Hugging Face's logo

aware-ai
/
wav2vec2-base-marian-german

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
speech-encoder-decoder
Inference Endpoints
Model card Files Metrics Community
1
flozi00 commited on
Commit
f651eff
·
1 Parent(s): b283002

Training in progress, step 1000

Browse files
Files changed (12) hide show
  1. .gitignore +1 -0
  2. config.json +285 -0
  3. preprocessor_config.json +10 -0
  4. pytorch_model.bin +3 -0
  5. runs/Sep01_14-11-59_aware/1662034350.1880262/events.out.tfevents.1662034350.aware.311129.1 +3 -0
  6. runs/Sep01_14-11-59_aware/events.out.tfevents.1662034350.aware.311129.0 +3 -0
  7. source.spm +0 -0
  8. special_tokens_map.json +5 -0
  9. target.spm +0 -0
  10. tokenizer_config.json +14 -0
  11. training_args.bin +3 -0
  12. vocab.json +0 -0
.gitignore ADDED
@@ -0,0 +1 @@
 
 
1
+ checkpoint-*/
config.json ADDED
@@ -0,0 +1,285 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "wav2vec2-base-marian-german",
3
+ "architectures": [
4
+ "SpeechEncoderDecoderModel"
5
+ ],
6
+ "decoder": {
7
+ "_name_or_path": "Helsinki-NLP/opus-mt-en-de",
8
+ "_num_labels": 3,
9
+ "activation_dropout": 0.0,
10
+ "activation_function": "swish",
11
+ "add_bias_logits": false,
12
+ "add_cross_attention": true,
13
+ "add_final_layer_norm": false,
14
+ "architectures": [
15
+ "MarianMTModel"
16
+ ],
17
+ "attention_dropout": 0.0,
18
+ "bad_words_ids": [
19
+ [
20
+ 58100
21
+ ]
22
+ ],
23
+ "bos_token_id": 0,
24
+ "chunk_size_feed_forward": 0,
25
+ "classif_dropout": 0.0,
26
+ "classifier_dropout": 0.0,
27
+ "cross_attention_hidden_size": null,
28
+ "d_model": 512,
29
+ "decoder_attention_heads": 8,
30
+ "decoder_ffn_dim": 2048,
31
+ "decoder_layerdrop": 0.0,
32
+ "decoder_layers": 6,
33
+ "decoder_start_token_id": 58100,
34
+ "decoder_vocab_size": 58101,
35
+ "diversity_penalty": 0.0,
36
+ "do_sample": false,
37
+ "dropout": 0.1,
38
+ "early_stopping": false,
39
+ "encoder_attention_heads": 8,
40
+ "encoder_ffn_dim": 2048,
41
+ "encoder_layerdrop": 0.0,
42
+ "encoder_layers": 6,
43
+ "encoder_no_repeat_ngram_size": 0,
44
+ "eos_token_id": 0,
45
+ "exponential_decay_length_penalty": null,
46
+ "finetuning_task": null,
47
+ "forced_bos_token_id": null,
48
+ "forced_eos_token_id": 0,
49
+ "gradient_checkpointing": false,
50
+ "id2label": {
51
+ "0": "LABEL_0",
52
+ "1": "LABEL_1",
53
+ "2": "LABEL_2"
54
+ },
55
+ "init_std": 0.02,
56
+ "is_decoder": true,
57
+ "is_encoder_decoder": false,
58
+ "label2id": {
59
+ "LABEL_0": 0,
60
+ "LABEL_1": 1,
61
+ "LABEL_2": 2
62
+ },
63
+ "length_penalty": 1.0,
64
+ "max_length": 512,
65
+ "max_position_embeddings": 512,
66
+ "min_length": 0,
67
+ "model_type": "marian",
68
+ "no_repeat_ngram_size": 0,
69
+ "normalize_before": false,
70
+ "normalize_embedding": false,
71
+ "num_beam_groups": 1,
72
+ "num_beams": 4,
73
+ "num_hidden_layers": 6,
74
+ "num_return_sequences": 1,
75
+ "output_attentions": false,
76
+ "output_hidden_states": false,
77
+ "output_scores": false,
78
+ "pad_token_id": 58100,
79
+ "prefix": null,
80
+ "problem_type": null,
81
+ "pruned_heads": {},
82
+ "remove_invalid_values": false,
83
+ "repetition_penalty": 1.0,
84
+ "return_dict": true,
85
+ "return_dict_in_generate": false,
86
+ "scale_embedding": true,
87
+ "sep_token_id": null,
88
+ "share_encoder_decoder_embeddings": true,
