Yasinjan99 commited on
Commit
8af493d
1 Parent(s): 8f60edc

End of training

Browse files
Files changed (5) hide show
  1. README.md +8 -17
  2. all_results.json +14 -0
  3. eval_results.json +9 -0
  4. train_results.json +8 -0
  5. trainer_state.json +910 -0
README.md CHANGED
@@ -1,7 +1,11 @@
1
  ---
 
 
2
  license: apache-2.0
3
  base_model: facebook/wav2vec2-xls-r-300m
4
  tags:
 
 
5
  - generated_from_trainer
6
  datasets:
7
  - common_voice_11_0
@@ -9,20 +13,7 @@ metrics:
9
  - wer
10
  model-index:
11
  - name: xls-r-uyghur-cv11
12
- results:
13
- - task:
14
- name: Automatic Speech Recognition
15
- type: automatic-speech-recognition
16
- dataset:
17
- name: common_voice_11_0
18
- type: common_voice_11_0
19
- config: ug
20
- split: test
21
- args: ug
22
- metrics:
23
- - name: Wer
24
- type: wer
25
- value: 0.32705699280295664
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -30,10 +21,10 @@ should probably proofread and complete it, then remove this comment. -->
30
 
31
  # xls-r-uyghur-cv11
32
 
33
- This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice_11_0 dataset.
34
  It achieves the following results on the evaluation set:
35
- - Loss: 0.2203
36
- - Wer: 0.3271
37
 
38
  ## Model description
39
 
 
1
  ---
2
+ language:
3
+ - ug
4
  license: apache-2.0
5
  base_model: facebook/wav2vec2-xls-r-300m
6
  tags:
7
+ - automatic-speech-recognition
8
+ - mozilla-foundation/common_voice_11_0
9
  - generated_from_trainer
10
  datasets:
11
  - common_voice_11_0
 
13
  - wer
14
  model-index:
15
  - name: xls-r-uyghur-cv11
16
+ results: []
 
 
 
 
 
 
 
 
 
 
 
 
 
17
  ---
18
 
19
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
21
 
22
  # xls-r-uyghur-cv11
23
 
24
+ This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_11_0 - UG dataset.
25
  It achieves the following results on the evaluation set:
26
+ - Loss: 0.2191
27
+ - Wer: 0.3257
28
 
