Hugging Face's logo

cpatonn
/
Qwen3-VL-32B-Thinking-AWQ-8bit

Image-Text-to-Text
Transformers
Safetensors
qwen3_vl
image-to-text
conversational
compressed-tensors
Model card Files
xet
Community
cpatonn commited on
Commit
2f1184d
·
verified ·
1 Parent(s): 550fb89

Upload folder using huggingface_hub

Browse files
Files changed (21) hide show
  1. .gitattributes +1 -0
  2. added_tokens.json +28 -0
  3. chat_template.jinja +110 -0
  4. config.json +214 -0
  5. generation_config.json +13 -0
  6. merges.txt +0 -0
  7. model-00001-of-00008.safetensors +3 -0
  8. model-00002-of-00008.safetensors +3 -0
  9. model-00003-of-00008.safetensors +3 -0
  10. model-00004-of-00008.safetensors +3 -0
  11. model-00005-of-00008.safetensors +3 -0
  12. model-00006-of-00008.safetensors +3 -0
  13. model-00007-of-00008.safetensors +3 -0
  14. model-00008-of-00008.safetensors +3 -0
  15. model.safetensors.index.json +0 -0
  16. processor_config.json +84 -0
  17. recipe.yaml +34 -0
  18. special_tokens_map.json +31 -0
  19. tokenizer.json +3 -0
  20. tokenizer_config.json +240 -0
  21. vocab.json +0 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
added_tokens.json ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</think>": 151668,
3
+ "</tool_call>": 151658,
4
+ "</tool_response>": 151666,
5
+ "<think>": 151667,
6
+ "<tool_call>": 151657,
7
+ "<tool_response>": 151665,
8
+ "<|box_end|>": 151649,
9
+ "<|box_start|>": 151648,
10
+ "<|endoftext|>": 151643,
11
+ "<|file_sep|>": 151664,
12
+ "<|fim_middle|>": 151660,
13
+ "<|fim_pad|>": 151662,
14
+ "<|fim_prefix|>": 151659,
15
+ "<|fim_suffix|>": 151661,
16
+ "<|im_end|>": 151645,
17
+ "<|im_start|>": 151644,
18
+ "<|image_pad|>": 151655,
19
+ "<|object_ref_end|>": 151647,
20
+ "<|object_ref_start|>": 151646,
21
+ "<|quad_end|>": 151651,
22
+ "<|quad_start|>": 151650,
23
+ "<|repo_name|>": 151663,
24
+ "<|video_pad|>": 151656,
25
+ "<|vision_end|>": 151653,
26
+ "<|vision_pad|>": 151654,
27
+ "<|vision_start|>": 151652
28
+ }
chat_template.jinja ADDED
@@ -0,0 +1,110 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- set image_count = namespace(value=0) %}
2
+ {%- set video_count = namespace(value=0) %}
3
+ {%- macro render_content(content, do_vision_count) %}
4
+ {%- if content is string %}
5
+ {{- content }}
6
+ {%- else %}
7
+ {%- for item in content %}
8
+ {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
9
+ {%- if do_vision_count %}
10
+ {%- set image_count.value = image_count.value + 1 %}
11
+ {%- endif %}
12
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
13
+ <|vision_start|><|image_pad|><|vision_end|>
14
+ {%- elif 'video' in item or item.type == 'video' %}
15
+ {%- if do_vision_count %}
16
+ {%- set video_count.value = video_count.value + 1 %}
17
+ {%- endif %}
18
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
19
+ <|vision_start|><|video_pad|><|vision_end|>
20
+ {%- elif 'text' in item %}
21
+ {{- item.text }}
22
+ {%- endif %}
23
+ {%- endfor %}
24
+ {%- endif %}
25
+ {%- endmacro %}
26
+ {%- if tools %}
27
+ {{- '<|im_start|>system\n' }}
28
+ {%- if messages[0].role == 'system' %}
29
+ {{- render_content(messages[0].content, false) + '\n\n' }}
30
+ {%- endif %}
31
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
32
+ {%- for tool in tools %}
33
+ {{- "\n" }}
34
+ {{- tool | tojson }}
35
+ {%- endfor %}
36
