stefan-it commited on
Commit
98ab713
1 Parent(s): 097dc95

model: add Basque model for XLSR Fine-Tuning week

Browse files
config.json ADDED
@@ -0,0 +1,76 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.1,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "Wav2Vec2ForCTC"
7
+ ],
8
+ "attention_dropout": 0.1,
9
+ "bos_token_id": 1,
10
+ "conv_bias": true,
11
+ "conv_dim": [
12
+ 512,
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512
19
+ ],
20
+ "conv_kernel": [
21
+ 10,
22
+ 3,
23
+ 3,
24
+ 3,
25
+ 3,
26
+ 2,
27
+ 2
28
+ ],
29
+ "conv_stride": [
30
+ 5,
31
+ 2,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2,
36
+ 2
37
+ ],
38
+ "ctc_loss_reduction": "mean",
39
+ "ctc_zero_infinity": true,
40
+ "do_stable_layer_norm": true,
41
+ "eos_token_id": 2,
42
+ "feat_extract_activation": "gelu",
43
+ "feat_extract_dropout": 0.0,
44
+ "feat_extract_norm": "layer",
45
+ "feat_proj_dropout": 0.0,
46
+ "final_dropout": 0.0,
47
+ "gradient_checkpointing": true,
48
+ "hidden_act": "gelu",
49
+ "hidden_dropout": 0.1,
50
+ "hidden_size": 1024,
51
+ "initializer_range": 0.02,
52
+ "intermediate_size": 4096,
53
+ "layer_norm_eps": 1e-05,
54
+ "layerdrop": 0.1,
55
+ "mask_channel_length": 10,
56
+ "mask_channel_min_space": 1,
57
+ "mask_channel_other": 0.0,
58
+ "mask_channel_prob": 0.0,
59
+ "mask_channel_selection": "static",
60
+ "mask_feature_length": 10,
61
+ "mask_feature_prob": 0.0,
62
+ "mask_time_length": 10,
63
+ "mask_time_min_space": 1,
64
+ "mask_time_other": 0.0,
65
+ "mask_time_prob": 0.05,
66
+ "mask_time_selection": "static",
67
+ "model_type": "wav2vec2",
68
+ "num_attention_heads": 16,
69
+ "num_conv_pos_embedding_groups": 16,
70
+ "num_conv_pos_embeddings": 128,
71
+ "num_feat_extract_layers": 7,
72
+ "num_hidden_layers": 24,
73
+ "pad_token_id": 30,
74
+ "transformers_version": "4.5.0.dev0",
75
+ "vocab_size": 31
76
+ }
preprocessor_config.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_size": 1,
4
+ "padding_side": "right",
5
+ "padding_value": 0.0,
6
+ "return_attention_mask": true,
7
+ "sampling_rate": 16000
8
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ea2d8092b25296935b087ef58109242b2f248265e7bd15f79e37e9ef62784ff
3
+ size 1262060951
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]"}
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "[UNK]", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "[PAD]", "do_lower_case": false, "word_delimiter_token": "|"}
trainer_state.json ADDED
@@ -0,0 +1,963 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 29.998738965952082,
5
+ "global_step": 11880,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.25,
12
+ "learning_rate": 5.9999999999999995e-05,
13
+ "loss": NaN,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 0.5,
18
+ "learning_rate": 0.00011999999999999999,
19
+ "loss": NaN,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 0.76,
24
+ "learning_rate": 0.00017999999999999998,
25
+ "loss": NaN,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 1.01,