29
  ## Model description
30
 
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 99.35,
3
+ "eval_loss": 0.21905453503131866,
4
+ "eval_runtime": 260.5584,
5
+ "eval_samples": 3270,
6
+ "eval_samples_per_second": 12.55,
7
+ "eval_steps_per_second": 1.57,
8
+ "eval_wer": 0.32569539000194514,
9
+ "train_loss": 1.3810767638892458,
10
+ "train_runtime": 73797.3063,
11
+ "train_samples": 7342,
12
+ "train_samples_per_second": 9.949,
13
+ "train_steps_per_second": 0.154
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 99.35,
3
+ "eval_loss": 0.21905453503131866,
4
+ "eval_runtime": 260.5584,
5
+ "eval_samples": 3270,
6
+ "eval_samples_per_second": 12.55,
7
+ "eval_steps_per_second": 1.57,
8
+ "eval_wer": 0.32569539000194514
9
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 99.35,
3
+ "train_loss": 1.3810767638892458,
4
+ "train_runtime": 73797.3063,
5
+ "train_samples": 7342,
6
+ "train_samples_per_second": 9.949,
7
+ "train_steps_per_second": 0.154
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,910 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 99.34640522875817,
5
+ "eval_steps": 500,
6
+ "global_step": 11400,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.87,
13
+ "learning_rate": 4.950000000000001e-06,
14
+ "loss": 14.1952,
15
+ "step": 100
16
+ },
17
+ {
18
+ "epoch": 1.74,
19
+ "learning_rate": 9.950000000000001e-06,
20
+ "loss": 6.4908,
21
+ "step": 200
22
+ },
23
+ {
24
+ "epoch": 2.61,
25
+ "learning_rate": 1.4950000000000001e-05,
26
+ "loss": 4.5428,
27
+ "step": 300
28
+ },
29
+ {
30
+ "epoch": 3.49,
31
+ "learning_rate": 1.995e-05,
32
+ "loss": 3.8598,
33
+ "step": 400
34
+ },
35
+ {
36
+ "epoch": 4.36,
37
+ "learning_rate": 2.495e-05,
38
+ "loss": 3.3958,
39
+ "step": 500
40
+ },
41
+ {
42
+ "epoch": 4.36,
43
+ "eval_loss": 3.3137385845184326,
44
+ "eval_runtime": 268.8924,
45
+ "eval_samples_per_second": 12.161,
46
+ "eval_steps_per_second": 1.521,
47
+ "eval_wer": 1.0,
48
+ "step": 500
49
+ },
50
+ {
51
+ "epoch": 5.23,
52
+ "learning_rate": 2.995e-05,
53
+ "loss": 3.2375,
54
+ "step": 600
55
+ },
56
+ {
57
+ "epoch": 6.1,
58
+ "learning_rate": 3.495e-05,
59
+ "loss": 3.1969,
60
+ "step": 700
61
+ },
62
+ {
63
+ "epoch": 6.97,
64
+ "learning_rate": 3.995e-05,
65
+ "loss": 3.151,
66
+ "step": 800
67
+ },
68
+ {
69
+ "epoch": 7.84,
70
+ "learning_rate": 4.495e-05,
71
+ "loss": 3.1114,
72
+ "step": 900
73
+ },
74
+ {
75
+ "epoch": 8.71,
76
+ "learning_rate": 4.995e-05,
77
+ "loss": 3.032,
78
+ "step": 1000
79
+ },
80
+ {
81
+ "epoch": 8.71,
82
+ "eval_loss": 2.858614921569824,
83
+ "eval_runtime": 245.2843,
84
+ "eval_samples_per_second": 13.331,
85
+ "eval_steps_per_second": 1.667,
86
+ "eval_wer": 0.9992608441937366,
87
+ "step": 1000
88
+ },
89
+ {
90
+ "epoch": 9.59,
91
+ "learning_rate": 5.495e-05,
92
+ "loss": 2.5279,
93
+ "step": 1100
94
+ },
95
+ {
96
+ "epoch": 10.46,
97
+ "learning_rate": 5.995000000000001e-05,
98
+ "loss": 1.8013,
99
+ "step": 1200
100
+ },
101
+ {
102
+ "epoch": 11.33,
103