+ {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
37
+ {%- else %}
38
+ {%- if messages[0].role == 'system' %}
39
+ {{- '<|im_start|>system\n' + render_content(messages[0].content, false) + '<|im_end|>\n' }}
40
+ {%- endif %}
41
+ {%- endif %}
42
+ {%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
43
+ {%- for message in messages[::-1] %}
44
+ {%- set index = (messages|length - 1) - loop.index0 %}
45
+ {%- if ns.multi_step_tool and message.role == "user" %}
46
+ {%- set content = render_content(message.content, false) %}
47
+ {%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
48
+ {%- set ns.multi_step_tool = false %}
49
+ {%- set ns.last_query_index = index %}
50
+ {%- endif %}
51
+ {%- endif %}
52
+ {%- endfor %}
53
+ {%- for message in messages %}
54
+ {%- set content = render_content(message.content, True) %}
55
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
56
+ {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
57
+ {%- elif message.role == "assistant" %}
58
+ {%- set reasoning_content = '' %}
59
+ {%- if message.reasoning_content is string %}
60
+ {%- set reasoning_content = message.reasoning_content %}
61
+ {%- else %}
62
+ {%- if '</think>' in content %}
63
+ {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
64
+ {%- set content = content.split('</think>')[-1].lstrip('\n') %}
65
+ {%- endif %}
66
+ {%- endif %}
67
+ {%- if loop.index0 > ns.last_query_index %}
68
+ {%- if loop.last or (not loop.last and reasoning_content) %}
69
+ {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
70
+ {%- else %}
71
+ {{- '<|im_start|>' + message.role + '\n' + content }}
72
+ {%- endif %}
73
+ {%- else %}
74
+ {{- '<|im_start|>' + message.role + '\n' + content }}
75
+ {%- endif %}
76
+ {%- if message.tool_calls %}
77
+ {%- for tool_call in message.tool_calls %}
78
+ {%- if (loop.first and content) or (not loop.first) %}
79
+ {{- '\n' }}
80
+ {%- endif %}
81
+ {%- if tool_call.function %}
82
+ {%- set tool_call = tool_call.function %}
83
+ {%- endif %}
84
+ {{- '<tool_call>\n{"name": "' }}
85
+ {{- tool_call.name }}
86
+ {{- '", "arguments": ' }}
87
+ {%- if tool_call.arguments is string %}
88
+ {{- tool_call.arguments }}
89
+ {%- else %}
90
+ {{- tool_call.arguments | tojson }}
91
+ {%- endif %}
92
+ {{- '}\n</tool_call>' }}
93
+ {%- endfor %}
94
+ {%- endif %}
95
+ {{- '<|im_end|>\n' }}
96
+ {%- elif message.role == "tool" %}
97
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
98
+ {{- '<|im_start|>user' }}
99
+ {%- endif %}
100
+ {{- '\n<tool_response>\n' }}
101
+ {{- content }}
102
+ {{- '\n</tool_response>' }}
103
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
104
+ {{- '<|im_end|>\n' }}
105
+ {%- endif %}
106
+ {%- endif %}
107
+ {%- endfor %}
108
+ {%- if add_generation_prompt %}
109
+ {{- '<|im_start|>assistant\n<think>\n' }}
110
+ {%- endif %}
config.json ADDED
@@ -0,0 +1,214 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3VLForConditionalGeneration"
4
+ ],
5
+ "dtype": "bfloat16",
6
+ "image_token_id": 151655,
7
+ "model_type": "qwen3_vl",
8
+ "quantization_config": {
9
+ "config_groups": {
10
+ "group_0": {
11
+ "format": "pack-quantized",
12
+ "input_activations": null,
13
+ "output_activations": null,
14
+ "targets": [
15
+ "Linear"
16
+ ],
17
+ "weights": {
18
+ "actorder": null,
19
+ "block_structure": null,
20
+ "dynamic": false,
21
+ "group_size": 32,
22
+ "num_bits": 8,
