speech31 commited on
Commit
068102d
1 Parent(s): f0c61a3

upload model

Browse files
config.json ADDED
@@ -0,0 +1,73 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/hubert-base-ls960",
3
+ "activation_dropout": 0.1,
4
+ "apply_spec_augment": true,
5
+ "architectures": [
6
+ "HubertForCTC"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 1,
10
+ "classifier_proj_size": 256,
11
+ "conv_bias": false,
12
+ "conv_dim": [
13
+ 512,
14
+ 512,
15
+ 512,
16
+ 512,
17
+ 512,
18
+ 512,
19
+ 512
20
+ ],
21
+ "conv_kernel": [
22
+ 10,
23
+ 3,
24
+ 3,
25
+ 3,
26
+ 3,
27
+ 2,
28
+ 2
29
+ ],
30
+ "conv_stride": [
31
+ 5,
32
+ 2,
33
+ 2,
34
+ 2,
35
+ 2,
36
+ 2,
37
+ 2
38
+ ],
39
+ "ctc_loss_reduction": "mean",
40
+ "ctc_zero_infinity": false,
41
+ "do_stable_layer_norm": false,
42
+ "eos_token_id": 2,
43
+ "feat_extract_activation": "gelu",
44
+ "feat_extract_dropout": 0.0,
45
+ "feat_extract_norm": "group",
46
+ "feat_proj_dropout": 0.0,
47
+ "final_dropout": 0.1,
48
+ "gradient_checkpointing": false,
49
+ "hidden_act": "gelu",
50
+ "hidden_dropout": 0.0,
51
+ "hidden_dropout_prob": 0.1,
52
+ "hidden_size": 768,
53
+ "initializer_range": 0.02,
54
+ "intermediate_size": 3072,
55
+ "layer_norm_eps": 1e-05,
56
+ "layerdrop": 0.0,
57
+ "mask_feature_length": 10,
58
+ "mask_feature_prob": 0.0,
59
+ "mask_time_length": 10,
60
+ "mask_time_prob": 0.05,
61
+ "model_type": "hubert",
62
+ "num_attention_heads": 12,
63
+ "num_conv_pos_embedding_groups": 16,
64
+ "num_conv_pos_embeddings": 128,
65
+ "num_feat_extract_layers": 7,
66
+ "num_hidden_layers": 12,
67
+ "pad_token_id": 42,
68
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
69
+ "torch_dtype": "float32",
70
+ "transformers_version": "4.11.3",
71
+ "use_weighted_layer_sum": false,
72
+ "vocab_size": 43
73
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd96f76c804dad5e562a9a7eaed4f72aeb3ce747a153a28636c9c27c327a269f
3
+ size 721752849
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
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e90a0cf2d07736816cf64424b0fbf982331caafacd6b7d853226434080b976ff
3
+ size 377702999
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c1d4db70ac7e94817857249df43d9d9b360b5975a0a719dfb90989a7a1a7187
3
+ size 14503
scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0571021828e8451967c6fe0e826edd76fab89ef42c17dd0201160f74f0534f6e
3
+ size 559
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:009d60cfe544495de99d20e1a4bcebd0a0d22527ec6618a0d78081ec93fa965e
3
+ size 623
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": "|", "tokenizer_class": "Wav2Vec2CTCTokenizer"}
trainer_state.json ADDED
@@ -0,0 +1,289 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.10831519216299057,
3
+ "best_model_checkpoint": "hubert-base-english-phoneme-v2/checkpoint-3500",
4
+ "epoch": 26.923076923076923,
5
+ "global_step": 3500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.77,
12
+ "learning_rate": 2.97e-05,
13
+ "loss": 4.5359,
14
+ "step": 100
15
+ },
16
+ {
17
+ "epoch": 1.54,
18
+ "learning_rate": 5.97e-05,
19
+ "loss": 2.2643,
20
+ "step": 200
21
+ },
22
+ {
23
+ "epoch": 2.31,
24
+ "learning_rate": 8.969999999999998e-05,
25
+ "loss": 2.2425,
26
