jonatasgrosman commited on
Commit
5e032ba
1 Parent(s): c281a97

jonatasgrosman/exp_w2v2r_de_vp-100k_age_teens-0_sixties-10_s278

Browse files
.gitattributes CHANGED
@@ -32,3 +32,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
32
  *.zip filter=lfs diff=lfs merge=lfs -text
33
  *.zst filter=lfs diff=lfs merge=lfs -text
34
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
32
  *.zip filter=lfs diff=lfs merge=lfs -text
33
  *.zst filter=lfs diff=lfs merge=lfs -text
34
  *tfevents* filter=lfs diff=lfs merge=lfs -text
35
+ cv7_transcriptions.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - de
4
+ license: apache-2.0
5
+ tags:
6
+ - automatic-speech-recognition
7
+ - de
8
+ datasets:
9
+ - mozilla-foundation/common_voice_7_0
10
+ ---
11
+ # exp_w2v2r_de_vp-100k_age_teens-0_sixties-10_s278
12
+
13
+ Fine-tuned [facebook/wav2vec2-large-100k-voxpopuli](https://huggingface.co/facebook/wav2vec2-large-100k-voxpopuli) for speech recognition using the train split of [Common Voice 7.0 (de)](https://huggingface.co/datasets/mozilla-foundation/common_voice_7_0).
14
+ When using this model, make sure that your speech input is sampled at 16kHz.
15
+
16
+ This model has been fine-tuned by the [HuggingSound](https://github.com/jonatasgrosman/huggingsound) tool.
17
+
config.json ADDED
@@ -0,0 +1,116 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-100k-voxpopuli",
3
+ "activation_dropout": 0.05,
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": 0.05,
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": 0.05,
53
+ "feat_quantizer_dropout": 0.0,
54
+ "final_dropout": 0.05,
55
+ "gradient_checkpointing": false,
56
+ "hidden_act": "gelu",
57
+ "hidden_dropout": 0.05,
58
+ "hidden_size": 1024,
59
+ "initializer_range": 0.02,
60
+ "intermediate_size": 4096,
61
+ "layer_norm_eps": 1e-05,
62
+ "layerdrop": 0.05,
63
+ "mask_channel_length": 10,
64
+ "mask_channel_min_space": 1,
65
+ "mask_channel_other": 0.0,
66
+ "mask_channel_prob": 0.0,
67
+ "mask_channel_selection": "static",
68
+ "mask_feature_length": 10,
69
+ "mask_feature_min_masks": 0,
70
+ "mask_feature_prob": 0.0,
71
+ "mask_time_length": 10,
72
+ "mask_time_min_masks": 2,
73
+ "mask_time_min_space": 1,
74
+ "mask_time_other": 0.0,
75
+ "mask_time_prob": 0.05,
76
+ "mask_time_selection": "static",
77
+ "model_type": "wav2vec2",
78
+ "num_adapter_layers": 3,
79
+ "num_attention_heads": 16,
80
+ "num_codevector_groups": 2,
81
+ "num_codevectors_per_group": 320,
82
+ "num_conv_pos_embedding_groups": 16,
83
+ "num_conv_pos_embeddings": 128,
84
+ "num_feat_extract_layers": 7,
85
+ "num_hidden_layers": 24,
86
+ "num_negatives": 100,
87
+ "output_hidden_size": 1024,
88
+ "pad_token_id": 0,
89
+ "proj_codevector_dim": 768,
90
+ "tdnn_dilation": [
91
+ 1,
92
+ 2,
93
+ 3,
94
+ 1,
95
+ 1
96
+ ],
97
+ "tdnn_dim": [
98
+ 512,
99
+ 512,
100
+ 512,
101
+ 512,
102
+ 1500
103
+ ],
104
+ "tdnn_kernel": [
105
+ 5,
106
+ 3,
107
+ 3,
108
+ 1,
109
+ 1
110
+ ],
111
+ "torch_dtype": "float32",
112
+ "transformers_version": "4.18.0",
113
+ "use_weighted_layer_sum": false,
114
+ "vocab_size": 38,
115
+ "xvector_output_dim": 512
116
+ }
cv7_evaluation.json ADDED
@@ -0,0 +1,396 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "validation": [
3
+ {
4
+ "subset_type": "all",
5
+ "subset": "all",
6