+ "learning_rate": 6.494999999999999e-05,
104
+ "loss": 1.5785,
105
+ "step": 1300
106
+ },
107
+ {
108
+ "epoch": 12.2,
109
+ "learning_rate": 6.995e-05,
110
+ "loss": 1.468,
111
+ "step": 1400
112
+ },
113
+ {
114
+ "epoch": 13.07,
115
+ "learning_rate": 7.495e-05,
116
+ "loss": 1.3977,
117
+ "step": 1500
118
+ },
119
+ {
120
+ "epoch": 13.07,
121
+ "eval_loss": 0.4785650372505188,
122
+ "eval_runtime": 252.0133,
123
+ "eval_samples_per_second": 12.976,
124
+ "eval_steps_per_second": 1.623,
125
+ "eval_wer": 0.637541334370745,
126
+ "step": 1500
127
+ },
128
+ {
129
+ "epoch": 13.94,
130
+ "learning_rate": 7.99e-05,
131
+ "loss": 1.3681,
132
+ "step": 1600
133
+ },
134
+ {
135
+ "epoch": 14.81,
136
+ "learning_rate": 8.49e-05,
137
+ "loss": 1.336,
138
+ "step": 1700
139
+ },
140
+ {
141
+ "epoch": 15.69,
142
+ "learning_rate": 8.99e-05,
143
+ "loss": 1.307,
144
+ "step": 1800
145
+ },
146
+ {
147
+ "epoch": 16.56,
148
+ "learning_rate": 9.49e-05,
149
+ "loss": 1.2934,
150
+ "step": 1900
151
+ },
152
+ {
153
+ "epoch": 17.43,
154
+ "learning_rate": 9.99e-05,
155
+ "loss": 1.2751,
156
+ "step": 2000
157
+ },
158
+ {
159
+ "epoch": 17.43,
160
+ "eval_loss": 0.3816487491130829,
161
+ "eval_runtime": 244.8161,
162
+ "eval_samples_per_second": 13.357,
163
+ "eval_steps_per_second": 1.671,
164
+ "eval_wer": 0.53931141801206,
165
+ "step": 2000
166
+ },
167
+ {
168
+ "epoch": 18.3,
169
+ "learning_rate": 9.895744680851065e-05,
170
+ "loss": 1.2545,
171
+ "step": 2100
172
+ },
173
+ {
174
+ "epoch": 19.17,
175
+ "learning_rate": 9.78936170212766e-05,
176
+ "loss": 1.2488,
177
+ "step": 2200
178
+ },
179
+ {
180
+ "epoch": 20.04,
181
+ "learning_rate": 9.682978723404255e-05,
182
+ "loss": 1.2315,
183
+ "step": 2300
184
+ },
185
+ {
186
+ "epoch": 20.92,
187
+ "learning_rate": 9.576595744680852e-05,
188
+ "loss": 1.2237,
189
+ "step": 2400
190
+ },
191
+ {
192
+ "epoch": 21.79,
193
+ "learning_rate": 9.470212765957447e-05,
194
+ "loss": 1.2113,
195
+ "step": 2500
196
+ },
197
+ {
198
+ "epoch": 21.79,
199
+ "eval_loss": 0.3450602889060974,
200
+ "eval_runtime": 260.0099,
201
+ "eval_samples_per_second": 12.576,
202
+ "eval_steps_per_second": 1.573,
203
+ "eval_wer": 0.509900797510212,
204
+ "step": 2500
205
+ },
206
+ {
207
+ "epoch": 22.66,
208
+ "learning_rate": 9.363829787234043e-05,
209
+ "loss": 1.204,
210
+ "step": 2600
211
+ },
212
+ {
213
+ "epoch": 23.53,
214
+ "learning_rate": 9.257446808510639e-05,
215
+ "loss": 1.1834,
216
+ "step": 2700
217
+ },
218
+ {
219
+ "epoch": 24.4,
220
+ "learning_rate": 9.151063829787234e-05,
221
+ "loss": 1.1782,
222
+ "step": 2800
223
+ },
224
+ {
225
+ "epoch": 25.27,
226
+ "learning_rate": 9.04468085106383e-05,
227
+ "loss": 1.1616,
228
+ "step": 2900
229
+ },
230
+ {
231
+ "epoch": 26.14,
232
+ "learning_rate": 8.938297872340426e-05,
233
+ "loss": 1.156,
234
+ "step": 3000
235
+ },
236
+ {
237
+ "epoch": 26.14,
238
+ "eval_loss": 0.3244517743587494,
239
+ "eval_runtime": 254.1562,
240
+ "eval_samples_per_second": 12.866,
241
+ "eval_steps_per_second": 1.609,
