vumichien commited on
Commit
cd1f931
1 Parent(s): 101a89a

End of training

Browse files
all_results.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 100.0,
3
+ "eval_cer": 0.09914388076368318,
4
+ "eval_loss": 0.21041299402713776,
5
+ "eval_runtime": 208.9129,
6
+ "eval_samples": 3676,
7
+ "eval_samples_per_second": 17.596,
8
+ "eval_steps_per_second": 2.202,
9
+ "eval_wer": 0.19410805775500697,
10
+ "train_loss": 0.804292679347349,
11
+ "train_runtime": 300809.9995,
12
+ "train_samples": 28516,
13
+ "train_samples_per_second": 9.48,
14
+ "train_steps_per_second": 0.148
15
+ }
eval_results.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 100.0,
3
+ "eval_cer": 0.09914388076368318,
4
+ "eval_loss": 0.21041299402713776,
5
+ "eval_runtime": 208.9129,
6
+ "eval_samples": 3676,
7
+ "eval_samples_per_second": 17.596,
8
+ "eval_steps_per_second": 2.202,
9
+ "eval_wer": 0.19410805775500697
10
+ }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:699b6967664875d97e73ee690d596268a4ec1e9705e1884e9007bda46339a861
3
  size 3850917361
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff591920abac5e239f21736ff7b503ef79a6b6da127464ccc75632ad744aaa5f
3
  size 3850917361
runs/Feb03_04-53-30_job-976da207-1032-42b1-8e4b-195231e136ae/events.out.tfevents.1643865655.job-976da207-1032-42b1-8e4b-195231e136ae.14574.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:17ac1180dee395489e6988818cf5e4cbcda648fc5acadb6933bd59a19b2ff2db
3
- size 9473
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:70690178dab1e011bf693d3d6068bc55dc63fb647eb9e8b5ae39fadb295e152f
3
+ size 20463
runs/Feb03_04-53-30_job-976da207-1032-42b1-8e4b-195231e136ae/events.out.tfevents.1644166682.job-976da207-1032-42b1-8e4b-195231e136ae.14574.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:328e79b7e615805bf7d231b74b03d52078e449c3040e96ba4ef114d743b5507a
3
+ size 412
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 100.0,
3
+ "train_loss": 0.804292679347349,
4
+ "train_runtime": 300809.9995,
5
+ "train_samples": 28516,
6
+ "train_samples_per_second": 9.48,
7
+ "train_steps_per_second": 0.148
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,489 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.21041299402713776,
3
+ "best_model_checkpoint": "wav2vec2-xls-r-1b-ja-dumy8/checkpoint-42000",
4
+ "epoch": 99.99831744251262,
5
+ "global_step": 44500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 3.37,
12
+ "learning_rate": 4.9427586206896555e-05,
13
+ "loss": 2.2896,
14
+ "step": 1500
15
+ },
16
+ {
17
+ "epoch": 3.37,
18
+ "eval_cer": 0.17672193185166804,
19
+ "eval_loss": 0.47477588057518005,
20
+ "eval_runtime": 210.4906,
21
+ "eval_samples_per_second": 17.464,
22
+ "eval_steps_per_second": 2.185,
23
+ "eval_wer": 0.4013274336283186,
24
+ "step": 1500
25
+ },
26
+ {
27
+ "epoch": 6.74,
28
+ "learning_rate": 4.770459770114943e-05,
29
+ "loss": 1.1608,
30
+ "step": 3000
31
+ },
32
+ {
33
+ "epoch": 6.74,
34
