Hugging Face's logo

RedHatAI
/
Voxtral-Mini-3B-2507-FP8-dynamic

Automatic Speech Recognition
Safetensors
voxtral
fp8
quantized
multimodal
conversational
text-generation-inference
automatic-speech-translation
audio-text-to-text
video-text-to-text
compressed-tensors
Model card Files
xet
Community
4
Voxtral-Mini-3B-2507-FP8-dynamic
File size: 17,778 Bytes 
02e3d99
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0bdcaf2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
02e3d99
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
 
{
  "dim": 3072,
  "n_layers": 30,
  "head_dim": 128,
  "hidden_dim": 8192,
  "n_heads": 32,
  "n_kv_heads": 8,
  "rope_theta": 100000000.0,
  "norm_eps": 1e-05,
  "vocab_size": 131072,
  "max_position_embeddings": 32768,
  "multimodal": {
    "whisper_model_args": {
      "encoder_args": {
        "dim": 1280,
        "n_layers": 32,
        "head_dim": 64,
        "hidden_dim": 5120,
        "n_heads": 20,
        "vocab_size": 51866,
        "max_source_positions": 1500,
        "audio_encoding_args": {
          "sampling_rate": 16000,
          "num_mel_bins": 128,
          "hop_length": 160,
          "window_size": 400
        }
      },
      "downsample_args": {
        "downsample_factor": 4
      }
    }
  },
  "quantization": {
    "config_groups": {
      "group_0": {
        "input_activations": {
          "actorder": null,
          "block_structure": null,
          "dynamic": true,
          "group_size": null,
          "num_bits": 8,
          "observer": null,
          "observer_kwargs": {},
          "strategy": "token",
          "symmetric": true,
          "type": "float"
        },
        "output_activations": null,
        "targets": [
          "Linear"
        ],
        "weights": {
          "actorder": null,
          "block_structure": null,
          "dynamic": false,
          "group_size": null,
          "num_bits": 8,
          "observer": "minmax",
          "observer_kwargs": {},
          "strategy": "channel",
          "symmetric": true,
          "type": "float"
        }
      }
    },
    "format": "float-quantized",
    "global_compression_ratio": null,
    "ignore": [
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.0.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.0.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.0.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.0.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.0.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.0.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.1.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.1.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.1.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.1.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.1.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.1.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.2.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.2.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.2.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.2.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.2.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.2.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.3.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.3.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.3.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.3.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.3.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.3.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.4.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.4.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.4.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.4.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.4.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.4.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.5.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.5.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.5.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.5.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.5.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.5.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.6.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.6.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.6.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.6.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.6.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.6.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.7.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.7.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.7.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.7.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.7.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.7.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.8.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.8.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.8.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.8.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.8.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.8.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.9.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.9.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.9.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.9.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.9.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.9.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.10.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.10.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.10.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.10.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.10.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.10.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.11.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.11.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.11.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.11.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.11.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.11.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.12.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.12.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.12.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.12.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.12.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.12.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.13.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.13.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.13.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.13.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.13.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.13.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.14.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.14.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.14.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.14.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.14.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.14.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.15.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.15.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.15.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.15.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.15.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.15.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.16.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.16.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.16.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.16.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.16.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.16.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.17.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.17.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.17.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.17.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.17.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.17.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.18.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.18.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.18.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.18.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.18.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.18.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.19.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.19.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.19.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.19.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.19.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.19.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.20.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.20.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.20.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.20.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.20.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.20.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.21.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.21.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.21.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.21.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.21.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.21.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.22.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.22.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.22.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.22.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.22.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.22.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.23.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.23.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.23.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.23.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.23.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.23.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.24.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.24.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.24.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.24.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.24.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.24.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.25.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.25.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.25.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.25.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.25.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.25.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.26.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.26.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.26.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.26.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.26.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.26.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.27.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.27.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.27.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.27.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.27.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.27.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.28.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.28.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.28.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.28.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.28.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.28.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.29.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.29.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.29.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.29.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.29.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.29.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.30.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.30.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.30.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.30.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.30.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.30.feed_forward.w2",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.31.attention.wk",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.31.attention.wv",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.31.attention.wq",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.31.attention.wo",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.31.feed_forward.w1",
      "mm_whisper_embeddings.whisper_encoder.transformer.layers.31.feed_forward.w2",
      "output",
      "mm_whisper_embeddings.audio_language_projection.0",
      "mm_whisper_embeddings.audio_language_projection.2"
    ],
    "kv_cache_scheme": null,
    "quant_method": "compressed-tensors",
    "quantization_status": "compressed"
  }
}