Hugging Face's logo

AngelSlim
/
Qwen2.5-VL-72B-Instruct-GPTQ

Safetensors
qwen2_5_vl
4-bit precision
gptq
Model card Files
xet
Community
woodchen7 commited on
Commit
d915301
·
verified ·
1 Parent(s): c0e6e2c

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +196 -0
config.json ADDED
@@ -0,0 +1,196 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2_5_VLForConditionalGeneration"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151645,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 8192,
10
+ "image_token_id": 151655,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 29568,
13
+ "max_position_embeddings": 128000,
14
+ "max_window_layers": 80,
15
+ "model_type": "qwen2_5_vl",
16
+ "num_attention_heads": 64,
17
+ "num_hidden_layers": 80,
18
+ "num_key_value_heads": 8,
19
+ "quantization_config": {
20
+ "bits": 4,
21
+ "checkpoint_format": "gptq",
22
+ "desc_act": true,
23
+ "group_size": 128,
24
+ "quant_method": "gptq",
25
+ "static_groups": true,
26
+ "sym": true,
27
+ "true_sequential": true
28
+ },
29
+ "rms_norm_eps": 1e-06,
30
+ "rope_scaling": {
31
+ "mrope_section": [
32
+ 16,
33
+ 24,
34
+ 24
35
+ ],
36
+ "rope_type": "default",
37
+ "type": "default"
38
+ },
39
+ "rope_theta": 1000000.0,
40
+ "sliding_window": 32768,
41
+ "text_config": {
42
+ "architectures": [
43
+ "Qwen2_5_VLForConditionalGeneration"
44
+ ],
45
+ "attention_dropout": 0.0,
46
+ "bos_token_id": 151643,
47
+ "eos_token_id": 151645,
48
+ "hidden_act": "silu",
49
+ "hidden_size": 8192,
50
+ "image_token_id": null,
51
+ "initializer_range": 0.02,
52
+ "intermediate_size": 29568,
53
+ "layer_types": [
54
+ "full_attention",
55
+ "full_attention",
56
+ "full_attention",
57
+ "full_attention",
58
+ "full_attention",
59
+ "full_attention",
60
+ "full_attention",
61
+ "full_attention",
62
+ "full_attention",
63
+ "full_attention",
64
+ "full_attention",
65
+ "full_attention",
66
+ "full_attention",
67
+ "full_attention",
68
+ "full_attention",
69
+ "full_attention",
70
+ "full_attention",
71
+ "full_attention",
72
+ "full_attention",
73
+ "full_attention",
74
+ "full_attention",
75
+ "full_attention",
76
+ "full_attention",
77
+ "full_attention",
78
+ "full_attention",
79
+ "full_attention",
80
+ "full_attention",
81
+ "full_attention",
82
+ "full_attention",
83
+ "full_attention",
84
+ "full_attention",
85
+ "full_attention",
86
+ "full_attention",
87
+ "full_attention",
88
+ "full_attention",
89
+ "full_attention",
90
+ "full_attention",
91
+ "full_attention",
92
+ "full_attention",
93
+ "full_attention",
94
+ "full_attention",
95
+ "full_attention",
96
+ "full_attention",
97
+ "full_attention",
98
+ "full_attention",
99
+ "full_attention",
100
+ "full_attention",
101
+ "full_attention",
102
+ "full_attention",
103
+ "full_attention",
104
+ "full_attention",
105
+ "full_attention",
106
+ "full_attention",
107
+ "full_attention",
108
+ "full_attention",
109
+ "full_attention",
110
+ "full_attention",
111
+ "full_attention",
112
+ "full_attention",
113
+ "full_attention",
114
+ "full_attention",
115
+ "full_attention",
116
+ "full_attention",
117
+ "full_attention",
118
+ "full_attention",
119
+ "full_attention",
120
+ "full_attention",
121
+ "full_attention",
122
+ "full_attention",
123
+ "full_attention",
124
+ "full_attention",
125
+ "full_attention",
126
+ "full_attention",
127
+ "full_attention",
128
+ "full_attention",
129
+ "full_attention",
130
+ "full_attention",
131
+ "full_attention",
132
+ "full_attention",
133
+ "full_attention"
134
+ ],
135
+ "max_position_embeddings": 128000,
136
+ "max_window_layers": 80,
137
+ "model_type": "qwen2_5_vl_text",
138
+ "num_attention_heads": 64,
139
+ "num_hidden_layers": 80,
140
+ "num_key_value_heads": 8,
141
+ "rms_norm_eps": 1e-06,
142
+ "rope_scaling": {
143
+ "mrope_section": [
144
+ 16,
145
+ 24,
146
+ 24
147
+ ],
148
+ "rope_type": "default",
149
+ "type": "default"
150
+ },
151
+ "rope_theta": 1000000.0,
152
+ "sliding_window": null,
153
+ "torch_dtype": "bfloat16",
154
+ "use_cache": true,
155
+ "use_sliding_window": false,
156
+ "video_token_id": null,
157
+ "vision_end_token_id": 151653,
158
+ "vision_start_token_id": 151652,
159
+ "vision_token_id": 151654,
160
+ "vocab_size": 152064
161
+ },
162
+ "tie_word_embeddings": false,
163
+ "torch_dtype": "bfloat16",
164
+ "transformers_version": "4.55.0.dev0",
165
+ "use_cache": false,
166
+ "use_sliding_window": false,
167
+ "video_token_id": 151656,
168
+ "vision_config": {
169
+ "depth": 32,
170
+ "fullatt_block_indexes": [
171
+ 7,
172
+ 15,
173
+ 23,
174
+ 31
175
+ ],
176
+ "hidden_act": "silu",
177
+ "hidden_size": 1280,
178
+ "in_channels": 3,
179
+ "in_chans": 3,
180
+ "initializer_range": 0.02,
181
+ "intermediate_size": 3456,
182
+ "model_type": "qwen2_5_vl",
183
+ "num_heads": 16,
184
+ "out_hidden_size": 8192,
185
+ "patch_size": 14,
186
+ "spatial_merge_size": 2,
187
+ "spatial_patch_size": 14,
188
+ "temporal_patch_size": 2,
189
+ "tokens_per_second": 2,
190
+ "window_size": 112
191
+ },
192
+ "vision_end_token_id": 151653,
193
+ "vision_start_token_id": 151652,
194
+ "vision_token_id": 151654,
195
+ "vocab_size": 152064
196
+ }