+ "cer": 0.06742130099167488,
7
+ "wer": 0.2605265303903447
8
+ },
9
+ {
10
+ "subset_type": "age",
11
+ "subset": "",
12
+ "cer": 0.06634671579367618,
13
+ "wer": 0.257145473770096
14
+ },
15
+ {
16
+ "subset_type": "age",
17
+ "subset": "thirties",
18
+ "cer": 0.06635223281485199,
19
+ "wer": 0.25082236842105265
20
+ },
21
+ {
22
+ "subset_type": "age",
23
+ "subset": "sixties",
24
+ "cer": 0.07129094412331406,
25
+ "wer": 0.276006272869838
26
+ },
27
+ {
28
+ "subset_type": "age",
29
+ "subset": "seventies",
30
+ "cer": 0.06776331942057451,
31
+ "wer": 0.27721088435374147
32
+ },
33
+ {
34
+ "subset_type": "age",
35
+ "subset": "fifties",
36
+ "cer": 0.06961059420075813,
37
+ "wer": 0.2663519892110587
38
+ },
39
+ {
40
+ "subset_type": "age",
41
+ "subset": "twenties",
42
+ "cer": 0.07271623560575402,
43
+ "wer": 0.2840481565086531
44
+ },
45
+ {
46
+ "subset_type": "age",
47
+ "subset": "eighties",
48
+ "cer": 0.0636215334420881,
49
+ "wer": 0.24390243902439024
50
+ },
51
+ {
52
+ "subset_type": "age",
53
+ "subset": "teens",
54
+ "cer": 0.07430408018739446,
55
+ "wer": 0.28321548195013024
56
+ },
57
+ {
58
+ "subset_type": "age",
59
+ "subset": "fourties",
60
+ "cer": 0.08567251461988304,
61
+ "wer": 0.3094766619519095
62
+ },
63
+ {
64
+ "subset_type": "gender",
65
+ "subset": "",
66
+ "cer": 0.06635088757254141,
67
+ "wer": 0.25713928898509697
68
+ },
69
+ {
70
+ "subset_type": "gender",
71
+ "subset": "male",
72
+ "cer": 0.07239016576833113,
73
+ "wer": 0.27654213727193744
74
+ },
75
+ {
76
+ "subset_type": "gender",
77
+ "subset": "other",
78
+ "cer": 0.1111111111111111,
79
+ "wer": 0.43902439024390244
80
+ },
81
+ {
82
+ "subset_type": "gender",
83
+ "subset": "female",
84
+ "cer": 0.06974030861874295,
85
+ "wer": 0.26283048211508553
86
+ },
87
+ {
88
+ "subset_type": "accent",
89
+ "subset": "",
90
+ "cer": 0.06691924472885617,
91
+ "wer": 0.25890539747844465
92
+ },
93
+ {
94
+ "subset_type": "accent",
95
+ "subset": "united_kingdom",
96
+ "cer": 0.11139896373056994,
97
+ "wer": 0.4188034188034188
98
+ },
99
+ {
100
+ "subset_type": "accent",
101
+ "subset": "france",
102
+ "cer": 0.037162162162162164,
103
+ "wer": 0.22916666666666666
104
+ },
105
+ {
106
+ "subset_type": "accent",
107
+ "subset": "poland",
108
+ "cer": 0.04390243902439024,
109
+ "wer": 0.18181818181818182
110
+ },
111
+ {
112
+ "subset_type": "accent",
113
+ "subset": "turkey",
114
+ "cer": 0.14496314496314497,
115
+ "wer": 0.4915254237288136
116
+ },
117
+ {
118
+ "subset_type": "accent",
119
+ "subset": "germany",
120
+ "cer": 0.06825976677780082,
121
+ "wer": 0.2622780092085069
122
+ },
123
+ {
124
+ "subset_type": "accent",
125
+ "subset": "austria",
126
+ "cer": 0.06972287735849056,
127
+ "wer": 0.26649746192893403
128
+ },
129
+ {
130
+ "subset_type": "accent",
131
+ "subset": "belgium",
132
+ "cer": 0.043243243243243246,
133
+ "wer": 0.23076923076923078
134
+ },
135
+ {
136
+ "subset_type": "accent",
137
+ "subset": "luxembourg",
138
+ "cer": 0.09927360774818401,
139
+ "wer": 0.32231404958677684
140
+ },
141
+ {
142
+ "subset_type": "accent",
143
+ "subset": "russia",
144
+ "cer": 0.09897610921501707,
145
+ "wer": 0.3422459893048128
146
+ },
147
+ {
148
+ "subset_type": "accent",
149
+ "subset": "czechia",
150
+ "cer": 0.0547945205479452,
151
