Hugging Face's logo

mesolitica
/
whisper-base-speakercount

Audio Classification
Transformers
Safetensors
whisper
Inference Endpoints
Model card Files Community
aisyahhrazak commited on
Commit
f6cba33
•
1 Parent(s): 5cf13f1

Upload WhisperForAudioClassification

Browse files
Files changed (2) hide show
  1. config.json +172 -0
  2. model.safetensors +3 -0
config.json ADDED
@@ -0,0 +1,172 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "whisper-base-speakercount/checkpoint-20000",
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": 512,
17
+ "decoder_attention_heads": 8,
18
+ "decoder_ffn_dim": 2048,
19
+ "decoder_layerdrop": 0.0,
20
+ "decoder_layers": 6,
21
+ "decoder_start_token_id": 50258,
22
+ "dropout": 0.0,
23
+ "encoder_attention_heads": 8,
24
+ "encoder_ffn_dim": 2048,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 6,
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": "0 speaker",
44
+ "1": "1 speaker",
45
+ "2": "2 speakers",
46
+ "3": "3 speakers",
47
+ "4": "4 speakers",
48
+ "5": "5 speakers",
49
+ "6": "more than 5 speakers"
50
+ },
51
+ "init_std": 0.02,
52
+ "is_encoder_decoder": true,
53
+ "label2id": {
54
+ "0 speaker": 0,
55
+ "1 speaker": 1,
56
+ "2 speakers": 2,
57
+ "3 speakers": 3,
58
+ "4 speakers": 4,
59
+ "5 speakers": 5,
60
+ "more than 5 speakers": 6
61
+ },
62
+ "mask_feature_length": 10,
63
+ "mask_feature_min_masks": 0,
64
+ "mask_feature_prob": 0.0,
65
+ "mask_time_length": 10,
66
+ "mask_time_min_masks": 2,
67
+ "mask_time_prob": 0.05,
68
+ "max_length": 448,
69
+ "max_source_positions": 1500,
70
+ "max_target_positions": 448,
71
+ "median_filter_width": 7,
72
+ "model_type": "whisper",
73
+ "num_hidden_layers": 6,
74
+ "num_mel_bins": 80,
75
+ "pad_token_id": 50257,
76
+ "scale_embedding": false,
77
+ "suppress_tokens": [
78
+ 1,
79
+ 2,
80
+ 7,
81
+ 8,
82
+ 9,
83
+ 10,
84
+ 14,
85
+ 25,
86
+ 26,
87
+ 27,
88
+ 28,
89
+ 29,
90
+ 31,
91
+ 58,
92
+ 59,
93
+ 60,
94
+ 61,
95
+ 62,
96
+ 63,
97
+ 90,
98
+ 91,
99
+ 92,
100
+ 93,
101
+ 359,
102
+ 503,
103
+ 522,
104
+ 542,
105
+ 873,
106
+ 893,
107
+ 902,
108
+ 918,
109
+ 922,
110
+ 931,
111
+ 1350,
112
+ 1853,
113
+ 1982,
114
+ 2460,
115
+ 2627,
116
+ 3246,
117
+ 3253,
118
+ 3268,
119
+ 3536,
120
+ 3846,
121
+ 3961,
122
+ 4183,
123
+ 4667,
124
+ 6585,
125
+ 6647,
126
+ 7273,
127
+ 9061,
128
+ 9383,
129
+ 10428,
130
+ 10929,
131
+ 11938,
132
+ 12033,
133
+ 12331,
134
+ 12562,
135
+ 13793,
136
+ 14157,
137
+ 14635,
138
+ 15265,
139
+ 15618,
140
+ 16553,
141
+ 16604,
142
+ 18362,
143
+ 18956,
144
+ 20075,
145
+ 21675,
146
+ 22520,
147
+ 26130,
148
+ 26161,
149
+ 26435,
150
+ 28279,
151
+ 29464,
152
+ 31650,
153
+ 32302,
154
+ 32470,
155
+ 36865,
156
+ 42863,
157
+ 47425,
158
+ 49870,
159
+ 50254,
160
+ 50258,
161
+ 50358,
162
+ 50359,
163
+ 50360,
164
+ 50361,
165
+ 50362
166
+ ],
167
+ "torch_dtype": "float32",
168
+ "transformers_version": "4.36.1",
169
+ "use_cache": true,
170
+ "use_weighted_layer_sum": false,
171
+ "vocab_size": 51865
172
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48d3f566cf047257fb8cb7df58c9bebaccfb640eb5a5c9bb3dfd302c7333b748
3
+ size 82905412