+ "eval_cer": 0.14555577958031546,
35
+ "eval_loss": 0.33504194021224976,
36
+ "eval_runtime": 209.7295,
37
+ "eval_samples_per_second": 17.527,
38
+ "eval_steps_per_second": 2.193,
39
+ "eval_wer": 0.315929203539823,
40
+ "step": 3000
41
+ },
42
+ {
43
+ "epoch": 10.11,
44
+ "learning_rate": 4.598045977011494e-05,
45
+ "loss": 1.1042,
46
+ "step": 4500
47
+ },
48
+ {
49
+ "epoch": 10.11,
50
+ "eval_cer": 0.1399956573661926,
51
+ "eval_loss": 0.31194448471069336,
52
+ "eval_runtime": 210.0423,
53
+ "eval_samples_per_second": 17.501,
54
+ "eval_steps_per_second": 2.19,
55
+ "eval_wer": 0.2970656730321379,
56
+ "step": 4500
57
+ },
58
+ {
59
+ "epoch": 13.48,
60
+ "learning_rate": 4.4257471264367814e-05,
61
+ "loss": 1.0494,
62
+ "step": 6000
63
+ },
64
+ {
65
+ "epoch": 13.48,
66
+ "eval_cer": 0.1353350807264606,
67
+ "eval_loss": 0.2974083721637726,
68
+ "eval_runtime": 209.3107,
69
+ "eval_samples_per_second": 17.562,
70
+ "eval_steps_per_second": 2.198,
71
+ "eval_wer": 0.2867489520260829,
72
+ "step": 6000
73
+ },
74
+ {
75
+ "epoch": 16.85,
76
+ "learning_rate": 4.253448275862069e-05,
77
+ "loss": 1.0061,
78
+ "step": 7500
79
+ },
80
+ {
81
+ "epoch": 16.85,
82
+ "eval_cer": 0.1299610713897979,
83
+ "eval_loss": 0.2802002429962158,
84
+ "eval_runtime": 210.2437,
85
+ "eval_samples_per_second": 17.484,
86
+ "eval_steps_per_second": 2.188,
87
+ "eval_wer": 0.2745924545877969,
88
+ "step": 7500
89
+ },
90
+ {
91
+ "epoch": 20.22,
92
+ "learning_rate": 4.0810344827586214e-05,
93
+ "loss": 0.9629,
94
+ "step": 9000
95
+ },
96
+ {
97
+ "epoch": 20.22,
98
+ "eval_cer": 0.13255889697101292,
99
+ "eval_loss": 0.2843669652938843,
100
+ "eval_runtime": 212.4442,
101
+ "eval_samples_per_second": 17.303,
102
+ "eval_steps_per_second": 2.165,
103
+ "eval_wer": 0.2776432231020028,
104
+ "step": 9000
105
+ },
106
+ {
107
+ "epoch": 23.59,
108
+ "learning_rate": 3.908735632183908e-05,
109
+ "loss": 0.9267,
110
+ "step": 10500
111
+ },
112
+ {
113
+ "epoch": 23.59,
114
+ "eval_cer": 0.12550211703398112,
115
+ "eval_loss": 0.2576734721660614,
116
+ "eval_runtime": 210.3131,
117
+ "eval_samples_per_second": 17.479,
118
+ "eval_steps_per_second": 2.187,
119
+ "eval_wer": 0.2603400093153237,
120
+ "step": 10500
121
+ },
122
+ {
123
+ "epoch": 26.96,
124
+ "learning_rate": 3.73632183908046e-05,
125
+ "loss": 0.8984,
126
+ "step": 12000
127
+ },
128
+ {
129
+ "epoch": 26.96,
130
+ "eval_cer": 0.12256308451075577,
131
+ "eval_loss": 0.2508338689804077,
132
+ "eval_runtime": 209.2521,
133
+ "eval_samples_per_second": 17.567,
134
+ "eval_steps_per_second": 2.198,
135
+ "eval_wer": 0.2530740568234746,
136
+ "step": 12000
137
+ },
138
+ {
139
+ "epoch": 30.34,
140
+ "learning_rate": 3.564022988505747e-05,
141
+ "loss": 0.8729,
142
+ "step": 13500
143
+ },
144
+ {
145
+ "epoch": 30.34,
146
+ "eval_cer": 0.12540906059525103,
147
+ "eval_loss": 0.26286962628364563,
148
+ "eval_runtime": 208.7235,
149
+ "eval_samples_per_second": 17.612,
