Hugging Face's logo

nihalbaig
/
wav2vec2-large-xlsr-bn

Automatic Speech Recognition
Transformers
PyTorch
wav2vec2
Inference Endpoints
Model card Files Community
nihalbaig commited on
Commit
e8053f9
•
1 Parent(s): c781bf7

adding training model

Browse files
Files changed (28) hide show
  1. all_results.json +9 -0
  2. checkpoint-18400/config.json +115 -0
  3. checkpoint-18400/optimizer.pt +3 -0
  4. checkpoint-18400/preprocessor_config.json +10 -0
  5. checkpoint-18400/pytorch_model.bin +3 -0
  6. checkpoint-18400/rng_state.pth +3 -0
  7. checkpoint-18400/scaler.pt +3 -0
  8. checkpoint-18400/scheduler.pt +3 -0
  9. checkpoint-18400/trainer_state.json +706 -0
  10. checkpoint-18400/training_args.bin +3 -0
  11. checkpoint-800/config.json +115 -0
  12. checkpoint-800/optimizer.pt +3 -0
  13. checkpoint-800/preprocessor_config.json +10 -0
  14. checkpoint-800/pytorch_model.bin +3 -0
  15. checkpoint-800/rng_state.pth +3 -0
  16. checkpoint-800/scaler.pt +3 -0
  17. checkpoint-800/scheduler.pt +3 -0
  18. checkpoint-800/trainer_state.json +46 -0
  19. checkpoint-800/training_args.bin +3 -0
  20. config.json +115 -0
  21. preprocessor_config.json +10 -0
  22. pytorch_model.bin +3 -0
  23. special_tokens_map.json +1 -0
  24. tokenizer_config.json +1 -0
  25. train_results.json +9 -0
  26. trainer_state.json +715 -0
  27. training_args.bin +3 -0
  28. vocab.json +1 -0
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 30.0,
3
+ "total_flos": 1.1136444297962675e+20,
4
+ "train_loss": 2.53877658203125,
5
+ "train_runtime": 56941.5798,
6
+ "train_samples": 20000,
7
+ "train_samples_per_second": 10.537,
8
+ "train_steps_per_second": 0.329
9
+ }
checkpoint-18400/config.json ADDED
@@ -0,0 +1,115 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.0,
4
+ "adapter_kernel_size": 3,
5
+ "adapter_stride": 2,
6
+ "add_adapter": false,
7
+ "apply_spec_augment": true,
8
+ "architectures": [
9
+ "Wav2Vec2ForCTC"
10
+ ],
11
+ "attention_dropout": 0.1,
12
+ "bos_token_id": 1,
13
+ "classifier_proj_size": 256,
14
+ "codevector_dim": 768,
15
+ "contrastive_logits_temperature": 0.1,
16
+ "conv_bias": true,
17
+ "conv_dim": [
18
+ 512,
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
+ ],
26
+ "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
+ ],
35
+ "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
+ ],
44
+ "ctc_loss_reduction": "mean",
45
+ "ctc_zero_infinity": false,
46
+ "diversity_loss_weight": 0.1,
47
+ "do_stable_layer_norm": true,
48
+ "eos_token_id": 2,
49
+ "feat_extract_activation": "gelu",
50
+ "feat_extract_dropout": 0.0,
51
+ "feat_extract_norm": "layer",
52
+ "feat_proj_dropout": 0.0,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.0,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
57
+ "hidden_size": 1024,
58
+ "initializer_range": 0.02,
59
+ "intermediate_size": 4096,
60
+ "layer_norm_eps": 1e-05,
61
+ "layerdrop": 0.1,
62
+ "mask_channel_length": 10,
63
+ "mask_channel_min_space": 1,
64
+ "mask_channel_other": 0.0,
65
+ "mask_channel_prob": 0.0,
66
+ "mask_channel_selection": "static",
67
+ "mask_feature_length": 10,
68
+ "mask_feature_min_masks": 0,
69
+ "mask_feature_prob": 0.0,
70
+ "mask_time_length": 10,
71
+ "mask_time_min_masks": 2,
72
+ "mask_time_min_space": 1,
73
+ "mask_time_other": 0.0,
74
+ "mask_time_prob": 0.05,
75
+ "mask_time_selection": "static",
76
+ "model_type": "wav2vec2",
77
+ "num_adapter_layers": 3,
78
+ "num_attention_heads": 16,
79
+ "num_codevector_groups": 2,
80
+ "num_codevectors_per_group": 320,
81
+ "num_conv_pos_embedding_groups": 16,
82
+ "num_conv_pos_embeddings": 128,
83
+ "num_feat_extract_layers": 7,
84
+ "num_hidden_layers": 24,
85
+ "num_negatives": 100,
86
+ "output_hidden_size": 1024,
87
+ "pad_token_id": 77,
88
+ "proj_codevector_dim": 768,
89
+ "tdnn_dilation": [
90
+ 1,
91
+ 2,
92
+ 3,
93
+ 1,
94
+ 1
95
+ ],
96
+ "tdnn_dim": [
97
+ 512,
98
+ 512,
99
+ 512,
100
+ 512,
101
+ 1500
102
+ ],
103
+ "tdnn_kernel": [
104
+ 5,
105
+ 3,
106
+ 3,
107
+ 1,
108
+ 1
109
+ ],
110
+ "torch_dtype": "float32",
111
+ "transformers_version": "4.18.0",
112
+ "use_weighted_layer_sum": false,
113
+ "vocab_size": 78,
114
+ "xvector_output_dim": 512
115
+ }
checkpoint-18400/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ba3803c4c7f28e0e007a987a5a8859daff5e38dfab4f7f8bdf645ede9d109090
3
+ size 2490698513
checkpoint-18400/preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
checkpoint-18400/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7f2bf907fe5979a91838d4c81bf39d8f47c5b6e2a596d8d0979eca6a95a818d
3
+ size 1262218481
checkpoint-18400/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ca5b70aad179e633220a2ca0500c0904432192968a39f457977dcffcedb9d5d
3
+ size 14567
checkpoint-18400/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5df0168a751b1257361283b5aa0c288d7c712011edfd861c618859611a2278ec
3
+ size 559
checkpoint-18400/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7ce64d0c518724da2926330a57f05a4589537588b4613fb8a65cd53821d2d5b
3
+ size 623
checkpoint-18400/trainer_state.json ADDED
@@ -0,0 +1,706 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8215018908698001,
3
+ "best_model_checkpoint": "/content/wav2vec2-large-xlsr-bn/checkpoint-800",
4
+ "epoch": 29.44,
5
+ "global_step": 18400,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.64,
12
+ "learning_rate": 0.00029405766150560596,
13
+ "loss": 4.1339,
14
+ "step": 400
15
+ },
16
+ {
17
+ "epoch": 0.64,
18
+ "eval_loss": 3.2776501178741455,
19
+ "eval_runtime": 74.0715,
20
+ "eval_samples_per_second": 13.5,
21
+ "eval_steps_per_second": 1.688,
22
+ "eval_wer": 1.0,
23
+ "step": 400
24
+ },
25
+ {
26
+ "epoch": 1.28,
27
+ "learning_rate": 0.00028774693005872927,
28
+ "loss": 2.0718,
29
+ "step": 800
30
+ },
31
+ {
32
+ "epoch": 1.28,
33
+ "eval_loss": 0.979073703289032,
34
+ "eval_runtime": 74.5153,