30
+ "learning_rate": 0.00023999999999999998,
31
+ "loss": NaN,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 1.01,
36
+ "eval_loss": 2.7731735706329346,
37
+ "eval_runtime": 603.0564,
38
+ "eval_samples_per_second": 8.576,
39
+ "eval_wer": 1.0,
40
+ "step": 400
41
+ },
42
+ {
43
+ "epoch": 1.26,
44
+ "learning_rate": 0.0003,
45
+ "loss": NaN,
46
+ "step": 500
47
+ },
48
+ {
49
+ "epoch": 1.51,
50
+ "learning_rate": 0.00029736379613356763,
51
+ "loss": NaN,
52
+ "step": 600
53
+ },
54
+ {
55
+ "epoch": 1.77,
56
+ "learning_rate": 0.0002947275922671353,
57
+ "loss": NaN,
58
+ "step": 700
59
+ },
60
+ {
61
+ "epoch": 2.02,
62
+ "learning_rate": 0.00029209138840070295,
63
+ "loss": NaN,
64
+ "step": 800
65
+ },
66
+ {
67
+ "epoch": 2.02,
68
+ "eval_loss": 0.2804177403450012,
69
+ "eval_runtime": 605.6208,
70
+ "eval_samples_per_second": 8.54,
71
+ "eval_wer": 0.516512825876631,
72
+ "step": 800
73
+ },
74
+ {
75
+ "epoch": 2.27,
76
+ "learning_rate": 0.0002894551845342706,
77
+ "loss": NaN,
78
+ "step": 900
79
+ },
80
+ {
81
+ "epoch": 2.52,
82
+ "learning_rate": 0.00028681898066783827,
83
+ "loss": NaN,
84
+ "step": 1000
85
+ },
86
+ {
87
+ "epoch": 2.78,
88
+ "learning_rate": 0.000284182776801406,
89
+ "loss": NaN,
90
+ "step": 1100
91
+ },
92
+ {
93
+ "epoch": 3.03,
94
+ "learning_rate": 0.00028154657293497364,
95
+ "loss": NaN,
96
+ "step": 1200
97
+ },
98
+ {
99
+ "epoch": 3.03,
100
+ "eval_loss": 0.19836615025997162,
101
+ "eval_runtime": 613.5068,
102
+ "eval_samples_per_second": 8.43,
103
+ "eval_wer": 0.41406270428575165,
104
+ "step": 1200
105
+ },
106
+ {
107
+ "epoch": 3.28,
108
+ "learning_rate": 0.00027891036906854124,
109
+ "loss": NaN,
110
+ "step": 1300
111
+ },
112
+ {
113
+ "epoch": 3.53,
114
+ "learning_rate": 0.00027627416520210896,
115
+ "loss": NaN,
116
+ "step": 1400
117
+ },
118
+ {
119
+ "epoch": 3.79,
120
+ "learning_rate": 0.0002736379613356766,
121
+ "loss": NaN,
122
+ "step": 1500
123
+ },
124
+ {
125
+ "epoch": 4.04,
126
+ "learning_rate": 0.00027100175746924427,
127
+ "loss": NaN,
128
+ "step": 1600
129
+ },
130
+ {
131
+ "epoch": 4.04,
132
+ "eval_loss": 0.18502908945083618,
133
+ "eval_runtime": 614.3947,
134
+ "eval_samples_per_second": 8.418,
135
+ "eval_wer": 0.3799911094840886,
136
+ "step": 1600
137
+ },
138
+ {
139
+ "epoch": 4.29,
140
+ "learning_rate": 0.00026836555360281193,
141
+ "loss": NaN,
142
+ "step": 1700
143
+ },
144
+ {
145
+ "epoch": 4.54,
146
+ "learning_rate": 0.0002657293497363796,
147
+ "loss": NaN,
148
+ "step": 1800
149
+ },
150
+ {
151
+ "epoch": 4.8,
152
+ "learning_rate": 0.00026309314586994725,
153
+ "loss": NaN,
154
+ "step": 1900
155
+ },
156
+ {
157
+ "epoch": 5.05,
158
+ "learning_rate": 0.0002604569420035149,
159
+ "loss": NaN,
160
+ "step": 2000
161
+ },
162
+ {
163
+ "epoch": 5.05,
164
+ "eval_loss": 0.16938243806362152,
165
+ "eval_runtime": 609.4739,
166
+ "eval_samples_per_second": 8.486,
167
+ "eval_wer": 0.36521716392542425,
168
