Hugging Face's logo

bkh6722
/
bach-arb

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
wav2vec2
Model card Files
xet
Metrics Community
bkh6722 commited on
Commit
9dc8d3d
·
1 Parent(s): 9c3f45f

add tokenizer

Browse files
Files changed (11) hide show
  1. runs/{May14_18-37-51_f8d43c95907a → wav2vec-vlbg-L}/1652553481.6213744/events.out.tfevents.1652553481.f8d43c95907a.2560.1 +0 -0
  2. runs/{May14_18-37-51_f8d43c95907a → wav2vec-vlbg-L}/events.out.tfevents.1652553481.f8d43c95907a.2560.0 +0 -0
  3. wav2vec-vlbg-L/config.json +80 -0
  4. wav2vec-vlbg-L/optimizer.pt +3 -0
  5. wav2vec-vlbg-L/preprocessor_config.json +9 -0
  6. wav2vec-vlbg-L/pytorch_model.bin +3 -0
  7. wav2vec-vlbg-L/rng_state.pth +3 -0
  8. wav2vec-vlbg-L/scaler.pt +3 -0
  9. wav2vec-vlbg-L/scheduler.pt +3 -0
  10. wav2vec-vlbg-L/trainer_state.json +256 -0
  11. wav2vec-vlbg-L/training_args.bin +3 -0
runs/{May14_18-37-51_f8d43c95907a → wav2vec-vlbg-L}/1652553481.6213744/events.out.tfevents.1652553481.f8d43c95907a.2560.1 RENAMED
File without changes
runs/{May14_18-37-51_f8d43c95907a → wav2vec-vlbg-L}/events.out.tfevents.1652553481.f8d43c95907a.2560.0 RENAMED
File without changes
wav2vec-vlbg-L/config.json ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-base-960h",
3
+ "activation_dropout": 0.1,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "Wav2Vec2ForCTC"
7
+ ],
8
+ "attention_dropout": 0.5,
9
+ "bos_token_id": 1,
10
+ "classifier_proj_size": 256,
11
+ "codevector_dim": 256,
12
+ "contrastive_logits_temperature": 0.1,
13
+ "conv_bias": false,
14
+ "conv_dim": [
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512,
19
+ 512,
20
+ 512,
21
+ 512
22
+ ],
23
+ "conv_kernel": [
24
+ 10,
25
+ 3,
26
+ 3,
27
+ 3,
28
+ 3,
29
+ 2,
30
+ 2
31
+ ],
32
+ "conv_stride": [
33
+ 5,
34
+ 2,
35
+ 2,
36
+ 2,
37
+ 2,
38
+ 2,
39
+ 2
40
+ ],
41
+ "ctc_loss_reduction": "mean",
42
+ "ctc_zero_infinity": false,
43
+ "diversity_loss_weight": 0.1,
44
+ "do_stable_layer_norm": false,
45
+ "eos_token_id": 2,
46
+ "feat_extract_activation": "gelu",
47
+ "feat_extract_dropout": 0.0,
48
+ "feat_extract_norm": "group",
49
+ "feat_proj_dropout": 0.0,
50
+ "feat_quantizer_dropout": 0.0,
51
+ "final_dropout": 0.1,
52
+ "gradient_checkpointing": false,
53
+ "hidden_act": "gelu",
54
+ "hidden_dropout": 0.0,
55
+ "hidden_dropout_prob": 0.1,
56
+ "hidden_size": 768,
57
+ "initializer_range": 0.02,
58
+ "intermediate_size": 3072,
59
+ "layer_norm_eps": 1e-05,
60
+ "layerdrop": 0.0,
61
+ "mask_feature_length": 10,
62
+ "mask_feature_prob": 0.0,
63
+ "mask_time_length": 10,
64
+ "mask_time_prob": 0.05,
65
+ "model_type": "wav2vec2",
66
+ "num_attention_heads": 12,
67
+ "num_codevector_groups": 2,
68
+ "num_codevectors_per_group": 320,
69
+ "num_conv_pos_embedding_groups": 16,
70
+ "num_conv_pos_embeddings": 128,
71
+ "num_feat_extract_layers": 7,
72
+ "num_hidden_layers": 12,
73
+ "num_negatives": 100,
74
+ "pad_token_id": 26,
75
+ "proj_codevector_dim": 256,
76
+ "torch_dtype": "float32",
77
+ "transformers_version": "4.11.3",
78
+ "use_weighted_layer_sum": false,
79
+ "vocab_size": 32
80
+ }
wav2vec-vlbg-L/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bb5689a0bc1b7a0df28cec228b14c9058a72f815c01e39967b55ad5ca81ad76c
3
+ size 721685265
wav2vec-vlbg-L/preprocessor_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0.0,
7
+ "return_attention_mask": true,
8
+ "sampling_rate": 16000
