AshtonLow commited on
Commit
671aed5
1 Parent(s): 276429d

Upload config

Browse files
Files changed (2) hide show
  1. README.md +2 -2
  2. generation_config.json +274 -76
README.md CHANGED
@@ -1,6 +1,8 @@
1
  ---
2
  language:
3
  - en
 
 
4
  tags:
5
  - audio
6
  - automatic-speech-recognition
@@ -11,8 +13,6 @@ widget:
11
  - example_title: LibriSpeech sample 2
12
  src: https://cdn-media.huggingface.co/speech_samples/sample2.flac
13
  pipeline_tag: automatic-speech-recognition
14
- license: mit
15
- library_name: transformers
16
  ---
17
 
18
  # Distil-Whisper: distil-medium.en
 
1
  ---
2
  language:
3
  - en
4
+ license: mit
5
+ library_name: transformers
6
  tags:
7
  - audio
8
  - automatic-speech-recognition
 
13
  - example_title: LibriSpeech sample 2
14
  src: https://cdn-media.huggingface.co/speech_samples/sample2.flac
15
  pipeline_tag: automatic-speech-recognition
 
 
16
  ---
17
 
18
  # Distil-Whisper: distil-medium.en
generation_config.json CHANGED
@@ -1,24 +1,222 @@
1
  {
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2
  "begin_suppress_tokens": [
3
  220,
4
- 50256
5
  ],
6
  "bos_token_id": 50257,
7
- "decoder_start_token_id": 50257,
8
- "eos_token_id": 50256,
9
  "forced_decoder_ids": [
10
  [
11
  1,
12
- 50362
 
 
 
 
13
  ]
14
  ],
15
- "is_multilingual": false,
16
- "language": "<|en|>",
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
17
  "max_initial_timestamp_index": 50,
18
  "max_length": 448,
19
- "no_timestamps_token_id": 50362,
20
- "pad_token_id": 50256,
21
- "prev_sot_token_id": 50360,
22
  "return_timestamps": false,
23
  "suppress_tokens": [
24
  1,
@@ -44,75 +242,75 @@
44
  91,
45
  92,
46
  93,
47
- 357,
48
- 366,
49
- 438,
50
- 532,
51
- 685,
52
- 705,
53
- 796,
54
- 930,
55
- 1058,
56
- 1220,
57
- 1267,
58
- 1279,
59
- 1303,
60
- 1343,
61
- 1377,
62
- 1391,
63
- 1635,
64
- 1782,
65
- 1875,
66
- 2162,
67
- 2361,
68
- 2488,
69
- 3467,
70
- 4008,
71
- 4211,
72
- 4600,
73
- 4808,
74
- 5299,
75
- 5855,
76
- 6329,
77
- 7203,
78
- 9609,
79
- 9959,
80
- 10563,
81
- 10786,
82
- 11420,
83
- 11709,
84
- 11907,
85
- 13163,
86
- 13697,
87
- 13700,
88
- 14808,
89
- 15306,
90
- 16410,
91
- 16791,
92
- 17992,
93
- 19203,
94
- 19510,
95
- 20724,
96
- 22305,
97
- 22935,
98
- 27007,
99
- 30109,
100
- 30420,
101
- 33409,
102
- 34949,
103
- 40283,
104
- 40493,
105
- 40549,
106
- 47282,
107
- 49146,
108
- 50257,
109
- 50357,
110
  50358,
111
  50359,
112
  50360,
113
- 50361
 
114
  ],
115
- "task": "transcribe",
116
- "transformers_version": "4.35.0.dev0",
117
- "use_scan": false
 