+ "step": 2000
169
+ },
170
+ {
171
+ "epoch": 5.3,
172
+ "learning_rate": 0.00025782073813708257,
173
+ "loss": NaN,
174
+ "step": 2100
175
+ },
176
+ {
177
+ "epoch": 5.55,
178
+ "learning_rate": 0.0002551845342706502,
179
+ "loss": NaN,
180
+ "step": 2200
181
+ },
182
+ {
183
+ "epoch": 5.81,
184
+ "learning_rate": 0.00025254833040421794,
185
+ "loss": NaN,
186
+ "step": 2300
187
+ },
188
+ {
189
+ "epoch": 6.06,
190
+ "learning_rate": 0.00024991212653778554,
191
+ "loss": NaN,
192
+ "step": 2400
193
+ },
194
+ {
195
+ "epoch": 6.06,
196
+ "eval_loss": 0.17296069860458374,
197
+ "eval_runtime": 608.666,
198
+ "eval_samples_per_second": 8.497,
199
+ "eval_wer": 0.35112308134822057,
200
+ "step": 2400
201
+ },
202
+ {
203
+ "epoch": 6.31,
204
+ "learning_rate": 0.00024727592267135325,
205
+ "loss": NaN,
206
+ "step": 2500
207
+ },
208
+ {
209
+ "epoch": 6.56,
210
+ "learning_rate": 0.0002446397188049209,
211
+ "loss": NaN,
212
+ "step": 2600
213
+ },
214
+ {
215
+ "epoch": 6.82,
216
+ "learning_rate": 0.00024200351493848854,
217
+ "loss": NaN,
218
+ "step": 2700
219
+ },
220
+ {
221
+ "epoch": 7.07,
222
+ "learning_rate": 0.00023936731107205623,
223
+ "loss": NaN,
224
+ "step": 2800
225
+ },
226
+ {
227
+ "epoch": 7.07,
228
+ "eval_loss": 0.17802052199840546,
229
+ "eval_runtime": 612.551,
230
+ "eval_samples_per_second": 8.443,
231
+ "eval_wer": 0.3466778233925163,
232
+ "step": 2800
233
+ },
234
+ {
235
+ "epoch": 7.32,
236
+ "learning_rate": 0.00023673110720562386,
237
+ "loss": NaN,
238
+ "step": 2900
239
+ },
240
+ {
241
+ "epoch": 7.58,
242
+ "learning_rate": 0.00023409490333919155,
243
+ "loss": NaN,
244
+ "step": 3000
245
+ },
246
+ {
247
+ "epoch": 7.83,
248
+ "learning_rate": 0.0002314586994727592,
249
+ "loss": NaN,
250
+ "step": 3100
251
+ },
252
+ {
253
+ "epoch": 8.08,
254
+ "learning_rate": 0.0002288224956063269,
255
+ "loss": NaN,
256
+ "step": 3200
257
+ },
258
+ {
259
+ "epoch": 8.08,
260
+ "eval_loss": 0.17137496173381805,
261
+ "eval_runtime": 621.5301,
262
+ "eval_samples_per_second": 8.321,
263
+ "eval_wer": 0.343304657061423,
264
+ "step": 3200
265
+ },
266
+ {
267
+ "epoch": 8.33,
268
+ "learning_rate": 0.00022618629173989452,
269
+ "loss": NaN,
270
+ "step": 3300
271
+ },
272
+ {
273
+ "epoch": 8.59,
274
+ "learning_rate": 0.00022355008787346218,
275
+ "loss": NaN,
276
+ "step": 3400
277
+ },
278
+ {
279
+ "epoch": 8.84,
280
+ "learning_rate": 0.00022091388400702987,
281
+ "loss": NaN,
282
+ "step": 3500
283
+ },
284
+ {
285
+ "epoch": 9.09,
286
+ "learning_rate": 0.0002182776801405975,
287
+ "loss": NaN,
288
+ "step": 3600
289
+ },
290
+ {
291
+ "epoch": 9.09,
292
+ "eval_loss": 0.16793519258499146,
293
+ "eval_runtime": 613.7107,
294
+ "eval_samples_per_second": 8.427,
295
+ "eval_wer": 0.3368721073137568,
296
+ "step": 3600
297
+ },
298
+ {
299
+ "epoch": 9.34,
300
+ "learning_rate": 0.00021564147627416518,
301
+ "loss": NaN,
302
+ "step": 3700
303
+ },
304
+ {
305
+ "epoch": 9.6,
306
