 0.1597,
841
  "step": 1060
842
  },
843
  {
844
- "epoch": 9.224137931034482,
845
- "grad_norm": 2.1262826919555664,
846
- "learning_rate": 9.484848484848486e-05,
847
- "loss": 0.2046,
848
  "step": 1070
849
  },
850
  {
851
- "epoch": 9.310344827586206,
852
- "grad_norm": 1.9750289916992188,
853
- "learning_rate": 9.40909090909091e-05,
854
- "loss": 0.2035,
855
  "step": 1080
856
  },
857
  {
858
- "epoch": 9.39655172413793,
859
- "grad_norm": 1.9659441709518433,
860
- "learning_rate": 9.333333333333334e-05,
861
- "loss": 0.1623,
862
  "step": 1090
863
  },
864
  {
865
- "epoch": 9.482758620689655,
866
- "grad_norm": 3.0013935565948486,
867
- "learning_rate": 9.257575757575758e-05,
868
- "loss": 0.2284,
869
  "step": 1100
870
  },
871
  {
872
- "epoch": 9.482758620689655,
873
- "eval_loss": 0.48344284296035767,
874
- "eval_runtime": 39.5675,
875
- "eval_samples_per_second": 33.967,
876
- "eval_steps_per_second": 33.967,
877
- "eval_wer": 0.48702540289538376,
878
  "step": 1100
879
  },
880
  {
881
- "epoch": 9.568965517241379,
882
- "grad_norm": 1.5623351335525513,
883
- "learning_rate": 9.181818181818183e-05,
884
- "loss": 0.182,
885
  "step": 1110
886
  },
887
  {
888
- "epoch": 9.655172413793103,
889
- "grad_norm": 1.5138722658157349,
890
- "learning_rate": 9.106060606060606e-05,
891
- "loss": 0.1741,
892
  "step": 1120
893
  },
894
  {
895
- "epoch": 9.741379310344827,
896
- "grad_norm": 1.3799549341201782,
897
- "learning_rate": 9.030303030303031e-05,
898
- "loss": 0.2227,
899
  "step": 1130
900
  },
901
  {
902
- "epoch": 9.827586206896552,
903
- "grad_norm": 1.6967511177062988,
904
- "learning_rate": 8.954545454545455e-05,
905
- "loss": 0.1617,
906
  "step": 1140
907
  },
908
  {
909
- "epoch": 9.913793103448276,
910
- "grad_norm": 2.4861340522766113,
911
- "learning_rate": 8.87878787878788e-05,
912
- "loss": 0.1712,
913
  "step": 1150
914
  },
915
  {
916
- "epoch": 10.0,
917
- "grad_norm": 4.324175834655762,
918
- "learning_rate": 8.803030303030304e-05,
919
- "loss": 0.2196,
920
  "step": 1160
921
  },
922
- {
923
- "epoch": 10.086206896551724,
924
- "grad_norm": 1.889611840248108,
925
- "learning_rate": 8.727272727272727e-05,
926
- "loss": 0.1392,
927
- "step": 1170
928
- },
929
- {
930
- "epoch": 10.172413793103448,
931
- "grad_norm": 1.7731642723083496,
932
- "learning_rate": 8.651515151515152e-05,
933
- "loss": 0.147,
934
- "step": 1180
935
- },
936
- {
937
- "epoch": 10.258620689655173,
938
- "grad_norm": 1.3258129358291626,
939
- "learning_rate": 8.575757575757576e-05,
940
- "loss": 0.1793,
941
- "step": 1190
942
- },
943
- {
944
- "epoch": 10.344827586206897,
945
- "grad_norm": 2.139970302581787,
946
- "learning_rate": 8.5e-05,
947
- "loss": 0.1372,
948
- "step": 1200
949
- },
950
- {
951
- "epoch": 10.344827586206897,
952
- "eval_loss": 0.43795427680015564,
953
- "eval_runtime": 39.566,
954
- "eval_samples_per_second": 33.969,
955
- "eval_steps_per_second": 33.969,
956
- "eval_wer": 0.4727305836292452,
957
- "step": 1200
958
- },
959
- {
960
- "epoch": 10.431034482758621,
961
- "grad_norm": 2.1916654109954834,
962
- "learning_rate": 8.424242424242424e-05,
963
- "loss": 0.1442,
964
- "step": 1210
965
- },
966
- {
967
- "epoch": 10.517241379310345,
968
- "grad_norm": 1.1983482837677002,
969
- "learning_rate": 8.348484848484849e-05,
970
- "loss": 0.1926,
971
- "step": 1220
972
- },
973
- {
974
- "epoch": 10.60344827586207,
975
- "grad_norm": 1.9001710414886475,
976
- "learning_rate": 8.272727272727273e-05,
977
- "loss": 0.1368,
978
- "step": 1230
979
- },
980
- {
981
- "epoch": 10.689655172413794,
982
- "grad_norm": 1.9396876096725464,
983
- "learning_rate": 8.196969696969698e-05,
984
- "loss": 0.131,
985
- "step": 1240
986
- },
987
- {
988
- "epoch": 10.775862068965518,
989
- "grad_norm": 1.2405284643173218,
990
- "learning_rate": 8.121212121212121e-05,
991
- "loss": 0.162,
992
- "step": 1250
993
- },
994
- {
995
- "epoch": 10.862068965517242,
996
- "grad_norm": 1.1841609477996826,
997
- "learning_rate": 8.045454545454546e-05,
998
- "loss": 0.1199,
999
- "step": 1260
1000
- },
1001
- {
1002
- "epoch": 10.948275862068966,
1003
- "grad_norm": 1.863655924797058,
1004
- "learning_rate": 7.96969696969697e-05,
1005
- "loss": 0.1571,
1006
- "step": 1270
1007
- },
1008
- {
1009
- "epoch": 11.03448275862069,
1010
- "grad_norm": 1.102363109588623,
1011
- "learning_rate": 7.893939393939395e-05,
1012
- "loss": 0.1672,
1013
- "step": 1280
1014
- },
1015
- {
1016
- "epoch": 11.120689655172415,
1017
- "grad_norm": 1.411650538444519,
1018
- "learning_rate": 7.818181818181818e-05,
1019
- "loss": 0.1004,
1020
- "step": 1290
1021
- },
1022
- {
1023
- "epoch": 11.206896551724139,
1024
- "grad_norm": 1.5993527173995972,
1025
- "learning_rate": 7.742424242424243e-05,
1026
- "loss": 0.1105,
1027
- "step": 1300
1028
- },
1029
- {
1030
- "epoch": 11.206896551724139,
1031
- "eval_loss": 0.45091673731803894,
1032
- "eval_runtime": 39.5057,
1033
- "eval_samples_per_second": 34.02,
1034
- "eval_steps_per_second": 34.02,
1035
- "eval_wer": 0.45943731220977874,
1036
- "step": 1300
1037
- },
1038
- {
1039
- "epoch": 11.293103448275861,
1040
- "grad_norm": 0.9610430598258972,
1041
- "learning_rate": 7.666666666666667e-05,
1042
- "loss": 0.1514,
1043
- "step": 1310
1044
- },
1045
- {
1046
- "epoch": 11.379310344827585,
1047
- "grad_norm": 1.1394908428192139,
1048
- "learning_rate": 7.59090909090909e-05,
1049
- "loss": 0.1087,
1050
- "step": 1320
1051
- },
1052
- {
1053
- "epoch": 11.46551724137931,
1054
- "grad_norm": 2.085495948791504,
1055
- "learning_rate": 7.515151515151515e-05,
1056
- "loss": 0.1328,
1057
- "step": 1330
1058
- },
1059
- {
1060
- "epoch": 11.551724137931034,
1061
- "grad_norm": 1.0012094974517822,
1062
- "learning_rate": 7.439393939393939e-05,
1063
- "loss": 0.1406,
1064
- "step": 1340
1065
- },
1066
- {
1067
- "epoch": 11.637931034482758,
1068
- "grad_norm": 1.3768068552017212,
1069
- "learning_rate": 7.363636363636364e-05,
1070
- "loss": 0.1068,
1071
- "step": 1350
1072
- },
1073
- {
1074
- "epoch": 11.724137931034482,
1075
- "grad_norm": 3.6142425537109375,
1076
- "learning_rate": 7.287878787878788e-05,
1077
- "loss": 0.1447,
1078
- "step": 1360
1079
- },
1080
- {
1081
- "epoch": 11.810344827586206,
1082
