Hugging Face's logo

foxxy-hm
/
wav2vec2-base-finetune-vi

Automatic Speech Recognition
Transformers
PyTorch
Vietnamese
wav2vec2
audio
Inference Endpoints
Model card Files Community
1
foxxy-hm commited on
Commit
4161986
·
1 Parent(s): e30cb17

Upload 9 files

Browse files
Files changed (9) hide show
  1. checkpoint-9500/config.json +108 -0
  2. checkpoint-9500/optimizer.pt +3 -0
  3. checkpoint-9500/preprocessor_config.json +10 -0
  4. checkpoint-9500/pytorch_model.bin +3 -0
  5. checkpoint-9500/rng_state.pth +3 -0
  6. checkpoint-9500/scaler.pt +3 -0
  7. checkpoint-9500/scheduler.pt +3 -0
  8. checkpoint-9500/trainer_state.json +301 -0
  9. checkpoint-9500/training_args.bin +3 -0
checkpoint-9500/config.json ADDED
@@ -0,0 +1,108 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "nguyenvulebinh/wav2vec2-base-vietnamese-250h",
3
+ "activation_dropout": 0.1,
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": 256,
15
+ "contrastive_logits_temperature": 0.1,
16
+ "conv_bias": false,
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": false,
48
+ "eos_token_id": 2,
49
+ "feat_extract_activation": "gelu",
50
+ "feat_extract_dropout": 0.0,
51
+ "feat_extract_norm": "group",
52
+ "feat_proj_dropout": 0.1,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.1,
55
+ "hidden_act": "gelu",
56
+ "hidden_dropout": 0.1,
57
+ "hidden_dropout_prob": 0.1,
58
+ "hidden_size": 768,
59
+ "initializer_range": 0.02,
60
+ "intermediate_size": 3072,
61
+ "layer_norm_eps": 1e-05,
62
+ "layerdrop": 0.1,
63
+ "mask_feature_length": 10,
64
+ "mask_feature_min_masks": 0,
65
+ "mask_feature_prob": 0.0,
66
+ "mask_time_length": 10,
67
+ "mask_time_min_masks": 2,
68
+ "mask_time_prob": 0.05,
69
+ "model_type": "wav2vec2",
70
+ "num_adapter_layers": 3,
71
+ "num_attention_heads": 12,
72
+ "num_codevector_groups": 2,
73
+ "num_codevectors_per_group": 320,
74
+ "num_conv_pos_embedding_groups": 16,
75
+ "num_conv_pos_embeddings": 128,
76
+ "num_feat_extract_layers": 7,
77
+ "num_hidden_layers": 12,
78
+ "num_negatives": 100,
79
+ "output_hidden_size": 768,
80
+ "pad_token_id": 107,
81
+ "proj_codevector_dim": 256,
82
+ "tdnn_dilation": [
83
+ 1,
84
+ 2,
85
+ 3,
86
+ 1,
87
+ 1
88
+ ],
89
+ "tdnn_dim": [
90
+ 512,
91
+ 512,
92
+ 512,
93
+ 512,
94
+ 1500
95
+ ],
96
+ "tdnn_kernel": [
97
+ 5,
98
+ 3,
99
+ 3,
100
+ 1,
101
+ 1
102
+ ],
103
+ "torch_dtype": "float32",
104
+ "transformers_version": "4.20.0",
105
+ "use_weighted_layer_sum": false,
106
+ "vocab_size": 110,
107
+ "xvector_output_dim": 512
108
+ }
checkpoint-9500/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69819c23f01b77d9c63b7b338595efe429cc99705a5032497d4d3b1a1bd04b8c
3
+ size 722166405
checkpoint-9500/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": false,
9
+ "sampling_rate": 16000
10
+ }
checkpoint-9500/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f130d659275463be357bce6d7ae08f5c4f98a6f4b2384ee68456ed8ccf509e2c
3
+ size 377898657
checkpoint-9500/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1990a5c4ded0f73f95570e0aaaf64573202b83065f1d0c544f25826d39399ca1
3
+ size 14639
checkpoint-9500/scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0396645081caca9a8e87b2fc203d6448c90f0c7640480c5ac0e5c60db3c169bb
3
+ size 557
checkpoint-9500/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb1950d337e6b1b8419fd4218ba925380620dd9415d10a2e1ea1235e6715eb62
3
+ size 627
checkpoint-9500/trainer_state.json ADDED
@@ -0,0 +1,301 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 12.683578104138851,
5
+ "global_step": 9500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.67,
12