242
+ "eval_wer": 0.49192764053686056,
243
+ "step": 3000
244
+ },
245
+ {
246
+ "epoch": 27.02,
247
+ "learning_rate": 8.831914893617022e-05,
248
+ "loss": 1.165,
249
+ "step": 3100
250
+ },
251
+ {
252
+ "epoch": 27.89,
253
+ "learning_rate": 8.725531914893618e-05,
254
+ "loss": 1.1559,
255
+ "step": 3200
256
+ },
257
+ {
258
+ "epoch": 28.76,
259
+ "learning_rate": 8.619148936170212e-05,
260
+ "loss": 1.1239,
261
+ "step": 3300
262
+ },
263
+ {
264
+ "epoch": 29.63,
265
+ "learning_rate": 8.51276595744681e-05,
266
+ "loss": 1.1219,
267
+ "step": 3400
268
+ },
269
+ {
270
+ "epoch": 30.5,
271
+ "learning_rate": 8.406382978723405e-05,
272
+ "loss": 1.1226,
273
+ "step": 3500
274
+ },
275
+ {
276
+ "epoch": 30.5,
277
+ "eval_loss": 0.2992089092731476,
278
+ "eval_runtime": 263.7476,
279
+ "eval_samples_per_second": 12.398,
280
+ "eval_steps_per_second": 1.551,
281
+ "eval_wer": 0.44411593075277184,
282
+ "step": 3500
283
+ },
284
+ {
285
+ "epoch": 31.37,
286
+ "learning_rate": 8.3e-05,
287
+ "loss": 1.1115,
288
+ "step": 3600
289
+ },
290
+ {
291
+ "epoch": 32.24,
292
+ "learning_rate": 8.194680851063831e-05,
293
+ "loss": 1.0983,
294
+ "step": 3700
295
+ },
296
+ {
297
+ "epoch": 33.12,
298
+ "learning_rate": 8.088297872340426e-05,
299
+ "loss": 1.0957,
300
+ "step": 3800
301
+ },
302
+ {
303
+ "epoch": 33.99,
304
+ "learning_rate": 7.981914893617022e-05,
305
+ "loss": 1.0874,
306
+ "step": 3900
307
+ },
308
+ {
309
+ "epoch": 34.86,
310
+ "learning_rate": 7.875531914893617e-05,
311
+ "loss": 1.0913,
312
+ "step": 4000
313
+ },
314
+ {
315
+ "epoch": 34.86,
316
+ "eval_loss": 0.2831476330757141,
317
+ "eval_runtime": 264.6688,
318
+ "eval_samples_per_second": 12.355,
319
+ "eval_steps_per_second": 1.545,
320
+ "eval_wer": 0.431472476171951,
321
+ "step": 4000
322
+ },
323
+ {
324
+ "epoch": 35.73,
325
+ "learning_rate": 7.769148936170213e-05,
326
+ "loss": 1.0896,
327
+ "step": 4100
328
+ },
329
+ {
330
+ "epoch": 36.6,
331
+ "learning_rate": 7.663829787234042e-05,
332
+ "loss": 1.071,
333
+ "step": 4200
334
+ },
335
+ {
336
+ "epoch": 37.47,
337
+ "learning_rate": 7.557446808510639e-05,
338
+ "loss": 1.0635,
339
+ "step": 4300
340
+ },
341
+ {
342
+ "epoch": 38.34,
343
+ "learning_rate": 7.451063829787235e-05,
344
+ "loss": 1.0689,
345
+ "step": 4400
346
+ },
347
+ {
348
+ "epoch": 39.22,
349
+ "learning_rate": 7.34468085106383e-05,
350
+ "loss": 1.0615,
351
+ "step": 4500
352
+ },
353
+ {
354
+ "epoch": 39.22,
355
+ "eval_loss": 0.2807982563972473,
356
+ "eval_runtime": 253.9269,
357
+ "eval_samples_per_second": 12.878,
358
+ "eval_steps_per_second": 1.611,
359
+ "eval_wer": 0.4339622641509434,
360
+ "step": 4500
361
+ },
362
+ {
363
+ "epoch": 40.09,
364
+ "learning_rate": 7.238297872340425e-05,
365
+ "loss": 1.0553,
366
+ "step": 4600
367
+ },
368
+ {
369
+ "epoch": 40.96,
370
+ "learning_rate": 7.131914893617023e-05,
371
+ "loss": 1.0533,
372
+ "step": 4700
373
+ },
374
+ {
375
+ "epoch": 41.83,
376
+ "learning_rate": 7.025531914893617e-05,
