Hugging Face's logo

hascrack
/
Minivoxtral-3-14B-Reasoning-2512_ASR

Any-to-Any
Safetensors
English
vllm
mistral3
multimodal
trimodal
audio
vision
asr
mistral
whisper
tool-calling
reasoning
4-bit precision
gptq
Model card Files
xet
 Community
hascrack commited on
Commit
f9e0efa
·
verified ·
1 Parent(s): 6653af4

Add config.json at root for download stats tracking

Browse files
Files changed (1) hide show
  1. config.json +114 -0
config.json ADDED
@@ -0,0 +1,114 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "MinivoxtralForConditionalGeneration"
4
+ ],
5
+ "dtype": "bfloat16",
6
+ "image_token_index": 10,
7
+ "model_type": "mistral3",
8
+ "multimodal_projector_bias": false,
9
+ "projector_hidden_act": "gelu",
10
+ "spatial_merge_size": 2,
11
+ "text_config": {
12
+ "attention_dropout": 0.0,
13
+ "head_dim": 128,
14
+ "hidden_act": "silu",
15
+ "hidden_size": 5120,
16
+ "initializer_range": 0.02,
17
+ "intermediate_size": 16384,
18
+ "max_position_embeddings": 262144,
19
+ "model_type": "ministral3",
20
+ "num_attention_heads": 32,
21
+ "num_hidden_layers": 40,
22
+ "num_key_value_heads": 8,
23
+ "rms_norm_eps": 1e-05,
24
+ "rope_parameters": {
25
+ "beta_fast": 32.0,
26
+ "beta_slow": 1.0,
27
+ "factor": 16.0,
28
+ "llama_4_scaling_beta": 0.1,
29
+ "mscale": 1.0,
30
+ "mscale_all_dim": 1.0,
31
+ "original_max_position_embeddings": 16384,
32
+ "rope_theta": 1000000000.0,
33
+ "rope_type": "yarn",
34
+ "type": "yarn"
35
+ },
36
+ "sliding_window": null,
37
+ "use_cache": true,
38
+ "vocab_size": 131072
39
+ },
40
+ "transformers_version": "5.0.0.dev0",
41
+ "vision_config": {
42
+ "attention_dropout": 0.0,
43
+ "head_dim": 64,
44
+ "hidden_act": "silu",
45
+ "hidden_size": 1024,
46
+ "image_size": 1540,
47
+ "initializer_range": 0.02,
48
+ "intermediate_size": 4096,
49
+ "model_type": "pixtral",
50
+ "num_attention_heads": 16,
51
+ "num_channels": 3,
52
+ "num_hidden_layers": 24,
53
+ "patch_size": 14,
54
+ "rope_parameters": {
55
+ "rope_theta": 10000.0,
56
+ "rope_type": "default"
57
+ },
58
+ "rope_theta": 10000.0
59
+ },
60
+ "vision_feature_layer": -1,
61
+ "audio_config": {
62
+ "activation_dropout": 0.0,
63
+ "activation_function": "gelu",
64
+ "attention_dropout": 0.0,
65
+ "dropout": 0.0,
66
+ "head_dim": 64,
67
+ "hidden_size": 1280,
68
+ "initializer_range": 0.02,
69
+ "intermediate_size": 5120,
70
+ "layerdrop": 0.0,
71
+ "max_source_positions": 1500,
72
+ "model_type": "voxtral_encoder",
73
+ "num_attention_heads": 20,
74
+ "num_hidden_layers": 32,
75
+ "num_key_value_heads": 20,
76
+ "num_mel_bins": 128,
77
+ "sampling_rate": 16000,
78
+ "hop_length": 160,
79
+ "window_size": 400,
80
+ "downsample_factor": 4,
81
+ "scale_embedding": false,
82
+ "vocab_size": 51866
83
+ },
84
+ "audio_token_id": 24,
85
+ "audio_downsample_factor": 4,
86
+ "audio_projector_hidden_act": "gelu",
87
+ "audio_multimodal_projector_bias": false,
88
+ "quantization_config": {
89
+ "bits": 4,
90
+ "checkpoint_format": "gptq",
91
+ "desc_act": false,
92
+ "format": "gptq",
93
+ "group_size": 64,
94
+ "lm_head": false,
95
+ "meta": {
96
+ "act_group_aware": true,
97
+ "damp_auto_increment": 0.01,
98
+ "damp_percent": 0.05,
99
+ "gptaq": false,
100
+ "gptaq_alpha": 0.25,
101
+ "mse": 0.0,
102
+ "quantizer": [
103
+ "gptqmodel:5.6.12"
104
+ ],
105
+ "static_groups": false,
106
+ "true_sequential": true,
107
+ "uri": "https://github.com/modelcloud/gptqmodel"
108
+ },
109
+ "pack_dtype": "int32",
110
+ "pack_impl": "cpu",
111
+ "quant_method": "gptq",
112
+ "sym": true
113
+ }
114
+ }