23
+ "observer": "minmax",
24
+ "observer_kwargs": {},
25
+ "strategy": "group",
26
+ "symmetric": true,
27
+ "type": "int"
28
+ }
29
+ }
30
+ },
31
+ "format": "pack-quantized",
32
+ "global_compression_ratio": null,
33
+ "ignore": [
34
+ "model.visual.blocks.0.attn.qkv",
35
+ "model.visual.blocks.0.attn.proj",
36
+ "model.visual.blocks.0.mlp.linear_fc1",
37
+ "model.visual.blocks.0.mlp.linear_fc2",
38
+ "model.visual.blocks.1.attn.qkv",
39
+ "model.visual.blocks.1.attn.proj",
40
+ "model.visual.blocks.1.mlp.linear_fc1",
41
+ "model.visual.blocks.1.mlp.linear_fc2",
42
+ "model.visual.blocks.2.attn.qkv",
43
+ "model.visual.blocks.2.attn.proj",
44
+ "model.visual.blocks.2.mlp.linear_fc1",
45
+ "model.visual.blocks.2.mlp.linear_fc2",
46
+ "model.visual.blocks.3.attn.qkv",
47
+ "model.visual.blocks.3.attn.proj",
48
+ "model.visual.blocks.3.mlp.linear_fc1",
49
+ "model.visual.blocks.3.mlp.linear_fc2",
50
+ "model.visual.blocks.4.attn.qkv",
51
+ "model.visual.blocks.4.attn.proj",
52
+ "model.visual.blocks.4.mlp.linear_fc1",
53
+ "model.visual.blocks.4.mlp.linear_fc2",
54
+ "model.visual.blocks.5.attn.qkv",
55
+ "model.visual.blocks.5.attn.proj",
56
+ "model.visual.blocks.5.mlp.linear_fc1",
57
+ "model.visual.blocks.5.mlp.linear_fc2",
58
+ "model.visual.blocks.6.attn.qkv",
59
+ "model.visual.blocks.6.attn.proj",
60
+ "model.visual.blocks.6.mlp.linear_fc1",
61
+ "model.visual.blocks.6.mlp.linear_fc2",
62
+ "model.visual.blocks.7.attn.qkv",
63
+ "model.visual.blocks.7.attn.proj",
64
+ "model.visual.blocks.7.mlp.linear_fc1",
65
+ "model.visual.blocks.7.mlp.linear_fc2",
66
+ "model.visual.blocks.8.attn.qkv",
67
+ "model.visual.blocks.8.attn.proj",
68
+ "model.visual.blocks.8.mlp.linear_fc1",
69
+ "model.visual.blocks.8.mlp.linear_fc2",
70
+ "model.visual.blocks.9.attn.qkv",
71
+ "model.visual.blocks.9.attn.proj",
72
+ "model.visual.blocks.9.mlp.linear_fc1",
73
+ "model.visual.blocks.9.mlp.linear_fc2",
74
+ "model.visual.blocks.10.attn.qkv",
75
+ "model.visual.blocks.10.attn.proj",
76
+ "model.visual.blocks.10.mlp.linear_fc1",
77
+ "model.visual.blocks.10.mlp.linear_fc2",
78
+ "model.visual.blocks.11.attn.qkv",
79
+ "model.visual.blocks.11.attn.proj",
80
+ "model.visual.blocks.11.mlp.linear_fc1",
81
+ "model.visual.blocks.11.mlp.linear_fc2",
82
+ "model.visual.blocks.12.attn.qkv",
83
+ "model.visual.blocks.12.attn.proj",
84
+ "model.visual.blocks.12.mlp.linear_fc1",
85
+ "model.visual.blocks.12.mlp.linear_fc2",
86
+ "model.visual.blocks.13.attn.qkv",
87
+ "model.visual.blocks.13.attn.proj",
88
+ "model.visual.blocks.13.mlp.linear_fc1",
89
+ "model.visual.blocks.13.mlp.linear_fc2",
90
+ "model.visual.blocks.14.attn.qkv",
91
+ "model.visual.blocks.14.attn.proj",
92
+ "model.visual.blocks.14.mlp.linear_fc1",
93
+ "model.visual.blocks.14.mlp.linear_fc2",
94
+ "model.visual.blocks.15.attn.qkv",
95
+ "model.visual.blocks.15.attn.proj",
96
+ "model.visual.blocks.15.mlp.linear_fc1",
97
+ "model.visual.blocks.15.mlp.linear_fc2",
98
+ "model.visual.blocks.16.attn.qkv",
99
+ "model.visual.blocks.16.attn.proj",
100
+ "model.visual.blocks.16.mlp.linear_fc1",
101
+ "model.visual.blocks.16.mlp.linear_fc2",
102
+ "model.visual.blocks.17.attn.qkv",
103
+ "model.visual.blocks.17.attn.proj",
104
+ "model.visual.blocks.17.mlp.linear_fc1",
105
+ "model.visual.blocks.17.mlp.linear_fc2",
106
+ "model.visual.blocks.18.attn.qkv",
107