35
+ "eval_samples_per_second": 13.42,
36
+ "eval_steps_per_second": 1.678,
37
+ "eval_wer": 0.8215018908698001,
38
+ "step": 800
39
+ },
40
+ {
41
+ "epoch": 1.92,
42
+ "learning_rate": 0.00028143619861185264,
43
+ "loss": 2.4619,
44
+ "step": 1200
45
+ },
46
+ {
47
+ "epoch": 1.92,
48
+ "eval_loss": 3.080274820327759,
49
+ "eval_runtime": 74.0712,
50
+ "eval_samples_per_second": 13.501,
51
+ "eval_steps_per_second": 1.688,
52
+ "eval_wer": 1.0,
53
+ "step": 1200
54
+ },
55
+ {
56
+ "epoch": 2.56,
57
+ "learning_rate": 0.000275157501334757,
58
+ "loss": 3.1703,
59
+ "step": 1600
60
+ },
61
+ {
62
+ "epoch": 2.56,
63
+ "eval_loss": 3.4948482513427734,
64
+ "eval_runtime": 74.092,
65
+ "eval_samples_per_second": 13.497,
66
+ "eval_steps_per_second": 1.687,
67
+ "eval_wer": 1.0,
68
+ "step": 1600
69
+ },
70
+ {
71
+ "epoch": 3.2,
72
+ "learning_rate": 0.00026886278697277094,
73
+ "loss": 5.5212,
74
+ "step": 2000
75
+ },
76
+ {
77
+ "epoch": 3.2,
78
+ "eval_loss": 6.071021556854248,
79
+ "eval_runtime": 74.4322,
80
+ "eval_samples_per_second": 13.435,
81
+ "eval_steps_per_second": 1.679,
82
+ "eval_wer": 1.0,
83
+ "step": 2000
84
+ },
85
+ {
86
+ "epoch": 3.84,
87
+ "learning_rate": 0.0002625680726107848,
88
+ "loss": 5.8864,
89
+ "step": 2400
90
+ },
91
+ {
92
+ "epoch": 3.84,
93
+ "eval_loss": 6.071021556854248,
94
+ "eval_runtime": 74.0924,
95
+ "eval_samples_per_second": 13.497,
96
+ "eval_steps_per_second": 1.687,
97
+ "eval_wer": 1.0,
98
+ "step": 2400
99
+ },
100
+ {
101
+ "epoch": 4.48,
102
+ "learning_rate": 0.00025628937533368924,
103
+ "loss": 5.9244,
104
+ "step": 2800
105
+ },
106
+ {
107
+ "epoch": 4.48,
108
+ "eval_loss": 6.071021556854248,
109
+ "eval_runtime": 71.9212,
110
+ "eval_samples_per_second": 13.904,
111
+ "eval_steps_per_second": 1.738,
112
+ "eval_wer": 1.0,
113
+ "step": 2800
114
+ },
115
+ {
116
+ "epoch": 5.12,
117
+ "learning_rate": 0.00024996262680192206,
118
+ "loss": 5.8882,
119
+ "step": 3200
120
+ },
121
+ {
122
+ "epoch": 5.12,
123
+ "eval_loss": 6.071021556854248,
124
+ "eval_runtime": 72.014,
125
+ "eval_samples_per_second": 13.886,
126
+ "eval_steps_per_second": 1.736,
127
+ "eval_wer": 1.0,
128
+ "step": 3200
129
+ },
130
+ {
131
+ "epoch": 5.76,
132
+ "learning_rate": 0.00024365189535504537,
133
+ "loss": 5.9011,
134
+ "step": 3600
135
+ },
136
+ {
137
+ "epoch": 5.76,
138
+ "eval_loss": 6.071021556854248,
139
+ "eval_runtime": 72.5441,
140
+ "eval_samples_per_second": 13.785,
141
+ "eval_steps_per_second": 1.723,
142
+ "eval_wer": 1.0,
143
+ "step": 3600
144
+ },
145
+ {
146
+ "epoch": 6.4,
147
+ "learning_rate": 0.00023734116390816868,
148
+ "loss": 5.8956,
149
+ "step": 4000
150
+ },
151
+ {
152
+ "epoch": 6.4,
153
+ "eval_loss": 6.071021556854248,
154
+ "eval_runtime": 72.2548,
155
+ "eval_samples_per_second": 13.84,
156
+ "eval_steps_per_second": 1.73,
157
+ "eval_wer": 1.0,
158
+ "step": 4000
159
+ },
160
+ {
161
+ "epoch": 7.04,
162
+ "learning_rate": 0.00023103043246129203,
163
+ "loss": 5.9103,
164
+ "step": 4400
165
+ },
166
+ {
167
+ "epoch": 7.04,
168
+ "eval_loss": 6.071021556854248,
169
+ "eval_runtime": 72.8707,
170
+ "eval_samples_per_second": 13.723,
171
+ "eval_steps_per_second": 1.715,
172
+ "eval_wer": 1.0,
173
+ "step": 4400
174
+ },
175
+ {
176
+ "epoch": 7.68,
177
+ "learning_rate": 0.0002247357180993059,
178
+ "loss": 5.9083,
179
+ "step": 4800
180
+ },
181
+ {
182
+ "epoch": 7.68,
183
+ "eval_loss": 6.071021556854248,
184
+ "eval_runtime": 72.091,
185
+ "eval_samples_per_second": 13.871,
186
+ "eval_steps_per_second": 1.734,
187
+ "eval_wer": 1.0,
188
+ "step": 4800
189
+ },
190
+ {
191
+ "epoch": 8.32,
192
+ "learning_rate": 0.0002184410037373198,
193
+ "loss": 5.9054,
194
+ "step": 5200
195
+ },
196
+ {
197
+ "epoch": 8.32,
198
+ "eval_loss": 6.071021556854248,
199
+ "eval_runtime": 74.6941,
200
+ "eval_samples_per_second": 13.388,
201
+ "eval_steps_per_second": 1.673,
202
+ "eval_wer": 1.0,
203
+ "step": 5200
204
+ },
205
+ {
206
+ "epoch": 8.96,
207
+ "learning_rate": 0.00021213027229044314,
208
+ "loss": 5.8996,
209
+ "step": 5600
210
+ },
211
+ {
212
+ "epoch": 8.96,
213
+ "eval_loss": 6.071021556854248,
214
+ "eval_runtime": 72.5093,
215
+ "eval_samples_per_second": 13.791,
216
+ "eval_steps_per_second": 1.724,
217
+ "eval_wer": 1.0,
218
+ "step": 5600
219
+ },
220
+ {
221
+ "epoch": 9.6,
222
+ "learning_rate": 0.00020581954084356645,
223
+ "loss": 5.8929,
224
+ "step": 6000
225
+ },
226
+ {
227
+ "epoch": 9.6,
228
+ "eval_loss": 6.071021556854248,
229
+ "eval_runtime": 73.1997,
230
+ "eval_samples_per_second": 13.661,
231
+ "eval_steps_per_second": 1.708,
232
+ "eval_wer": 1.0,
233
+ "step": 6000
234
+ },
235
+ {
236
+ "epoch": 10.24,
237
+ "learning_rate": 0.0001995408435664709,
238
+ "loss": 5.9021,
239
+ "step": 6400
240
+ },
241
+ {
242
+ "epoch": 10.24,
243
+ "eval_loss": 6.071021556854248,
244
+ "eval_runtime": 72.6358,
245
+ "eval_samples_per_second": 13.767,
246
+ "eval_steps_per_second": 1.721,
247
+ "eval_wer": 1.0,
248
+ "step": 6400
249
+ },
250
+ {
251
+ "epoch": 10.88,
252
+ "learning_rate": 0.00019323011211959423,
253
+ "loss": 5.8853,
254
+ "step": 6800
255
+ },
256
+ {
257
+ "epoch": 10.88,
258
+ "eval_loss": 6.071021556854248,
259
+ "eval_runtime": 72.5502,
260
+ "eval_samples_per_second": 13.784,
261
+ "eval_steps_per_second": 1.723,
262
+ "eval_wer": 1.0,
263
+ "step": 6800
264
+ },
265
+ {
266
+ "epoch": 11.52,
267
+ "learning_rate": 0.000186903363587827,
268
+ "loss": 5.8968,
269
+ "step": 7200
270
+ },
271
+ {
272
+ "epoch": 11.52,
273
+ "eval_loss": 6.071021556854248,
274
+ "eval_runtime": 72.2966,
275
+ "eval_samples_per_second": 13.832,
276
+ "eval_steps_per_second": 1.729,
277
+ "eval_wer": 1.0,
278
+ "step": 7200
279
+ },
280
+ {
281
+ "epoch": 12.16,
282
+ "learning_rate": 0.00018060864922584088,
283
