chickenrice0721 commited on
Commit
0006c7d
1 Parent(s): 47fcd1a
Files changed (6) hide show
  1. README.md +103 -3
  2. config.json +291 -0
  3. model.bin +3 -0
  4. preprocessor_config.json +14 -0
  5. tokenizer.json +0 -0
  6. vocabulary.json +0 -0
README.md CHANGED
@@ -1,3 +1,103 @@
1
- ---
2
- license: apache-2.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: openai/whisper-large-v2
4
+ tags:
5
+ - audio
6
+ - automatic-speech-recognition
7
+ - translate
8
+ - generated_from_trainer
9
+ language:
10
+ - zh
11
+ metrics:
12
+ - cer
13
+ - wer
14
+ model-index:
15
+ - name: whisper-large-v2-translate-zh-v0.1-lt-ct2
16
+ results: []
17
+ ---
18
+
19
+ # whisper-large-v2-translate-zh-v0.1-lt-ct2
20
+
21
+ This model is a fine-tuned version of [openai/whisper-large-v2](https://huggingface.co/openai/whisper-large-v2).
22
+
23
+ ## Model description
24
+
25
+ 3500小时 (日语音频,中文字幕) 数据微调, 翻译模式直出中文
26
+
27
+ CTranslate2 版
28
+
29
+ ## Usage
30
+
31
+ task='translate', language='ja'
32
+
33
+ ## Training procedure
34
+
35
+ ### Training hyperparameters
36
+
37
+ The following hyperparameters were used during training:
38
+ - learning_rate: 1e-05
39
+ - train_batch_size: 32
40
+ - eval_batch_size: 32
41
+ - seed: 42
42
+ - gradient_accumulation_steps: 2
43
+ - total_train_batch_size: 64
44
+ - optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-08
45
+ - lr_scheduler_type: linear
46
+ - lr_scheduler_warmup_steps: 4000
47
+ - dropout: 0.1
48
+ - mask_time_prob: 0.05
49
+ - mask_feature_prob: 0.2
50
+ - condition_on_previous_text_rate: 0.5
51
+
52
+ ### Training results
53
+
54
+ | Training Loss | Epoch | Step | Validation Loss | Cer | Wer |
55
+ |:-------------:|:------:|:-----:|:---------------:|:------:|:------:|
56
+ | 1.743 | 0.0740 | 1000 | 1.5631 | 0.8223 | 1.4517 |
57
+ | 1.6014 | 0.1479 | 2000 | 1.4808 | 0.6775 | 1.0950 |
58
+ | 1.5549 | 0.2219 | 3000 | 1.4381 | 0.6756 | 1.1158 |
59
+ | 1.5283 | 0.2958 | 4000 | 1.4174 | 0.6992 | 1.1137 |
60
+ | 1.474 | 0.3698 | 5000 | 1.3849 | 0.6570 | 1.1369 |
61
+ | 1.4193 | 0.4437 | 6000 | 1.3657 | 0.6544 | 1.1339 |
62
+ | 1.4148 | 0.5177 | 7000 | 1.3477 | 0.6386 | 1.1647 |
63
+ | 1.3754 | 0.5916 | 8000 | 1.3392 | 0.6228 | 1.0461 |
64
+ | 1.3441 | 0.6656 | 9000 | 1.3362 | 0.6196 | 1.0609 |
65
+ | 1.3545 | 0.7395 | 10000 | 1.3176 | 0.6354 | 1.2138 |
66
+ | 1.3498 | 0.8135 | 11000 | 1.3236 | 0.6631 | 1.2232 |
67
+ | 1.31 | 0.8874 | 12000 | 1.3020 | 0.6199 | 1.0018 |
68
+ | 1.3213 | 0.9614 | 13000 | 1.2966 | 0.5922 | 1.0021 |
69
+ | 1.2375 | 1.0353 | 14000 | 1.2900 | 0.6097 | 1.0639 |
70
+ | 1.2334 | 1.1093 | 15000 | 1.2963 | 0.6150 | 1.0920 |
71
+ | 1.2277 | 1.1832 | 16000 | 1.2888 | 0.6077 | 1.0929 |
72
+ | 1.2087 | 1.2572 | 17000 | 1.2779 | 0.5954 | 1.0012 |
73