+ "model.visual.blocks.18.attn.proj",
108
+ "model.visual.blocks.18.mlp.linear_fc1",
109
+ "model.visual.blocks.18.mlp.linear_fc2",
110
+ "model.visual.blocks.19.attn.qkv",
111
+ "model.visual.blocks.19.attn.proj",
112
+ "model.visual.blocks.19.mlp.linear_fc1",
113
+ "model.visual.blocks.19.mlp.linear_fc2",
114
+ "model.visual.blocks.20.attn.qkv",
115
+ "model.visual.blocks.20.attn.proj",
116
+ "model.visual.blocks.20.mlp.linear_fc1",
117
+ "model.visual.blocks.20.mlp.linear_fc2",
118
+ "model.visual.blocks.21.attn.qkv",
119
+ "model.visual.blocks.21.attn.proj",
120
+ "model.visual.blocks.21.mlp.linear_fc1",
121
+ "model.visual.blocks.21.mlp.linear_fc2",
122
+ "model.visual.blocks.22.attn.qkv",
123
+ "model.visual.blocks.22.attn.proj",
124
+ "model.visual.blocks.22.mlp.linear_fc1",
125
+ "model.visual.blocks.22.mlp.linear_fc2",
126
+ "model.visual.blocks.23.attn.qkv",
127
+ "model.visual.blocks.23.attn.proj",
128
+ "model.visual.blocks.23.mlp.linear_fc1",
129
+ "model.visual.blocks.23.mlp.linear_fc2",
130
+ "model.visual.blocks.24.attn.qkv",
131
+ "model.visual.blocks.24.attn.proj",
132
+ "model.visual.blocks.24.mlp.linear_fc1",
133
+ "model.visual.blocks.24.mlp.linear_fc2",
134
+ "model.visual.blocks.25.attn.qkv",
135
+ "model.visual.blocks.25.attn.proj",
136
+ "model.visual.blocks.25.mlp.linear_fc1",
137
+ "model.visual.blocks.25.mlp.linear_fc2",
138
+ "model.visual.blocks.26.attn.qkv",
139
+ "model.visual.blocks.26.attn.proj",
140
+ "model.visual.blocks.26.mlp.linear_fc1",
141
+ "model.visual.blocks.26.mlp.linear_fc2",
142
+ "model.visual.merger.linear_fc1",
143
+ "model.visual.merger.linear_fc2",
144
+ "model.visual.deepstack_merger_list.0.linear_fc1",
145
+ "model.visual.deepstack_merger_list.0.linear_fc2",
146
+ "model.visual.deepstack_merger_list.1.linear_fc1",
147
+ "model.visual.deepstack_merger_list.1.linear_fc2",
148
+ "model.visual.deepstack_merger_list.2.linear_fc1",
149
+ "model.visual.deepstack_merger_list.2.linear_fc2",
150
+ "lm_head"
151
+ ],
152
+ "kv_cache_scheme": null,
153
+ "quant_method": "compressed-tensors",
154
+ "quantization_status": "compressed",
155
+ "sparsity_config": {},
156
+ "transform_config": {},
157
+ "version": "0.12.1.dev0"
158
+ },
159
+ "text_config": {
160
+ "attention_bias": false,
161
+ "attention_dropout": 0.0,
162
+ "bos_token_id": 151643,
163
+ "dtype": "bfloat16",
164
+ "eos_token_id": 151645,
165
+ "head_dim": 128,
166
+ "hidden_act": "silu",
167
+ "hidden_size": 5120,
168
+ "initializer_range": 0.02,
169
+ "intermediate_size": 25600,
170
+ "max_position_embeddings": 262144,
171
+ "model_type": "qwen3_vl_text",
172
+ "num_attention_heads": 64,
173
+ "num_hidden_layers": 64,
174
+ "num_key_value_heads": 8,
175
+ "rms_norm_eps": 1e-06,
176
+ "rope_scaling": {
177
+ "mrope_interleaved": true,
178
+ "mrope_section": [
179
+ 24,
180
+ 20,
181
+ 20
182
+ ],
183
+ "rope_type": "default"
184
+ },
185
+ "rope_theta": 5000000,
186
+ "use_cache": true,
187
+ "vocab_size": 151936
188
+ },
189
+ "tie_word_embeddings": false,
190
+ "transformers_version": "5.0.0.dev0",
191
+ "video_token_id": 151656,
192
+ "vision_config": {
193
+ "deepstack_visual_indexes": [
194
+ 8,
195
+ 16,
196
+ 24
197
+ ],
198
+ "depth": 27,
199
+ "hidden_act": "gelu_pytorch_tanh",
200
+ "hidden_size": 1152,
201
+ "in_channels": 3,
202
+ "initializer_range": 0.02,
203
+ "intermediate_size": 4304,
204