- "grad_norm": 1.16116201877594,
1083
- "learning_rate": 7.212121212121213e-05,
1084
- "loss": 0.1099,
1085
- "step": 1370
1086
- },
1087
- {
1088
- "epoch": 11.89655172413793,
1089
- "grad_norm": 1.4963595867156982,
1090
- "learning_rate": 7.136363636363636e-05,
1091
- "loss": 0.098,
1092
- "step": 1380
1093
- },
1094
- {
1095
- "epoch": 11.982758620689655,
1096
- "grad_norm": 1.1868886947631836,
1097
- "learning_rate": 7.060606060606061e-05,
1098
- "loss": 0.1603,
1099
- "step": 1390
1100
- },
1101
- {
1102
- "epoch": 12.068965517241379,
1103
- "grad_norm": 1.6679799556732178,
1104
- "learning_rate": 6.984848484848485e-05,
1105
- "loss": 0.0992,
1106
- "step": 1400
1107
- },
1108
- {
1109
- "epoch": 12.068965517241379,
1110
- "eval_loss": 0.41960862278938293,
1111
- "eval_runtime": 39.5319,
1112
- "eval_samples_per_second": 33.998,
1113
- "eval_steps_per_second": 33.998,
1114
- "eval_wer": 0.4544295729764181,
1115
- "step": 1400
1116
- },
1117
- {
1118
- "epoch": 12.155172413793103,
1119
- "grad_norm": 0.9624470472335815,
1120
- "learning_rate": 6.90909090909091e-05,
1121
- "loss": 0.092,
1122
- "step": 1410
1123
- },
1124
- {
1125
- "epoch": 12.241379310344827,
1126
- "grad_norm": 3.408127546310425,
1127
- "learning_rate": 6.833333333333333e-05,
1128
- "loss": 0.1716,
1129
- "step": 1420
1130
- },
1131
- {
1132
- "epoch": 12.327586206896552,
1133
- "grad_norm": 1.4163497686386108,
1134
- "learning_rate": 6.757575757575758e-05,
1135
- "loss": 0.0885,
1136
- "step": 1430
1137
- },
1138
- {
1139
- "epoch": 12.413793103448276,
1140
- "grad_norm": 1.494627594947815,
1141
- "learning_rate": 6.681818181818183e-05,
1142
- "loss": 0.0891,
1143
- "step": 1440
1144
- },
1145
- {
1146
- "epoch": 12.5,
1147
- "grad_norm": 1.1766793727874756,
1148
- "learning_rate": 6.606060606060607e-05,
1149
- "loss": 0.138,
1150
- "step": 1450
1151
- },
1152
- {
1153
- "epoch": 12.586206896551724,
1154
- "grad_norm": 1.2334414720535278,
1155
- "learning_rate": 6.530303030303032e-05,
1156
- "loss": 0.0946,
1157
- "step": 1460
1158
- },
1159
- {
1160
- "epoch": 12.672413793103448,
1161
- "grad_norm": 2.576019763946533,
1162
- "learning_rate": 6.454545454545455e-05,
1163
- "loss": 0.1043,
1164
- "step": 1470
1165
- },
1166
- {
1167
- "epoch": 12.758620689655173,
1168
- "grad_norm": 0.9484899640083313,
1169
- "learning_rate": 6.37878787878788e-05,
1170
- "loss": 0.1155,
1171
- "step": 1480
1172
- },
1173
- {
1174
- "epoch": 12.844827586206897,
1175
- "grad_norm": 1.0187716484069824,
1176
- "learning_rate": 6.303030303030302e-05,
1177
- "loss": 0.0828,
1178
- "step": 1490
1179
- },
1180
- {
1181
- "epoch": 12.931034482758621,
1182
- "grad_norm": 1.0715346336364746,
1183
- "learning_rate": 6.227272727272727e-05,
1184
- "loss": 0.1226,
1185
- "step": 1500
1186
- },
1187
- {
1188
- "epoch": 12.931034482758621,
1189
- "eval_loss": 0.4237208664417267,
1190
- "eval_runtime": 39.6767,
1191
- "eval_samples_per_second": 33.874,
1192
- "eval_steps_per_second": 33.874,
1193
- "eval_wer": 0.4321223709369025,
1194
- "step": 1500
1195
- },
1196
- {
1197
- "epoch": 13.017241379310345,
1198
- "grad_norm": 0.8969095945358276,
1199
- "learning_rate": 6.151515151515151e-05,
1200
- "loss": 0.1272,
1201
- "step": 1510
1202
- },
1203
- {
1204
- "epoch": 13.10344827586207,
1205
- "grad_norm": 1.4273531436920166,
1206
- "learning_rate": 6.075757575757576e-05,
1207
- "loss": 0.0804,
1208
- "step": 1520
1209
- },
1210
- {
1211
- "epoch": 13.189655172413794,
1212
- "grad_norm": 1.5268211364746094,
1213
- "learning_rate": 6e-05,
1214
- "loss": 0.0907,
1215
- "step": 1530
1216
- },
1217
- {
1218
- "epoch": 13.275862068965518,
1219
- "grad_norm": 1.1891446113586426,
1220
- "learning_rate": 5.9242424242424244e-05,
1221
- "loss": 0.122,
1222
- "step": 1540
1223
- },
1224
- {
1225
- "epoch": 13.362068965517242,
1226
- "grad_norm": 1.2826952934265137,
1227
- "learning_rate": 5.848484848484849e-05,
1228
- "loss": 0.079,
1229
- "step": 1550
1230
- },
1231
- {
1232
- "epoch": 13.448275862068966,
1233
- "grad_norm": 1.4693151712417603,
1234
- "learning_rate": 5.772727272727273e-05,
1235
- "loss": 0.0953,
1236
- "step": 1560
1237
- },
1238
- {
1239
- "epoch": 13.53448275862069,
1240
- "grad_norm": 1.1638069152832031,
1241
- "learning_rate": 5.696969696969697e-05,
1242
- "loss": 0.1076,
1243
- "step": 1570
1244
- },
1245
- {
1246
- "epoch": 13.620689655172415,
1247
- "grad_norm": 2.33617901802063,
1248
- "learning_rate": 5.6212121212121215e-05,
1249
- "loss": 0.0712,
1250
- "step": 1580
1251
- },
1252
- {
1253
- "epoch": 13.706896551724139,
1254
- "grad_norm": 2.3127667903900146,
1255
- "learning_rate": 5.545454545454546e-05,
1256
- "loss": 0.0947,
1257
- "step": 1590
1258
- },
1259
- {
1260
- "epoch": 13.793103448275861,
1261
- "grad_norm": 0.9676732420921326,
1262
- "learning_rate": 5.46969696969697e-05,
1263
- "loss": 0.1013,
1264
- "step": 1600
1265
- },
1266
- {
1267
- "epoch": 13.793103448275861,
1268
- "eval_loss": 0.4112694263458252,
1269
- "eval_runtime": 39.48,
1270
- "eval_samples_per_second": 34.043,
1271
- "eval_steps_per_second": 34.043,
1272
- "eval_wer": 0.4298461258308295,
1273
- "step": 1600
1274
- },
1275
- {
1276
- "epoch": 13.879310344827585,
1277
- "grad_norm": 1.8281100988388062,
1278
- "learning_rate": 5.393939393939394e-05,
1279
- "loss": 0.0696,
1280
- "step": 1610
1281
- },
1282
- {
1283
- "epoch": 13.96551724137931,
1284
- "grad_norm": 2.354693651199341,
1285
- "learning_rate": 5.3181818181818186e-05,
1286
- "loss": 0.1129,
1287
- "step": 1620
1288
- },
1289
- {
1290
- "epoch": 14.051724137931034,
1291
- "grad_norm": 1.0465914011001587,
1292
- "learning_rate": 5.242424242424243e-05,
1293
- "loss": 0.0744,
1294
- "step": 1630
1295
- },
1296
- {
1297
- "epoch": 14.137931034482758,
1298
- "grad_norm": 1.0246185064315796,
1299
- "learning_rate": 5.166666666666667e-05,
1300
- "loss": 0.0656,
1301
- "step": 1640
1302
- },
1303
- {
1304
- "epoch": 14.224137931034482,
1305
- "grad_norm": 2.083202838897705,
1306
- "learning_rate": 5.090909090909091e-05,
1307
- "loss": 0.0869,
1308
- "step": 1650
1309
- },
1310
- {
1311
- "epoch": 14.310344827586206,
1312
- "grad_norm": 1.176715612411499,
1313
- "learning_rate": 5.015151515151515e-05,
1314
- "loss": 0.0821,
1315
- "step": 1660
1316
- },