377
+ "loss": 1.0546,
378
+ "step": 4800
379
+ },
380
+ {
381
+ "epoch": 42.7,
382
+ "learning_rate": 6.919148936170213e-05,
383
+ "loss": 1.0517,
384
+ "step": 4900
385
+ },
386
+ {
387
+ "epoch": 43.57,
388
+ "learning_rate": 6.812765957446809e-05,
389
+ "loss": 1.0455,
390
+ "step": 5000
391
+ },
392
+ {
393
+ "epoch": 43.57,
394
+ "eval_loss": 0.27128756046295166,
395
+ "eval_runtime": 251.6872,
396
+ "eval_samples_per_second": 12.992,
397
+ "eval_steps_per_second": 1.625,
398
+ "eval_wer": 0.4087531608636452,
399
+ "step": 5000
400
+ },
401
+ {
402
+ "epoch": 44.44,
403
+ "learning_rate": 6.706382978723405e-05,
404
+ "loss": 1.0386,
405
+ "step": 5100
406
+ },
407
+ {
408
+ "epoch": 45.32,
409
+ "learning_rate": 6.6e-05,
410
+ "loss": 1.0287,
411
+ "step": 5200
412
+ },
413
+ {
414
+ "epoch": 46.19,
415
+ "learning_rate": 6.493617021276595e-05,
416
+ "loss": 1.0251,
417
+ "step": 5300
418
+ },
419
+ {
420
+ "epoch": 47.06,
421
+ "learning_rate": 6.387234042553192e-05,
422
+ "loss": 1.0335,
423
+ "step": 5400
424
+ },
425
+ {
426
+ "epoch": 47.93,
427
+ "learning_rate": 6.280851063829788e-05,
428
+ "loss": 1.0228,
429
+ "step": 5500
430
+ },
431
+ {
432
+ "epoch": 47.93,
433
+ "eval_loss": 0.2622440457344055,
434
+ "eval_runtime": 249.5189,
435
+ "eval_samples_per_second": 13.105,
436
+ "eval_steps_per_second": 1.639,
437
+ "eval_wer": 0.39599299747130906,
438
+ "step": 5500
439
+ },
440
+ {
441
+ "epoch": 48.8,
442
+ "learning_rate": 6.174468085106383e-05,
443
+ "loss": 1.0184,
444
+ "step": 5600
445
+ },
446
+ {
447
+ "epoch": 49.67,
448
+ "learning_rate": 6.068085106382979e-05,
449
+ "loss": 1.0174,
450
+ "step": 5700
451
+ },
452
+ {
453
+ "epoch": 50.54,
454
+ "learning_rate": 5.9617021276595755e-05,
455
+ "loss": 1.0133,
456
+ "step": 5800
457
+ },
458
+ {
459
+ "epoch": 51.42,
460
+ "learning_rate": 5.85531914893617e-05,
461
+ "loss": 1.0108,
462
+ "step": 5900
463
+ },
464
+ {
465
+ "epoch": 52.29,
466
+ "learning_rate": 5.7489361702127666e-05,
467
+ "loss": 0.9936,
468
+ "step": 6000
469
+ },
470
+ {
471
+ "epoch": 52.29,
472
+ "eval_loss": 0.25247836112976074,
473
+ "eval_runtime": 249.2066,
474
+ "eval_samples_per_second": 13.122,
475
+ "eval_steps_per_second": 1.641,
476
+ "eval_wer": 0.37957595798482785,
477
+ "step": 6000
478
+ },
479
+ {
480
+ "epoch": 53.16,
481
+ "learning_rate": 5.642553191489362e-05,
482
+ "loss": 1.0015,
483
+ "step": 6100
484
+ },
485
+ {
486
+ "epoch": 54.03,
487
+ "learning_rate": 5.5361702127659576e-05,
488
+ "loss": 1.0003,
489
+ "step": 6200
490
+ },
491
+ {
492
+ "epoch": 54.9,
493
+ "learning_rate": 5.429787234042554e-05,
494
+ "loss": 0.988,
495
+ "step": 6300
496
+ },
497
+ {
498
+ "epoch": 55.77,
499
+ "learning_rate": 5.3234042553191486e-05,
500
+ "loss": 0.9845,
501
+ "step": 6400
502
+ },
503
+ {
504
+ "epoch": 56.64,
505
+ "learning_rate": 5.217021276595745e-05,
506
+ "loss": 0.968,
507
+ "step": 6500
508
+ },
509
+ {
510
+ "epoch": 56.64,
511
+ "eval_loss": 0.2505591809749603,
512
+ "eval_runtime": 253.8437,
