Hugging Face's logo

octanove
/
mosla-whisper-langid-zho

Audio Classification
Transformers
PyTorch
whisper
Inference Endpoints
Model card Files Community
Jotanner commited on
Commit
707ed27
•
1 Parent(s): fac03b6

Upload WhisperForAudioClassification

Browse files
Files changed (2) hide show
  1. config.json +161 -0
  2. pytorch_model.bin +3 -0
config.json ADDED
@@ -0,0 +1,161 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "whisper-lang-id-zho/checkpoint-876",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
+ "architectures": [
7
+ "WhisperForAudioClassification"
8
+ ],
9
+ "attention_dropout": 0.0,
10
+ "begin_suppress_tokens": [
11
+ 220,
12
+ 50257
13
+ ],
14
+ "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
+ "d_model": 1280,
17
+ "decoder_attention_heads": 20,
18
+ "decoder_ffn_dim": 5120,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 32,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 20,
24
+ "encoder_ffn_dim": 5120,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 32,
27
+ "eos_token_id": 50257,
28
+ "forced_decoder_ids": [
29
+ [
30
+ 1,
31
+ 50259
32
+ ],
33
+ [
34
+ 2,
35
+ 50359
36
+ ],
37
+ [
38
+ 3,
39
+ 50363
40
+ ]
41
+ ],
42
+ "id2label": {
43
+ "0": "eng",
44
+ "1": "zho"
45
+ },
46
+ "init_std": 0.02,
47
+ "is_encoder_decoder": true,
48
+ "label2id": {
49
+ "eng": 0,
50
+ "zho": 1
51
+ },
52
+ "mask_feature_length": 10,
53
+ "mask_feature_min_masks": 0,
54
+ "mask_feature_prob": 0.0,
55
+ "mask_time_length": 10,
56
+ "mask_time_min_masks": 2,
57
+ "mask_time_prob": 0.05,
58
+ "max_length": 448,
59
+ "max_source_positions": 1500,
60
+ "max_target_positions": 448,
61
+ "model_type": "whisper",
62
+ "num_hidden_layers": 32,
63
+ "num_mel_bins": 80,
64
+ "pad_token_id": 50257,
65
+ "scale_embedding": false,
66
+ "suppress_tokens": [
67
+ 1,
68
+ 2,
69
+ 7,
70
+ 8,
71
+ 9,
72
+ 10,
73
+ 14,
74
+ 25,
75
+ 26,
76
+ 27,
77
+ 28,
78
+ 29,
79
+ 31,
80
+ 58,
81
+ 59,
82
+ 60,
83
+ 61,
84
+ 62,
85
+ 63,
86
+ 90,
87
+ 91,
88
+ 92,
89
+ 93,
90
+ 359,
91
+ 503,
92
+ 522,
93
+ 542,
94
+ 873,
95
+ 893,
96
+ 902,
97
+ 918,
98
+ 922,
99
+ 931,
100
+ 1350,
101
+ 1853,
102
+ 1982,
103
+ 2460,
104
+ 2627,
105
+ 3246,
106
+ 3253,
107
+ 3268,
108
+ 3536,
109
+ 3846,
110
+ 3961,
111
+ 4183,
112
+ 4667,
113
+ 6585,
114
+ 6647,
115
+ 7273,
116
+ 9061,
117
+ 9383,
118
+ 10428,
119
+ 10929,
120
+ 11938,
121
+ 12033,
122
+ 12331,
123
+ 12562,
124
+ 13793,
125
+ 14157,
126
+ 14635,
127
+ 15265,
128
+ 15618,
129
+ 16553,
130
+ 16604,
131
+ 18362,
132
+ 18956,
133
+ 20075,
134
+ 21675,
135
+ 22520,
136
+ 26130,
137
+ 26161,
138
+ 26435,
139
+ 28279,
140
+ 29464,
141
+ 31650,
142
+ 32302,
143
+ 32470,
144
+ 36865,
145
+ 42863,
146
+ 47425,
147
+ 49870,
148
+ 50254,
149
+ 50258,
150
+ 50358,
151
+ 50359,
152
+ 50360,
153
+ 50361,
154
+ 50362
155
+ ],
156
+ "torch_dtype": "float32",
157
+ "transformers_version": "4.27.4",
158
+ "use_cache": false,
159
+ "use_weighted_layer_sum": false,
160
+ "vocab_size": 51865
161
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ee18167609e3d3a54ede90ebcbfc065b04e949534b45caa8e1cdd55b1c8f85d
3
+ size 2548605483