1317
- {
1318
- "epoch": 14.39655172413793,
1319
- "grad_norm": 1.6385036706924438,
1320
- "learning_rate": 4.93939393939394e-05,
1321
- "loss": 0.0695,
1322
- "step": 1670
1323
- },
1324
- {
1325
- "epoch": 14.482758620689655,
1326
- "grad_norm": 2.735548973083496,
1327
- "learning_rate": 4.863636363636364e-05,
1328
- "loss": 0.0875,
1329
- "step": 1680
1330
- },
1331
- {
1332
- "epoch": 14.568965517241379,
1333
- "grad_norm": 1.3907177448272705,
1334
- "learning_rate": 4.787878787878788e-05,
1335
- "loss": 0.0773,
1336
- "step": 1690
1337
- },
1338
- {
1339
- "epoch": 14.655172413793103,
1340
- "grad_norm": 1.582645297050476,
1341
- "learning_rate": 4.712121212121212e-05,
1342
- "loss": 0.0661,
1343
- "step": 1700
1344
- },
1345
- {
1346
- "epoch": 14.655172413793103,
1347
- "eval_loss": 0.4037941098213196,
1348
- "eval_runtime": 39.716,
1349
- "eval_samples_per_second": 33.84,
1350
- "eval_steps_per_second": 33.84,
1351
- "eval_wer": 0.42756988072475643,
1352
- "step": 1700
1353
- },
1354
- {
1355
- "epoch": 14.741379310344827,
1356
- "grad_norm": 1.1024818420410156,
1357
- "learning_rate": 4.636363636363636e-05,
1358
- "loss": 0.1285,
1359
- "step": 1710
1360
- },
1361
- {
1362
- "epoch": 14.827586206896552,
1363
- "grad_norm": 1.0617228746414185,
1364
- "learning_rate": 4.5606060606060606e-05,
1365
- "loss": 0.0638,
1366
- "step": 1720
1367
- },
1368
- {
1369
- "epoch": 14.913793103448276,
1370
- "grad_norm": 1.8958903551101685,
1371
- "learning_rate": 4.484848484848485e-05,
1372
- "loss": 0.0772,
1373
- "step": 1730
1374
- },
1375
- {
1376
- "epoch": 15.0,
1377
- "grad_norm": 2.5858068466186523,
1378
- "learning_rate": 4.409090909090909e-05,
1379
- "loss": 0.094,
1380
- "step": 1740
1381
- },
1382
- {
1383
- "epoch": 15.086206896551724,
1384
- "grad_norm": 0.6185240745544434,
1385
- "learning_rate": 4.3333333333333334e-05,
1386
- "loss": 0.0506,
1387
- "step": 1750
1388
- },
1389
- {
1390
- "epoch": 15.172413793103448,
1391
- "grad_norm": 1.9594557285308838,
1392
- "learning_rate": 4.257575757575758e-05,
1393
- "loss": 0.0667,
1394
- "step": 1760
1395
- },
1396
- {
1397
- "epoch": 15.258620689655173,
1398
- "grad_norm": 0.6401175260543823,
1399
- "learning_rate": 4.181818181818182e-05,
1400
- "loss": 0.0909,
1401
- "step": 1770
1402
- },
1403
- {
1404
- "epoch": 15.344827586206897,
1405
- "grad_norm": 1.2541193962097168,
1406
- "learning_rate": 4.106060606060606e-05,
1407
- "loss": 0.0573,
1408
- "step": 1780
1409
- },
1410
- {
1411
- "epoch": 15.431034482758621,
1412
- "grad_norm": 1.74609375,
1413
- "learning_rate": 4.0303030303030305e-05,
1414
- "loss": 0.0666,
1415
- "step": 1790
1416
- },
1417
- {
1418
- "epoch": 15.517241379310345,
1419
- "grad_norm": 1.3555305004119873,
1420
- "learning_rate": 3.954545454545455e-05,
1421
- "loss": 0.0901,
1422
- "step": 1800
1423
- },
1424
- {
1425
- "epoch": 15.517241379310345,
1426
- "eval_loss": 0.43212294578552246,
1427
- "eval_runtime": 40.1884,
1428
- "eval_samples_per_second": 33.443,
1429
- "eval_steps_per_second": 33.443,
1430
- "eval_wer": 0.42247109168715286,
1431
- "step": 1800
1432
- },
1433
- {
1434
- "epoch": 15.60344827586207,
1435
- "grad_norm": 0.8425617218017578,
1436
- "learning_rate": 3.878787878787879e-05,
1437
- "loss": 0.0567,
1438
- "step": 1810
1439
- },
1440
- {
1441
- "epoch": 15.689655172413794,
1442
- "grad_norm": 1.145479679107666,
1443
- "learning_rate": 3.803030303030303e-05,
1444
- "loss": 0.0682,
1445
- "step": 1820
1446
- },
1447
- {
1448
- "epoch": 15.775862068965518,
1449
- "grad_norm": 1.0084822177886963,
1450
- "learning_rate": 3.7272727272727276e-05,
1451
- "loss": 0.0847,
1452
- "step": 1830
1453
- },
1454
- {
1455
- "epoch": 15.862068965517242,
1456
- "grad_norm": 0.986998438835144,
1457
- "learning_rate": 3.651515151515152e-05,
1458
- "loss": 0.0623,
1459
- "step": 1840
1460
- },
1461
- {
1462
- "epoch": 15.948275862068966,
1463
- "grad_norm": 1.5837907791137695,
1464
- "learning_rate": 3.575757575757576e-05,
1465
- "loss": 0.0719,
1466
- "step": 1850
1467
- },
1468
- {
1469
- "epoch": 16.03448275862069,
1470
- "grad_norm": 0.8865498900413513,
1471
- "learning_rate": 3.5e-05,
1472
- "loss": 0.0766,
1473
- "step": 1860
1474
- },
1475
- {
1476
- "epoch": 16.120689655172413,
1477
- "grad_norm": 1.3217542171478271,
1478
- "learning_rate": 3.424242424242424e-05,
1479
- "loss": 0.0598,
1480
- "step": 1870
1481
- },
1482
- {
1483
- "epoch": 16.20689655172414,
1484
- "grad_norm": 1.5297330617904663,
1485
- "learning_rate": 3.348484848484848e-05,
1486
- "loss": 0.0632,
1487
- "step": 1880
1488
- },
1489
- {
1490
- "epoch": 16.29310344827586,
1491
- "grad_norm": 0.6415010094642639,
1492
- "learning_rate": 3.272727272727273e-05,
1493
- "loss": 0.0732,
1494
- "step": 1890
1495
- },
1496
- {
1497
- "epoch": 16.379310344827587,
1498
- "grad_norm": 2.088172435760498,
1499
- "learning_rate": 3.1969696969696974e-05,
1500
- "loss": 0.053,
1501
- "step": 1900
1502
- },
1503
- {
1504
- "epoch": 16.379310344827587,
1505
- "eval_loss": 0.4076109826564789,
1506
- "eval_runtime": 40.2143,
1507
- "eval_samples_per_second": 33.421,
1508
- "eval_steps_per_second": 33.421,
1509
- "eval_wer": 0.42356368933806793,
1510
- "step": 1900
1511
- },
1512
- {
1513
- "epoch": 16.46551724137931,
1514
- "grad_norm": 1.5730756521224976,
1515
- "learning_rate": 3.121212121212122e-05,
1516
- "loss": 0.0722,
1517
- "step": 1910
1518
- },
1519
- {
1520
- "epoch": 16.551724137931036,
1521
- "grad_norm": 0.7452704906463623,
1522
- "learning_rate": 3.0454545454545456e-05,
1523
- "loss": 0.0743,
1524
- "step": 1920
1525
- },
1526
- {
1527
- "epoch": 16.637931034482758,
1528
- "grad_norm": 0.890084445476532,
1529
- "learning_rate": 2.96969696969697e-05,
1530
- "loss": 0.0462,
1531
- "step": 1930
1532
- },
1533
- {
1534
- "epoch": 16.724137931034484,
1535
- "grad_norm": 2.400942325592041,
1536
- "learning_rate": 2.893939393939394e-05,
1537
- "loss": 0.076,
1538
- "step": 1940
1539
- },
1540
- {
1541
- "epoch": 16.810344827586206,
1542
- "grad_norm": 1.0481088161468506,
1543
- "learning_rate": 2.818181818181818e-05,
1544
- "loss": 0.0577,
1545
- "step": 1950
1546
- },
1547
- {
1548
- "epoch": 16.896551724137932,
1549
- "grad_norm": 1.7372900247573853,
1550
- "learning_rate": 2.7424242424242424e-05,
