Hugging Face's logo

zeromodels
/
whisper_medium

Automatic Speech Recognition
KerasFormers
Keras
PyTorch
JAX
TensorFlow
whisper
audio
multilingual
Model card Files
xet
 Community
IMvision12 commited on
Commit
78092d2
·
verified ·
1 Parent(s): 610c32e

Upload kf_config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. kf_config.json +157 -5
kf_config.json CHANGED
@@ -21,6 +21,157 @@
21
  "layer_norm_eps": 1e-05,
22
  "scale_embedding": false,
23
  "generate_args": {
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
24
  "suppress_tokens": [
25
  1,
26
  2,
@@ -105,14 +256,15 @@
105
  49870,
106
  50254,
107
  50258,
 
 
108
  50360,
109
  50361,
110
  50362
111
  ],
112
- "begin_suppress_tokens": [
113
- 220,
114
- 50257
115
- ],
116
- "max_new_tokens": 224
117
  }
118
  }
 
21
  "layer_norm_eps": 1e-05,
22
  "scale_embedding": false,
23
  "generate_args": {
24
+ "alignment_heads": [
25
+ [
26
+ 13,
27
+ 15
28
+ ],
29
+ [
30
+ 15,
31
+ 4
32
+ ],
33
+ [
34
+ 15,
35
+ 15
36
+ ],
37
+ [
38
+ 16,
39
+ 1
40
+ ],
41
+ [
42
+ 20,
43
+ 0
44
+ ],
45
+ [
46
+ 23,
47
+ 4
48
+ ]
49
+ ],
50
+ "begin_suppress_tokens": [
51
+ 220,
52
+ 50257
53
+ ],
54
+ "bos_token_id": 50257,
55
+ "decoder_start_token_id": 50258,
56
+ "eos_token_id": 50257,
57
+ "forced_decoder_ids": [
58
+ [
59
+ 1,
60
+ null
61
+ ],
62
+ [
63
+ 2,
64
+ 50359
65
+ ]
66
+ ],
67
+ "is_multilingual": true,
68
+ "lang_to_id": {
69
+ "<|af|>": 50327,
70
+ "<|am|>": 50334,
71
+ "<|ar|>": 50272,
72
+ "<|as|>": 50350,
73
+ "<|az|>": 50304,
74
+ "<|ba|>": 50355,
75
+ "<|be|>": 50330,
76
+ "<|bg|>": 50292,
77
+ "<|bn|>": 50302,
78
+ "<|bo|>": 50347,
79
+ "<|br|>": 50309,
80
+ "<|bs|>": 50315,
81
+ "<|ca|>": 50270,
82
+ "<|cs|>": 50283,
83
+ "<|cy|>": 50297,
84
+ "<|da|>": 50285,
85
+ "<|de|>": 50261,
86
+ "<|el|>": 50281,
87
+ "<|en|>": 50259,
88
+ "<|es|>": 50262,
89
+ "<|et|>": 50307,
90
+ "<|eu|>": 50310,
91
+ "<|fa|>": 50300,
92
+ "<|fi|>": 50277,
93
+ "<|fo|>": 50338,
94
+ "<|fr|>": 50265,
95
+ "<|gl|>": 50319,
96
+ "<|gu|>": 50333,
97
+ "<|haw|>": 50352,
98
+ "<|ha|>": 50354,
99
+ "<|he|>": 50279,
100
+ "<|hi|>": 50276,
101
+ "<|hr|>": 50291,
102
+ "<|ht|>": 50339,
103
+ "<|hu|>": 50286,
104
+ "<|hy|>": 50312,
105
+ "<|id|>": 50275,
106
+ "<|is|>": 50311,
107
+ "<|it|>": 50274,
108
+ "<|ja|>": 50266,
109
+ "<|jw|>": 50356,
110
+ "<|ka|>": 50329,
111
+ "<|kk|>": 50316,
112
+ "<|km|>": 50323,
113
+ "<|kn|>": 50306,
114
+ "<|ko|>": 50264,
115
+ "<|la|>": 50294,
116
+ "<|lb|>": 50345,
117
+ "<|ln|>": 50353,
118
+ "<|lo|>": 50336,
119
+ "<|lt|>": 50293,
120
+ "<|lv|>": 50301,
121
+ "<|mg|>": 50349,
122
+ "<|mi|>": 50295,
123
+ "<|mk|>": 50308,
124
+ "<|ml|>": 50296,
125
+ "<|mn|>": 50314,
126
+ "<|mr|>": 50320,
127
+ "<|ms|>": 50282,
128
+ "<|mt|>": 50343,
129
+ "<|my|>": 50346,
130
+ "<|ne|>": 50313,
131
+ "<|nl|>": 50271,
132
+ "<|nn|>": 50342,
133
+ "<|no|>": 50288,
134
+ "<|oc|>": 50328,
135
+ "<|pa|>": 50321,
136
+ "<|pl|>": 50269,
137
+ "<|ps|>": 50340,
138
+ "<|pt|>": 50267,
139
+ "<|ro|>": 50284,
140
+ "<|ru|>": 50263,
141
+ "<|sa|>": 50344,
142
+ "<|sd|>": 50332,
143
+ "<|si|>": 50322,
144
+ "<|sk|>": 50298,
145
+ "<|sl|>": 50305,
146
+ "<|sn|>": 50324,
147
+ "<|so|>": 50326,
148
+ "<|sq|>": 50317,
149
+ "<|sr|>": 50303,
150
+ "<|su|>": 50357,
151
+ "<|sv|>": 50273,
152
+ "<|sw|>": 50318,
153
+ "<|ta|>": 50287,
154
+ "<|te|>": 50299,
155
+ "<|tg|>": 50331,
156
+ "<|th|>": 50289,
157
+ "<|tk|>": 50341,
158
+ "<|tl|>": 50348,
159
+ "<|tr|>": 50268,
160
+ "<|tt|>": 50351,
161
+ "<|uk|>": 50280,
162
+ "<|ur|>": 50290,
163
+ "<|uz|>": 50337,
164
+ "<|vi|>": 50278,
165
+ "<|yi|>": 50335,
166
+ "<|yo|>": 50325,
167
+ "<|zh|>": 50260
168
+ },
169
+ "max_initial_timestamp_index": 50,
170
+ "max_length": 448,
171
+ "no_timestamps_token_id": 50363,
172
+ "pad_token_id": 50257,
173
+ "prev_sot_token_id": 50361,
174
+ "return_timestamps": false,
175
  "suppress_tokens": [
176
  1,
177
  2,
 
256
  49870,
257
  50254,
258
  50258,
259
+ 50358,
260
+ 50359,
261
  50360,
262
  50361,
263
  50362
264
  ],
265
+ "task_to_id": {
266
+ "transcribe": 50359,
267
+ "translate": 50358
268
+ }
 
269
  }
270
  }