+ "step": 300
27
+ },
28
+ {
29
+ "epoch": 3.08,
30
+ "learning_rate": 0.0001197,
31
+ "loss": 2.2134,
32
+ "step": 400
33
+ },
34
+ {
35
+ "epoch": 3.85,
36
+ "learning_rate": 0.00014969999999999998,
37
+ "loss": 1.7074,
38
+ "step": 500
39
+ },
40
+ {
41
+ "epoch": 3.85,
42
+ "eval_cer": 0.709044908285895,
43
+ "eval_loss": 0.9500909447669983,
44
+ "eval_runtime": 24.8403,
45
+ "eval_samples_per_second": 18.599,
46
+ "eval_steps_per_second": 2.335,
47
+ "step": 500
48
+ },
49
+ {
50
+ "epoch": 4.62,
51
+ "learning_rate": 0.00017969999999999998,
52
+ "loss": 0.7708,
53
+ "step": 600
54
+ },
55
+ {
56
+ "epoch": 5.38,
57
+ "learning_rate": 0.00020969999999999997,
58
+ "loss": 0.4614,
59
+ "step": 700
60
+ },
61
+ {
62
+ "epoch": 6.15,
63
+ "learning_rate": 0.0002397,
64
+ "loss": 0.3393,
65
+ "step": 800
66
+ },
67
+ {
68
+ "epoch": 6.92,
69
+ "learning_rate": 0.0002697,
70
+ "loss": 0.2852,
71
+ "step": 900
72
+ },
73
+ {
74
+ "epoch": 7.69,
75
+ "learning_rate": 0.00029969999999999997,
76
+ "loss": 0.2549,
77
+ "step": 1000
78
+ },
79
+ {
80
+ "epoch": 7.69,
81
+ "eval_cer": 0.15725806451612903,
82
+ "eval_loss": 0.2167779505252838,
83
+ "eval_runtime": 24.1643,
84
+ "eval_samples_per_second": 19.119,
85
+ "eval_steps_per_second": 2.4,
86
+ "step": 1000
87
+ },
88
+ {
89
+ "epoch": 8.46,
90
+ "learning_rate": 0.00028975862068965515,
91
+ "loss": 0.2373,
92
+ "step": 1100
93
+ },
94
+ {
95
+ "epoch": 9.23,
96
+ "learning_rate": 0.00027941379310344826,
97
+ "loss": 0.2001,
98
+ "step": 1200
99
+ },
100
+ {
101
+ "epoch": 10.0,
102
+ "learning_rate": 0.0002690689655172414,
103
+ "loss": 0.184,
104
+ "step": 1300
105
+ },
106
+ {
107
+ "epoch": 10.77,
108
+ "learning_rate": 0.00025872413793103443,
109
+ "loss": 0.1479,
110
+ "step": 1400
111
+ },
112
+ {
113
+ "epoch": 11.54,
114
+ "learning_rate": 0.00024837931034482754,
115
+ "loss": 0.1328,
116
+ "step": 1500
117
+ },
118
+ {
119
+ "epoch": 11.54,
120
+ "eval_cer": 0.13432953826691968,
121
+ "eval_loss": 0.1408727467060089,
122
+ "eval_runtime": 23.8901,
123
+ "eval_samples_per_second": 19.339,
124
+ "eval_steps_per_second": 2.428,
125
+ "step": 1500
126
+ },
127
+ {
128
+ "epoch": 12.31,
129
+ "learning_rate": 0.00023803448275862068,
130
+ "loss": 0.1223,
131
+ "step": 1600
132
+ },
133
+ {
134
+ "epoch": 13.08,
135
+ "learning_rate": 0.00022768965517241377,
136
+ "loss": 0.115,
137
+ "step": 1700
138
+ },
139
+ {
140
+ "epoch": 13.85,
141
+ "learning_rate": 0.00021734482758620688,
142
+ "loss": 0.1048,
143
+ "step": 1800
144
+ },
145
+ {
146
+ "epoch": 14.62,
147
+ "learning_rate": 0.00020699999999999996,
148
+ "loss": 0.0944,
149
+ "step": 1900
150
+ },
151
+ {
152
+ "epoch": 15.38,
153
+ "learning_rate": 0.00019665517241379307,
154
+ "loss": 0.0867,
155
+ "step": 2000
156
+ },
157
+ {
158
+ "epoch": 15.38,
159
+ "eval_cer": 0.10468058191018342,
160
+ "eval_loss": 0.13811400532722473,
161
+ "eval_runtime": 23.9414,
162
+ "eval_samples_per_second": 19.297,
163
+ "eval_steps_per_second": 2.423,
164
+ "step": 2000
165
+ },
166
+ {
167
+ "epoch": 16.15,
168
+ "learning_rate": 0.0001863103448275862,
169
+ "loss": 0.0738,
170
+ "step": 2100
171
+ },
172
+ {
173
+ "epoch": 16.92,
174