89
+ "static_position_embeddings": true,
90
+ "task_specific_params": null,
91
+ "temperature": 1.0,
92
+ "tf_legacy_loss": false,
93
+ "tie_encoder_decoder": false,
94
+ "tie_word_embeddings": true,
95
+ "tokenizer_class": null,
96
+ "top_k": 50,
97
+ "top_p": 1.0,
98
+ "torch_dtype": null,
99
+ "torchscript": false,
100
+ "transformers_version": "4.21.1",
101
+ "typical_p": 1.0,
102
+ "use_bfloat16": false,
103
+ "use_cache": true,
104
+ "vocab_size": 58101
105
+ },
106
+ "decoder_start_token_id": 0,
107
+ "encoder": {
108
+ "_name_or_path": "oliverguhr/wav2vec2-base-german-cv9",
109
+ "activation_dropout": 0.05,
110
+ "adapter_kernel_size": 3,
111
+ "adapter_stride": 2,
112
+ "add_adapter": true,
113
+ "add_cross_attention": false,
114
+ "apply_spec_augment": true,
115
+ "architectures": [
116
+ "Wav2Vec2ForCTC"
117
+ ],
118
+ "attention_dropout": 0.1,
119
+ "bad_words_ids": null,
120
+ "bos_token_id": 1,
121
+ "chunk_size_feed_forward": 0,
122
+ "classifier_proj_size": 256,
123
+ "codevector_dim": 256,
124
+ "contrastive_logits_temperature": 0.1,
125
+ "conv_bias": false,
126
+ "conv_dim": [
127
+ 512,
128
+ 512,
129
+ 512,
130
+ 512,
131
+ 512,
132
+ 512,
133
+ 512
134
+ ],
135
+ "conv_kernel": [
136
+ 10,
137
+ 3,
138
+ 3,
139
+ 3,
140
+ 3,
141
+ 2,
142
+ 2
143
+ ],
144
+ "conv_stride": [
145
+ 5,
146
+ 2,
147
+ 2,
148
+ 2,
149
+ 2,
150
+ 2,
151
+ 2
152
+ ],
153
+ "cross_attention_hidden_size": null,
154
+ "ctc_loss_reduction": "mean",
155
+ "ctc_zero_infinity": false,
156
+ "decoder_start_token_id": null,
157
+ "diversity_loss_weight": 0.1,
158
+ "diversity_penalty": 0.0,
159
+ "do_sample": false,
160
+ "do_stable_layer_norm": false,
161
+ "early_stopping": false,
162
+ "encoder_no_repeat_ngram_size": 0,
163
+ "eos_token_id": 2,
164
+ "exponential_decay_length_penalty": null,
165
+ "feat_extract_activation": "gelu",
166
+ "feat_extract_norm": "group",
167
+ "feat_proj_dropout": 0.0,
168
+ "feat_quantizer_dropout": 0.0,
169
+ "final_dropout": 0.0,
170
+ "finetuning_task": null,
171
+ "forced_bos_token_id": null,
172
+ "forced_eos_token_id": null,
173
+ "freeze_feat_extract_train": true,
174
+ "hidden_act": "gelu",
175
+ "hidden_dropout": 0.1,
176
+ "hidden_size": 768,
177
+ "id2label": {
178
+ "0": "LABEL_0",
179
+ "1": "LABEL_1"
180
+ },
181
+ "initializer_range": 0.02,
182
+ "intermediate_size": 3072,
183
+ "is_decoder": false,
184
+ "is_encoder_decoder": false,
185
+ "label2id": {
186
+ "LABEL_0": 0,
187
+ "LABEL_1": 1
188
+ },
189
+ "layer_norm_eps": 1e-05,
190
+ "layerdrop": 0.0,
191
+ "length_penalty": 1.0,
192
+ "mask_channel_length": 10,
193
+ "mask_channel_min_space": 1,
194
+ "mask_channel_other": 0.0,
195
+ "mask_channel_prob": 0.0,
196
+ "mask_channel_selection": "static",
197
+ "mask_feature_length": 10,
198
+ "mask_feature_min_masks": 0,
199
+ "mask_feature_prob": 0.0,
200
+ "mask_time_length": 10,
201
+ "mask_time_min_masks": 2,
202
+ "mask_time_min_space": 1,
203
+ "mask_time_other": 0.0,
204
+ "mask_time_prob": 0.05,
205
+ "mask_time_selection": "static",
206
+ "max_length": 20,
207
+ "min_length": 0,
208
+ "model_type": "wav2vec2",
209
+ "no_mask_channel_overlap": false,
210
+ "no_mask_time_overlap": false,
211
+ "no_repeat_ngram_size": 0,
212
+ "num_adapter_layers": 3,
213
+ "num_attention_heads": 12,
214