+ "learning_rate": 4.9500000000000004e-05,
13
+ "loss": 13.7697,
14
+ "step": 500
15
+ },
16
+ {
17
+ "epoch": 0.67,
18
+ "eval_loss": 3.130753755569458,
19
+ "eval_runtime": 142.0908,
20
+ "eval_samples_per_second": 10.543,
21
+ "eval_steps_per_second": 1.323,
22
+ "eval_wer": 1.0547889135773594,
23
+ "step": 500
24
+ },
25
+ {
26
+ "epoch": 1.34,
27
+ "learning_rate": 9.95e-05,
28
+ "loss": 2.5213,
29
+ "step": 1000
30
+ },
31
+ {
32
+ "epoch": 1.34,
33
+ "eval_loss": 1.1151297092437744,
34
+ "eval_runtime": 142.3636,
35
+ "eval_samples_per_second": 10.522,
36
+ "eval_steps_per_second": 1.321,
37
+ "eval_wer": 0.5353175067243597,
38
+ "step": 1000
39
+ },
40
+ {
41
+ "epoch": 2.0,
42
+ "learning_rate": 9.478178368121443e-05,
43
+ "loss": 1.2366,
44
+ "step": 1500
45
+ },
46
+ {
47
+ "epoch": 2.0,
48
+ "eval_loss": 0.6335494518280029,
49
+ "eval_runtime": 141.3238,
50
+ "eval_samples_per_second": 10.6,
51
+ "eval_steps_per_second": 1.33,
52
+ "eval_wer": 0.32534206525552567,
53
+ "step": 1500
54
+ },
55
+ {
56
+ "epoch": 2.67,
57
+ "learning_rate": 8.951085810668353e-05,
58
+ "loss": 0.8513,
59
+ "step": 2000
60
+ },
61
+ {
62
+ "epoch": 2.67,
63
+ "eval_loss": 0.4479542076587677,
64
+ "eval_runtime": 141.529,
65
+ "eval_samples_per_second": 10.584,
66
+ "eval_steps_per_second": 1.328,
67
+ "eval_wer": 0.2673956262425447,
68
+ "step": 2000
69
+ },
70
+ {
71
+ "epoch": 3.34,
72
+ "learning_rate": 8.423993253215266e-05,
73
+ "loss": 0.6485,
74
+ "step": 2500
75
+ },
76
+ {
77
+ "epoch": 3.34,
78
+ "eval_loss": 0.3942696750164032,
79
+ "eval_runtime": 142.7432,
80
+ "eval_samples_per_second": 10.494,
81
+ "eval_steps_per_second": 1.317,
82
+ "eval_wer": 0.24038124196000468,
83
+ "step": 2500
84
+ },
85
+ {
86
+ "epoch": 4.01,
87
+ "learning_rate": 7.896900695762176e-05,
88
+ "loss": 0.5742,
89
+ "step": 3000
90
+ },
91
+ {
92
+ "epoch": 4.01,
93
+ "eval_loss": 0.3576195538043976,
94
+ "eval_runtime": 141.7609,
95
+ "eval_samples_per_second": 10.567,
96
+ "eval_steps_per_second": 1.326,
97
+ "eval_wer": 0.22254707051806807,
98
+ "step": 3000
99
+ },
100
+ {
101
+ "epoch": 4.67,
102
+ "learning_rate": 7.369808138309088e-05,
103
+ "loss": 0.4864,
104
+ "step": 3500
105
+ },
106
+ {
107
+ "epoch": 4.67,
108
+ "eval_loss": 0.3353235721588135,
109
+ "eval_runtime": 141.2544,
110
+ "eval_samples_per_second": 10.605,
111
+ "eval_steps_per_second": 1.331,
112
+ "eval_wer": 0.2099169687755818,
113
+ "step": 3500
114
+ },
115
+ {
116
+ "epoch": 5.34,
117
+ "learning_rate": 6.842715580855998e-05,
118
+ "loss": 0.4255,
119
+ "step": 4000
120
+ },
121
+ {
122
+ "epoch": 5.34,
123
+ "eval_loss": 0.30898207426071167,
124
+ "eval_runtime": 141.0567,
125
+ "eval_samples_per_second": 10.62,
126
+ "eval_steps_per_second": 1.333,
127
+ "eval_wer": 0.1951818500760145,
128
+ "step": 4000
129
+ },
130
+ {
131
+ "epoch": 6.01,
132
+ "learning_rate": 6.316677208517816e-05,
133
+ "loss": 0.4415,
134
+ "step": 4500
135
+ },
136
+ {
137
+ "epoch": 6.01,
138
+ "eval_loss": 0.28900569677352905,
139
+ "eval_runtime": 141.7282,
140
+ "eval_samples_per_second": 10.57,
141
+ "eval_steps_per_second": 1.326,
142
+ "eval_wer": 0.18986083499005965,
143
+ "step": 4500
144
+ },
145
+ {
146
+ "epoch": 6.68,
147
+ "learning_rate": 5.790638836179634e-05,
148
+ "loss": 0.3936,
149
+ "step": 5000
150
+ },
151
+ {
152
+ "epoch": 6.68,
153
+ "eval_loss": 0.29400354623794556,
154
+ "eval_runtime": 141.7038,
155
+ "eval_samples_per_second": 10.571,
156
+ "eval_steps_per_second": 1.327,
157
+ "eval_wer": 0.17980353175067243,
158
+ "step": 5000
159
+ },
160
+ {
161
+ "epoch": 7.34,
162
+ "learning_rate": 5.2635462787265446e-05,