+ | 1.2131 | 1.3311 | 18000 | 1.2722 | 0.5776 | 1.0075 |
74
+ | 1.2012 | 1.4051 | 19000 | 1.2716 | 0.5726 | 1.0211 |
75
+ | 1.1912 | 1.4790 | 20000 | 1.2707 | 0.6007 | 1.1538 |
76
+ | 1.2127 | 1.5530 | 21000 | 1.2749 | 0.6086 | 1.0742 |
77
+ | 1.1789 | 1.6269 | 22000 | 1.2797 | 0.5765 | 1.0072 |
78
+ | 1.1527 | 1.7009 | 23000 | 1.2761 | 0.5855 | 1.0588 |
79
+ | 1.1693 | 1.7748 | 24000 | 1.2701 | 0.5635 | 0.9928 |
80
+ | 1.1709 | 1.8488 | 25000 | 1.2662 | 0.5980 | 1.0697 |
81
+ | 1.1637 | 1.9227 | 26000 | 1.2749 | 0.5872 | 1.0392 |
82
+ | 1.1562 | 1.9967 | 27000 | 1.2587 | 0.5651 | 1.0121 |
83
+ | 1.0929 | 2.0706 | 28000 | 1.2668 | 0.5857 | 1.0139 |
84
+ | 1.1232 | 2.1446 | 29000 | 1.2710 | 0.5742 | 0.9997 |
85
+ | 1.1045 | 2.2185 | 30000 | 1.2656 | 0.5643 | 0.9897 |
86
+ | 1.0841 | 2.2925 | 31000 | 1.2695 | 0.5835 | 1.0181 |
87
+ | 1.0868 | 2.3664 | 32000 | 1.2707 | 0.5673 | 0.9964 |
88
+ | 1.0938 | 2.4404 | 33000 | 1.2644 | 0.5712 | 0.9928 |
89
+ | 1.0938 | 2.5143 | 34000 | 1.2662 | 0.5750 | 1.0109 |
90
+ | 1.0848 | 2.5883 | 35000 | 1.2677 | 0.5841 | 1.0832 |
91
+ | 1.0914 | 2.6622 | 36000 | 1.2638 | 0.5801 | 1.0299 |
92
+ | 1.0688 | 2.7362 | 37000 | 1.2587 | 0.5694 | 1.0072 |
93
+ | 1.0856 | 2.8101 | 38000 | 1.2581 | 0.5646 | 1.0057 |
94
+ | 1.1037 | 2.8841 | 39000 | 1.2557 | 0.5771 | 1.0262 |
95
+ | 1.0652 | 2.9580 | 40000 | 1.2566 | 0.5634 | 0.9979 |
96
+
97
+
98
+ ### Framework versions
99
+
100
+ - Transformers 4.41.1
101
+ - Pytorch 2.3.1+cu121
102
+ - Datasets 2.19.1
103
+ - Tokenizers 0.19.1
config.json ADDED
@@ -0,0 +1,291 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alignment_heads": [
3
+ [
4
+ 10,
5
+ 12
6
+ ],
7
+ [
8
+ 13,
9
+ 17
10
+ ],
11
+ [
12
+ 16,
13
+ 11
14
+ ],
15
+ [
16
+ 16,
17
+ 12
18
+ ],
19
+ [
20
+ 16,
21
+ 13
22
+ ],
23
+ [
24
+ 17,
25
+ 15
26
+ ],
27
+ [
28
+ 17,
29
+ 16
30
+ ],
31
+ [
32
+ 18,
33
+ 4
34
+ ],
35
+ [
36
+ 18,
37
+ 11
38
+ ],
39
+ [
40
+ 18,
41
+ 19
42
+ ],
43
+ [
44
+ 19,
45
+ 11
46
+ ],
47
+ [
48
+ 21,
49
+ 2
50
+ ],
51
+ [
52
+ 21,
53
+ 3
54
+ ],
55
+ [
56
+ 22,
57
+ 3
58
+ ],
59
+ [
60
+ 22,
61
+ 9
62
+ ],
63
+ [
64
+ 22,
65
+ 12
66
+ ],
67
+ [
68
+ 23,
69
+ 5
70
+ ],
71
+ [
72
+ 23,
73
+ 7
74
+ ],
75
+ [
76
+ 23,
77
+ 13
78
+ ],
79
+ [
80
+ 25,
81
+ 5
82
+ ],
83
+ [
84
+ 26,
85
+ 1
86
+ ],
87
+ [
88
+ 26,
89
+ 12
90
+ ],
91
+ [
92
+ 27,
93
+ 15
94
+ ]
95
+ ],
96
+ "lang_ids": [
97
+ 50259,
98
+ 50260,
99
+ 50261,
100
+ 50262,
101
+ 50263,
102
+ 50264,
103
+ 50265,
104
+ 50266,
105
+ 50267,
106
+ 50268,
107
+ 50269,
108
+ 50270,
109
+ 50271,
110
+ 50272,
111
+ 50273,
112
+ 50274,
113
+ 50275,
114
+ 50276,
115
+ 50277,
116
+ 50278,
117
+ 50279,
118
+ 50280,
119
+ 50281,
120
+ 50282,
121
+ 50283,
122
+ 50284,
123
+ 50285,
124
+ 50286,
125
+ 50287,