9
+ }
wav2vec-vlbg-L/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6451949ee5edd6f48492fd7780c5059e130c8bf561b796103e7611fcc34056b0
3
+ size 377670039
wav2vec-vlbg-L/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c4b5a071eb06400f757526c1afd5c7195b447d0cf77237f0a05a9b8d3f7b215
3
+ size 14567
wav2vec-vlbg-L/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13ee4be8816a2c5e4b438bdad3867fb6f6d7b7cf0aeac89f281c59f40cfa18de
3
+ size 559
wav2vec-vlbg-L/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0721280043c43c04fe53223087d0aa4e538969c3ad45a776336f40d9ba30e223
3
+ size 623
wav2vec-vlbg-L/trainer_state.json ADDED
@@ -0,0 +1,256 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 320.0,
5
+ "global_step": 1600,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 20.0,
12
+ "learning_rate": 5.82e-05,
13
+ "loss": 5.6466,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 20.0,
18
+ "eval_loss": 3.2344367504119873,
19
+ "eval_runtime": 24.0305,
20
+ "eval_samples_per_second": 2.538,
21
+ "eval_steps_per_second": 0.333,
22
+ "eval_wer": 1.0,
23
+ "step": 100
24
+ },
25
+ {
26
+ "epoch": 40.0,
27
+ "learning_rate": 0.0001182,
28
+ "loss": 3.0849,
29
+ "step": 200
30
+ },
31
+ {
32
+ "epoch": 40.0,
33
+ "eval_loss": 3.034189224243164,
34
+ "eval_runtime": 1.6121,
35
+ "eval_samples_per_second": 37.84,
36
+ "eval_steps_per_second": 4.963,
37
+ "eval_wer": 1.0,
38
+ "step": 200
39
+ },
40
+ {
41
+ "epoch": 60.0,
42
+ "learning_rate": 0.00017819999999999997,
43
+ "loss": 2.9201,
44
+ "step": 300
45
+ },
46
+ {
47
+ "epoch": 60.0,
48
+ "eval_loss": 2.925170421600342,
49
+ "eval_runtime": 1.622,
50
+ "eval_samples_per_second": 37.608,
51
+ "eval_steps_per_second": 4.932,
52
+ "eval_wer": 1.0,
53
+ "step": 300
54
+ },
55
+ {
56
+ "epoch": 80.0,
57
+ "learning_rate": 0.0002382,
58
+ "loss": 2.8738,
59
+ "step": 400
60
+ },
61
+ {
62
+ "epoch": 80.0,
63
+ "eval_loss": 2.8987646102905273,
64
+ "eval_runtime": 1.6209,
65
+ "eval_samples_per_second": 37.634,
66
+ "eval_steps_per_second": 4.936,
67
+ "eval_wer": 1.0,
68
+ "step": 400
69
+ },
70
+ {
71
+ "epoch": 100.0,
72
+ "learning_rate": 0.0002982,
73
+ "loss": 2.311,
74
+ "step": 500
75
+ },
76
+ {
77
+ "epoch": 100.0,
78
+ "eval_loss": 1.823543906211853,
79
+ "eval_runtime": 1.6203,
80
+ "eval_samples_per_second": 37.648,
81
+ "eval_steps_per_second": 4.937,
82
+ "eval_wer": 0.968299711815562,
83
+ "step": 500
84
+ },
85
+ {
86
+ "epoch": 120.0,
87
+ "learning_rate": 0.00027354545454545454,
88
+ "loss": 1.0145,
89
+ "step": 600
90
+ },
91
+ {
92
+ "epoch": 120.0,
93
+ "eval_loss": 1.8741708993911743,
94
+ "eval_runtime": 1.6089,
95
+ "eval_samples_per_second": 37.915,
96
+ "eval_steps_per_second": 4.972,
97
+ "eval_wer": 0.8126801152737753,
98
+ "step": 600
99
+ },
100
+ {
101
+ "epoch": 140.0,
102
+ "learning_rate": 0.00024627272727272725,
103
+ "loss": 0.4719,
104
+ "step": 700
105
+ },
106
+ {
107
+ "epoch": 140.0,
108
+ "eval_loss": 2.4738245010375977,
109
+ "eval_runtime": 1.621,
110
+ "eval_samples_per_second": 37.63,
111
+ "eval_steps_per_second": 4.935,
112
+ "eval_wer": 0.7838616714697406,
113
+ "step": 700
114
+ },
115
+ {
116
+ "epoch": 160.0,
117
+ "learning_rate": 0.00021899999999999998,
118
+ "loss": 0.2909,
119
+ "step": 800
120
+ },
121
+ {
122
+ "epoch": 160.0,
123
+ "eval_loss": 2.7554588317871094,
124
+ "eval_runtime": 1.598,
125