+ "loss": 5.9139,
284
+ "step": 7600
285
+ },
286
+ {
287
+ "epoch": 12.16,
288
+ "eval_loss": 6.071021556854248,
289
+ "eval_runtime": 72.4188,
290
+ "eval_samples_per_second": 13.809,
291
+ "eval_steps_per_second": 1.726,
292
+ "eval_wer": 1.0,
293
+ "step": 7600
294
+ },
295
+ {
296
+ "epoch": 12.8,
297
+ "learning_rate": 0.00017431393486385476,
298
+ "loss": 5.8976,
299
+ "step": 8000
300
+ },
301
+ {
302
+ "epoch": 12.8,
303
+ "eval_loss": 6.071021556854248,
304
+ "eval_runtime": 72.6699,
305
+ "eval_samples_per_second": 13.761,
306
+ "eval_steps_per_second": 1.72,
307
+ "eval_wer": 1.0,
308
+ "step": 8000
309
+ },
310
+ {
311
+ "epoch": 13.44,
312
+ "learning_rate": 0.00016798718633208754,
313
+ "loss": 5.8909,
314
+ "step": 8400
315
+ },
316
+ {
317
+ "epoch": 13.44,
318
+ "eval_loss": 6.071021556854248,
319
+ "eval_runtime": 72.1967,
320
+ "eval_samples_per_second": 13.851,
321
+ "eval_steps_per_second": 1.731,
322
+ "eval_wer": 1.0,
323
+ "step": 8400
324
+ },
325
+ {
326
+ "epoch": 14.08,
327
+ "learning_rate": 0.00016202883075280296,
328
+ "loss": 7.2473,
329
+ "step": 8800
330
+ },
331
+ {
332
+ "epoch": 14.08,
333
+ "eval_loss": NaN,
334
+ "eval_runtime": 74.9323,
335
+ "eval_samples_per_second": 13.345,
336
+ "eval_steps_per_second": 1.668,
337
+ "eval_wer": 1.0,
338
+ "step": 8800
339
+ },
340
+ {
341
+ "epoch": 14.72,
342
+ "learning_rate": 0.000155621996796583,
343
+ "loss": 0.0,
344
+ "step": 9200
345
+ },
346
+ {
347
+ "epoch": 14.72,
348
+ "eval_loss": NaN,
349
+ "eval_runtime": 72.0418,
350
+ "eval_samples_per_second": 13.881,
351
+ "eval_steps_per_second": 1.735,
352
+ "eval_wer": 1.0,
353
+ "step": 9200
354
+ },
355
+ {
356
+ "epoch": 15.36,
357
+ "learning_rate": 0.00014921516284036303,
358
+ "loss": 0.0,
359
+ "step": 9600
360
+ },
361
+ {
362
+ "epoch": 15.36,
363
+ "eval_loss": NaN,
364
+ "eval_runtime": 72.991,
365
+ "eval_samples_per_second": 13.7,
366
+ "eval_steps_per_second": 1.713,
367
+ "eval_wer": 1.0,
368
+ "step": 9600
369
+ },
370
+ {
371
+ "epoch": 16.0,
372
+ "learning_rate": 0.00014280832888414308,
373
+ "loss": 0.0,
374
+ "step": 10000
375
+ },
376
+ {
377
+ "epoch": 16.0,
378
+ "eval_loss": NaN,
379
+ "eval_runtime": 74.3256,
380
+ "eval_samples_per_second": 13.454,
381
+ "eval_steps_per_second": 1.682,
382
+ "eval_wer": 1.0,
383
+ "step": 10000
384
+ },
385
+ {
386
+ "epoch": 16.64,
387
+ "learning_rate": 0.0001364014949279231,
388
+ "loss": 0.0,
389
+ "step": 10400
390
+ },
391
+ {
392
+ "epoch": 16.64,
393
+ "eval_loss": NaN,
394
+ "eval_runtime": 85.1556,
395
+ "eval_samples_per_second": 11.743,
396
+ "eval_steps_per_second": 1.468,
397
+ "eval_wer": 1.0,
398
+ "step": 10400
399
+ },
400
+ {
401
+ "epoch": 17.28,
402
+ "learning_rate": 0.00012999466097170315,
403
+ "loss": 0.0,
404
+ "step": 10800
405
+ },
406
+ {
407
+ "epoch": 17.28,
408
+ "eval_loss": NaN,
409
+ "eval_runtime": 74.297,
410
+ "eval_samples_per_second": 13.46,
411
+ "eval_steps_per_second": 1.682,
412
+ "eval_wer": 1.0,
413
+ "step": 10800
414
+ },
415
+ {
416
+ "epoch": 17.92,
417
+ "learning_rate": 0.00012358782701548318,
418
+ "loss": 0.0,
419
+ "step": 11200
420
+ },
421
+ {
422
+ "epoch": 17.92,
423
+ "eval_loss": NaN,
424
+ "eval_runtime": 71.8705,
425
+ "eval_samples_per_second": 13.914,
426
+ "eval_steps_per_second": 1.739,
427
+ "eval_wer": 1.0,
428
+ "step": 11200
429
+ },
430
+ {
431
+ "epoch": 18.56,
432
+ "learning_rate": 0.00011718099305926321,
433
+ "loss": 0.0,
434
+ "step": 11600
435
+ },
436
+ {
437
+ "epoch": 18.56,
438
+ "eval_loss": NaN,
439
+ "eval_runtime": 72.0682,
440
+ "eval_samples_per_second": 13.876,
441
+ "eval_steps_per_second": 1.734,
442
+ "eval_wer": 1.0,
443
+ "step": 11600
444
+ },
445
+ {
446
+ "epoch": 19.2,
447
+ "learning_rate": 0.00011077415910304323,
448
+ "loss": 0.0,
449
+ "step": 12000
450
+ },
451
+ {
452
+ "epoch": 19.2,
453
+ "eval_loss": NaN,
454
+ "eval_runtime": 86.6422,
455
+ "eval_samples_per_second": 11.542,
456
+ "eval_steps_per_second": 1.443,
457
+ "eval_wer": 1.0,
458
+ "step": 12000
459
+ },
460
+ {
461
+ "epoch": 19.84,
462
+ "learning_rate": 0.00010436732514682327,
463
+ "loss": 0.0,
464
+ "step": 12400
465
+ },
466
+ {
467
+ "epoch": 19.84,
468
+ "eval_loss": NaN,
469
+ "eval_runtime": 74.8452,
470
+ "eval_samples_per_second": 13.361,
471
+ "eval_steps_per_second": 1.67,
472
+ "eval_wer": 1.0,
473
+ "step": 12400
474
+ },
475
+ {
476
+ "epoch": 20.48,
477
+ "learning_rate": 9.79604911906033e-05,
478
+ "loss": 0.0,
479
+ "step": 12800
480
+ },
481
+ {
482
+ "epoch": 20.48,
483
+ "eval_loss": NaN,
484
+ "eval_runtime": 73.8255,
485
+ "eval_samples_per_second": 13.545,
486
+ "eval_steps_per_second": 1.693,
487
+ "eval_wer": 1.0,
488
+ "step": 12800
489
+ },
490
+ {
491
+ "epoch": 21.12,
492
+ "learning_rate": 9.155365723438334e-05,
493
+ "loss": 0.0,
494
+ "step": 13200
495
+ },
496
+ {
497
+ "epoch": 21.12,
498
+ "eval_loss": NaN,
499
+ "eval_runtime": 77.5777,
500
+ "eval_samples_per_second": 12.89,
501
+ "eval_steps_per_second": 1.611,
502
+ "eval_wer": 1.0,
503
+ "step": 13200
504
+ },
505
+ {
506
+ "epoch": 21.76,
507
+ "learning_rate": 8.514682327816338e-05,
508
+ "loss": 0.0,
509
+ "step": 13600
510
+ },
511
+ {
512
+ "epoch": 21.76,
513
+ "eval_loss": NaN,
514
+ "eval_runtime": 78.2086,
515
+ "eval_samples_per_second": 12.786,
516
+ "eval_steps_per_second": 1.598,
517
+ "eval_wer": 1.0,
518
+ "step": 13600
519
+ },
520
+ {
521
+ "epoch": 22.4,
522
+ "learning_rate": 7.87399893219434e-05,
523
+ "loss": 0.0,
524
+ "step": 14000
525
+ },
526
+ {
527
+ "epoch": 22.4,
528
+ "eval_loss": NaN,
529
+ "eval_runtime": 81.9911,
530
+ "eval_samples_per_second": 12.196,
531
+ "eval_steps_per_second": 1.525,
532
+ "eval_wer": 1.0,
533
+ "step": 14000
534
+ },
535
+ {
536
+ "epoch": 23.04,
537
+ "learning_rate": 7.233315536572343e-05,
538
+ "loss": 0.0,
539