+ "model_type": "qwen3_vl",
205
+ "num_heads": 16,
206
+ "num_position_embeddings": 2304,
207
+ "out_hidden_size": 5120,
208
+ "patch_size": 16,
209
+ "spatial_merge_size": 2,
210
+ "temporal_patch_size": 2
211
+ },
212
+ "vision_end_token_id": 151653,
213
+ "vision_start_token_id": 151652
214
+ }
generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "temperature": 0.8,
10
+ "top_k": 20,
11
+ "top_p": 0.95,
12
+ "transformers_version": "5.0.0.dev0"
13
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model-00001-of-00008.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f48736b9cf38e1f95e8884e38fed5c9ddefe5a4747d72a416365651882cb2d93
3
+ size 4919016576
model-00002-of-00008.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3298fd7e542a2a6d6d30637b57cec83f7b95020381b64768ffb83f0f78a34ae5
3
+ size 4941306472
model-00003-of-00008.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0da5cfb5c09f47a29f45a81add8b87c27674bcdfdcbeda825d5da5c51aafc61c
3
+ size 4902335448
model-00004-of-00008.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9208e7ddf3e6270c6e0c88199dcb4fd1e4ba472483682373fa960de56b66d8d
3
+ size 4941306616
model-00005-of-00008.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5aa5eb19cee08e9b6c1b2d222ab53938b69f7d48f38874956008fbd384c1cc09
3
+ size 4902335448
model-00006-of-00008.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfcead3d9187c3b70ecc6ed3908c8f952738db29fc9c05fadcf86b5cad3d04f2
3
+ size 4941306616
model-00007-of-00008.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b49abb76489b2a02cd73145a37bfcc24dd50c70c0c405f7a2cb487ac628a9b0a
3
+ size 4902335448
model-00008-of-00008.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b2ac9b3edf65c3d8273580d508ccb283f61f8e03800050b9c4208c23838261e
3
+ size 3009821744
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
processor_config.json ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "crop_size": null,
4
+ "data_format": "channels_first",
5
+ "device": null,
6
+ "disable_grouping": null,
7
+ "do_center_crop": null,
8
+ "do_convert_rgb": true,
9
+ "do_normalize": true,
10
+ "do_pad": null,
11
+ "do_rescale": true,
12
+ "do_resize": true,
13
+ "image_mean": [
14
+ 0.5,
15
+ 0.5,
16
+ 0.5
17
+ ],
18
+ "image_processor_type": "Qwen2VLImageProcessorFast",
19
+ "image_std": [
20
+ 0.5,
21
+ 0.5,
22
+ 0.5
23
+ ],
24
+ "input_data_format": null,
25
+ "max_pixels": null,
26
+ "merge_size": 2,
27
+ "min_pixels": null,
28
+ "pad_size": null,
29
+ "patch_size": 16,
30
+ "processor_class": "Qwen3VLProcessor",
31
+ "resample": 3,
32
+ "rescale_factor": 0.00392156862745098,
33
+ "return_tensors": null,
34
+ "size": {
35
+ "longest_edge": 16777216,
36
+ "shortest_edge": 65536
37
+ },
38
+ "temporal_patch_size": 2
39
+ },
40
+ "processor_class": "Qwen3VLProcessor",
41
+ "video_processor": {
42
+ "crop_size": null,
43
+ "data_format": "channels_first",
44
+ "default_to_square": true,
45
+ "device": null,
46
+ "do_center_crop": null,
47
+ "do_convert_rgb": true,
48
+ "do_normalize": true,
49
+ "do_pad": null,
50
+ "do_rescale": true,
51
+ "do_resize": true,
52
+ "do_sample_frames": true,
53
+ "fps": 2,
54
+ "image_mean": [
55
+ 0.5,
56
+ 0.5,
57
+ 0.5
58
+ ],
59
+ "image_std": [
60
+ 0.5,
61
+ 0.5,
62
+ 0.5
63
+ ],
64
+ "input_data_format": null,
65
+ "max_frames": 768,
66
+ "merge_size": 2,
67
+ "min_frames": 4,
68
+ "num_frames": null,