+ "learning_rate": 0.00021300527240773284,
307
+ "loss": NaN,
308
+ "step": 3800
309
+ },
310
+ {
311
+ "epoch": 9.85,
312
+ "learning_rate": 0.00021036906854130053,
313
+ "loss": NaN,
314
+ "step": 3900
315
+ },
316
+ {
317
+ "epoch": 10.1,
318
+ "learning_rate": 0.00020773286467486816,
319
+ "loss": NaN,
320
+ "step": 4000
321
+ },
322
+ {
323
+ "epoch": 10.1,
324
+ "eval_loss": 0.165544793009758,
325
+ "eval_runtime": 616.3457,
326
+ "eval_samples_per_second": 8.391,
327
+ "eval_wer": 0.3317731349528018,
328
+ "step": 4000
329
+ },
330
+ {
331
+ "epoch": 10.35,
332
+ "learning_rate": 0.00020509666080843582,
333
+ "loss": NaN,
334
+ "step": 4100
335
+ },
336
+ {
337
+ "epoch": 10.61,
338
+ "learning_rate": 0.0002024604569420035,
339
+ "loss": NaN,
340
+ "step": 4200
341
+ },
342
+ {
343
+ "epoch": 10.86,
344
+ "learning_rate": 0.00019982425307557116,
345
+ "loss": NaN,
346
+ "step": 4300
347
+ },
348
+ {
349
+ "epoch": 11.11,
350
+ "learning_rate": 0.00019718804920913882,
351
+ "loss": NaN,
352
+ "step": 4400
353
+ },
354
+ {
355
+ "epoch": 11.11,
356
+ "eval_loss": 0.17516674101352692,
357
+ "eval_runtime": 616.3789,
358
+ "eval_samples_per_second": 8.391,
359
+ "eval_wer": 0.32222890463614257,
360
+ "step": 4400
361
+ },
362
+ {
363
+ "epoch": 11.36,
364
+ "learning_rate": 0.00019455184534270648,
365
+ "loss": NaN,
366
+ "step": 4500
367
+ },
368
+ {
369
+ "epoch": 11.62,
370
+ "learning_rate": 0.00019191564147627417,
371
+ "loss": NaN,
372
+ "step": 4600
373
+ },
374
+ {
375
+ "epoch": 11.87,
376
+ "learning_rate": 0.0001892794376098418,
377
+ "loss": NaN,
378
+ "step": 4700
379
+ },
380
+ {
381
+ "epoch": 12.12,
382
+ "learning_rate": 0.00018664323374340946,
383
+ "loss": NaN,
384
+ "step": 4800
385
+ },
386
+ {
387
+ "epoch": 12.12,
388
+ "eval_loss": 0.18037526309490204,
389
+ "eval_runtime": 612.0676,
390
+ "eval_samples_per_second": 8.45,
391
+ "eval_wer": 0.3176790523755981,
392
+ "step": 4800
393
+ },
394
+ {
395
+ "epoch": 12.37,
396
+ "learning_rate": 0.00018400702987697714,
397
+ "loss": NaN,
398
+ "step": 4900
399
+ },
400
+ {
401
+ "epoch": 12.63,
402
+ "learning_rate": 0.0001813708260105448,
403
+ "loss": NaN,
404
+ "step": 5000
405
+ },
406
+ {
407
+ "epoch": 12.88,
408
+ "learning_rate": 0.00017873462214411246,
409
+ "loss": NaN,
410
+ "step": 5100
411
+ },
412
+ {
413
+ "epoch": 13.13,
414
+ "learning_rate": 0.00017609841827768012,
415
+ "loss": NaN,
416
+ "step": 5200
417
+ },
418
+ {
419
+ "epoch": 13.13,
420
+ "eval_loss": 0.1690378040075302,
421
+ "eval_runtime": 616.3617,
422
+ "eval_samples_per_second": 8.391,
423
+ "eval_wer": 0.31728682373244776,
424
+ "step": 5200
425
+ },
426
+ {
427
+ "epoch": 13.38,
428
+ "learning_rate": 0.0001734622144112478,
429
+ "loss": NaN,
430
+ "step": 5300
431
+ },
432
+ {
433
+ "epoch": 13.64,
434
+ "learning_rate": 0.00017082601054481546,
435
+ "loss": NaN,
436
+ "step": 5400
437
+ },
438
+ {
439
+ "epoch": 13.89,
440
+ "learning_rate": 0.0001681898066783831,
441
+ "loss": NaN,
442
+ "step": 5500