+ "step": 14400
540
+ },
541
+ {
542
+ "epoch": 23.04,
543
+ "eval_loss": NaN,
544
+ "eval_runtime": 74.0428,
545
+ "eval_samples_per_second": 13.506,
546
+ "eval_steps_per_second": 1.688,
547
+ "eval_wer": 1.0,
548
+ "step": 14400
549
+ },
550
+ {
551
+ "epoch": 23.68,
552
+ "learning_rate": 6.592632140950347e-05,
553
+ "loss": 0.0,
554
+ "step": 14800
555
+ },
556
+ {
557
+ "epoch": 23.68,
558
+ "eval_loss": NaN,
559
+ "eval_runtime": 73.6487,
560
+ "eval_samples_per_second": 13.578,
561
+ "eval_steps_per_second": 1.697,
562
+ "eval_wer": 1.0,
563
+ "step": 14800
564
+ },
565
+ {
566
+ "epoch": 24.32,
567
+ "learning_rate": 5.95194874532835e-05,
568
+ "loss": 0.0,
569
+ "step": 15200
570
+ },
571
+ {
572
+ "epoch": 24.32,
573
+ "eval_loss": NaN,
574
+ "eval_runtime": 73.0072,
575
+ "eval_samples_per_second": 13.697,
576
+ "eval_steps_per_second": 1.712,
577
+ "eval_wer": 1.0,
578
+ "step": 15200
579
+ },
580
+ {
581
+ "epoch": 24.96,
582
+ "learning_rate": 5.3112653497063527e-05,
583
+ "loss": 0.0,
584
+ "step": 15600
585
+ },
586
+ {
587
+ "epoch": 24.96,
588
+ "eval_loss": NaN,
589
+ "eval_runtime": 79.0121,
590
+ "eval_samples_per_second": 12.656,
591
+ "eval_steps_per_second": 1.582,
592
+ "eval_wer": 1.0,
593
+ "step": 15600
594
+ },
595
+ {
596
+ "epoch": 25.6,
597
+ "learning_rate": 4.670581954084356e-05,
598
+ "loss": 0.0,
599
+ "step": 16000
600
+ },
601
+ {
602
+ "epoch": 25.6,
603
+ "eval_loss": NaN,
604
+ "eval_runtime": 82.7961,
605
+ "eval_samples_per_second": 12.078,
606
+ "eval_steps_per_second": 1.51,
607
+ "eval_wer": 1.0,
608
+ "step": 16000
609
+ },
610
+ {
611
+ "epoch": 26.24,
612
+ "learning_rate": 4.02989855846236e-05,
613
+ "loss": 0.0,
614
+ "step": 16400
615
+ },
616
+ {
617
+ "epoch": 26.24,
618
+ "eval_loss": NaN,
619
+ "eval_runtime": 83.3819,
620
+ "eval_samples_per_second": 11.993,
621
+ "eval_steps_per_second": 1.499,
622
+ "eval_wer": 1.0,
623
+ "step": 16400
624
+ },
625
+ {
626
+ "epoch": 26.88,
627
+ "learning_rate": 3.3892151628403626e-05,
628
+ "loss": 0.0,
629
+ "step": 16800
630
+ },
631
+ {
632
+ "epoch": 26.88,
633
+ "eval_loss": NaN,
634
+ "eval_runtime": 79.5109,
635
+ "eval_samples_per_second": 12.577,
636
+ "eval_steps_per_second": 1.572,
637
+ "eval_wer": 1.0,
638
+ "step": 16800
639
+ },
640
+ {
641
+ "epoch": 27.52,
642
+ "learning_rate": 2.748531767218366e-05,
643
+ "loss": 0.0,
644
+ "step": 17200
645
+ },
646
+ {
647
+ "epoch": 27.52,
648
+ "eval_loss": NaN,
649
+ "eval_runtime": 76.3506,
650
+ "eval_samples_per_second": 13.097,
651
+ "eval_steps_per_second": 1.637,
652
+ "eval_wer": 1.0,
653
+ "step": 17200
654
+ },
655
+ {
656
+ "epoch": 28.16,
657
+ "learning_rate": 2.107848371596369e-05,
658
+ "loss": 0.0,
659
+ "step": 17600
660
+ },
661
+ {
662
+ "epoch": 28.16,
663
+ "eval_loss": NaN,
664
+ "eval_runtime": 72.6042,
665
+ "eval_samples_per_second": 13.773,
666
+ "eval_steps_per_second": 1.722,
667
+ "eval_wer": 1.0,
668
+ "step": 17600
669
+ },
670
+ {
671
+ "epoch": 28.8,
672
+ "learning_rate": 1.4671649759743725e-05,
673
+ "loss": 0.0,
674
+ "step": 18000
675
+ },
676
+ {
677
+ "epoch": 28.8,
678
+ "eval_loss": NaN,
679
+ "eval_runtime": 83.2188,
680
+ "eval_samples_per_second": 12.017,
681
+ "eval_steps_per_second": 1.502,
682
+ "eval_wer": 1.0,
683
+ "step": 18000
684
+ },
685
+ {
686
+ "epoch": 29.44,
687
+ "learning_rate": 8.264815803523758e-06,
688
+ "loss": 0.0,
689
+ "step": 18400
690
+ },
691
+ {
692
+ "epoch": 29.44,
693
+ "eval_loss": NaN,
694
+ "eval_runtime": 76.8024,
695
+ "eval_samples_per_second": 13.02,
696
+ "eval_steps_per_second": 1.628,
697
+ "eval_wer": 1.0,
698
+ "step": 18400
699
+ }
700
+ ],
701
+ "max_steps": 18750,
702
+ "num_train_epochs": 30,
703
+ "total_flos": 1.0932289159179166e+20,
704
+ "trial_name": null,
705
+ "trial_params": null
706
+ }
checkpoint-18400/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:77e7b4d47fda2b77a45550d4ff0fe73a284067ebbe69000322d16d11dc546be2
3
+ size 3119
checkpoint-800/config.json ADDED
@@ -0,0 +1,115 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.0,
4
+ "adapter_kernel_size": 3,
5
+ "adapter_stride": 2,
6
+ "add_adapter": false,
7
+ "apply_spec_augment": true,
8
+ "architectures": [
9
+ "Wav2Vec2ForCTC"
10
+ ],
11
+ "attention_dropout": 0.1,
12
+ "bos_token_id": 1,
13
+ "classifier_proj_size": 256,
14
+ "codevector_dim": 768,
15
+ "contrastive_logits_temperature": 0.1,
16
+ "conv_bias": true,
17
+ "conv_dim": [
18
+ 512,
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
+ ],
26
+ "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
+ ],
35
+ "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
+ ],
44
+ "ctc_loss_reduction": "mean",
45
+ "ctc_zero_infinity": false,
46
+ "diversity_loss_weight": 0.1,
47
+ "do_stable_layer_norm": true,
48
+ "eos_token_id": 2,
49
+ "feat_extract_activation": "gelu",
50
+ "feat_extract_dropout": 0.0,
51
+ "feat_extract_norm": "layer",
52
+ "feat_proj_dropout": 0.0,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.0,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
57
+ "hidden_size": 1024,
58
+ "initializer_range": 0.02,
59
+ "intermediate_size": 4096,
60
+ "layer_norm_eps": 1e-05,
61
+ "layerdrop": 0.1,
62
+ "mask_channel_length": 10,
63
+ "mask_channel_min_space": 1,
64
+ "mask_channel_other": 0.0,
65
+ "mask_channel_prob": 0.0,
66
+ "mask_channel_selection": "static",
67
+ "mask_feature_length": 10,
68
+ "mask_feature_min_masks": 0,
69
+ "mask_feature_prob": 0.0,
70
+ "mask_time_length": 10,
71
+ "mask_time_min_masks": 2,
72
+ "mask_time_min_space": 1,
73
+ "mask_time_other": 0.0,
74
+ "mask_time_prob": 0.05,
75
+ "mask_time_selection": "static",
76
+ "model_type": "wav2vec2",
77
+ "num_adapter_layers": 3,
78
+ "num_attention_heads": 16,
79
+ "num_codevector_groups": 2,
80
+ "num_codevectors_per_group": 320,
81
+ "num_conv_pos_embedding_groups": 16,