150
+ "eval_steps_per_second": 2.204,
151
+ "eval_wer": 0.2605728924080112,
152
+ "step": 13500
153
+ },
154
+ {
155
+ "epoch": 33.71,
156
+ "learning_rate": 3.3917241379310346e-05,
157
+ "loss": 0.8546,
158
+ "step": 15000
159
+ },
160
+ {
161
+ "epoch": 33.71,
162
+ "eval_cer": 0.11932161856165764,
163
+ "eval_loss": 0.2401771992444992,
164
+ "eval_runtime": 208.4946,
165
+ "eval_samples_per_second": 17.631,
166
+ "eval_steps_per_second": 2.206,
167
+ "eval_wer": 0.24473684210526317,
168
+ "step": 15000
169
+ },
170
+ {
171
+ "epoch": 37.08,
172
+ "learning_rate": 3.219310344827586e-05,
173
+ "loss": 0.8304,
174
+ "step": 16500
175
+ },
176
+ {
177
+ "epoch": 37.08,
178
+ "eval_cer": 0.12086480450393164,
179
+ "eval_loss": 0.25319021940231323,
180
+ "eval_runtime": 209.1002,
181
+ "eval_samples_per_second": 17.58,
182
+ "eval_steps_per_second": 2.2,
183
+ "eval_wer": 0.2472286911970191,
184
+ "step": 16500
185
+ },
186
+ {
187
+ "epoch": 40.45,
188
+ "learning_rate": 3.046896551724138e-05,
189
+ "loss": 0.8075,
190
+ "step": 18000
191
+ },
192
+ {
193
+ "epoch": 40.45,
194
+ "eval_cer": 0.11977139134885308,
195
+ "eval_loss": 0.2439287155866623,
196
+ "eval_runtime": 209.0991,
197
+ "eval_samples_per_second": 17.58,
198
+ "eval_steps_per_second": 2.2,
199
+ "eval_wer": 0.2468793665579879,
200
+ "step": 18000
201
+ },
202
+ {
203
+ "epoch": 43.82,
204
+ "learning_rate": 2.8747126436781612e-05,
205
+ "loss": 0.7827,
206
+ "step": 19500
207
+ },
208
+ {
209
+ "epoch": 43.82,
210
+ "eval_cer": 0.11672379298044264,
211
+ "eval_loss": 0.23866486549377441,
212
+ "eval_runtime": 208.9618,
213
+ "eval_samples_per_second": 17.592,
214
+ "eval_steps_per_second": 2.201,
215
+ "eval_wer": 0.23721471821145784,
216
+ "step": 19500
217
+ },
218
+ {
219
+ "epoch": 47.19,
220
+ "learning_rate": 2.7022988505747126e-05,
221
+ "loss": 0.7627,
222
+ "step": 21000
223
+ },
224
+ {
225
+ "epoch": 47.19,
226
+ "eval_cer": 0.11469981543806318,
227
+ "eval_loss": 0.23443765938282013,
228
+ "eval_runtime": 209.1358,
229
+ "eval_samples_per_second": 17.577,
230
+ "eval_steps_per_second": 2.2,
231
+ "eval_wer": 0.23306939916162087,
232
+ "step": 21000
233
+ },
234
+ {
235
+ "epoch": 50.56,
236
+ "learning_rate": 2.5300000000000002e-05,
237
+ "loss": 0.7402,
238
+ "step": 22500
239
+ },
240
+ {
241
+ "epoch": 50.56,
242
+ "eval_cer": 0.11345906292166198,
243
+ "eval_loss": 0.23135580122470856,
244
+ "eval_runtime": 210.5917,
245
+ "eval_samples_per_second": 17.456,
246
+ "eval_steps_per_second": 2.184,
247
+ "eval_wer": 0.22994876571960876,
248
+ "step": 22500
249
+ },
250
+ {
251
+ "epoch": 53.93,
252
+ "learning_rate": 2.357586206896552e-05,
253
+ "loss": 0.718,
254
+ "step": 24000
255
+ },
256
+ {
257
+ "epoch": 53.93,
258
+ "eval_cer": 0.1114040665663725,
259
+ "eval_loss": 0.22568760812282562,
260
+ "eval_runtime": 208.402,
261
+ "eval_samples_per_second": 17.639,
262
+ "eval_steps_per_second": 2.207,