513
+ "eval_samples_per_second": 12.882,
514
+ "eval_steps_per_second": 1.611,
515
+ "eval_wer": 0.3798093756078584,
516
+ "step": 6500
517
+ },
518
+ {
519
+ "epoch": 57.52,
520
+ "learning_rate": 5.110638297872341e-05,
521
+ "loss": 0.9852,
522
+ "step": 6600
523
+ },
524
+ {
525
+ "epoch": 58.39,
526
+ "learning_rate": 5.004255319148936e-05,
527
+ "loss": 0.9737,
528
+ "step": 6700
529
+ },
530
+ {
531
+ "epoch": 59.26,
532
+ "learning_rate": 4.897872340425532e-05,
533
+ "loss": 0.9752,
534
+ "step": 6800
535
+ },
536
+ {
537
+ "epoch": 60.13,
538
+ "learning_rate": 4.792553191489362e-05,
539
+ "loss": 0.9717,
540
+ "step": 6900
541
+ },
542
+ {
543
+ "epoch": 61.0,
544
+ "learning_rate": 4.686170212765958e-05,
545
+ "loss": 0.9704,
546
+ "step": 7000
547
+ },
548
+ {
549
+ "epoch": 61.0,
550
+ "eval_loss": 0.24809595942497253,
551
+ "eval_runtime": 257.0988,
552
+ "eval_samples_per_second": 12.719,
553
+ "eval_steps_per_second": 1.591,
554
+ "eval_wer": 0.37346819684886207,
555
+ "step": 7000
556
+ },
557
+ {
558
+ "epoch": 61.87,
559
+ "learning_rate": 4.579787234042554e-05,
560
+ "loss": 0.9603,
561
+ "step": 7100
562
+ },
563
+ {
564
+ "epoch": 62.75,
565
+ "learning_rate": 4.474468085106383e-05,
566
+ "loss": 0.967,
567
+ "step": 7200
568
+ },
569
+ {
570
+ "epoch": 63.62,
571
+ "learning_rate": 4.368085106382979e-05,
572
+ "loss": 0.9623,
573
+ "step": 7300
574
+ },
575
+ {
576
+ "epoch": 64.49,
577
+ "learning_rate": 4.261702127659575e-05,
578
+ "loss": 0.9506,
579
+ "step": 7400
580
+ },
581
+ {
582
+ "epoch": 65.36,
583
+ "learning_rate": 4.15531914893617e-05,
584
+ "loss": 0.9552,
585
+ "step": 7500
586
+ },
587
+ {
588
+ "epoch": 65.36,
589
+ "eval_loss": 0.23935528099536896,
590
+ "eval_runtime": 264.2571,
591
+ "eval_samples_per_second": 12.374,
592
+ "eval_steps_per_second": 1.548,
593
+ "eval_wer": 0.36432600661349934,
594
+ "step": 7500
595
+ },
596
+ {
597
+ "epoch": 66.23,
598
+ "learning_rate": 4.048936170212766e-05,
599
+ "loss": 0.9449,
600
+ "step": 7600
601
+ },
602
+ {
603
+ "epoch": 67.1,
604
+ "learning_rate": 3.9425531914893624e-05,
605
+ "loss": 0.9411,
606
+ "step": 7700
607
+ },
608
+ {
609
+ "epoch": 67.97,
610
+ "learning_rate": 3.8361702127659576e-05,
611
+ "loss": 0.9468,
612
+ "step": 7800
613
+ },
614
+ {
615
+ "epoch": 68.85,
616
+ "learning_rate": 3.7297872340425534e-05,
617
+ "loss": 0.9448,
618
+ "step": 7900
619
+ },
620
+ {
621
+ "epoch": 69.72,
622
+ "learning_rate": 3.6234042553191486e-05,
623
+ "loss": 0.9417,
624
+ "step": 8000
625
+ },
626
+ {
627
+ "epoch": 69.72,
628
+ "eval_loss": 0.23499037325382233,
629
+ "eval_runtime": 254.8702,
630
+ "eval_samples_per_second": 12.83,
631
+ "eval_steps_per_second": 1.605,
632
+ "eval_wer": 0.35366660182843807,
633
+ "step": 8000
634
+ },
635
+ {
636
+ "epoch": 70.59,
637
+ "learning_rate": 3.517021276595745e-05,
638
+ "loss": 0.9342,
639
+ "step": 8100
640
+ },
641
+ {
642
+ "epoch": 71.46,
643
+ "learning_rate": 3.41063829787234e-05,
644
+ "loss": 0.9338,
645