82
+ "num_conv_pos_embeddings": 128,
83
+ "num_feat_extract_layers": 7,
84
+ "num_hidden_layers": 24,
85
+ "num_negatives": 100,
86
+ "output_hidden_size": 1024,
87
+ "pad_token_id": 77,
88
+ "proj_codevector_dim": 768,
89
+ "tdnn_dilation": [
90
+ 1,
91
+ 2,
92
+ 3,
93
+ 1,
94
+ 1
95
+ ],
96
+ "tdnn_dim": [
97
+ 512,
98
+ 512,
99
+ 512,
100
+ 512,
101
+ 1500
102
+ ],
103
+ "tdnn_kernel": [
104
+ 5,
105
+ 3,
106
+ 3,
107
+ 1,
108
+ 1
109
+ ],
110
+ "torch_dtype": "float32",
111
+ "transformers_version": "4.18.0",
112
+ "use_weighted_layer_sum": false,
113
+ "vocab_size": 78,
114
+ "xvector_output_dim": 512
115
+ }
checkpoint-800/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:83d079fc0f49c691b08efa2d6775c41dbd9ff89c275277dea2051ca00ce7bdb6
3
+ size 2490698513
checkpoint-800/preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
checkpoint-800/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38d1ac9002755a5a14da021f92fd8802c1aa24a3acc48909947aaeb8075b6a54
3
+ size 1262218481
checkpoint-800/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a12b1e37ff4151edc2649d24a5f484f04d3d46d5dadaaa2f8c1e150d2e46ea1f
3
+ size 14567
checkpoint-800/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51c8cfdeef4508cfcc185be955c9fff1f1e0dc871cf9c97111be1b428bd1aaff
3
+ size 559
checkpoint-800/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d189e7d30e61f22cf97caf3dd6874c0fe519179c7faef8c25fc23bd57fe8d92f
3
+ size 623
checkpoint-800/trainer_state.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8215018908698001,
3
+ "best_model_checkpoint": "/content/wav2vec2-large-xlsr-bn/checkpoint-800",
4
+ "epoch": 1.28,
5
+ "global_step": 800,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.64,
12
+ "learning_rate": 0.00029405766150560596,
13
+ "loss": 4.1339,
14
+ "step": 400
15
+ },
16
+ {
17
+ "epoch": 0.64,
18
+ "eval_loss": 3.2776501178741455,
19
+ "eval_runtime": 74.0715,
20
+ "eval_samples_per_second": 13.5,
21
+ "eval_steps_per_second": 1.688,
22
+ "eval_wer": 1.0,
23
+ "step": 400
24
+ },
25
+ {
26
+ "epoch": 1.28,
27
+ "learning_rate": 0.00028774693005872927,
28
+ "loss": 2.0718,
29
+ "step": 800
30
+ },
31
+ {
32
+ "epoch": 1.28,
33
+ "eval_loss": 0.979073703289032,
34
+ "eval_runtime": 74.5153,
35
+ "eval_samples_per_second": 13.42,
36
+ "eval_steps_per_second": 1.678,
37
+ "eval_wer": 0.8215018908698001,
38
+ "step": 800
39
+ }
40
+ ],
41
+ "max_steps": 18750,
42
+ "num_train_epochs": 30,
43
+ "total_flos": 4.78415493023146e+18,
44
+ "trial_name": null,
45
+ "trial_params": null
46
+ }
checkpoint-800/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:77e7b4d47fda2b77a45550d4ff0fe73a284067ebbe69000322d16d11dc546be2
3
+ size 3119
config.json ADDED
@@ -0,0 +1,115 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.0,
4
+ "adapter_kernel_size": 3,
5
+ "adapter_stride": 2,
6
+ "add_adapter": false,
7
+ "apply_spec_augment": true,
8
+ "architectures": [
9
+ "Wav2Vec2ForCTC"
10
+ ],
11
+ "attention_dropout": 0.1,
12
+ "bos_token_id": 1,
13
+ "classifier_proj_size": 256,
14
+ "codevector_dim": 768,
15
+ "contrastive_logits_temperature": 0.1,
16
+ "conv_bias": true,
17
+ "conv_dim": [
18
+ 512,
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512
25
+ ],
26
+ "conv_kernel": [
27
+ 10,
28
+ 3,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 2,
33
+ 2
34
+ ],
35
+ "conv_stride": [
36
+ 5,
37
+ 2,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2
43
+ ],
44
+ "ctc_loss_reduction": "mean",
45
+ "ctc_zero_infinity": false,
46
+ "diversity_loss_weight": 0.1,
47
+ "do_stable_layer_norm": true,
48
+ "eos_token_id": 2,
49
+ "feat_extract_activation": "gelu",
50
+ "feat_extract_dropout": 0.0,
51
+ "feat_extract_norm": "layer",
52
+ "feat_proj_dropout": 0.0,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.0,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
57
+ "hidden_size": 1024,
58
+ "initializer_range": 0.02,
59
+ "intermediate_size": 4096,
60
+ "layer_norm_eps": 1e-05,
61
+ "layerdrop": 0.1,
62
+ "mask_channel_length": 10,
63
+ "mask_channel_min_space": 1,
64
+ "mask_channel_other": 0.0,
65
+ "mask_channel_prob": 0.0,
66
+ "mask_channel_selection": "static",
67
+ "mask_feature_length": 10,
68
+ "mask_feature_min_masks": 0,
69
+ "mask_feature_prob": 0.0,
70
+ "mask_time_length": 10,
71
+ "mask_time_min_masks": 2,
72
+ "mask_time_min_space": 1,
73
+ "mask_time_other": 0.0,
74
+ "mask_time_prob": 0.05,
75
+ "mask_time_selection": "static",
76
+ "model_type": "wav2vec2",
77
+ "num_adapter_layers": 3,
78
+ "num_attention_heads": 16,
79
+ "num_codevector_groups": 2,
80
+ "num_codevectors_per_group": 320,
81
+ "num_conv_pos_embedding_groups": 16,
82
+ "num_conv_pos_embeddings": 128,
83
+ "num_feat_extract_layers": 7,
84
+ "num_hidden_layers": 24,
85
+ "num_negatives": 100,
86
+ "output_hidden_size": 1024,
87
+ "pad_token_id": 77,
88
+ "proj_codevector_dim": 768,
89
+ "tdnn_dilation": [
90
+ 1,
91
+ 2,
92
+ 3,
93
+ 1,
94
+ 1
95
+ ],
96
+ "tdnn_dim": [
97
+ 512,
98
+ 512,
99
+ 512,
100
+ 512,
101
+ 1500
102
+ ],
103
+ "tdnn_kernel": [
104
+ 5,
105
+ 3,
106
+ 3,
107
+ 1,
108
+ 1
109
+ ],
110
+ "torch_dtype": "float32",
111
+ "transformers_version": "4.18.0",
112
+ "use_weighted_layer_sum": false,
113
+ "vocab_size": 78,
114
+ "xvector_output_dim": 512
115
+ }
preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38d1ac9002755a5a14da021f92fd8802c1aa24a3acc48909947aaeb8075b6a54
3
+ size 1262218481
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": "|", "replace_word_delimiter_char": " ", "tokenizer_class": "Wav2Vec2CTCTokenizer", "processor_class": "Wav2Vec2Processor"}
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 30.0,
3
+ "total_flos": 1.1136444297962675e+20,
4
+ "train_loss": 2.53877658203125,
5
+ "train_runtime": 56941.5798,
6
+ "train_samples": 20000,
7