443
+ },
444
+ {
445
+ "epoch": 14.14,
446
+ "learning_rate": 0.00016555360281195078,
447
+ "loss": NaN,
448
+ "step": 5600
449
+ },
450
+ {
451
+ "epoch": 14.14,
452
+ "eval_loss": 0.17649142444133759,
453
+ "eval_runtime": 619.1476,
454
+ "eval_samples_per_second": 8.353,
455
+ "eval_wer": 0.31932641267682976,
456
+ "step": 5600
457
+ },
458
+ {
459
+ "epoch": 14.39,
460
+ "learning_rate": 0.00016291739894551844,
461
+ "loss": NaN,
462
+ "step": 5700
463
+ },
464
+ {
465
+ "epoch": 14.65,
466
+ "learning_rate": 0.00016028119507908612,
467
+ "loss": NaN,
468
+ "step": 5800
469
+ },
470
+ {
471
+ "epoch": 14.9,
472
+ "learning_rate": 0.00015764499121265376,
473
+ "loss": NaN,
474
+ "step": 5900
475
+ },
476
+ {
477
+ "epoch": 15.15,
478
+ "learning_rate": 0.00015500878734622144,
479
+ "loss": NaN,
480
+ "step": 6000
481
+ },
482
+ {
483
+ "epoch": 15.15,
484
+ "eval_loss": 0.17734038829803467,
485
+ "eval_runtime": 622.5524,
486
+ "eval_samples_per_second": 8.308,
487
+ "eval_wer": 0.3148288575687054,
488
+ "step": 6000
489
+ },
490
+ {
491
+ "epoch": 15.4,
492
+ "learning_rate": 0.0001523725834797891,
493
+ "loss": NaN,
494
+ "step": 6100
495
+ },
496
+ {
497
+ "epoch": 15.66,
498
+ "learning_rate": 0.00014973637961335676,
499
+ "loss": NaN,
500
+ "step": 6200
501
+ },
502
+ {
503
+ "epoch": 15.91,
504
+ "learning_rate": 0.00014710017574692442,
505
+ "loss": NaN,
506
+ "step": 6300
507
+ },
508
+ {
509
+ "epoch": 16.16,
510
+ "learning_rate": 0.00014446397188049208,
511
+ "loss": NaN,
512
+ "step": 6400
513
+ },
514
+ {
515
+ "epoch": 16.16,
516
+ "eval_loss": 0.187224343419075,
517
+ "eval_runtime": 615.2372,
518
+ "eval_samples_per_second": 8.407,
519
+ "eval_wer": 0.3136260230630442,
520
+ "step": 6400
521
+ },
522
+ {
523
+ "epoch": 16.41,
524
+ "learning_rate": 0.00014182776801405973,
525
+ "loss": NaN,
526
+ "step": 6500
527
+ },
528
+ {
529
+ "epoch": 16.67,
530
+ "learning_rate": 0.0001391915641476274,
531
+ "loss": NaN,
532
+ "step": 6600
533
+ },
534
+ {
535
+ "epoch": 16.92,
536
+ "learning_rate": 0.00013655536028119505,
537
+ "loss": NaN,
538
+ "step": 6700
539
+ },
540
+ {
541
+ "epoch": 17.17,
542
+ "learning_rate": 0.00013391915641476274,
543
+ "loss": NaN,
544
+ "step": 6800
545
+ },
546
+ {
547
+ "epoch": 17.17,
548
+ "eval_loss": 0.1843162626028061,
549
+ "eval_runtime": 620.0086,
550
+ "eval_samples_per_second": 8.342,
551
+ "eval_wer": 0.30912846795491983,
552
+ "step": 6800
553
+ },
554
+ {
555
+ "epoch": 17.42,
556
+ "learning_rate": 0.0001312829525483304,
557
+ "loss": NaN,
558
+ "step": 6900
559
+ },
560
+ {
561
+ "epoch": 17.68,
562
+ "learning_rate": 0.00012864674868189805,
563
+ "loss": NaN,
564
+ "step": 7000
565
+ },
566
+ {
567
+ "epoch": 17.93,
568
+ "learning_rate": 0.0001260105448154657,
569
+ "loss": NaN,
570
+ "step": 7100
571
+ },
572
+ {
573
+ "epoch": 18.18,
574
+ "learning_rate": 0.00012337434094903337,
575
+ "loss": NaN,
576
+ "step": 7200
577
+ },
578
+ {
579
+ "epoch": 18.18,
580