 
118
  }
 
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
  "begin_suppress_tokens": [
97
  220,
98
+ 50257
99
  ],
100
  "bos_token_id": 50257,
101
+ "decoder_start_token_id": 50258,
102
+ "eos_token_id": 50257,
103
  "forced_decoder_ids": [
104
  [
105
  1,
106
+ null
107
+ ],
108
+ [
109
+ 2,
110
+ 50359
111
  ]
112
  ],
113
+ "is_multilingual": true,
114
+ "lang_to_id": {
115
+ "<|af|>": 50327,
116
+ "<|am|>": 50334,
117
+ "<|ar|>": 50272,
118
+ "<|as|>": 50350,
119
+ "<|az|>": 50304,
120
+ "<|ba|>": 50355,
121
+ "<|be|>": 50330,
122
+ "<|bg|>": 50292,
123
+ "<|bn|>": 50302,
124
+ "<|bo|>": 50347,
125
+ "<|br|>": 50309,
126
+ "<|bs|>": 50315,
127
+ "<|ca|>": 50270,
128
+ "<|cs|>": 50283,
129
+ "<|cy|>": 50297,
130
+ "<|da|>": 50285,
131
+ "<|de|>": 50261,
132
+ "<|el|>": 50281,
133
+ "<|en|>": 50259,
134
+ "<|es|>": 50262,
135
+ "<|et|>": 50307,
136
+ "<|eu|>": 50310,
137
+ "<|fa|>": 50300,
138
+ "<|fi|>": 50277,
139
+ "<|fo|>": 50338,
140
+ "<|fr|>": 50265,
141
+ "<|gl|>": 50319,
142
+ "<|gu|>": 50333,
143
+ "<|haw|>": 50352,
144
+ "<|ha|>": 50354,
145
+ "<|he|>": 50279,
146
+ "<|hi|>": 50276,
147
+ "<|hr|>": 50291,
148
+ "<|ht|>": 50339,
149
+ "<|hu|>": 50286,
150
+ "<|hy|>": 50312,
151
+ "<|id|>": 50275,
152
+ "<|is|>": 50311,
153
+ "<|it|>": 50274,
154
+ "<|ja|>": 50266,
155
+ "<|jw|>": 50356,
156
+ "<|ka|>": 50329,
157
+ "<|kk|>": 50316,
158
+ "<|km|>": 50323,
159
+ "<|kn|>": 50306,
160
+ "<|ko|>": 50264,
161
+ "<|la|>": 50294,
162
+ "<|lb|>": 50345,
163
+ "<|ln|>": 50353,
164
+ "<|lo|>": 50336,
165
+ "<|lt|>": 50293,
166
+ "<|lv|>": 50301,
167
+ "<|mg|>": 50349,
168
+ "<|mi|>": 50295,
169
+ "<|mk|>": 50308,
170
+ "<|ml|>": 50296,
171
+ "<|mn|>": 50314,
172
+ "<|mr|>": 50320,
173
+ "<|ms|>": 50282,
174
+ "<|mt|>": 50343,
175
+ "<|my|>": 50346,
176
+ "<|ne|>": 50313,
177
+ "<|nl|>": 50271,
178
+ "<|nn|>": 50342,
179
+ "<|no|>": 50288,
180
+ "<|oc|>": 50328,
181
+ "<|pa|>": 50321,
182
+ "<|pl|>": 50269,
183
+ "<|ps|>": 50340,
184
+ "<|pt|>": 50267,
185
+ "<|ro|>": 50284,
186
+ "<|ru|>": 50263,
187
+ "<|sa|>": 50344,
188
+ "<|sd|>": 50332,
189
+ "<|si|>": 50322,
190
+ "<|sk|>": 50298,
191
+ "<|sl|>": 50305,
192
+ "<|sn|>": 50324,
193
+ "<|so|>": 50326,
194
+ "<|sq|>": 50317,
195
+ "<|sr|>": 50303,
196
+ "<|su|>": 50357,
197
+ "<|sv|>": 50273,
198
+ "<|sw|>": 50318,
199
+ "<|ta|>": 50287,
200
+ "<|te|>": 50299,
201
+ "<|tg|>": 50331,
202
+ "<|th|>": 50289,
203
+ "<|tk|>": 50341,
204
+ "<|tl|>": 50348,
205
+ "<|tr|>": 50268,
206
+ "<|tt|>": 50351,
207
+ "<|uk|>": 50280,
208
+ "<|ur|>": 50290,
209
+ "<|uz|>": 50337,
210
+ "<|vi|>": 50278,
211
+ "<|yi|>": 50335,
212
+ "<|yo|>": 50325,
213
+ "<|zh|>": 50260
214
+ },
215
  "max_initial_timestamp_index": 50,
216
  "max_length": 448,
217
+ "no_timestamps_token_id": 50363,
218
+ "pad_token_id": 50257,
219
+ "prev_sot_token_id": 50361,
220
  "return_timestamps": false,
221
  "suppress_tokens": [
222
  1,
 
242
  91,
243
  92,
244
  93,
245
+ 359,
246
+ 503,
247
+ 522,
248
+ 542,
249
+ 873,
250
+ 893,
251
+ 902,
252
+ 918,
253
+ 922,
254
+ 931,
255
+ 1350,
256
+ 1853,
257
+ 1982,
258
+ 2460,
259
+ 2627,
260
+ 3246,
261
+ 3253,
262
+ 3268,
263
+ 3536,
264
+ 3846,
265
+ 3961,
266
+ 4183,
267
+ 4667,
268
+ 6585,
269
+ 6647,
270
+ 7273,
271
+ 9061,
272
+ 9383,
273
+ 10428,
274
+ 10929,
275
+ 11938,
276
+ 12033,
277
+ 12331,
278
+ 12562,
279
+ 13793,
280
+ 14157,
281
+ 14635,
282
+ 15265,
283
+ 15618,
284
+ 16553,
285
+ 16604,
286
+ 18362,
287
+ 18956,
288
+ 20075,
289
+ 21675,
290
+ 22520,
291
+ 26130,
292
+ 26161,
293
+ 26435,
294
+ 28279,
295
+ 29464,
296
+ 31650,
297
+ 32302,
298
+ 32470,
299
+ 36865,
300
+ 42863,
301
+ 47425,
302
+ 49870,
303
+ 50254,
304
+ 50258,
 
 
 
305
  50358,
306
  50359,
307
  50360,
308
+ 50361,
309
+ 50362
310
  ],
311
+ "task_to_id": {
312
+ "transcribe": 50359,
313
+ "translate": 50358
314
+ },
315
+ "transformers_version": "4.38.2"
316
  }