+ "train_samples_per_second": 10.537,
8
+ "train_steps_per_second": 0.329
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,715 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8215018908698001,
3
+ "best_model_checkpoint": "/content/wav2vec2-large-xlsr-bn/checkpoint-800",
4
+ "epoch": 30.0,
5
+ "global_step": 18750,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.64,
12
+ "learning_rate": 0.00029405766150560596,
13
+ "loss": 4.1339,
14
+ "step": 400
15
+ },
16
+ {
17
+ "epoch": 0.64,
18
+ "eval_loss": 3.2776501178741455,
19
+ "eval_runtime": 74.0715,
20
+ "eval_samples_per_second": 13.5,
21
+ "eval_steps_per_second": 1.688,
22
+ "eval_wer": 1.0,
23
+ "step": 400
24
+ },
25
+ {
26
+ "epoch": 1.28,
27
+ "learning_rate": 0.00028774693005872927,
28
+ "loss": 2.0718,
29
+ "step": 800
30
+ },
31
+ {
32
+ "epoch": 1.28,
33
+ "eval_loss": 0.979073703289032,
34
+ "eval_runtime": 74.5153,
35
+ "eval_samples_per_second": 13.42,
36
+ "eval_steps_per_second": 1.678,
37
+ "eval_wer": 0.8215018908698001,
38
+ "step": 800
39
+ },
40
+ {
41
+ "epoch": 1.92,
42
+ "learning_rate": 0.00028143619861185264,
43
+ "loss": 2.4619,
44
+ "step": 1200
45
+ },
46
+ {
47
+ "epoch": 1.92,
48
+ "eval_loss": 3.080274820327759,
49
+ "eval_runtime": 74.0712,
50
+ "eval_samples_per_second": 13.501,
51
+ "eval_steps_per_second": 1.688,
52
+ "eval_wer": 1.0,
53
+ "step": 1200
54
+ },
55
+ {
56
+ "epoch": 2.56,
57
+ "learning_rate": 0.000275157501334757,
58
+ "loss": 3.1703,
59
+ "step": 1600
60
+ },
61
+ {
62
+ "epoch": 2.56,
63
+ "eval_loss": 3.4948482513427734,
64
+ "eval_runtime": 74.092,
65
+ "eval_samples_per_second": 13.497,
66
+ "eval_steps_per_second": 1.687,
67
+ "eval_wer": 1.0,
68
+ "step": 1600
69
+ },
70
+ {
71
+ "epoch": 3.2,
72
+ "learning_rate": 0.00026886278697277094,
73
+ "loss": 5.5212,
74
+ "step": 2000
75
+ },
76
+ {
77
+ "epoch": 3.2,
78
+ "eval_loss": 6.071021556854248,
79
+ "eval_runtime": 74.4322,
80
+ "eval_samples_per_second": 13.435,
81
+ "eval_steps_per_second": 1.679,
82
+ "eval_wer": 1.0,
83
+ "step": 2000
84
+ },
85
+ {
86
+ "epoch": 3.84,
87
+ "learning_rate": 0.0002625680726107848,
88
+ "loss": 5.8864,
89
+ "step": 2400
90
+ },
91
+ {
92
+ "epoch": 3.84,
93
+ "eval_loss": 6.071021556854248,
94
+ "eval_runtime": 74.0924,
95
+ "eval_samples_per_second": 13.497,
96
+ "eval_steps_per_second": 1.687,
97
+ "eval_wer": 1.0,
98
+ "step": 2400
99
+ },
100
+ {
101
+ "epoch": 4.48,
102
+ "learning_rate": 0.00025628937533368924,
103
+ "loss": 5.9244,
104
+ "step": 2800
105
+ },
106
+ {
107
+ "epoch": 4.48,
108
+ "eval_loss": 6.071021556854248,
109
+ "eval_runtime": 71.9212,
110
+ "eval_samples_per_second": 13.904,
111
+ "eval_steps_per_second": 1.738,
112
+ "eval_wer": 1.0,
113
+ "step": 2800
114
+ },
115
+ {
116
+ "epoch": 5.12,
117
+ "learning_rate": 0.00024996262680192206,
118
+ "loss": 5.8882,
119
+ "step": 3200
120
+ },
121
+ {
122
+ "epoch": 5.12,
123
+ "eval_loss": 6.071021556854248,
124
+ "eval_runtime": 72.014,
125
+ "eval_samples_per_second": 13.886,
126
+ "eval_steps_per_second": 1.736,
127
+ "eval_wer": 1.0,
128
+ "step": 3200
129
+ },
130
+ {
131
+ "epoch": 5.76,
132
+ "learning_rate": 0.00024365189535504537,
133
+ "loss": 5.9011,
134
+ "step": 3600
135
+ },
136
+ {
137
+ "epoch": 5.76,
138
+ "eval_loss": 6.071021556854248,
139
+ "eval_runtime": 72.5441,
140
+ "eval_samples_per_second": 13.785,
141
+ "eval_steps_per_second": 1.723,
142
+ "eval_wer": 1.0,
143
+ "step": 3600
144
+ },
145
+ {
146
+ "epoch": 6.4,
147
+ "learning_rate": 0.00023734116390816868,
148
+ "loss": 5.8956,
149
+ "step": 4000
150
+ },
151
+ {
152
+ "epoch": 6.4,
153
+ "eval_loss": 6.071021556854248,
154
+ "eval_runtime": 72.2548,
155
+ "eval_samples_per_second": 13.84,
156
+ "eval_steps_per_second": 1.73,
157
+ "eval_wer": 1.0,
158
+ "step": 4000
159
+ },
160
+ {
161
+ "epoch": 7.04,
162
+ "learning_rate": 0.00023103043246129203,
163
+ "loss": 5.9103,
164
+ "step": 4400
165
+ },
166
+ {
167
+ "epoch": 7.04,
168
+ "eval_loss": 6.071021556854248,
169
+ "eval_runtime": 72.8707,
170
+ "eval_samples_per_second": 13.723,
171
+ "eval_steps_per_second": 1.715,
172
+ "eval_wer": 1.0,
173
+ "step": 4400
174
+ },
175
+ {
176
+ "epoch": 7.68,
177
+ "learning_rate": 0.0002247357180993059,
178
+ "loss": 5.9083,
179
+ "step": 4800
180
+ },
181
+ {
182
+ "epoch": 7.68,
183
+ "eval_loss": 6.071021556854248,
184
+ "eval_runtime": 72.091,
185
+ "eval_samples_per_second": 13.871,
186
+ "eval_steps_per_second": 1.734,
187
+ "eval_wer": 1.0,
188
+ "step": 4800
189
+ },
190
+ {
191
+ "epoch": 8.32,
192
+ "learning_rate": 0.0002184410037373198,
193
+ "loss": 5.9054,
194
+ "step": 5200
195
+ },
196
+ {
197
+ "epoch": 8.32,
198
+ "eval_loss": 6.071021556854248,
199
+ "eval_runtime": 74.6941,
200
+ "eval_samples_per_second": 13.388,
201
+ "eval_steps_per_second": 1.673,
202
+ "eval_wer": 1.0,
203
+ "step": 5200
204
+ },
205
+ {
206
+ "epoch": 8.96,
207
+ "learning_rate": 0.00021213027229044314,
208
+ "loss": 5.8996,
209
+ "step": 5600
210
+ },
211
+ {
212
+ "epoch": 8.96,
213
+ "eval_loss": 6.071021556854248,
214
+ "eval_runtime": 72.5093,
215
+ "eval_samples_per_second": 13.791,
216
+ "eval_steps_per_second": 1.724,
217
+ "eval_wer": 1.0,
218
+ "step": 5600
219
+ },
220
+ {
221
+ "epoch": 9.6,
222
+ "learning_rate": 0.00020581954084356645,
223
+ "loss": 5.8929,
224
+ "step": 6000
225
+ },
226
+ {
227
+ "epoch": 9.6,
228
+ "eval_loss": 6.071021556854248,
229
+ "eval_runtime": 73.1997,
230
+ "eval_samples_per_second": 13.661,
231
+ "eval_steps_per_second": 1.708,
232
+ "eval_wer": 1.0,
233
+ "step": 6000
234
+ },
235
+ {
236
+ "epoch": 10.24,
237
+ "learning_rate": 0.0001995408435664709,
238
+ "loss": 5.9021,
239
+ "step": 6400
240
+ },
241
+ {
242
+ "epoch": 10.24,
243
+ "eval_loss": 6.071021556854248,
244
+ "eval_runtime": 72.6358,
245
+ "eval_samples_per_second": 13.767,
246