263
+ "eval_wer": 0.22666511411271542,
264
+ "step": 24000
265
+ },
266
+ {
267
+ "epoch": 57.3,
268
+ "learning_rate": 2.1851724137931033e-05,
269
+ "loss": 0.7016,
270
+ "step": 25500
271
+ },
272
+ {
273
+ "epoch": 57.3,
274
+ "eval_cer": 0.10889929742388758,
275
+ "eval_loss": 0.22044427692890167,
276
+ "eval_runtime": 208.5673,
277
+ "eval_samples_per_second": 17.625,
278
+ "eval_steps_per_second": 2.206,
279
+ "eval_wer": 0.2183977643223102,
280
+ "step": 25500
281
+ },
282
+ {
283
+ "epoch": 60.67,
284
+ "learning_rate": 2.012873563218391e-05,
285
+ "loss": 0.6804,
286
+ "step": 27000
287
+ },
288
+ {
289
+ "epoch": 60.67,
290
+ "eval_cer": 0.10851931696573971,
291
+ "eval_loss": 0.22273367643356323,
292
+ "eval_runtime": 208.2294,
293
+ "eval_samples_per_second": 17.654,
294
+ "eval_steps_per_second": 2.209,
295
+ "eval_wer": 0.21811830461108522,
296
+ "step": 27000
297
+ },
298
+ {
299
+ "epoch": 64.04,
300
+ "learning_rate": 1.8405747126436782e-05,
301
+ "loss": 0.6625,
302
+ "step": 28500
303
+ },
304
+ {
305
+ "epoch": 64.04,
306
+ "eval_cer": 0.10578966142965708,
307
+ "eval_loss": 0.21378228068351746,
308
+ "eval_runtime": 209.3597,
309
+ "eval_samples_per_second": 17.558,
310
+ "eval_steps_per_second": 2.197,
311
+ "eval_wer": 0.21115510013972985,
312
+ "step": 28500
313
+ },
314
+ {
315
+ "epoch": 67.42,
316
+ "learning_rate": 1.66816091954023e-05,
317
+ "loss": 0.6465,
318
+ "step": 30000
319
+ },
320
+ {
321
+ "epoch": 67.42,
322
+ "eval_cer": 0.10442483366161577,
323
+ "eval_loss": 0.2140830010175705,
324
+ "eval_runtime": 208.6345,
325
+ "eval_samples_per_second": 17.619,
326
+ "eval_steps_per_second": 2.205,
327
+ "eval_wer": 0.20808104331625524,
328
+ "step": 30000
329
+ },
330
+ {
331
+ "epoch": 70.79,
332
+ "learning_rate": 1.4958620689655173e-05,
333
+ "loss": 0.6238,
334
+ "step": 31500
335
+ },
336
+ {
337
+ "epoch": 70.79,
338
+ "eval_cer": 0.1049754175907688,
339
+ "eval_loss": 0.21720194816589355,
340
+ "eval_runtime": 210.4781,
341
+ "eval_samples_per_second": 17.465,
342
+ "eval_steps_per_second": 2.186,
343
+ "eval_wer": 0.2082207731718677,
344
+ "step": 31500
345
+ },
346
+ {
347
+ "epoch": 74.16,
348
+ "learning_rate": 1.323448275862069e-05,
349
+ "loss": 0.6062,
350
+ "step": 33000
351
+ },
352
+ {
353
+ "epoch": 74.16,
354
+ "eval_cer": 0.10427749430029312,
355
+ "eval_loss": 0.2174130082130432,
356
+ "eval_runtime": 209.3158,
357
+ "eval_samples_per_second": 17.562,
358
+ "eval_steps_per_second": 2.198,
359
+ "eval_wer": 0.20582207731718677,
360
+ "step": 33000
361
+ },
362
+ {
363
+ "epoch": 77.53,
364
+ "learning_rate": 1.1510344827586207e-05,
365
+ "loss": 0.588,
366
+ "step": 34500
367
+ },
368
+ {
369
+ "epoch": 77.53,
370
+ "eval_cer": 0.1026955348418816,
371
+ "eval_loss": 0.21560049057006836,
372
+ "eval_runtime": 208.8439,
373
+ "eval_samples_per_second": 17.602,
374
+ "eval_steps_per_second": 2.203,
375
+ "eval_wer": 0.20344666977177456,