126
+ 50288,
127
+ 50289,
128
+ 50290,
129
+ 50291,
130
+ 50292,
131
+ 50293,
132
+ 50294,
133
+ 50295,
134
+ 50296,
135
+ 50297,
136
+ 50298,
137
+ 50299,
138
+ 50300,
139
+ 50301,
140
+ 50302,
141
+ 50303,
142
+ 50304,
143
+ 50305,
144
+ 50306,
145
+ 50307,
146
+ 50308,
147
+ 50309,
148
+ 50310,
149
+ 50311,
150
+ 50312,
151
+ 50313,
152
+ 50314,
153
+ 50315,
154
+ 50316,
155
+ 50317,
156
+ 50318,
157
+ 50319,
158
+ 50320,
159
+ 50321,
160
+ 50322,
161
+ 50323,
162
+ 50324,
163
+ 50325,
164
+ 50326,
165
+ 50327,
166
+ 50328,
167
+ 50329,
168
+ 50330,
169
+ 50331,
170
+ 50332,
171
+ 50333,
172
+ 50334,
173
+ 50335,
174
+ 50336,
175
+ 50337,
176
+ 50338,
177
+ 50339,
178
+ 50340,
179
+ 50341,
180
+ 50342,
181
+ 50343,
182
+ 50344,
183
+ 50345,
184
+ 50346,
185
+ 50347,
186
+ 50348,
187
+ 50349,
188
+ 50350,
189
+ 50351,
190
+ 50352,
191
+ 50353,
192
+ 50354,
193
+ 50355,
194
+ 50356,
195
+ 50357
196
+ ],
197
+ "suppress_ids": [
198
+ 1,
199
+ 2,
200
+ 7,
201
+ 8,
202
+ 9,
203
+ 10,
204
+ 14,
205
+ 25,
206
+ 26,
207
+ 27,
208
+ 28,
209
+ 29,
210
+ 31,
211
+ 58,
212
+ 59,
213
+ 60,
214
+ 61,
215
+ 62,
216
+ 63,
217
+ 90,
218
+ 91,
219
+ 92,
220
+ 93,
221
+ 359,
222
+ 503,
223
+ 522,
224
+ 542,
225
+ 873,
226
+ 893,
227
+ 902,
228
+ 918,
229
+ 922,
230
+ 931,
231
+ 1350,
232
+ 1853,
233
+ 1982,
234
+ 2460,
235
+ 2627,
236
+ 3246,
237
+ 3253,
238
+ 3268,
239
+ 3536,
240
+ 3846,
241
+ 3961,
242
+ 4183,
243
+ 4667,
244
+ 6585,
245
+ 6647,
246
+ 7273,
247
+ 9061,
248
+ 9383,
249
+ 10428,
250
+ 10929,
251
+ 11938,
252
+ 12033,
253
+ 12331,
254
+ 12562,
255
+ 13793,
256
+ 14157,
257
+ 14635,
258
+ 15265,
259
+ 15618,
260
+ 16553,
261
+ 16604,
262
+ 18362,
263
+ 18956,
264
+ 20075,
265
+ 21675,
266
+ 22520,
267
+ 26130,
268
+ 26161,
269
+ 26435,
270
+ 28279,
271
+ 29464,
272
+ 31650,
273
+ 32302,
274
+ 32470,
275
+ 36865,
276
+ 42863,
277
+ 47425,
278
+ 49870,
279
+ 50254,
280
+ 50258,
281
+ 50358,
282
+ 50359,
283
+ 50360,
284
+ 50361,
285
+ 50362
286
+ ],
287
+ "suppress_ids_begin": [
288
+ 220,
289
+ 50257
290
+ ]
291
+ }
model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b2868edaad88c79bae0bbb228ae2a0a19e4b5cc820d430e42e5c84f80478c63
3
+ size 3086913076
preprocessor_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chunk_length": 30,
3
+ "feature_extractor_type": "WhisperFeatureExtractor",
4
+ "feature_size": 80,
5
+ "hop_length": 160,
6
+ "n_fft": 400,
7
+ "n_samples": 480000,
8
+ "nb_max_frames": 3000,
9
+ "padding_side": "right",
10
+ "padding_value": 0.0,
11
+ "processor_class": "WhisperProcessor",
12
+ "return_attention_mask": false,
13
+ "sampling_rate": 16000
14
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
vocabulary.json ADDED
The diff for this file is too large to render. See raw diff