1551
- "loss": 0.0539,
1552
- "step": 1960
1553
- },
1554
- {
1555
- "epoch": 16.982758620689655,
1556
- "grad_norm": 0.7519413828849792,
1557
- "learning_rate": 2.6666666666666667e-05,
1558
- "loss": 0.076,
1559
- "step": 1970
1560
- },
1561
- {
1562
- "epoch": 17.06896551724138,
1563
- "grad_norm": 1.1846277713775635,
1564
- "learning_rate": 2.590909090909091e-05,
1565
- "loss": 0.0553,
1566
- "step": 1980
1567
- },
1568
- {
1569
- "epoch": 17.155172413793103,
1570
- "grad_norm": 0.7746219038963318,
1571
- "learning_rate": 2.5151515151515155e-05,
1572
- "loss": 0.0617,
1573
- "step": 1990
1574
- },
1575
- {
1576
- "epoch": 17.24137931034483,
1577
- "grad_norm": 1.951456904411316,
1578
- "learning_rate": 2.4393939393939395e-05,
1579
- "loss": 0.0805,
1580
- "step": 2000
1581
- },
1582
- {
1583
- "epoch": 17.24137931034483,
1584
- "eval_loss": 0.43357911705970764,
1585
- "eval_runtime": 40.2275,
1586
- "eval_samples_per_second": 33.41,
1587
- "eval_steps_per_second": 33.41,
1588
- "eval_wer": 0.4156423563689338,
1589
- "step": 2000
1590
- },
1591
- {
1592
- "epoch": 17.32758620689655,
1593
- "grad_norm": 0.5796881318092346,
1594
- "learning_rate": 2.3636363636363637e-05,
1595
- "loss": 0.0466,
1596
- "step": 2010
1597
- },
1598
- {
1599
- "epoch": 17.413793103448278,
1600
- "grad_norm": 1.2703924179077148,
1601
- "learning_rate": 2.287878787878788e-05,
1602
- "loss": 0.0557,
1603
- "step": 2020
1604
- },
1605
- {
1606
- "epoch": 17.5,
1607
- "grad_norm": 1.242013692855835,
1608
- "learning_rate": 2.2121212121212123e-05,
1609
- "loss": 0.0719,
1610
- "step": 2030
1611
- },
1612
- {
1613
- "epoch": 17.586206896551722,
1614
- "grad_norm": 1.1339091062545776,
1615
- "learning_rate": 2.1363636363636362e-05,
1616
- "loss": 0.0488,
1617
- "step": 2040
1618
- },
1619
- {
1620
- "epoch": 17.67241379310345,
1621
- "grad_norm": 1.4959948062896729,
1622
- "learning_rate": 2.0606060606060608e-05,
1623
- "loss": 0.0522,
1624
- "step": 2050
1625
- },
1626
- {
1627
- "epoch": 17.75862068965517,
1628
- "grad_norm": 0.9912744760513306,
1629
- "learning_rate": 1.984848484848485e-05,
1630
- "loss": 0.0777,
1631
- "step": 2060
1632
- },
1633
- {
1634
- "epoch": 17.844827586206897,
1635
- "grad_norm": 1.1574323177337646,
1636
- "learning_rate": 1.9090909090909094e-05,
1637
- "loss": 0.0517,
1638
- "step": 2070
1639
- },
1640
- {
1641
- "epoch": 17.93103448275862,
1642
- "grad_norm": 1.5049959421157837,
1643
- "learning_rate": 1.8333333333333333e-05,
1644
- "loss": 0.0565,
1645
- "step": 2080
1646
- },
1647
- {
1648
- "epoch": 18.017241379310345,
1649
- "grad_norm": 0.6796301603317261,
1650
- "learning_rate": 1.7575757575757576e-05,
1651
- "loss": 0.0618,
1652
- "step": 2090
1653
- },
1654
- {
1655
- "epoch": 18.103448275862068,
1656
- "grad_norm": 0.8808640837669373,
1657
- "learning_rate": 1.6818181818181818e-05,
1658
- "loss": 0.049,
1659
- "step": 2100
1660
- },
1661
- {
1662
- "epoch": 18.103448275862068,
1663
- "eval_loss": 0.4193398654460907,
1664
- "eval_runtime": 40.1393,
1665
- "eval_samples_per_second": 33.483,
1666
- "eval_steps_per_second": 33.483,
1667
- "eval_wer": 0.41136301556951654,
1668
- "step": 2100
1669
- },
1670
- {
1671
- "epoch": 18.189655172413794,
1672
- "grad_norm": 1.1478749513626099,
1673
- "learning_rate": 1.606060606060606e-05,
1674
- "loss": 0.0467,
1675
- "step": 2110
1676
- },
1677
- {
1678
- "epoch": 18.275862068965516,
1679
- "grad_norm": 0.9449758529663086,
1680
- "learning_rate": 1.5303030303030304e-05,
1681
- "loss": 0.074,
1682
- "step": 2120
1683
- },
1684
- {
1685
- "epoch": 18.362068965517242,
1686
- "grad_norm": 0.7945724725723267,
1687
- "learning_rate": 1.4545454545454545e-05,
1688
- "loss": 0.0415,
1689
- "step": 2130
1690
- },
1691
- {
1692
- "epoch": 18.448275862068964,
1693
- "grad_norm": 1.1694364547729492,
1694
- "learning_rate": 1.3787878787878789e-05,
1695
- "loss": 0.0547,
1696
- "step": 2140
1697
- },
1698
- {
1699
- "epoch": 18.53448275862069,
1700
- "grad_norm": 0.7521646022796631,
1701
- "learning_rate": 1.3030303030303032e-05,
1702
- "loss": 0.0544,
1703
- "step": 2150
1704
- },
1705
- {
1706
- "epoch": 18.620689655172413,
1707
- "grad_norm": 0.9526033997535706,
1708
- "learning_rate": 1.2272727272727273e-05,
1709
- "loss": 0.0507,
1710
- "step": 2160
1711
- },
1712
- {
1713
- "epoch": 18.70689655172414,
1714
- "grad_norm": 1.0075325965881348,
1715
- "learning_rate": 1.1515151515151517e-05,
1716
- "loss": 0.0595,
1717
- "step": 2170
1718
- },
1719
- {
1720
- "epoch": 18.79310344827586,
1721
- "grad_norm": 0.6182655096054077,
1722
- "learning_rate": 1.0757575757575758e-05,
1723
- "loss": 0.0593,
1724
- "step": 2180
1725
- },
1726
- {
1727
- "epoch": 18.879310344827587,
1728
- "grad_norm": 0.5009527206420898,
1729
- "learning_rate": 1e-05,
1730
- "loss": 0.0415,
1731
- "step": 2190
1732
- },
1733
- {
1734
- "epoch": 18.96551724137931,
1735
- "grad_norm": 2.459338426589966,
1736
- "learning_rate": 9.242424242424244e-06,
1737
- "loss": 0.0717,
1738
- "step": 2200
1739
- },
1740
- {
1741
- "epoch": 18.96551724137931,
1742
- "eval_loss": 0.4138683080673218,
1743
- "eval_runtime": 40.2569,
1744
- "eval_samples_per_second": 33.386,
1745
- "eval_steps_per_second": 33.386,
1746
- "eval_wer": 0.4090867704634435,
1747
- "step": 2200
1748
- },
1749
- {
1750
- "epoch": 19.051724137931036,
1751
- "grad_norm": 1.2094837427139282,
1752
- "learning_rate": 8.484848484848486e-06,
1753
- "loss": 0.0486,
1754
- "step": 2210
1755
- },
1756
- {
1757
- "epoch": 19.137931034482758,
1758
- "grad_norm": 1.1685006618499756,
1759
- "learning_rate": 7.727272727272727e-06,
1760
- "loss": 0.0376,
1761
- "step": 2220
1762
- },
1763
- {
1764
- "epoch": 19.224137931034484,
1765
- "grad_norm": 0.764283299446106,
1766
- "learning_rate": 6.969696969696971e-06,
1767
- "loss": 0.0598,
1768
- "step": 2230
1769
- },
1770
- {
1771
- "epoch": 19.310344827586206,
1772
- "grad_norm": 0.6357612013816833,
1773
- "learning_rate": 6.212121212121212e-06,
1774
- "loss": 0.0494,
1775
- "step": 2240
1776
- },
1777
- {
1778
- "epoch": 19.396551724137932,
1779
- "grad_norm": 2.4395949840545654,
1780
- "learning_rate": 5.4545454545454545e-06,
1781
- "loss": 0.0395,
1782
- "step": 2250
1783
- },
1784
- {
1785
- "epoch": 19.482758620689655,