+ "step": 8200
646
+ },
647
+ {
648
+ "epoch": 72.33,
649
+ "learning_rate": 3.304255319148936e-05,
650
+ "loss": 0.9298,
651
+ "step": 8300
652
+ },
653
+ {
654
+ "epoch": 73.2,
655
+ "learning_rate": 3.1978723404255326e-05,
656
+ "loss": 0.9239,
657
+ "step": 8400
658
+ },
659
+ {
660
+ "epoch": 74.07,
661
+ "learning_rate": 3.091489361702128e-05,
662
+ "loss": 0.9215,
663
+ "step": 8500
664
+ },
665
+ {
666
+ "epoch": 74.07,
667
+ "eval_loss": 0.23256558179855347,
668
+ "eval_runtime": 242.8597,
669
+ "eval_samples_per_second": 13.465,
670
+ "eval_steps_per_second": 1.684,
671
+ "eval_wer": 0.35070997860338454,
672
+ "step": 8500
673
+ },
674
+ {
675
+ "epoch": 74.95,
676
+ "learning_rate": 2.9851063829787236e-05,
677
+ "loss": 0.9233,
678
+ "step": 8600
679
+ },
680
+ {
681
+ "epoch": 75.82,
682
+ "learning_rate": 2.878723404255319e-05,
683
+ "loss": 0.9193,
684
+ "step": 8700
685
+ },
686
+ {
687
+ "epoch": 76.69,
688
+ "learning_rate": 2.7723404255319153e-05,
689
+ "loss": 0.9186,
690
+ "step": 8800
691
+ },
692
+ {
693
+ "epoch": 77.56,
694
+ "learning_rate": 2.6659574468085108e-05,
695
+ "loss": 0.9064,
696
+ "step": 8900
697
+ },
698
+ {
699
+ "epoch": 78.43,
700
+ "learning_rate": 2.5595744680851063e-05,
701
+ "loss": 0.9097,
702
+ "step": 9000
703
+ },
704
+ {
705
+ "epoch": 78.43,
706
+ "eval_loss": 0.227674201130867,
707
+ "eval_runtime": 258.8243,
708
+ "eval_samples_per_second": 12.634,
709
+ "eval_steps_per_second": 1.58,
710
+ "eval_wer": 0.348725928807625,
711
+ "step": 9000
712
+ },
713
+ {
714
+ "epoch": 79.3,
715
+ "learning_rate": 2.453191489361702e-05,
716
+ "loss": 0.9029,
717
+ "step": 9100
718
+ },
719
+ {
720
+ "epoch": 80.17,
721
+ "learning_rate": 2.346808510638298e-05,
722
+ "loss": 0.9137,
723
+ "step": 9200
724
+ },
725
+ {
726
+ "epoch": 81.05,
727
+ "learning_rate": 2.2404255319148938e-05,
728
+ "loss": 0.9045,
729
+ "step": 9300
730
+ },
731
+ {
732
+ "epoch": 81.92,
733
+ "learning_rate": 2.1340425531914897e-05,
734
+ "loss": 0.905,
735
+ "step": 9400
736
+ },
737
+ {
738
+ "epoch": 82.79,
739
+ "learning_rate": 2.027659574468085e-05,
740
+ "loss": 0.9003,
741
+ "step": 9500
742
+ },
743
+ {
744
+ "epoch": 82.79,
745
+ "eval_loss": 0.2229508012533188,
746
+ "eval_runtime": 256.366,
747
+ "eval_samples_per_second": 12.755,
748
+ "eval_steps_per_second": 1.595,
749
+ "eval_wer": 0.33623808597549115,
750
+ "step": 9500
751
+ },
752
+ {
753
+ "epoch": 83.66,
754
+ "learning_rate": 1.921276595744681e-05,
755
+ "loss": 0.8983,
756
+ "step": 9600
757
+ },
758
+ {
759
+ "epoch": 84.53,
760
+ "learning_rate": 1.8148936170212765e-05,
761
+ "loss": 0.9026,
762
+ "step": 9700
763
+ },
764
+ {
765
+ "epoch": 85.4,
766
+ "learning_rate": 1.7095744680851063e-05,
767
+ "loss": 0.8921,
768
+ "step": 9800
769
+ },
770
+ {
771
+ "epoch": 86.27,
772
+ "learning_rate": 1.603191489361702e-05,
773
+ "loss": 0.895,
774
+ "step": 9900
775
+ },
776
+ {
777
+ "epoch": 87.15,
778
+ "learning_rate": 1.4968085106382978e-05,
779
+ "loss": 0.8857,
780
+ "step": 10000