+ "num_beam_groups": 1,
215
+ "num_beams": 1,
216
+ "num_codevector_groups": 2,
217
+ "num_codevectors_per_group": 320,
218
+ "num_conv_pos_embedding_groups": 16,
219
+ "num_conv_pos_embeddings": 128,
220
+ "num_feat_extract_layers": 7,
221
+ "num_hidden_layers": 12,
222
+ "num_negatives": 100,
223
+ "num_return_sequences": 1,
224
+ "output_attentions": false,
225
+ "output_hidden_size": 768,
226
+ "output_hidden_states": false,
227
+ "output_scores": false,
228
+ "pad_token_id": 32,
229
+ "prefix": null,
230
+ "problem_type": null,
231
+ "proj_codevector_dim": 256,
232
+ "pruned_heads": {},
233
+ "remove_invalid_values": false,
234
+ "repetition_penalty": 1.0,
235
+ "return_dict": true,
236
+ "return_dict_in_generate": false,
237
+ "sep_token_id": null,
238
+ "task_specific_params": null,
239
+ "tdnn_dilation": [
240
+ 1,
241
+ 2,
242
+ 3,
243
+ 1,
244
+ 1
245
+ ],
246
+ "tdnn_dim": [
247
+ 512,
248
+ 512,
249
+ 512,
250
+ 512,
251
+ 1500
252
+ ],
253
+ "tdnn_kernel": [
254
+ 5,
255
+ 3,
256
+ 3,
257
+ 1,
258
+ 1
259
+ ],
260
+ "temperature": 1.0,
261
+ "tf_legacy_loss": false,
262
+ "tie_encoder_decoder": false,
263
+ "tie_word_embeddings": true,
264
+ "tokenizer_class": null,
265
+ "top_k": 50,
266
+ "top_p": 1.0,
267
+ "torch_dtype": "float32",
268
+ "torchscript": false,
269
+ "transformers_version": "4.21.1",
270
+ "typical_p": 1.0,
271
+ "use_bfloat16": false,
272
+ "use_weighted_layer_sum": false,
273
+ "vocab_size": 35,
274
+ "xvector_output_dim": 512
275
+ },
276
+ "eos_token_id": 0,
277
+ "is_encoder_decoder": true,
278
+ "max_length": 200,
279
+ "model_type": "speech-encoder-decoder",
280
+ "num_beams": 5,
281
+ "pad_token_id": 58100,
282
+ "tie_word_embeddings": false,
283
+ "torch_dtype": "float32",
284
+ "transformers_version": null
285
+ }
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
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:385bac37afe4bb06a8d1c1424a44d602f11c306a3d3cffd1f0acbdc2b417ae8e
3
+ size 642611599
runs/Sep01_14-11-59_aware/1662034350.1880262/events.out.tfevents.1662034350.aware.311129.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:425b7b963e2079c2b189fe1e0dea8f1a2036f73d07fb39bc87f9b499ea09f566
3
+ size 5487
runs/Sep01_14-11-59_aware/events.out.tfevents.1662034350.aware.311129.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f177e63adc48e10011be9d26175c7c50f5467d0e0e413dac08518474a5d001d6
3
+ size 166792
source.spm ADDED
Binary file (768 kB). View file
 
special_tokens_map.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "eos_token": "</s>",
3
+ "pad_token": "<pad>",
4
+ "unk_token": "<unk>"
5
+ }
target.spm ADDED
Binary file (797 kB). View file
 
tokenizer_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "eos_token": "</s>",
3
+ "model_max_length": 512,
4
+ "name_or_path": "wav2vec2-base-marian-german",
5
+ "pad_token": "<pad>",
6
+ "processor_class": "Wav2Vec2Processor",
7
+ "separate_vocabs": false,
8
+ "source_lang": "en",
9
+ "sp_model_kwargs": {},
10
+ "special_tokens_map_file": null,
11
+ "target_lang": "de",
12
+ "tokenizer_class": "MarianTokenizer",
13
+ "unk_token": "<unk>"
14
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:934b3c55768b893eadb286d0bbe8da0f74e24a203122740fa27fea4d08a4146c
3
+ size 3439
vocab.json ADDED
The diff for this file is too large to render. See raw diff