1786
- "grad_norm": 2.8990790843963623,
1787
- "learning_rate": 4.696969696969697e-06,
1788
- "loss": 0.0643,
1789
- "step": 2260
1790
- },
1791
- {
1792
- "epoch": 19.56896551724138,
1793
- "grad_norm": 0.5647782683372498,
1794
- "learning_rate": 3.939393939393939e-06,
1795
- "loss": 0.0458,
1796
- "step": 2270
1797
- },
1798
- {
1799
- "epoch": 19.655172413793103,
1800
- "grad_norm": 0.6177101731300354,
1801
- "learning_rate": 3.1818181818181817e-06,
1802
- "loss": 0.0521,
1803
- "step": 2280
1804
- },
1805
- {
1806
- "epoch": 19.74137931034483,
1807
- "grad_norm": 0.9529480934143066,
1808
- "learning_rate": 2.4242424242424244e-06,
1809
- "loss": 0.0497,
1810
- "step": 2290
1811
- },
1812
- {
1813
- "epoch": 19.82758620689655,
1814
- "grad_norm": 0.7542730569839478,
1815
- "learning_rate": 1.6666666666666667e-06,
1816
- "loss": 0.0389,
1817
- "step": 2300
1818
- },
1819
- {
1820
- "epoch": 19.82758620689655,
1821
- "eval_loss": 0.4216426610946655,
1822
- "eval_runtime": 39.4724,
1823
- "eval_samples_per_second": 34.049,
1824
- "eval_steps_per_second": 34.049,
1825
- "eval_wer": 0.4087225712464718,
1826
- "step": 2300
1827
- },
1828
- {
1829
- "epoch": 19.913793103448278,
1830
- "grad_norm": 0.9613261222839355,
1831
- "learning_rate": 9.09090909090909e-07,
1832
- "loss": 0.0483,
1833
- "step": 2310
1834
- },
1835
- {
1836
- "epoch": 20.0,
1837
- "grad_norm": 1.2119308710098267,
1838
- "learning_rate": 1.5151515151515152e-07,
1839
- "loss": 0.0557,
1840
- "step": 2320
1841
- },
1842
  {
1843
  "epoch": 20.0,
1844
- "step": 2320,
1845
- "total_flos": 2.1476719263248095e+18,
1846
- "train_loss": 0.8590125822430027,
1847
- "train_runtime": 3151.1477,
1848
- "train_samples_per_second": 23.458,
1849
- "train_steps_per_second": 0.736
1850
  }
1851
  ],
1852
  "logging_steps": 10,
1853
- "max_steps": 2320,
1854
  "num_input_tokens_seen": 0,
1855
  "num_train_epochs": 20,
1856
  "save_steps": 400,
@@ -1866,8 +946,8 @@
1866
  "attributes": {}
1867
  }
1868
  },
1869
- "total_flos": 2.1476719263248095e+18,
1870
- "train_batch_size": 32,
1871
  "trial_name": null,
1872
  "trial_params": null
1873
  }
 
3
  "best_model_checkpoint": null,
4
  "epoch": 20.0,
5
  "eval_steps": 100,
6
+ "global_step": 1160,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.1724137931034483,
13
+ "grad_norm": 9.273743629455566,
14
  "learning_rate": 9e-07,
15
+ "loss": 9.21,
16
  "step": 10
17
  },
18
  {
19
+ "epoch": 0.3448275862068966,
20
+ "grad_norm": 10.659689903259277,
21
  "learning_rate": 1.9e-06,
22
+ "loss": 9.0596,
23
  "step": 20
24
  },
25
  {
26
+ "epoch": 0.5172413793103449,
27
+ "grad_norm": 13.765968322753906,
28
  "learning_rate": 2.8000000000000003e-06,
29
+ "loss": 8.6308,
30
  "step": 30
31
  },
32
  {
33
+ "epoch": 0.6896551724137931,
34
+ "grad_norm": 13.644205093383789,
35
  "learning_rate": 3.8e-06,
36
+ "loss": 7.9066,
37
  "step": 40
38
  },
39
  {
40
+ "epoch": 0.8620689655172413,
41
+ "grad_norm": 18.20844078063965,
42
  "learning_rate": 4.800000000000001e-06,
43
+ "loss": 7.8077,
44
  "step": 50
45
  },
46
  {
47
+ "epoch": 1.0344827586206897,
48
+ "grad_norm": 18.099712371826172,
49
  "learning_rate": 5.8e-06,
50
+ "loss": 6.6628,
51
  "step": 60
52
  },
53
  {
54
+ "epoch": 1.206896551724138,
55
+ "grad_norm": 10.263163566589355,
56
  "learning_rate": 6.800000000000001e-06,
57
+ "loss": 4.6321,
58
  "step": 70
59
  },
60
  {
61
+ "epoch": 1.3793103448275863,
62
+ "grad_norm": 4.330440998077393,
63
  "learning_rate": 7.8e-06,
64
+ "loss": 3.9137,
65
  "step": 80
66
  },
67
  {
68
+ "epoch": 1.5517241379310345,
69
+ "grad_norm": 1.776586651802063,
70
  "learning_rate": 8.8e-06,
71
+ "loss": 3.3387,
72
  "step": 90
73
  },
74
  {
75
+ "epoch": 1.7241379310344827,
76
+ "grad_norm": 4.104515075683594,
77
  "learning_rate": 9.800000000000001e-06,
78
+ "loss": 3.158,
79
  "step": 100
80
  },
81
  {
82
+ "epoch": 1.7241379310344827,
83
+ "eval_loss": 3.6802542209625244,
84
+ "eval_runtime": 14.7124,
85
+ "eval_samples_per_second": 91.352,
86
+ "eval_steps_per_second": 2.855,
87
  "eval_wer": 1.0,
88
  "step": 100
89
  },
90
  {
91
+ "epoch": 1.896551724137931,
92
+ "grad_norm": 0.8036434650421143,
93
  "learning_rate": 1.08e-05,
94
+ "loss": 3.0952,
95
  "step": 110
96
  },
97
  {
98
+ "epoch": 2.0689655172413794,
99
+ "grad_norm": 0.9299842119216919,
100
  "learning_rate": 1.18e-05,
101
+ "loss": 3.0977,
102
  "step": 120
103
  },
104
  {
105
+ "epoch": 2.2413793103448274,
106
+ "grad_norm": 2.4941225051879883,
107
  "learning_rate": 1.2800000000000001e-05,
108
+ "loss": 3.0263,
109
  "step": 130
110
  },
111
  {
112
+ "epoch": 2.413793103448276,
113
+ "grad_norm": 1.423953652381897,
114
  "learning_rate": 1.3800000000000002e-05,
115
+ "loss": 3.0224,
116
  "step": 140
117
  },
118
  {
119
+ "epoch": 2.586206896551724,
120
+ "grad_norm": 1.3000880479812622,
121
  "learning_rate": 1.48e-05,
122
+ "loss": 3.0211,
123
  "step": 150
124
  },
125
  {
126
+ "epoch": 2.7586206896551726,
127
+ "grad_norm": 1.0403013229370117,
128
  "learning_rate": 1.58e-05,
129
+ "loss": 3.0063,
130
  "step": 160
131
  },
132
  {
133
+ "epoch": 2.9310344827586206,
134
+ "grad_norm": 0.8328425884246826,
135
  "learning_rate": 1.6800000000000002e-05,
136
+ "loss": 2.9927,
137
  "step": 170
138
  },
139
  {
140
+ "epoch": 3.103448275862069,
141
+ "grad_norm": 1.010672926902771,
142
  "learning_rate": 1.78e-05,
143
+ "loss": 3.0028,
144
  "step": 180
145
  },
146
  {
147
+ "epoch": 3.2758620689655173,
148
+ "grad_norm": 2.069948434829712,
149
  "learning_rate": 1.88e-05,
150
+ "loss": 2.9785,
151
  "step": 190
152
  },
153
  {
154
+ "epoch": 3.4482758620689653,
155
+ "grad_norm": 1.1152503490447998,
156
  "learning_rate": 1.9800000000000004e-05,
157
+ "loss": 2.9744,
158
  "step": 200
159
  },
160
  {
161
+ "epoch": 3.4482758620689653,
162
+ "eval_loss": 3.1165249347686768,
163
+ "eval_runtime": 14.3806,
164
+ "eval_samples_per_second": 93.459,
165
+ "eval_steps_per_second": 2.921,
166
  "eval_wer": 1.0,
167
  "step": 200
168
  },
169
  {
170
+ "epoch": 3.6206896551724137,
171
+ "grad_norm": 1.179228663444519,
172
  "learning_rate": 2.08e-05,
173
+ "loss": 2.9759,
174
  "step": 210
175
  },
176
  {
177
+ "epoch": 3.793103448275862,
178