376
+ "step": 34500
377
+ },
378
+ {
379
+ "epoch": 80.9,
380
+ "learning_rate": 9.786206896551724e-06,
381
+ "loss": 0.5722,
382
+ "step": 36000
383
+ },
384
+ {
385
+ "epoch": 80.9,
386
+ "eval_cer": 0.10293593064193433,
387
+ "eval_loss": 0.21624404191970825,
388
+ "eval_runtime": 208.6816,
389
+ "eval_samples_per_second": 17.615,
390
+ "eval_steps_per_second": 2.204,
391
+ "eval_wer": 0.2032137866790871,
392
+ "step": 36000
393
+ },
394
+ {
395
+ "epoch": 84.27,
396
+ "learning_rate": 8.064367816091953e-06,
397
+ "loss": 0.5585,
398
+ "step": 37500
399
+ },
400
+ {
401
+ "epoch": 84.27,
402
+ "eval_cer": 0.10209066799013601,
403
+ "eval_loss": 0.21560421586036682,
404
+ "eval_runtime": 211.1243,
405
+ "eval_samples_per_second": 17.412,
406
+ "eval_steps_per_second": 2.179,
407
+ "eval_wer": 0.20216581276199347,
408
+ "step": 37500
409
+ },
410
+ {
411
+ "epoch": 87.64,
412
+ "learning_rate": 6.341379310344828e-06,
413
+ "loss": 0.5456,
414
+ "step": 39000
415
+ },
416
+ {
417
+ "epoch": 87.64,
418
+ "eval_cer": 0.10090419839632737,
419
+ "eval_loss": 0.21256230771541595,
420
+ "eval_runtime": 208.9966,
421
+ "eval_samples_per_second": 17.589,
422
+ "eval_steps_per_second": 2.201,
423
+ "eval_wer": 0.19930135072193758,
424
+ "step": 39000
425
+ },
426
+ {
427
+ "epoch": 91.01,
428
+ "learning_rate": 4.618390804597701e-06,
429
+ "loss": 0.5325,
430
+ "step": 40500
431
+ },
432
+ {
433
+ "epoch": 91.01,
434
+ "eval_cer": 0.10033810506071933,
435
+ "eval_loss": 0.2121460884809494,
436
+ "eval_runtime": 208.4119,
437
+ "eval_samples_per_second": 17.638,
438
+ "eval_steps_per_second": 2.207,
439
+ "eval_wer": 0.19659990684676293,
440
+ "step": 40500
441
+ },
442
+ {
443
+ "epoch": 94.38,
444
+ "learning_rate": 2.8942528735632185e-06,
445
+ "loss": 0.5229,
446
+ "step": 42000
447
+ },
448
+ {
449
+ "epoch": 94.38,
450
+ "eval_cer": 0.09914388076368318,
451
+ "eval_loss": 0.21041299402713776,
452
+ "eval_runtime": 208.604,
453
+ "eval_samples_per_second": 17.622,
454
+ "eval_steps_per_second": 2.205,
455
+ "eval_wer": 0.19410805775500697,
456
+ "step": 42000
457
+ },
458
+ {
459
+ "epoch": 97.75,
460
+ "learning_rate": 1.1701149425287358e-06,
461
+ "loss": 0.5134,
462
+ "step": 43500
463
+ },
464
+ {
465
+ "epoch": 97.75,
466
+ "eval_cer": 0.0991593901701382,
467
+ "eval_loss": 0.21077032387256622,
468
+ "eval_runtime": 208.8814,
469
+ "eval_samples_per_second": 17.599,
470
+ "eval_steps_per_second": 2.202,
471
+ "eval_wer": 0.19482999534233814,
472
+ "step": 43500
473
+ },
474
+ {
475
+ "epoch": 100.0,
476
+ "step": 44500,
477
+ "total_flos": 1.4850172866055105e+21,
478
+ "train_loss": 0.804292679347349,
479
+ "train_runtime": 300809.9995,
480
+ "train_samples_per_second": 9.48,
481
+ "train_steps_per_second": 0.148
482
+ }
483
+ ],
484
+ "max_steps": 44500,
485
+ "num_train_epochs": 100,
486
+ "total_flos": 1.4850172866055105e+21,
487
+ "trial_name": null,
488
+ "trial_params": null
489
+ }