+ "eval_samples_per_second": 38.172,
126
+ "eval_steps_per_second": 5.006,
127
+ "eval_wer": 0.7463976945244957,
128
+ "step": 800
129
+ },
130
+ {
131
+ "epoch": 180.0,
132
+ "learning_rate": 0.0001917272727272727,
133
+ "loss": 0.1995,
134
+ "step": 900
135
+ },
136
+ {
137
+ "epoch": 180.0,
138
+ "eval_loss": 2.6272172927856445,
139
+ "eval_runtime": 1.6286,
140
+ "eval_samples_per_second": 37.456,
141
+ "eval_steps_per_second": 4.912,
142
+ "eval_wer": 0.7723342939481268,
143
+ "step": 900
144
+ },
145
+ {
146
+ "epoch": 200.0,
147
+ "learning_rate": 0.00016472727272727271,
148
+ "loss": 0.1518,
149
+ "step": 1000
150
+ },
151
+ {
152
+ "epoch": 200.0,
153
+ "eval_loss": 2.9259815216064453,
154
+ "eval_runtime": 1.5854,
155
+ "eval_samples_per_second": 38.477,
156
+ "eval_steps_per_second": 5.046,
157
+ "eval_wer": 0.7377521613832853,
158
+ "step": 1000
159
+ },
160
+ {
161
+ "epoch": 220.0,
162
+ "learning_rate": 0.00013745454545454545,
163
+ "loss": 0.1195,
164
+ "step": 1100
165
+ },
166
+ {
167
+ "epoch": 220.0,
168
+ "eval_loss": 2.876216173171997,
169
+ "eval_runtime": 1.6072,
170
+ "eval_samples_per_second": 37.954,
171
+ "eval_steps_per_second": 4.978,
172
+ "eval_wer": 0.7089337175792507,
173
+ "step": 1100
174
+ },
175
+ {
176
+ "epoch": 240.0,
177
+ "learning_rate": 0.00011018181818181816,
178
+ "loss": 0.082,
179
+ "step": 1200
180
+ },
181
+ {
182
+ "epoch": 240.0,
183
+ "eval_loss": 2.799006462097168,
184
+ "eval_runtime": 1.5997,
185
+ "eval_samples_per_second": 38.132,
186
+ "eval_steps_per_second": 5.001,
187
+ "eval_wer": 0.6858789625360231,
188
+ "step": 1200
189
+ },
190
+ {
191
+ "epoch": 260.0,
192
+ "learning_rate": 8.29090909090909e-05,
193
+ "loss": 0.0613,
194
+ "step": 1300
195
+ },
196
+ {
197
+ "epoch": 260.0,
198
+ "eval_loss": 3.1452391147613525,
199
+ "eval_runtime": 1.6033,
200
+ "eval_samples_per_second": 38.047,
201
+ "eval_steps_per_second": 4.99,
202
+ "eval_wer": 0.6974063400576369,
203
+ "step": 1300
204
+ },
205
+ {
206
+ "epoch": 280.0,
207
+ "learning_rate": 5.563636363636363e-05,
208
+ "loss": 0.0453,
209
+ "step": 1400
210
+ },
211
+ {
212
+ "epoch": 280.0,
213
+ "eval_loss": 3.2089500427246094,
214
+ "eval_runtime": 1.61,
215
+ "eval_samples_per_second": 37.888,
216
+ "eval_steps_per_second": 4.969,
217
+ "eval_wer": 0.69164265129683,
218
+ "step": 1400
219
+ },
220
+ {
221
+ "epoch": 300.0,
222
+ "learning_rate": 2.836363636363636e-05,
223
+ "loss": 0.0366,
224
+ "step": 1500
225
+ },
226
+ {
227
+ "epoch": 300.0,
228
+ "eval_loss": 3.0762720108032227,
229
+ "eval_runtime": 1.6163,
230
+ "eval_samples_per_second": 37.741,
231
+ "eval_steps_per_second": 4.95,
232
+ "eval_wer": 0.6801152737752162,
233
+ "step": 1500
234
+ },
235
+ {
236
+ "epoch": 320.0,
237
+ "learning_rate": 1.0909090909090908e-06,
238
+ "loss": 0.0306,
239
+ "step": 1600
240
+ },
241
+ {
242
+ "epoch": 320.0,
243
+ "eval_loss": 3.160898447036743,
244
+ "eval_runtime": 1.6091,
245
+ "eval_samples_per_second": 37.909,
246
+ "eval_steps_per_second": 4.972,
247
+ "eval_wer": 0.6829971181556196,
248
+ "step": 1600
249
+ }
250
+ ],
251
+ "max_steps": 1600,
252
+ "num_train_epochs": 320,
253
+ "total_flos": 1.990148569882203e+18,
254
+ "trial_name": null,
255
+ "trial_params": null
256
+ }
wav2vec-vlbg-L/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a963f3b525bb1a2693e0bad9d35ab995eb16054b17eb1ddfa78f9004ca1d399
3
+ size 2799