+ "eval_steps_per_second": 1.721,
247
+ "eval_wer": 1.0,
248
+ "step": 6400
249
+ },
250
+ {
251
+ "epoch": 10.88,
252
+ "learning_rate": 0.00019323011211959423,
253
+ "loss": 5.8853,
254
+ "step": 6800
255
+ },
256
+ {
257
+ "epoch": 10.88,
258
+ "eval_loss": 6.071021556854248,
259
+ "eval_runtime": 72.5502,
260
+ "eval_samples_per_second": 13.784,
261
+ "eval_steps_per_second": 1.723,
262
+ "eval_wer": 1.0,
263
+ "step": 6800
264
+ },
265
+ {
266
+ "epoch": 11.52,
267
+ "learning_rate": 0.000186903363587827,
268
+ "loss": 5.8968,
269
+ "step": 7200
270
+ },
271
+ {
272
+ "epoch": 11.52,
273
+ "eval_loss": 6.071021556854248,
274
+ "eval_runtime": 72.2966,
275
+ "eval_samples_per_second": 13.832,
276
+ "eval_steps_per_second": 1.729,
277
+ "eval_wer": 1.0,
278
+ "step": 7200
279
+ },
280
+ {
281
+ "epoch": 12.16,
282
+ "learning_rate": 0.00018060864922584088,
283
+ "loss": 5.9139,
284
+ "step": 7600
285
+ },
286
+ {
287
+ "epoch": 12.16,
288
+ "eval_loss": 6.071021556854248,
289
+ "eval_runtime": 72.4188,
290
+ "eval_samples_per_second": 13.809,
291
+ "eval_steps_per_second": 1.726,
292
+ "eval_wer": 1.0,
293
+ "step": 7600
294
+ },
295
+ {
296
+ "epoch": 12.8,
297
+ "learning_rate": 0.00017431393486385476,
298
+ "loss": 5.8976,
299
+ "step": 8000
300
+ },
301
+ {
302
+ "epoch": 12.8,
303
+ "eval_loss": 6.071021556854248,
304
+ "eval_runtime": 72.6699,
305
+ "eval_samples_per_second": 13.761,
306
+ "eval_steps_per_second": 1.72,
307
+ "eval_wer": 1.0,
308
+ "step": 8000
309
+ },
310
+ {
311
+ "epoch": 13.44,
312
+ "learning_rate": 0.00016798718633208754,
313
+ "loss": 5.8909,
314
+ "step": 8400
315
+ },
316
+ {
317
+ "epoch": 13.44,
318
+ "eval_loss": 6.071021556854248,
319
+ "eval_runtime": 72.1967,
320
+ "eval_samples_per_second": 13.851,
321
+ "eval_steps_per_second": 1.731,
322
+ "eval_wer": 1.0,
323
+ "step": 8400
324
+ },
325
+ {
326
+ "epoch": 14.08,
327
+ "learning_rate": 0.00016202883075280296,
328
+ "loss": 7.2473,
329
+ "step": 8800
330
+ },
331
+ {
332
+ "epoch": 14.08,
333
+ "eval_loss": NaN,
334
+ "eval_runtime": 74.9323,
335
+ "eval_samples_per_second": 13.345,
336
+ "eval_steps_per_second": 1.668,
337
+ "eval_wer": 1.0,
338
+ "step": 8800
339
+ },
340
+ {
341
+ "epoch": 14.72,
342
+ "learning_rate": 0.000155621996796583,
343
+ "loss": 0.0,
344
+ "step": 9200
345
+ },
346
+ {
347
+ "epoch": 14.72,
348
+ "eval_loss": NaN,
349
+ "eval_runtime": 72.0418,
350
+ "eval_samples_per_second": 13.881,
351
+ "eval_steps_per_second": 1.735,
352
+ "eval_wer": 1.0,
353
+ "step": 9200
354
+ },
355
+ {
356
+ "epoch": 15.36,
357
+ "learning_rate": 0.00014921516284036303,
358
+ "loss": 0.0,
359
+ "step": 9600
360
+ },
361
+ {
362
+ "epoch": 15.36,
363
+ "eval_loss": NaN,
364
+ "eval_runtime": 72.991,
365
+ "eval_samples_per_second": 13.7,
366
+ "eval_steps_per_second": 1.713,
367
+ "eval_wer": 1.0,
368
+ "step": 9600
369
+ },
370
+ {
371
+ "epoch": 16.0,
372
+ "learning_rate": 0.00014280832888414308,
373
+ "loss": 0.0,
374
+ "step": 10000
375
+ },
376
+ {
377
+ "epoch": 16.0,
378
+ "eval_loss": NaN,
379
+ "eval_runtime": 74.3256,
380
+ "eval_samples_per_second": 13.454,
381
+ "eval_steps_per_second": 1.682,
382
+ "eval_wer": 1.0,
383
+ "step": 10000
384
+ },
385
+ {
386
+ "epoch": 16.64,
387
+ "learning_rate": 0.0001364014949279231,
388
+ "loss": 0.0,
389
+ "step": 10400
390
+ },
391
+ {
392
+ "epoch": 16.64,
393
+ "eval_loss": NaN,
394
+ "eval_runtime": 85.1556,
395
+ "eval_samples_per_second": 11.743,
396
+ "eval_steps_per_second": 1.468,
397
+ "eval_wer": 1.0,
398
+ "step": 10400
399
+ },
400
+ {
401
+ "epoch": 17.28,
402
+ "learning_rate": 0.00012999466097170315,
403
+ "loss": 0.0,
404
+ "step": 10800
405
+ },
406
+ {
407
+ "epoch": 17.28,
408
+ "eval_loss": NaN,
409
+ "eval_runtime": 74.297,
410
+ "eval_samples_per_second": 13.46,
411
+ "eval_steps_per_second": 1.682,
412
+ "eval_wer": 1.0,
413
+ "step": 10800
414
+ },
415
+ {
416
+ "epoch": 17.92,
417
+ "learning_rate": 0.00012358782701548318,
418
+ "loss": 0.0,
419
+ "step": 11200
420
+ },
421
+ {
422
+ "epoch": 17.92,
423
+ "eval_loss": NaN,
424
+ "eval_runtime": 71.8705,
425
+ "eval_samples_per_second": 13.914,
426
+ "eval_steps_per_second": 1.739,
427
+ "eval_wer": 1.0,
428
+ "step": 11200
429
+ },
430
+ {
431
+ "epoch": 18.56,
432
+ "learning_rate": 0.00011718099305926321,
433
+ "loss": 0.0,
434
+ "step": 11600
435
+ },
436
+ {
437
+ "epoch": 18.56,
438
+ "eval_loss": NaN,
439
+ "eval_runtime": 72.0682,
440
+ "eval_samples_per_second": 13.876,
441
+ "eval_steps_per_second": 1.734,
442
+ "eval_wer": 1.0,
443
+ "step": 11600
444
+ },
445
+ {
446
+ "epoch": 19.2,
447
+ "learning_rate": 0.00011077415910304323,
448
+ "loss": 0.0,
449
+ "step": 12000
450
+ },
451
+ {
452
+ "epoch": 19.2,
453
+ "eval_loss": NaN,
454
+ "eval_runtime": 86.6422,
455
+ "eval_samples_per_second": 11.542,
456
+ "eval_steps_per_second": 1.443,
457
+ "eval_wer": 1.0,
458
+ "step": 12000
459
+ },
460
+ {
461
+ "epoch": 19.84,
462
+ "learning_rate": 0.00010436732514682327,
463
+ "loss": 0.0,
464
+ "step": 12400
465
+ },
466
+ {
467
+ "epoch": 19.84,
468
+ "eval_loss": NaN,
469
+ "eval_runtime": 74.8452,
470
+ "eval_samples_per_second": 13.361,
471
+ "eval_steps_per_second": 1.67,
472
+ "eval_wer": 1.0,
473
+ "step": 12400
474
+ },
475
+ {
476
+ "epoch": 20.48,
477
+ "learning_rate": 9.79604911906033e-05,
478
+ "loss": 0.0,
479
+ "step": 12800
480
+ },
481
+ {
482
+ "epoch": 20.48,
483
+ "eval_loss": NaN,
484
+ "eval_runtime": 73.8255,
485
+ "eval_samples_per_second": 13.545,
486
+ "eval_steps_per_second": 1.693,
487
+ "eval_wer": 1.0,
488
+ "step": 12800
489
+ },
490
+ {
491
+ "epoch": 21.12,
492
+ "learning_rate": 9.155365723438334e-05,
493
+ "loss": 0.0,
494
+ "step": 13200
495
+ },
496
+ {
497
+ "epoch": 21.12,
498
+ "eval_loss": NaN,
499
+ "eval_runtime": 77.5777,
500
+ "eval_samples_per_second": 12.89,