163
+ "loss": 0.3539,
164
+ "step": 5500
165
+ },
166
+ {
167
+ "epoch": 7.34,
168
+ "eval_loss": 0.28268352150917053,
169
+ "eval_runtime": 143.4799,
170
+ "eval_samples_per_second": 10.44,
171
+ "eval_steps_per_second": 1.31,
172
+ "eval_wer": 0.17477488013097883,
173
+ "step": 5500
174
+ },
175
+ {
176
+ "epoch": 8.01,
177
+ "learning_rate": 4.736453721273456e-05,
178
+ "loss": 0.3731,
179
+ "step": 6000
180
+ },
181
+ {
182
+ "epoch": 8.01,
183
+ "eval_loss": 0.27123069763183594,
184
+ "eval_runtime": 144.4362,
185
+ "eval_samples_per_second": 10.371,
186
+ "eval_steps_per_second": 1.302,
187
+ "eval_wer": 0.16851830195298795,
188
+ "step": 6000
189
+ },
190
+ {
191
+ "epoch": 8.68,
192
+ "learning_rate": 4.209361163820367e-05,
193
+ "loss": 0.3003,
194
+ "step": 6500
195
+ },
196
+ {
197
+ "epoch": 8.68,
198
+ "eval_loss": 0.2683442533016205,
199
+ "eval_runtime": 144.0393,
200
+ "eval_samples_per_second": 10.4,
201
+ "eval_steps_per_second": 1.305,
202
+ "eval_wer": 0.16927844696526723,
203
+ "step": 6500
204
+ },
205
+ {
206
+ "epoch": 9.35,
207
+ "learning_rate": 3.6822686063672785e-05,
208
+ "loss": 0.2858,
209
+ "step": 7000
210
+ },
211
+ {
212
+ "epoch": 9.35,
213
+ "eval_loss": 0.2772638499736786,
214
+ "eval_runtime": 144.5546,
215
+ "eval_samples_per_second": 10.363,
216
+ "eval_steps_per_second": 1.301,
217
+ "eval_wer": 0.16477604958484388,
218
+ "step": 7000
219
+ },
220
+ {
221
+ "epoch": 10.01,
222
+ "learning_rate": 3.15517604891419e-05,
223
+ "loss": 0.3133,
224
+ "step": 7500
225
+ },
226
+ {
227
+ "epoch": 10.01,
228
+ "eval_loss": 0.25764620304107666,
229
+ "eval_runtime": 142.0521,
230
+ "eval_samples_per_second": 10.545,
231
+ "eval_steps_per_second": 1.323,
232
+ "eval_wer": 0.16325575956028535,
233
+ "step": 7500
234
+ },
235
+ {
236
+ "epoch": 10.68,
237
+ "learning_rate": 2.628083491461101e-05,
238
+ "loss": 0.2803,
239
+ "step": 8000
240
+ },
241
+ {
242
+ "epoch": 10.68,
243
+ "eval_loss": 0.2704939544200897,
244
+ "eval_runtime": 143.5154,
245
+ "eval_samples_per_second": 10.438,
246
+ "eval_steps_per_second": 1.31,
247
+ "eval_wer": 0.16050754297742953,
248
+ "step": 8000
249
+ },
250
+ {
251
+ "epoch": 11.35,
252
+ "learning_rate": 2.1009909340080118e-05,
253
+ "loss": 0.2787,
254
+ "step": 8500
255
+ },
256
+ {
257
+ "epoch": 11.35,
258
+ "eval_loss": 0.2567004859447479,
259
+ "eval_runtime": 143.3412,
260
+ "eval_samples_per_second": 10.451,
261
+ "eval_steps_per_second": 1.312,
262
+ "eval_wer": 0.16062448836393403,
263
+ "step": 8500
264
+ },
265
+ {
266
+ "epoch": 12.02,
267
+ "learning_rate": 1.573898376554923e-05,
268
+ "loss": 0.2302,
269
+ "step": 9000
270
+ },
271
+ {
272
+ "epoch": 12.02,
273
+ "eval_loss": 0.24703864753246307,
274
+ "eval_runtime": 143.5724,
275
+ "eval_samples_per_second": 10.434,
276
+ "eval_steps_per_second": 1.309,
277
+ "eval_wer": 0.15822710794059175,
278
+ "step": 9000
279
+ },
280
+ {
281
+ "epoch": 12.68,
282
+ "learning_rate": 1.0468058191018343e-05,
283
+ "loss": 0.2147,
284
+ "step": 9500
285
+ },
286
+ {
287
+ "epoch": 12.68,
288
+ "eval_loss": 0.24555708467960358,
289
+ "eval_runtime": 142.5225,
290
+ "eval_samples_per_second": 10.511,
291
+ "eval_steps_per_second": 1.319,
292
+ "eval_wer": 0.15717459946205123,
293
+ "step": 9500
294
+ }
295
+ ],
296
+ "max_steps": 10486,
297
+ "num_train_epochs": 14,
298
+ "total_flos": 4.4296766058875546e+18,
299
+ "trial_name": null,
300
+ "trial_params": null
301
+ }
checkpoint-9500/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:045f9b5c540095c0c180207709fb8c16c3b9e75203de6a5090c126f7a0ff2334
3
+ size 3323