+ "eval_loss": 0.1754978895187378,
581
+ "eval_runtime": 616.3908,
582
+ "eval_samples_per_second": 8.391,
583
+ "eval_wer": 0.3102005595795309,
584
+ "step": 7200
585
+ },
586
+ {
587
+ "epoch": 18.43,
588
+ "learning_rate": 0.00012073813708260104,
589
+ "loss": NaN,
590
+ "step": 7300
591
+ },
592
+ {
593
+ "epoch": 18.69,
594
+ "learning_rate": 0.0001181019332161687,
595
+ "loss": NaN,
596
+ "step": 7400
597
+ },
598
+ {
599
+ "epoch": 18.94,
600
+ "learning_rate": 0.00011546572934973637,
601
+ "loss": NaN,
602
+ "step": 7500
603
+ },
604
+ {
605
+ "epoch": 19.19,
606
+ "learning_rate": 0.00011282952548330403,
607
+ "loss": NaN,
608
+ "step": 7600
609
+ },
610
+ {
611
+ "epoch": 19.19,
612
+ "eval_loss": 0.18895429372787476,
613
+ "eval_runtime": 623.3336,
614
+ "eval_samples_per_second": 8.297,
615
+ "eval_wer": 0.30520618152341605,
616
+ "step": 7600
617
+ },
618
+ {
619
+ "epoch": 19.44,
620
+ "learning_rate": 0.0001101933216168717,
621
+ "loss": NaN,
622
+ "step": 7700
623
+ },
624
+ {
625
+ "epoch": 19.7,
626
+ "learning_rate": 0.00010755711775043936,
627
+ "loss": NaN,
628
+ "step": 7800
629
+ },
630
+ {
631
+ "epoch": 19.95,
632
+ "learning_rate": 0.00010492091388400701,
633
+ "loss": NaN,
634
+ "step": 7900
635
+ },
636
+ {
637
+ "epoch": 20.2,
638
+ "learning_rate": 0.00010228471001757468,
639
+ "loss": NaN,
640
+ "step": 8000
641
+ },
642
+ {
643
+ "epoch": 20.2,
644
+ "eval_loss": 0.19008345901966095,
645
+ "eval_runtime": 614.7127,
646
+ "eval_samples_per_second": 8.414,
647
+ "eval_wer": 0.30334963261250425,
648
+ "step": 8000
649
+ },
650
+ {
651
+ "epoch": 20.45,
652
+ "learning_rate": 9.964850615114234e-05,
653
+ "loss": NaN,
654
+ "step": 8100
655
+ },
656
+ {
657
+ "epoch": 20.71,
658
+ "learning_rate": 9.701230228471001e-05,
659
+ "loss": NaN,
660
+ "step": 8200
661
+ },
662
+ {
663
+ "epoch": 20.96,
664
+ "learning_rate": 9.437609841827767e-05,
665
+ "loss": NaN,
666
+ "step": 8300
667
+ },
668
+ {
669
+ "epoch": 21.21,
670
+ "learning_rate": 9.173989455184534e-05,
671
+ "loss": NaN,
672
+ "step": 8400
673
+ },
674
+ {
675
+ "epoch": 21.21,
676
+ "eval_loss": 0.1935020387172699,
677
+ "eval_runtime": 619.5234,
678
+ "eval_samples_per_second": 8.348,
679
+ "eval_wer": 0.307193473315378,
680
+ "step": 8400
681
+ },
682
+ {
683
+ "epoch": 21.46,
684
+ "learning_rate": 8.9103690685413e-05,
685
+ "loss": NaN,
686
+ "step": 8500
687
+ },
688
+ {
689
+ "epoch": 21.72,
690
+ "learning_rate": 8.646748681898065e-05,
691
+ "loss": NaN,
692
+ "step": 8600
693
+ },
694
+ {
695
+ "epoch": 21.97,
696
+ "learning_rate": 8.383128295254832e-05,
697
+ "loss": NaN,
698
+ "step": 8700
699
+ },
700
+ {
701
+ "epoch": 22.22,
702
+ "learning_rate": 8.119507908611598e-05,
703
+ "loss": NaN,
704
+ "step": 8800
705
+ },
706
+ {
707
+ "epoch": 22.22,
708
+ "eval_loss": 0.19641530513763428,
709
+ "eval_runtime": 611.6485,
710
+ "eval_samples_per_second": 8.456,
711
+ "eval_wer": 0.30502314148994586,
712
+ "step": 8800
713
+ },
714
+ {
715