+ "learning_rate": 0.00017596551724137927,
175
+ "loss": 0.0749,
176
+ "step": 2200
177
+ },
178
+ {
179
+ "epoch": 17.69,
180
+ "learning_rate": 0.0001656206896551724,
181
+ "loss": 0.0642,
182
+ "step": 2300
183
+ },
184
+ {
185
+ "epoch": 18.46,
186
+ "learning_rate": 0.00015527586206896552,
187
+ "loss": 0.0643,
188
+ "step": 2400
189
+ },
190
+ {
191
+ "epoch": 19.23,
192
+ "learning_rate": 0.0001449310344827586,
193
+ "loss": 0.0579,
194
+ "step": 2500
195
+ },
196
+ {
197
+ "epoch": 19.23,
198
+ "eval_cer": 0.1101359898798229,
199
+ "eval_loss": 0.13352537155151367,
200
+ "eval_runtime": 24.2513,
201
+ "eval_samples_per_second": 19.051,
202
+ "eval_steps_per_second": 2.392,
203
+ "step": 2500
204
+ },
205
+ {
206
+ "epoch": 20.0,
207
+ "learning_rate": 0.00013458620689655172,
208
+ "loss": 0.0564,
209
+ "step": 2600
210
+ },
211
+ {
212
+ "epoch": 20.77,
213
+ "learning_rate": 0.00012424137931034483,
214
+ "loss": 0.0478,
215
+ "step": 2700
216
+ },
217
+ {
218
+ "epoch": 21.54,
219
+ "learning_rate": 0.00011389655172413792,
220
+ "loss": 0.044,
221
+ "step": 2800
222
+ },
223
+ {
224
+ "epoch": 22.31,
225
+ "learning_rate": 0.00010355172413793101,
226
+ "loss": 0.0417,
227
+ "step": 2900
228
+ },
229
+ {
230
+ "epoch": 23.08,
231
+ "learning_rate": 9.320689655172414e-05,
232
+ "loss": 0.0415,
233
+ "step": 3000
234
+ },
235
+ {
236
+ "epoch": 23.08,
237
+ "eval_cer": 0.09875079063883618,
238
+ "eval_loss": 0.11283650994300842,
239
+ "eval_runtime": 24.3063,
240
+ "eval_samples_per_second": 19.007,
241
+ "eval_steps_per_second": 2.386,
242
+ "step": 3000
243
+ },
244
+ {
245
+ "epoch": 23.85,
246
+ "learning_rate": 8.286206896551724e-05,
247
+ "loss": 0.0411,
248
+ "step": 3100
249
+ },
250
+ {
251
+ "epoch": 24.62,
252
+ "learning_rate": 7.251724137931034e-05,
253
+ "loss": 0.033,
254
+ "step": 3200
255
+ },
256
+ {
257
+ "epoch": 25.38,
258
+ "learning_rate": 6.217241379310345e-05,
259
+ "loss": 0.0318,
260
+ "step": 3300
261
+ },
262
+ {
263
+ "epoch": 26.15,
264
+ "learning_rate": 5.1827586206896546e-05,
265
+ "loss": 0.0282,
266
+ "step": 3400
267
+ },
268
+ {
269
+ "epoch": 26.92,
270
+ "learning_rate": 4.148275862068966e-05,
271
+ "loss": 0.0274,
272
+ "step": 3500
273
+ },
274
+ {
275
+ "epoch": 26.92,
276
+ "eval_cer": 0.09930423782416192,
277
+ "eval_loss": 0.10831519216299057,
278
+ "eval_runtime": 24.5094,
279
+ "eval_samples_per_second": 18.85,
280
+ "eval_steps_per_second": 2.366,
281
+ "step": 3500
282
+ }
283
+ ],
284
+ "max_steps": 3900,
285
+ "num_train_epochs": 30,
286
+ "total_flos": 3.19030343120722e+18,
287
+ "trial_name": null,
288
+ "trial_params": null
289
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51cde43985a0b309ff16c2f2c073f43e57ffce5c1e19efd52038a40ce294424f
3
+ size 2863
vocab.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"ɑ": 0, "aj": 2, "m": 3, "f": 4, "æ": 5, "ə": 6, "z": 7, "p": 8, "ɪ": 9, "n": 10, "ʒ": 11, "k": 12, "h": 13, "j": 14, "ð": 15, "θ": 16, "ʊ": 17, "ow": 18, "ʌ": 19, "ɔ": 20, "ɹ̩": 21, "d": 22, "ŋ": 23, "i": 24, "ɹ": 25, "t": 26, "v": 27, "b": 28, "d͡ʒ": 29, "ɛ": 30, "oj": 31, "l": 32, "u": 33, "s": 34, "w": 35, "ɡ": 36, "ʃ": 37, "aw": 38, "t͡ʃ": 39, "ej": 40, "|": 1, "[UNK]": 41, "[PAD]": 42}