+ "grad_norm": 0.9685543179512024,
179
  "learning_rate": 2.18e-05,
180
+ "loss": 2.9684,
181
  "step": 220
182
  },
183
  {
184
+ "epoch": 3.9655172413793105,
185
+ "grad_norm": 0.710486114025116,
186
  "learning_rate": 2.2800000000000002e-05,
187
+ "loss": 2.9692,
188
  "step": 230
189
  },
190
  {
191
+ "epoch": 4.137931034482759,
192
+ "grad_norm": 0.4123358726501465,
193
  "learning_rate": 2.38e-05,
194
+ "loss": 2.9609,
195
  "step": 240
196
  },
197
  {
198
+ "epoch": 4.310344827586207,
199
+ "grad_norm": 1.9703031778335571,
200
  "learning_rate": 2.48e-05,
201
+ "loss": 2.9573,
202
  "step": 250
203
  },
204
  {
205
+ "epoch": 4.482758620689655,
206
+ "grad_norm": 1.264237403869629,
207
  "learning_rate": 2.58e-05,
208
+ "loss": 2.952,
209
  "step": 260
210
  },
211
  {
212
+ "epoch": 4.655172413793103,
213
+ "grad_norm": 0.4067547619342804,
214
  "learning_rate": 2.6800000000000004e-05,
215
+ "loss": 2.9432,
216
  "step": 270
217
  },
218
  {
219
+ "epoch": 4.827586206896552,
220
+ "grad_norm": 0.5488032698631287,
221
  "learning_rate": 2.7800000000000005e-05,
222
+ "loss": 2.9385,
223
  "step": 280
224
  },
225
  {
226
+ "epoch": 5.0,
227
+ "grad_norm": 1.3628854751586914,
228
  "learning_rate": 2.88e-05,
229
+ "loss": 2.9384,
230
  "step": 290
231
  },
232
  {
233
+ "epoch": 5.172413793103448,
234
+ "grad_norm": 0.5179504752159119,
235
  "learning_rate": 2.98e-05,
236
+ "loss": 2.9266,
237
  "step": 300
238
  },
239
  {
240
+ "epoch": 5.172413793103448,
241
+ "eval_loss": 3.017453908920288,
242
+ "eval_runtime": 14.4537,
243
+ "eval_samples_per_second": 92.987,
244
+ "eval_steps_per_second": 2.906,
245
  "eval_wer": 1.0,
246
  "step": 300
247
  },
248
  {
249
+ "epoch": 5.344827586206897,
250
+ "grad_norm": 1.5788885354995728,
251
  "learning_rate": 3.08e-05,
252
+ "loss": 2.9229,
253
  "step": 310
254
  },
255
  {
256
+ "epoch": 5.517241379310345,
257
+ "grad_norm": 1.777976155281067,
258
  "learning_rate": 3.18e-05,
259
+ "loss": 2.8996,
260
  "step": 320
261
  },
262
  {
263
+ "epoch": 5.689655172413794,
264
+ "grad_norm": 0.7765872478485107,
265
  "learning_rate": 3.2800000000000004e-05,
266
  "loss": 2.8698,
267
  "step": 330
268
  },
269
  {
270
+ "epoch": 5.862068965517241,
271
+ "grad_norm": 1.5203367471694946,
272
  "learning_rate": 3.38e-05,
273
+ "loss": 2.8412,
274
  "step": 340
275
  },
276
  {
277
+ "epoch": 6.0344827586206895,
278
+ "grad_norm": 3.4122729301452637,
279
  "learning_rate": 3.48e-05,
280
+ "loss": 2.7877,
281
  "step": 350
282
  },
283
  {
284
+ "epoch": 6.206896551724138,
285
+ "grad_norm": 1.2957384586334229,
286
  "learning_rate": 3.58e-05,
287
+ "loss": 2.6514,
288
  "step": 360
289
  },
290
  {
291
+ "epoch": 6.379310344827586,
292
+ "grad_norm": 1.9932421445846558,
293
  "learning_rate": 3.68e-05,
294
+ "loss": 2.5492,
295
  "step": 370
296
  },
297
  {
298
+ "epoch": 6.551724137931035,
299
+ "grad_norm": 2.7136945724487305,
300
  "learning_rate": 3.7800000000000004e-05,
301
+ "loss": 2.3894,
302
  "step": 380
303
  },
304
  {
305
+ "epoch": 6.724137931034483,
306
+ "grad_norm": 3.479182720184326,
307
  "learning_rate": 3.88e-05,
308
+ "loss": 2.2604,
309
  "step": 390
310
  },
311
  {
312
+ "epoch": 6.896551724137931,
313
+ "grad_norm": 1.5610332489013672,
314
  "learning_rate": 3.9800000000000005e-05,
315
+ "loss": 2.1336,
316
  "step": 400
317
  },
318
  {
319
+ "epoch": 6.896551724137931,
320
+ "eval_loss": 2.2134621143341064,
321
+ "eval_runtime": 14.4845,
322
+ "eval_samples_per_second": 92.789,
323
+ "eval_steps_per_second": 2.9,
324
+ "eval_wer": 1.011654374943094,
325
  "step": 400
326
  },
327
  {
328
+ "epoch": 7.068965517241379,
329
+ "grad_norm": 1.7063915729522705,
330
  "learning_rate": 4.08e-05,
331
+ "loss": 2.0117,
332
  "step": 410
333
  },
334
  {
335
+ "epoch": 7.241379310344827,
336
+ "grad_norm": 2.4562268257141113,
337
  "learning_rate": 4.18e-05,
338
+ "loss": 1.827,
339
  "step": 420
340
  },
341
  {
342
+ "epoch": 7.413793103448276,
343
+ "grad_norm": 2.533010482788086,
344
  "learning_rate": 4.2800000000000004e-05,
345
+ "loss": 1.6851,
346
  "step": 430
347
  },
348
  {
349
+ "epoch": 7.586206896551724,
350
+ "grad_norm": 3.3842196464538574,
351
  "learning_rate": 4.38e-05,
352
+ "loss": 1.5765,
353
  "step": 440
354
  },
355
  {
356
+ "epoch": 7.758620689655173,
357
+ "grad_norm": 4.379862308502197,
358
  "learning_rate": 4.4800000000000005e-05,
359
+ "loss": 1.4748,
360
  "step": 450
361
  },
362
  {
363
+ "epoch": 7.931034482758621,
364
+ "grad_norm": 2.6355247497558594,
365
  "learning_rate": 4.58e-05,
366
+ "loss": 1.3422,
367
  "step": 460
368
  },
369
  {
370
+ "epoch": 8.10344827586207,
371
+ "grad_norm": 3.2931857109069824,
372
  "learning_rate": 4.6800000000000006e-05,
373
+ "loss": 1.2681,
374
  "step": 470
375
  },
376
  {
377
+ "epoch": 8.275862068965518,
378
+ "grad_norm": 1.8078395128250122,
379
  "learning_rate": 4.78e-05,
380
+ "loss": 1.1503,
381
  "step": 480
382
  },
383
  {
384
+ "epoch": 8.448275862068966,
385
+ "grad_norm": 2.258090019226074,
386
  "learning_rate": 4.88e-05,
387
+ "loss": 1.0564,
388
  "step": 490
389
  },
390
  {
391
+ "epoch": 8.620689655172415,
392
+ "grad_norm": 2.1892685890197754,
393
  "learning_rate": 4.9800000000000004e-05,
394
+ "loss": 1.0119,
395
  "step": 500
396
  },
397
  {
398
+ "epoch": 8.620689655172415,
399
+ "eval_loss": 1.022745966911316,
400
+ "eval_runtime": 14.558,
401
+ "eval_samples_per_second": 92.321,
402
+ "eval_steps_per_second": 2.885,
403
+ "eval_wer": 0.825093326049349,
404
  "step": 500
405
  },
406
  {
407
+ "epoch": 8.793103448275861,
408
+ "grad_norm": 1.7447644472122192,
409
  "learning_rate": 5.08e-05,
410
+ "loss": 0.9426,
411
  "step": 510
412
  },
413
  {
414
+ "epoch": 8.96551724137931,
415
+ "grad_norm": 2.7097115516662598,
416
  "learning_rate": 5.1800000000000005e-05,
417
+ "loss": 0.8773,
418
  "step": 520
419
  },
420
  {
421
+ "epoch": 9.137931034482758,
422
+ "grad_norm": 1.6821448802947998,
423
  "learning_rate": 5.28e-05,
424
+ "loss": 0.7822,
425
  "step": 530
426
  },
427
  {
428
+ "epoch": 9.310344827586206,
429
+ "grad_norm": 1.5076041221618652,
430
  "learning_rate": 5.380000000000001e-05,
431
+ "loss": 0.7334,
432
  "step": 540
433
  },
434
  {
435
+ "epoch": 9.482758620689655,