+ "epoch": 22.47,
716
+ "learning_rate": 7.855887521968365e-05,
717
+ "loss": NaN,
718
+ "step": 8900
719
+ },
720
+ {
721
+ "epoch": 22.73,
722
+ "learning_rate": 7.592267135325131e-05,
723
+ "loss": NaN,
724
+ "step": 9000
725
+ },
726
+ {
727
+ "epoch": 22.98,
728
+ "learning_rate": 7.328646748681898e-05,
729
+ "loss": NaN,
730
+ "step": 9100
731
+ },
732
+ {
733
+ "epoch": 23.23,
734
+ "learning_rate": 7.065026362038664e-05,
735
+ "loss": NaN,
736
+ "step": 9200
737
+ },
738
+ {
739
+ "epoch": 23.23,
740
+ "eval_loss": 0.19613772630691528,
741
+ "eval_runtime": 613.3141,
742
+ "eval_samples_per_second": 8.433,
743
+ "eval_wer": 0.3008132207201318,
744
+ "step": 9200
745
+ },
746
+ {
747
+ "epoch": 23.48,
748
+ "learning_rate": 6.80140597539543e-05,
749
+ "loss": NaN,
750
+ "step": 9300
751
+ },
752
+ {
753
+ "epoch": 23.74,
754
+ "learning_rate": 6.537785588752197e-05,
755
+ "loss": NaN,
756
+ "step": 9400
757
+ },
758
+ {
759
+ "epoch": 23.99,
760
+ "learning_rate": 6.274165202108963e-05,
761
+ "loss": NaN,
762
+ "step": 9500
763
+ },
764
+ {
765
+ "epoch": 24.24,
766
+ "learning_rate": 6.0105448154657286e-05,
767
+ "loss": NaN,
768
+ "step": 9600
769
+ },
770
+ {
771
+ "epoch": 24.24,
772
+ "eval_loss": 0.18839870393276215,
773
+ "eval_runtime": 615.6693,
774
+ "eval_samples_per_second": 8.401,
775
+ "eval_wer": 0.29623721988337737,
776
+ "step": 9600
777
+ },
778
+ {
779
+ "epoch": 24.49,
780
+ "learning_rate": 5.746924428822495e-05,
781
+ "loss": NaN,
782
+ "step": 9700
783
+ },
784
+ {
785
+ "epoch": 24.75,
786
+ "learning_rate": 5.483304042179262e-05,
787
+ "loss": NaN,
788
+ "step": 9800
789
+ },
790
+ {
791
+ "epoch": 25.0,
792
+ "learning_rate": 5.2196836555360276e-05,
793
+ "loss": NaN,
794
+ "step": 9900
795
+ },
796
+ {
797
+ "epoch": 25.25,
798
+ "learning_rate": 4.956063268892794e-05,
799
+ "loss": NaN,
800
+ "step": 10000
801
+ },
802
+ {
803
+ "epoch": 25.25,
804
+ "eval_loss": 0.19206073880195618,
805
+ "eval_runtime": 610.6418,
806
+ "eval_samples_per_second": 8.47,
807
+ "eval_wer": 0.2963418141882175,
808
+ "step": 10000
809
+ },
810
+ {
811
+ "epoch": 25.5,
812
+ "learning_rate": 4.6924428822495606e-05,
813
+ "loss": NaN,
814
+ "step": 10100
815
+ },
816
+ {
817
+ "epoch": 25.76,
818
+ "learning_rate": 4.4288224956063265e-05,
819
+ "loss": NaN,
820
+ "step": 10200
821
+ },
822
+ {
823
+ "epoch": 26.01,
824
+ "learning_rate": 4.165202108963093e-05,
825
+ "loss": NaN,
826
+ "step": 10300
827
+ },
828
+ {
829
+ "epoch": 26.26,
830
+ "learning_rate": 3.901581722319859e-05,
831
+ "loss": NaN,
832
+ "step": 10400
833
+ },
834
+ {
835
+ "epoch": 26.26,
836
+ "eval_loss": 0.19253745675086975,
837
+ "eval_runtime": 622.5244,
838
+ "eval_samples_per_second": 8.308,
839
+ "eval_wer": 0.2921841905708234,
840
+ "step": 10400
841
+ },
842
+ {
843
+ "epoch": 26.51,
844
+ "learning_rate": 3.6379613356766254e-05,
845
+ "loss": NaN,
846
+ "step": 10500
847
+ },
848
+ {
849
+ "epoch": 26.77,
850
+ "learning_rate": 3.374340949033391e-05,