+ "wer": 0.22424242424242424
152
+ },
153
+ {
154
+ "subset_type": "accent",
155
+ "subset": "other",
156
+ "cer": 0.09580838323353294,
157
+ "wer": 0.32
158
+ },
159
+ {
160
+ "subset_type": "accent",
161
+ "subset": "slovakia",
162
+ "cer": 0.06111111111111111,
163
+ "wer": 0.2631578947368421
164
+ },
165
+ {
166
+ "subset_type": "accent",
167
+ "subset": "united_states",
168
+ "cer": 0.10982658959537572,
169
+ "wer": 0.4098939929328622
170
+ },
171
+ {
172
+ "subset_type": "accent",
173
+ "subset": "netherlands",
174
+ "cer": 0.07744107744107744,
175
+ "wer": 0.21739130434782608
176
+ },
177
+ {
178
+ "subset_type": "accent",
179
+ "subset": "switzerland",
180
+ "cer": 0.07848764661841777,
181
+ "wer": 0.3126079447322971
182
+ },
183
+ {
184
+ "subset_type": "accent",
185
+ "subset": "canada",
186
+ "cer": 0.10661764705882353,
187
+ "wer": 0.425
188
+ }
189
+ ],
190
+ "test": [
191
+ {
192
+ "subset_type": "all",
193
+ "subset": "all",
194
+ "cer": 0.07679478470187219,
195
+ "wer": 0.2818758757108712
196
+ },
197
+ {
198
+ "subset_type": "age",
199
+ "subset": "",
200
+ "cer": 0.0758498150684626,
201
+ "wer": 0.2796152224734603
202
+ },
203
+ {
204
+ "subset_type": "age",
205
+ "subset": "thirties",
206
+ "cer": 0.07463367873282513,
207
+ "wer": 0.2688564476885645
208
+ },
209
+ {
210
+ "subset_type": "age",
211
+ "subset": "seventies",
212
+ "cer": 0.097096852890949,
213
+ "wer": 0.3197969543147208
214
+ },
215
+ {
216
+ "subset_type": "age",
217
+ "subset": "sixties",
218
+ "cer": 0.10182666148464826,
219
+ "wer": 0.33923705722070846
220
+ },
221
+ {
222
+ "subset_type": "age",
223
+ "subset": "fifties",
224
+ "cer": 0.0666905700968089,
225
+ "wer": 0.26104218362282877
226
+ },
227
+ {
228
+ "subset_type": "age",
229
+ "subset": "twenties",
230
+ "cer": 0.0811146101727264,
231
+ "wer": 0.30157701447396845
232
+ },
233
+ {
234
+ "subset_type": "age",
235
+ "subset": "eighties",
236
+ "cer": 0.1706422018348624,
237
+ "wer": 0.4931506849315068
238
+ },
239
+ {
240
+ "subset_type": "age",
241
+ "subset": "teens",
242
+ "cer": 0.09201733127708012,
243
+ "wer": 0.3251503006012024
244
+ },
245
+ {
246
+ "subset_type": "age",
247
+ "subset": "fourties",
248
+ "cer": 0.11175368575384866,
249
+ "wer": 0.34958448753462606
250
+ },
251
+ {
252
+ "subset_type": "gender",
253
+ "subset": "",
254
+ "cer": 0.07597557684974365,
255
+ "wer": 0.2798550347089201
256
+ },
257
+ {
258
+ "subset_type": "gender",
259
+ "subset": "male",
260
+ "cer": 0.08009040186453846,
261
+ "wer": 0.28926346061671193
262
+ },
263
+ {
264
+ "subset_type": "gender",
265
+ "subset": "other",
266
+ "cer": 0.11872909698996656,
267
+ "wer": 0.3068181818181818
268
+ },
269
+ {
270
+ "subset_type": "gender",
271
+ "subset": "female",
272
+ "cer": 0.1012895804322819,
273
+ "wer": 0.35175879396984927
274
+ },
275
+ {
276
+ "subset_type": "accent",
277
+ "subset": "",
278
+ "cer": 0.07672884606451527,
279
+ "wer": 0.2818792456270991
280
+ },
281
+ {
282
+ "subset_type": "accent",
283
+ "subset": "united_kingdom",
284
+ "cer": 0.14576802507836992,
285
+ "wer": 0.49019607843137253
286
+ },
287
+ {
288
+ "subset_type": "accent",
289
+ "subset": "russia",
290
+ "cer": 0.26936026936026936,
291
+ "wer": 0.6888888888888889
292
+ },
293
+ {
294
+ "subset_type": "accent",
295
+ "subset": "bulgaria",