436
+ "grad_norm": 2.951249361038208,
437
  "learning_rate": 5.4800000000000004e-05,
438
+ "loss": 0.6852,
439
  "step": 550
440
  },
441
  {
442
+ "epoch": 9.655172413793103,
443
+ "grad_norm": 1.5876814126968384,
444
  "learning_rate": 5.580000000000001e-05,
445
+ "loss": 0.653,
446
  "step": 560
447
  },
448
  {
449
+ "epoch": 9.827586206896552,
450
+ "grad_norm": 2.462275743484497,
451
  "learning_rate": 5.68e-05,
452
+ "loss": 0.6415,
453
  "step": 570
454
  },
455
  {
456
+ "epoch": 10.0,
457
+ "grad_norm": 1.9461971521377563,
458
  "learning_rate": 5.7799999999999995e-05,
459
+ "loss": 0.6092,
460
  "step": 580
461
  },
462
  {
463
+ "epoch": 10.172413793103448,
464
+ "grad_norm": 1.474002480506897,
465
  "learning_rate": 5.88e-05,
466
+ "loss": 0.5086,
467
  "step": 590
468
  },
469
  {
470
+ "epoch": 10.344827586206897,
471
+ "grad_norm": 1.8599964380264282,
472
  "learning_rate": 5.9800000000000003e-05,
473
+ "loss": 0.4995,
474
  "step": 600
475
  },
476
  {
477
+ "epoch": 10.344827586206897,
478
+ "eval_loss": 0.7699851989746094,
479
+ "eval_runtime": 14.4827,
480
+ "eval_samples_per_second": 92.8,
481
+ "eval_steps_per_second": 2.9,
482
+ "eval_wer": 0.6573795866338887,
483
  "step": 600
484
  },
485
  {
486
+ "epoch": 10.517241379310345,
487
+ "grad_norm": 1.4216625690460205,
488
  "learning_rate": 6.08e-05,
489
+ "loss": 0.4884,
490
  "step": 610
491
  },
492
  {
493
+ "epoch": 10.689655172413794,
494
+ "grad_norm": 2.1162352561950684,
495
  "learning_rate": 6.18e-05,
496
+ "loss": 0.4398,
497
  "step": 620
498
  },
499
  {
500
+ "epoch": 10.862068965517242,
501
+ "grad_norm": 1.4033368825912476,
502
  "learning_rate": 6.280000000000001e-05,
503
+ "loss": 0.4458,
504
  "step": 630
505
  },
506
  {
507
+ "epoch": 11.03448275862069,
508
+ "grad_norm": 1.6064072847366333,
509
  "learning_rate": 6.38e-05,
510
+ "loss": 0.4414,
511
  "step": 640
512
  },
513
  {
514
+ "epoch": 11.206896551724139,
515
+ "grad_norm": 2.22825288772583,
516
  "learning_rate": 6.48e-05,
517
+ "loss": 0.3528,
518
  "step": 650
519
  },
520
  {
521
+ "epoch": 11.379310344827585,
522
+ "grad_norm": 1.4636731147766113,
523
  "learning_rate": 6.58e-05,
524
+ "loss": 0.3699,
525
  "step": 660
526
  },
527
  {
528
+ "epoch": 11.551724137931034,
529
+ "grad_norm": 1.3389075994491577,
530
  "learning_rate": 6.680000000000001e-05,
531
+ "loss": 0.3621,
532
  "step": 670
533
  },
534
  {
535
+ "epoch": 11.724137931034482,
536
+ "grad_norm": 2.6080307960510254,
537
  "learning_rate": 6.780000000000001e-05,
538
+ "loss": 0.338,
539
  "step": 680
540
  },
541
  {
542
+ "epoch": 11.89655172413793,
543
+ "grad_norm": 1.3124964237213135,
544
  "learning_rate": 6.879999999999999e-05,
545
+ "loss": 0.3143,
546
  "step": 690
547
  },
548
  {
549
+ "epoch": 12.068965517241379,
550
+ "grad_norm": 1.1012139320373535,
551
  "learning_rate": 6.98e-05,
552
+ "loss": 0.3233,
553
  "step": 700
554
  },
555
  {
556
+ "epoch": 12.068965517241379,
557
+ "eval_loss": 0.49695342779159546,
558
+ "eval_runtime": 14.4504,
559
+ "eval_samples_per_second": 93.008,
560
+ "eval_steps_per_second": 2.906,
561
+ "eval_wer": 0.5240826732222525,
562
  "step": 700
563
  },
564
  {
565
+ "epoch": 12.241379310344827,
566
+ "grad_norm": 2.3211851119995117,
567
  "learning_rate": 7.08e-05,
568
+ "loss": 0.2827,
569
  "step": 710
570
  },
571
  {
572
+ "epoch": 12.413793103448276,
573
+ "grad_norm": 1.2446343898773193,
574
  "learning_rate": 7.18e-05,
575
+ "loss": 0.2478,
576
  "step": 720
577
  },
578
  {
579
+ "epoch": 12.586206896551724,
580
+ "grad_norm": 1.1632747650146484,
581
  "learning_rate": 7.280000000000001e-05,
582
+ "loss": 0.2761,
583
  "step": 730
584
  },
585
  {
586
+ "epoch": 12.758620689655173,
587
+ "grad_norm": 1.39242684841156,
588
  "learning_rate": 7.38e-05,
589
+ "loss": 0.2893,
590
  "step": 740
591
  },
592
  {
593
+ "epoch": 12.931034482758621,
594
+ "grad_norm": 1.546246886253357,
595
  "learning_rate": 7.48e-05,
596
+ "loss": 0.2382,
597
  "step": 750
598
  },
599
  {
600
+ "epoch": 13.10344827586207,
601
+ "grad_norm": 1.10295832157135,
602
  "learning_rate": 7.58e-05,
603
+ "loss": 0.2563,
604
  "step": 760
605
  },
606
  {
607
+ "epoch": 13.275862068965518,
608
+ "grad_norm": 1.3670897483825684,
609
  "learning_rate": 7.680000000000001e-05,
610
+ "loss": 0.2438,
611
  "step": 770
612
  },
613
  {
614
+ "epoch": 13.448275862068966,
615
+ "grad_norm": 1.2440053224563599,
616
  "learning_rate": 7.780000000000001e-05,
617
+ "loss": 0.2149,
618
  "step": 780
619
  },
620
  {
621
+ "epoch": 13.620689655172415,
622
+ "grad_norm": 1.1088565587997437,
623
  "learning_rate": 7.88e-05,
624
+ "loss": 0.2341,
625
  "step": 790
626
  },
627
  {
628
+ "epoch": 13.793103448275861,
629
+ "grad_norm": 1.2869517803192139,
630
  "learning_rate": 7.98e-05,
631
+ "loss": 0.2452,
632
  "step": 800
633
  },
634
  {
635
+ "epoch": 13.793103448275861,
636
+ "eval_loss": 0.4584949016571045,
637
+ "eval_runtime": 14.4542,
638
+ "eval_samples_per_second": 92.983,
639
+ "eval_steps_per_second": 2.906,
640
+ "eval_wer": 0.49084949467358646,
641
  "step": 800
642
  },
643
  {
644
+ "epoch": 13.96551724137931,
645
+ "grad_norm": 1.8122053146362305,
646
  "learning_rate": 8.080000000000001e-05,
647
+ "loss": 0.2153,
648
  "step": 810
649
  },
650
  {
651
+ "epoch": 14.137931034482758,
652
+ "grad_norm": 1.5954018831253052,
653
  "learning_rate": 8.18e-05,
654
+ "loss": 0.1834,
655
  "step": 820
656
  },
657
  {
658
+ "epoch": 14.310344827586206,
659
+ "grad_norm": 1.2168644666671753,
660
  "learning_rate": 8.28e-05,
661
+ "loss": 0.2001,
662
  "step": 830
663
  },
664
  {
665
+ "epoch": 14.482758620689655,
666
+ "grad_norm": 2.418687343597412,
667
  "learning_rate": 8.38e-05,
668
+ "loss": 0.2027,
669
  "step": 840
670
  },
671
  {
672
+ "epoch": 14.655172413793103,
673
+ "grad_norm": 4.489363670349121,
674
  "learning_rate": 8.48e-05,
675
+ "loss": 0.1739,
676
  "step": 850
677
  },
678
  {
679
+ "epoch": 14.827586206896552,
680
+ "grad_norm": 1.143477201461792,
681
  "learning_rate": 8.58e-05,
682
+ "loss": 0.2029,
683
  "step": 860
684
  },
685
  {
686
+ "epoch": 15.0,
687
+ "grad_norm": 1.1375572681427002,
688
  "learning_rate": 8.680000000000001e-05,
689
+ "loss": 0.2061,