69
+ "pad_size": null,
70
+ "patch_size": 16,
71
+ "processor_class": "Qwen3VLProcessor",
72
+ "resample": 3,
73
+ "rescale_factor": 0.00392156862745098,
74
+ "return_metadata": false,
75
+ "return_tensors": null,
76
+ "size": {
77
+ "longest_edge": 25165824,
78
+ "shortest_edge": 4096
79
+ },
80
+ "temporal_patch_size": 2,
81
+ "video_metadata": null,
82
+ "video_processor_type": "Qwen3VLVideoProcessor"
83
+ }
84
+ }
recipe.yaml ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ default_stage:
2
+ default_modifiers:
3
+ AWQModifier:
4
+ config_groups:
5
+ group_0:
6
+ targets: [Linear]
7
+ weights:
8
+ num_bits: 8
9
+ type: int
10
+ symmetric: true
11
+ group_size: 32
12
+ strategy: group
13
+ block_structure: null
14
+ dynamic: false
15
+ actorder: null
16
+ observer: minmax
17
+ observer_kwargs: {}
18
+ input_activations: null
19
+ output_activations: null
20
+ format: null
21
+ targets: [Linear]
22
+ ignore: ['re:.*embed_tokens', 're:.*input_layernorm$', 're:.*mlp[.]gate$', 're:.*post_attention_layernorm$',
23
+ 're:.*norm$', 're:model[.]visual.*', lm_head]
24
+ mappings:
25
+ - smooth_layer: re:.*input_layernorm$
26
+ balance_layers: ['re:.*q_proj$', 're:.*k_proj$', 're:.*v_proj$']
27
+ - smooth_layer: re:.*v_proj$
28
+ balance_layers: ['re:.*o_proj$']
29
+ - smooth_layer: re:.*post_attention_layernorm$
30
+ balance_layers: ['re:.*gate_proj$', 're:.*up_proj$']
31
+ - smooth_layer: re:.*up_proj$
32
+ balance_layers: ['re:.*down_proj$']
33
+ offload_device: !!python/object/apply:torch.device [cpu]
34
+ duo_scaling: true
special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|endoftext|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
3
+ size 11422654
tokenizer_config.json ADDED
@@ -0,0 +1,240 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ },
181
+ "151665": {
182
+ "content": "<tool_response>",
183
+ "lstrip": false,
184
+ "normalized": false,
185
+ "rstrip": false,
186
+ "single_word": false,
187
+ "special": false
188
+ },
189
+ "151666": {
190
+ "content": "</tool_response>",
191
+ "lstrip": false,
192
+ "normalized": false,
193
+ "rstrip": false,
194
+ "single_word": false,
195
+ "special": false
196
+ },
197
+ "151667": {
198
+ "content": "<think>",
199
+ "lstrip": false,
200
+ "normalized": false,
201
+ "rstrip": false,
202
+ "single_word": false,
203
+ "special": false
204
+ },
205
+ "151668": {
206
+ "content": "</think>",
207
+ "lstrip": false,
208
+ "normalized": false,
209
+ "rstrip": false,
210
+ "single_word": false,
211
+ "special": false
212
+ }
213
+ },
214
+ "additional_special_tokens": [
215
+ "<|im_start|>",
216
+ "<|im_end|>",
217
+ "<|object_ref_start|>",
218
+ "<|object_ref_end|>",
219
+ "<|box_start|>",
220
+ "<|box_end|>",
221
+ "<|quad_start|>",
222
+ "<|quad_end|>",
223
+ "<|vision_start|>",
224
+ "<|vision_end|>",
225
+ "<|vision_pad|>",
226
+ "<|image_pad|>",
227
+ "<|video_pad|>"
228
+ ],
229
+ "bos_token": null,
230
+ "clean_up_tokenization_spaces": false,
231
+ "eos_token": "<|im_end|>",
232
+ "errors": "replace",
233
+ "extra_special_tokens": {},
234
+ "model_max_length": 262144,
235
+ "pad_token": "<|endoftext|>",
236
+ "processor_class": "Qwen3VLProcessor",
237
+ "split_special_tokens": false,
238
+ "tokenizer_class": "Qwen2Tokenizer",
239
+ "unk_token": null
240
+ }
vocab.json ADDED
The diff for this file is too large to render. See raw diff