296
+ "cer": 0.0,
297
+ "wer": 0.0
298
+ },
299
+ {
300
+ "subset_type": "accent",
301
+ "subset": "czechia",
302
+ "cer": 0.21951219512195122,
303
+ "wer": 0.7547169811320755
304
+ },
305
+ {
306
+ "subset_type": "accent",
307
+ "subset": "other",
308
+ "cer": 0.021739130434782608,
309
+ "wer": 0.07142857142857142
310
+ },
311
+ {
312
+ "subset_type": "accent",
313
+ "subset": "slovakia",
314
+ "cer": 0.05303030303030303,
315
+ "wer": 0.25862068965517243
316
+ },
317
+ {
318
+ "subset_type": "accent",
319
+ "subset": "united_states",
320
+ "cer": 0.07376058041112454,
321
+ "wer": 0.26556016597510373
322
+ },
323
+ {
324
+ "subset_type": "accent",
325
+ "subset": "switzerland",
326
+ "cer": 0.0909275421148772,
327
+ "wer": 0.3487394957983193
328
+ },
329
+ {
330
+ "subset_type": "accent",
331
+ "subset": "turkey",
332
+ "cer": 0.21505376344086022,
333
+ "wer": 0.6
334
+ },
335
+ {
336
+ "subset_type": "accent",
337
+ "subset": "poland",
338
+ "cer": 0.13879598662207357,
339
+ "wer": 0.43478260869565216
340
+ },
341
+ {
342
+ "subset_type": "accent",
343
+ "subset": "austria",
344
+ "cer": 0.08104469574582661,
345
+ "wer": 0.2871652816251154
346
+ },
347
+ {
348
+ "subset_type": "accent",
349
+ "subset": "belgium",
350
+ "cer": 0.06598984771573604,
351
+ "wer": 0.3333333333333333
352
+ },
353
+ {
354
+ "subset_type": "accent",
355
+ "subset": "luxembourg",
356
+ "cer": 0.04929577464788732,
357
+ "wer": 0.2631578947368421
358
+ },
359
+ {
360
+ "subset_type": "accent",
361
+ "subset": "hungary",
362
+ "cer": 0.11282051282051282,
363
+ "wer": 0.42857142857142855
364
+ },
365
+ {
366
+ "subset_type": "accent",
367
+ "subset": "france",
368
+ "cer": 0.0979631425800194,
369
+ "wer": 0.37333333333333335
370
+ },
371
+ {
372
+ "subset_type": "accent",
373
+ "subset": "germany",
374
+ "cer": 0.07328505037933454,
375
+ "wer": 0.2677122769064359
376
+ },
377
+ {
378
+ "subset_type": "accent",
379
+ "subset": "denmark",
380
+ "cer": 0.058823529411764705,
381
+ "wer": 0.2222222222222222
382
+ },
383
+ {
384
+ "subset_type": "accent",
385
+ "subset": "greece",
386
+ "cer": 0.09615384615384616,
387
+ "wer": 0.25925925925925924
388
+ },
389
+ {
390
+ "subset_type": "accent",
391
+ "subset": "netherlands",
392
+ "cer": 0.09606656580937972,
393
+ "wer": 0.35751295336787564
394
+ }
395
+ ]
396
+ }
cv7_transcriptions.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:98a734dca949edce8fcfc671ed7204aeea83edae8ffafc95ab31d5d45d8745ae
3
+ size 137502576
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,
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 sha256:8ae6e05cb9f5f9338268776c0e589f2a80ac72c7396b0a55c37c9c63c8ac634a
3
+ size 1262054513
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": " ", "special_tokens_map_file": null, "name_or_path": "/tmp/tmpsedmsc3k", "tokenizer_class": "Wav2Vec2CTCTokenizer", "processor_class": "Wav2Vec2Processor"}
vocab.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"<pad>": 0, "<s>": 1, "</s>": 2, "<unk>": 3, "|": 4, "'": 5, "-": 6, "a": 7, "b": 8, "c": 9, "d": 10, "e": 11, "f": 12, "g": 13, "h": 14, "i": 15, "j": 16, "k": 17, "l": 18, "m": 19, "n": 20, "o": 21, "p": 22, "q": 23, "r": 24, "s": 25, "t": 26, "u": 27, "v": 28, "w": 29, "x": 30, "y": 31, "z": 32, "ß": 33, "á": 34, "ä": 35, "ö": 36, "ü": 37}