851
+ "loss": NaN,
852
+ "step": 10600
853
+ },
854
+ {
855
+ "epoch": 27.02,
856
+ "learning_rate": 3.110720562390158e-05,
857
+ "loss": NaN,
858
+ "step": 10700
859
+ },
860
+ {
861
+ "epoch": 27.27,
862
+ "learning_rate": 2.847100175746924e-05,
863
+ "loss": NaN,
864
+ "step": 10800
865
+ },
866
+ {
867
+ "epoch": 27.27,
868
+ "eval_loss": 0.19466331601142883,
869
+ "eval_runtime": 617.6765,
870
+ "eval_samples_per_second": 8.373,
871
+ "eval_wer": 0.292785607823654,
872
+ "step": 10800
873
+ },
874
+ {
875
+ "epoch": 27.52,
876
+ "learning_rate": 2.5834797891036906e-05,
877
+ "loss": NaN,
878
+ "step": 10900
879
+ },
880
+ {
881
+ "epoch": 27.78,
882
+ "learning_rate": 2.3198594024604568e-05,
883
+ "loss": NaN,
884
+ "step": 11000
885
+ },
886
+ {
887
+ "epoch": 28.03,
888
+ "learning_rate": 2.0562390158172233e-05,
889
+ "loss": NaN,
890
+ "step": 11100
891
+ },
892
+ {
893
+ "epoch": 28.28,
894
+ "learning_rate": 1.792618629173989e-05,
895
+ "loss": NaN,
896
+ "step": 11200
897
+ },
898
+ {
899
+ "epoch": 28.28,
900
+ "eval_loss": 0.1931692361831665,
901
+ "eval_runtime": 617.2527,
902
+ "eval_samples_per_second": 8.379,
903
+ "eval_wer": 0.29189655623251315,
904
+ "step": 11200
905
+ },
906
+ {
907
+ "epoch": 28.53,
908
+ "learning_rate": 1.5289982425307557e-05,
909
+ "loss": NaN,
910
+ "step": 11300
911
+ },
912
+ {
913
+ "epoch": 28.79,
914
+ "learning_rate": 1.2653778558875219e-05,
915
+ "loss": NaN,
916
+ "step": 11400
917
+ },
918
+ {
919
+ "epoch": 29.04,
920
+ "learning_rate": 1.0017574692442883e-05,
921
+ "loss": NaN,
922
+ "step": 11500
923
+ },
924
+ {
925
+ "epoch": 29.29,
926
+ "learning_rate": 7.381370826010544e-06,
927
+ "loss": NaN,
928
+ "step": 11600
929
+ },
930
+ {
931
+ "epoch": 29.29,
932
+ "eval_loss": 0.1954003870487213,
933
+ "eval_runtime": 617.8209,
934
+ "eval_samples_per_second": 8.371,
935
+ "eval_wer": 0.2904322359647517,
936
+ "step": 11600
937
+ },
938
+ {
939
+ "epoch": 29.54,
940
+ "learning_rate": 4.745166959578207e-06,
941
+ "loss": NaN,
942
+ "step": 11700
943
+ },
944
+ {
945
+ "epoch": 29.8,
946
+ "learning_rate": 2.1089630931458696e-06,
947
+ "loss": NaN,
948
+ "step": 11800
949
+ },
950
+ {
951
+ "epoch": 30.0,
952
+ "step": 11880,
953
+ "total_flos": 6.1923487227453555e+19,
954
+ "train_runtime": 56662.4324,
955
+ "train_samples_per_second": 0.21
956
+ }
957
+ ],
958
+ "max_steps": 11880,
959
+ "num_train_epochs": 30,
960
+ "total_flos": 6.1923487227453555e+19,
961
+ "trial_name": null,
962
+ "trial_params": null
963
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50fab4116f25a030f496149d1ebf5bba2adfc5863fd95a909a4c78b400cd9155
3
+ size 2351
vocab.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"i": 0, "t": 1, "b": 2, "n": 3, "q": 4, "a": 5, "d": 6, "o": 7, "r": 8, "h": 9, "x": 10, "y": 11, "ñ": 12, "f": 14, "í": 15, "e": 16, "z": 17, "g": 18, "j": 19, "v": 20, "p": 21, "l": 22, "m": 23, "s": 24, "c": 25, "w": 26, "k": 27, "u": 28, "|": 13, "[UNK]": 29, "[PAD]": 30}