501
+ "eval_steps_per_second": 1.611,
502
+ "eval_wer": 1.0,
503
+ "step": 13200
504
+ },
505
+ {
506
+ "epoch": 21.76,
507
+ "learning_rate": 8.514682327816338e-05,
508
+ "loss": 0.0,
509
+ "step": 13600
510
+ },
511
+ {
512
+ "epoch": 21.76,
513
+ "eval_loss": NaN,
514
+ "eval_runtime": 78.2086,
515
+ "eval_samples_per_second": 12.786,
516
+ "eval_steps_per_second": 1.598,
517
+ "eval_wer": 1.0,
518
+ "step": 13600
519
+ },
520
+ {
521
+ "epoch": 22.4,
522
+ "learning_rate": 7.87399893219434e-05,
523
+ "loss": 0.0,
524
+ "step": 14000
525
+ },
526
+ {
527
+ "epoch": 22.4,
528
+ "eval_loss": NaN,
529
+ "eval_runtime": 81.9911,
530
+ "eval_samples_per_second": 12.196,
531
+ "eval_steps_per_second": 1.525,
532
+ "eval_wer": 1.0,
533
+ "step": 14000
534
+ },
535
+ {
536
+ "epoch": 23.04,
537
+ "learning_rate": 7.233315536572343e-05,
538
+ "loss": 0.0,
539
+ "step": 14400
540
+ },
541
+ {
542
+ "epoch": 23.04,
543
+ "eval_loss": NaN,
544
+ "eval_runtime": 74.0428,
545
+ "eval_samples_per_second": 13.506,
546
+ "eval_steps_per_second": 1.688,
547
+ "eval_wer": 1.0,
548
+ "step": 14400
549
+ },
550
+ {
551
+ "epoch": 23.68,
552
+ "learning_rate": 6.592632140950347e-05,
553
+ "loss": 0.0,
554
+ "step": 14800
555
+ },
556
+ {
557
+ "epoch": 23.68,
558
+ "eval_loss": NaN,
559
+ "eval_runtime": 73.6487,
560
+ "eval_samples_per_second": 13.578,
561
+ "eval_steps_per_second": 1.697,
562
+ "eval_wer": 1.0,
563
+ "step": 14800
564
+ },
565
+ {
566
+ "epoch": 24.32,
567
+ "learning_rate": 5.95194874532835e-05,
568
+ "loss": 0.0,
569
+ "step": 15200
570
+ },
571
+ {
572
+ "epoch": 24.32,
573
+ "eval_loss": NaN,
574
+ "eval_runtime": 73.0072,
575
+ "eval_samples_per_second": 13.697,
576
+ "eval_steps_per_second": 1.712,
577
+ "eval_wer": 1.0,
578
+ "step": 15200
579
+ },
580
+ {
581
+ "epoch": 24.96,
582
+ "learning_rate": 5.3112653497063527e-05,
583
+ "loss": 0.0,
584
+ "step": 15600
585
+ },
586
+ {
587
+ "epoch": 24.96,
588
+ "eval_loss": NaN,
589
+ "eval_runtime": 79.0121,
590
+ "eval_samples_per_second": 12.656,
591
+ "eval_steps_per_second": 1.582,
592
+ "eval_wer": 1.0,
593
+ "step": 15600
594
+ },
595
+ {
596
+ "epoch": 25.6,
597
+ "learning_rate": 4.670581954084356e-05,
598
+ "loss": 0.0,
599
+ "step": 16000
600
+ },
601
+ {
602
+ "epoch": 25.6,
603
+ "eval_loss": NaN,
604
+ "eval_runtime": 82.7961,
605
+ "eval_samples_per_second": 12.078,
606
+ "eval_steps_per_second": 1.51,
607
+ "eval_wer": 1.0,
608
+ "step": 16000
609
+ },
610
+ {
611
+ "epoch": 26.24,
612
+ "learning_rate": 4.02989855846236e-05,
613
+ "loss": 0.0,
614
+ "step": 16400
615
+ },
616
+ {
617
+ "epoch": 26.24,
618
+ "eval_loss": NaN,
619
+ "eval_runtime": 83.3819,
620
+ "eval_samples_per_second": 11.993,
621
+ "eval_steps_per_second": 1.499,
622
+ "eval_wer": 1.0,
623
+ "step": 16400
624
+ },
625
+ {
626
+ "epoch": 26.88,
627
+ "learning_rate": 3.3892151628403626e-05,
628
+ "loss": 0.0,
629
+ "step": 16800
630
+ },
631
+ {
632
+ "epoch": 26.88,
633
+ "eval_loss": NaN,
634
+ "eval_runtime": 79.5109,
635
+ "eval_samples_per_second": 12.577,
636
+ "eval_steps_per_second": 1.572,
637
+ "eval_wer": 1.0,
638
+ "step": 16800
639
+ },
640
+ {
641
+ "epoch": 27.52,
642
+ "learning_rate": 2.748531767218366e-05,
643
+ "loss": 0.0,
644
+ "step": 17200
645
+ },
646
+ {
647
+ "epoch": 27.52,
648
+ "eval_loss": NaN,
649
+ "eval_runtime": 76.3506,
650
+ "eval_samples_per_second": 13.097,
651
+ "eval_steps_per_second": 1.637,
652
+ "eval_wer": 1.0,
653
+ "step": 17200
654
+ },
655
+ {
656
+ "epoch": 28.16,
657
+ "learning_rate": 2.107848371596369e-05,
658
+ "loss": 0.0,
659
+ "step": 17600
660
+ },
661
+ {
662
+ "epoch": 28.16,
663
+ "eval_loss": NaN,
664
+ "eval_runtime": 72.6042,
665
+ "eval_samples_per_second": 13.773,
666
+ "eval_steps_per_second": 1.722,
667
+ "eval_wer": 1.0,
668
+ "step": 17600
669
+ },
670
+ {
671
+ "epoch": 28.8,
672
+ "learning_rate": 1.4671649759743725e-05,
673
+ "loss": 0.0,
674
+ "step": 18000
675
+ },
676
+ {
677
+ "epoch": 28.8,
678
+ "eval_loss": NaN,
679
+ "eval_runtime": 83.2188,
680
+ "eval_samples_per_second": 12.017,
681
+ "eval_steps_per_second": 1.502,
682
+ "eval_wer": 1.0,
683
+ "step": 18000
684
+ },
685
+ {
686
+ "epoch": 29.44,
687
+ "learning_rate": 8.264815803523758e-06,
688
+ "loss": 0.0,
689
+ "step": 18400
690
+ },
691
+ {
692
+ "epoch": 29.44,
693
+ "eval_loss": NaN,
694
+ "eval_runtime": 76.8024,
695
+ "eval_samples_per_second": 13.02,
696
+ "eval_steps_per_second": 1.628,
697
+ "eval_wer": 1.0,
698
+ "step": 18400
699
+ },
700
+ {
701
+ "epoch": 30.0,
702
+ "step": 18750,
703
+ "total_flos": 1.1136444297962675e+20,
704
+ "train_loss": 2.53877658203125,
705
+ "train_runtime": 56941.5798,
706
+ "train_samples_per_second": 10.537,
707
+ "train_steps_per_second": 0.329
708
+ }
709
+ ],
710
+ "max_steps": 18750,
711
+ "num_train_epochs": 30,
712
+ "total_flos": 1.1136444297962675e+20,
713
+ "trial_name": null,
714
+ "trial_params": null
715
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:77e7b4d47fda2b77a45550d4ff0fe73a284067ebbe69000322d16d11dc546be2
3
+ size 3119
vocab.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"=": 1, "।": 2, "ঁ": 3, "ং": 4, "ঃ": 5, "অ": 6, "আ": 7, "ই": 8, "ঈ": 9, "উ": 10, "ঊ": 11, "ঋ": 12, "এ": 13, "ঐ": 14, "ও": 15, "ঔ": 16, "ক": 17, "খ": 18, "গ": 19, "ঘ": 20, "ঙ": 21, "চ": 22, "ছ": 23, "জ": 24, "ঝ": 25, "ঞ": 26, "ট": 27, "ঠ": 28, "ড": 29, "ঢ": 30, "ণ": 31, "ত": 32, "থ": 33, "দ": 34, "ধ": 35, "ন": 36, "প": 37, "ফ": 38, "ব": 39, "ভ": 40, "ম": 41, "য": 42, "র": 43, "ল": 44, "শ": 45, "ষ": 46, "স": 47, "হ": 48, "া": 49, "ি": 50, "ী": 51, "ু": 52, "ূ": 53, "ৃ": 54, "ে": 55, "ৈ": 56, "ো": 57, "ৌ": 58, "্": 59, "ৎ": 60, "ড়": 61, "ঢ়": 62, "য়": 63, "০": 64, "১": 65, "২": 66, "৩": 67, "৪": 68, "৫": 69, "৬": 70, "৭": 71, "৮": 72, "৯": 73, "‍": 74, "—": 75, "|": 0, "[UNK]": 76, "[PAD]": 77}