690
  "step": 870
691
  },
692
  {
693
+ "epoch": 15.172413793103448,
694
+ "grad_norm": 1.596352458000183,
695
  "learning_rate": 8.78e-05,
696
+ "loss": 0.1517,
697
  "step": 880
698
  },
699
  {
700
+ "epoch": 15.344827586206897,
701
+ "grad_norm": 1.0544416904449463,
702
  "learning_rate": 8.88e-05,
703
+ "loss": 0.1687,
704
  "step": 890
705
  },
706
  {
707
+ "epoch": 15.517241379310345,
708
+ "grad_norm": 0.9486349821090698,
709
  "learning_rate": 8.98e-05,
710
+ "loss": 0.181,
711
  "step": 900
712
  },
713
  {
714
+ "epoch": 15.517241379310345,
715
+ "eval_loss": 0.4625524878501892,
716
+ "eval_runtime": 14.6126,
717
+ "eval_samples_per_second": 91.975,
718
+ "eval_steps_per_second": 2.874,
719
+ "eval_wer": 0.4813803150323227,
720
  "step": 900
721
  },
722
  {
723
+ "epoch": 15.689655172413794,
724
+ "grad_norm": 2.3056046962738037,
725
  "learning_rate": 9.080000000000001e-05,
726
+ "loss": 0.1855,
727
  "step": 910
728
  },
729
  {
730
+ "epoch": 15.862068965517242,
731
+ "grad_norm": 1.6048717498779297,
732
  "learning_rate": 9.180000000000001e-05,
733
+ "loss": 0.1783,
734
  "step": 920
735
  },
736
  {
737
+ "epoch": 16.03448275862069,
738
+ "grad_norm": 0.8028109073638916,
739
  "learning_rate": 9.28e-05,
740
+ "loss": 0.1877,
741
  "step": 930
742
  },
743
  {
744
+ "epoch": 16.20689655172414,
745
+ "grad_norm": 1.15639066696167,
746
  "learning_rate": 9.38e-05,
747
+ "loss": 0.1278,
748
  "step": 940
749
  },
750
  {
751
+ "epoch": 16.379310344827587,
752
+ "grad_norm": 1.2636216878890991,
753
  "learning_rate": 9.48e-05,
754
+ "loss": 0.1544,
755
  "step": 950
756
  },
757
  {
758
+ "epoch": 16.551724137931036,
759
+ "grad_norm": 1.0739763975143433,
760
  "learning_rate": 9.58e-05,
761
+ "loss": 0.1659,
762
  "step": 960
763
  },
764
  {
765
+ "epoch": 16.724137931034484,
766
+ "grad_norm": 1.9558533430099487,
767
  "learning_rate": 9.680000000000001e-05,
768
+ "loss": 0.1622,
769
  "step": 970
770
  },
771
  {
772
+ "epoch": 16.896551724137932,
773
+ "grad_norm": 1.494171142578125,
774
  "learning_rate": 9.78e-05,
775
+ "loss": 0.1448,
776
  "step": 980
777
  },
778
  {
779
+ "epoch": 17.06896551724138,
780
+ "grad_norm": 0.8110372424125671,
781
  "learning_rate": 9.88e-05,
782
+ "loss": 0.1522,
783
  "step": 990
784
  },
785
  {
786
+ "epoch": 17.24137931034483,
787
+ "grad_norm": 1.800034761428833,
788
  "learning_rate": 9.98e-05,
789
+ "loss": 0.1419,
790
  "step": 1000
791
  },
792
  {
793
+ "epoch": 17.24137931034483,
794
+ "eval_loss": 0.4916817247867584,
795
+ "eval_runtime": 14.4094,
796
+ "eval_samples_per_second": 93.273,
797
+ "eval_steps_per_second": 2.915,
798
+ "eval_wer": 0.4774651734498771,
799
  "step": 1000
800
  },
801
  {
802
+ "epoch": 17.413793103448278,
803
+ "grad_norm": 1.428348422050476,
804
+ "learning_rate": 9.5e-05,
805
+ "loss": 0.1284,
806
  "step": 1010
807
  },
808
  {
809
+ "epoch": 17.586206896551722,
810
+ "grad_norm": 1.0604817867279053,
811
+ "learning_rate": 8.875e-05,
812
+ "loss": 0.1525,
813
  "step": 1020
814
  },
815
  {
816
+ "epoch": 17.75862068965517,
817
+ "grad_norm": 0.9464378952980042,
818
+ "learning_rate": 8.25e-05,
819
+ "loss": 0.1326,
820
  "step": 1030
821
  },
822
  {
823
+ "epoch": 17.93103448275862,
824
+ "grad_norm": 1.1999770402908325,
825
+ "learning_rate": 7.625e-05,
826
+ "loss": 0.1266,
827
  "step": 1040
828
  },
829
  {
830
+ "epoch": 18.103448275862068,
831
+ "grad_norm": 0.815773606300354,
832
+ "learning_rate": 7e-05,
833
+ "loss": 0.118,
834
  "step": 1050
835
  },
836
  {
837
+ "epoch": 18.275862068965516,
838
+ "grad_norm": 0.8608018755912781,
839
+ "learning_rate": 6.375e-05,
840
+ "loss": 0.1203,
841
  "step": 1060
842
  },
843
  {
844
+ "epoch": 18.448275862068964,
845
+ "grad_norm": 1.4489638805389404,
846
+ "learning_rate": 5.7499999999999995e-05,
847
+ "loss": 0.1175,
848
  "step": 1070
849
  },
850
  {
851
+ "epoch": 18.620689655172413,
852
+ "grad_norm": 0.9969916939735413,
853
+ "learning_rate": 5.125e-05,
854
+ "loss": 0.1075,
855
  "step": 1080
856
  },
857
  {
858
+ "epoch": 18.79310344827586,
859
+ "grad_norm": 1.1026653051376343,
860
+ "learning_rate": 4.5e-05,
861
+ "loss": 0.109,
862
  "step": 1090
863
  },
864
  {
865
+ "epoch": 18.96551724137931,
866
+ "grad_norm": 1.557403802871704,
867
+ "learning_rate": 3.875e-05,
868
+ "loss": 0.1175,
869
  "step": 1100
870
  },
871
  {
872
+ "epoch": 18.96551724137931,
873
+ "eval_loss": 0.42788705229759216,
874
+ "eval_runtime": 14.5391,
875
+ "eval_samples_per_second": 92.441,
876
+ "eval_steps_per_second": 2.889,
877
+ "eval_wer": 0.4359464627151052,
878
  "step": 1100
879
  },
880
  {
881
+ "epoch": 19.137931034482758,
882
+ "grad_norm": 0.6065557599067688,
883
+ "learning_rate": 3.2500000000000004e-05,
884
+ "loss": 0.0727,
885
  "step": 1110
886
  },
887
  {
888
+ "epoch": 19.310344827586206,
889
+ "grad_norm": 0.7714155912399292,
890
+ "learning_rate": 2.625e-05,
891
+ "loss": 0.0907,
892
  "step": 1120
893
  },
894
  {
895
+ "epoch": 19.482758620689655,
896
+ "grad_norm": 1.672736644744873,
897
+ "learning_rate": 2e-05,
898
+ "loss": 0.0953,
899
  "step": 1130
900
  },
901
  {
902
+ "epoch": 19.655172413793103,
903
+ "grad_norm": 0.798320472240448,
904
+ "learning_rate": 1.3750000000000002e-05,
905
+ "loss": 0.084,
906
  "step": 1140
907
  },
908
  {
909
+ "epoch": 19.82758620689655,
910
+ "grad_norm": 0.8147380352020264,
911
+ "learning_rate": 7.5e-06,
912
+ "loss": 0.0907,
913
  "step": 1150
914
  },
915
  {
916
+ "epoch": 20.0,
917
+ "grad_norm": 1.1494929790496826,
918
+ "learning_rate": 1.25e-06,
919
+ "loss": 0.0897,
920
  "step": 1160
921
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
922
  {
923
  "epoch": 20.0,
924
+ "step": 1160,
925
+ "total_flos": 2.2493941126178012e+18,
926
+ "train_loss": 1.5806312146885642,
927
+ "train_runtime": 1021.9437,
928
+ "train_samples_per_second": 72.333,
929
+ "train_steps_per_second": 1.135
930
  }
931
  ],
932
  "logging_steps": 10,
933
+ "max_steps": 1160,
934
  "num_input_tokens_seen": 0,
935
  "num_train_epochs": 20,
936
  "save_steps": 400,
 
946
  "attributes": {}
947
  }
948
  },
949
+ "total_flos": 2.2493941126178012e+18,
950
+ "train_batch_size": 64,
951
  "trial_name": null,
952
  "trial_params": null
953
  }