781
+ },
782
+ {
783
+ "epoch": 87.15,
784
+ "eval_loss": 0.2246018499135971,
785
+ "eval_runtime": 255.6354,
786
+ "eval_samples_per_second": 12.792,
787
+ "eval_steps_per_second": 1.6,
788
+ "eval_wer": 0.3361602801011476,
789
+ "step": 10000
790
+ },
791
+ {
792
+ "epoch": 88.02,
793
+ "learning_rate": 1.3904255319148937e-05,
794
+ "loss": 0.8916,
795
+ "step": 10100
796
+ },
797
+ {
798
+ "epoch": 88.89,
799
+ "learning_rate": 1.2840425531914893e-05,
800
+ "loss": 0.8808,
801
+ "step": 10200
802
+ },
803
+ {
804
+ "epoch": 89.76,
805
+ "learning_rate": 1.1776595744680852e-05,
806
+ "loss": 0.8817,
807
+ "step": 10300
808
+ },
809
+ {
810
+ "epoch": 90.63,
811
+ "learning_rate": 1.0712765957446808e-05,
812
+ "loss": 0.8805,
813
+ "step": 10400
814
+ },
815
+ {
816
+ "epoch": 91.5,
817
+ "learning_rate": 9.648936170212767e-06,
818
+ "loss": 0.882,
819
+ "step": 10500
820
+ },
821
+ {
822
+ "epoch": 91.5,
823
+ "eval_loss": 0.2236073762178421,
824
+ "eval_runtime": 260.0762,
825
+ "eval_samples_per_second": 12.573,
826
+ "eval_steps_per_second": 1.573,
827
+ "eval_wer": 0.3315308305777086,
828
+ "step": 10500
829
+ },
830
+ {
831
+ "epoch": 92.37,
832
+ "learning_rate": 8.585106382978724e-06,
833
+ "loss": 0.8768,
834
+ "step": 10600
835
+ },
836
+ {
837
+ "epoch": 93.25,
838
+ "learning_rate": 7.521276595744681e-06,
839
+ "loss": 0.8711,
840
+ "step": 10700
841
+ },
842
+ {
843
+ "epoch": 94.12,
844
+ "learning_rate": 6.457446808510638e-06,
845
+ "loss": 0.8774,
846
+ "step": 10800
847
+ },
848
+ {
849
+ "epoch": 94.99,
850
+ "learning_rate": 5.393617021276596e-06,
851
+ "loss": 0.8781,
852
+ "step": 10900
853
+ },
854
+ {
855
+ "epoch": 95.86,
856
+ "learning_rate": 4.329787234042553e-06,
857
+ "loss": 0.8719,
858
+ "step": 11000
859
+ },
860
+ {
861
+ "epoch": 95.86,
862
+ "eval_loss": 0.22029033303260803,
863
+ "eval_runtime": 256.8466,
864
+ "eval_samples_per_second": 12.731,
865
+ "eval_steps_per_second": 1.592,
866
+ "eval_wer": 0.32705699280295664,
867
+ "step": 11000
868
+ },
869
+ {
870
+ "epoch": 96.73,
871
+ "learning_rate": 3.265957446808511e-06,
872
+ "loss": 0.8687,
873
+ "step": 11100
874
+ },
875
+ {
876
+ "epoch": 97.6,
877
+ "learning_rate": 2.2021276595744685e-06,
878
+ "loss": 0.8771,
879
+ "step": 11200
880
+ },
881
+ {
882
+ "epoch": 98.47,
883
+ "learning_rate": 1.1382978723404256e-06,
884
+ "loss": 0.8633,
885
+ "step": 11300
886
+ },
887
+ {
888
+ "epoch": 99.35,
889
+ "learning_rate": 7.446808510638299e-08,
890
+ "loss": 0.8779,
891
+ "step": 11400
892
+ },
893
+ {
894
+ "epoch": 99.35,
895
+ "step": 11400,
896
+ "total_flos": 1.353536402091675e+20,
897
+ "train_loss": 1.3810767638892458,
898
+ "train_runtime": 73797.3063,
899
+ "train_samples_per_second": 9.949,
900
+ "train_steps_per_second": 0.154
901
+ }
902
+ ],
903
+ "logging_steps": 100,
904
+ "max_steps": 11400,
905
+ "num_train_epochs": 100,
906
+ "save_steps": 500,
907
+ "total_flos": 1.353536402091675e+20,
908
+ "trial_name": null,
909
+ "trial_params": null
910
+ }