Hugging Face's logo

LeaderboardModel1
/
Qwen3.5-27B-Writer-V2-AutoRound-W4A16-Tuning

Text Generation
Safetensors
qwen3_5
quantized
w4a16
tuning
low-bit-open-llm-leaderboard
conversational
4-bit precision
auto-round
Model card Files
xet
 Community
INC4AI commited on
Commit
d28b5ff
·
verified ·
1 Parent(s): e7f399a

Upload quantized model Qwen3.5-27B-Writer-V2-AutoRound-W4A16-Tuning

Browse files
Files changed (16) hide show
  1. .gitattributes +1 -0
  2. README.md +178 -0
  3. chat_template.jinja +154 -0
  4. config.json +538 -0
  5. generation_config.json +13 -0
  6. model-00001-of-00005.safetensors +3 -0
  7. model-00002-of-00005.safetensors +3 -0
  8. model-00003-of-00005.safetensors +3 -0
  9. model-00004-of-00005.safetensors +3 -0
  10. model-00005-of-00005.safetensors +3 -0
  11. model.safetensors.index.json +0 -0
  12. preprocessor_config.json +27 -0
  13. processor_config.json +63 -0
  14. quantization_config.json +397 -0
  15. tokenizer.json +3 -0
  16. tokenizer_config.json +33 -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
README.md ADDED
@@ -0,0 +1,178 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model:
3
+ - ConicCat/Qwen3.5-27B-Writer-V2
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - quantized
7
+ - w4a16
8
+ - tuning
9
+ - low-bit-open-llm-leaderboard
10
+ ---
11
+
12
+ # Qwen3.5-27B-Writer-V2-AutoRound-W4A16-Tuning
13
+
14
+ ## Model Details
15
+
16
+ This model is a int4 weight-only quantization with group_size 128 and symmetric quantization of [ConicCat/Qwen3.5-27B-Writer-V2](https://huggingface.co/ConicCat/Qwen3.5-27B-Writer-V2) generated by TUNING. Please follow the license of the original model.
17
+
18
+ ## Quantization Details
19
+
20
+ | Attribute | Value |
21
+ |-----------|-------|
22
+ | Base Model | [ConicCat/Qwen3.5-27B-Writer-V2](https://huggingface.co/ConicCat/Qwen3.5-27B-Writer-V2) |
23
+ | Quantization Tool | TUNING |
24
+ | Quantization Scheme | W4A16 |
25
+ | Quantized Size | 17832 MB |
26
+
27
+ ## Evaluation Results
28
+
29
+ | Task | Accuracy |
30
+ |------|----------|
31
+ | hellaswag | 0.6712 |
32
+ | mmlu | 0.8552 |
33
+ | mmlu_abstract_algebra | 0.7700 |
34
+ | mmlu_anatomy | 0.8296 |
35
+ | mmlu_astronomy | 0.9605 |
36
+ | mmlu_business_ethics | 0.8500 |
37
+ | mmlu_clinical_knowledge | 0.9057 |
38
+ | mmlu_college_biology | 0.9653 |
39
+ | mmlu_college_chemistry | 0.6900 |
40
+ | mmlu_college_computer_science | 0.8400 |
41
+ | mmlu_college_mathematics | 0.7200 |
42
+ | mmlu_college_medicine | 0.8439 |
43
+ | mmlu_college_physics | 0.7941 |
44
+ | mmlu_computer_security | 0.8700 |
45
+ | mmlu_conceptual_physics | 0.9447 |
46
+ | mmlu_econometrics | 0.8246 |
47
+ | mmlu_electrical_engineering | 0.8483 |
48
+ | mmlu_elementary_mathematics | 0.9127 |
49
+ | mmlu_formal_logic | 0.7381 |
50
+ | mmlu_global_facts | 0.6200 |
51
+ | mmlu_high_school_biology | 0.9484 |
52
+ | mmlu_high_school_chemistry | 0.8571 |
53
+ | mmlu_high_school_computer_science | 0.9300 |
54
+ | mmlu_high_school_european_history | 0.8848 |
55
+ | mmlu_high_school_geography | 0.9545 |
56
+ | mmlu_high_school_government_and_politics | 0.9845 |
57
+ | mmlu_high_school_macroeconomics | 0.9231 |
58
+ | mmlu_high_school_mathematics | 0.6593 |
59
+ | mmlu_high_school_microeconomics | 0.9706 |
60
+ | mmlu_high_school_physics | 0.8344 |
61
+ | mmlu_high_school_psychology | 0.9523 |
62
+ | mmlu_high_school_statistics | 0.8657 |
63
+ | mmlu_high_school_us_history | 0.9510 |
64
+ | mmlu_high_school_world_history | 0.9409 |
65
+ | mmlu_human_aging | 0.8475 |
66
+ | mmlu_human_sexuality | 0.9160 |
67
+ | mmlu_humanities | 0.8028 |
68
+ | mmlu_international_law | 0.9339 |
69
+ | mmlu_jurisprudence | 0.9167 |
70
+ | mmlu_logical_fallacies | 0.9080 |
71
+ | mmlu_machine_learning | 0.7946 |
72
+ | mmlu_management | 0.8932 |
73
+ | mmlu_marketing | 0.9573 |
74
+ | mmlu_medical_genetics | 0.9600 |
75
+ | mmlu_miscellaneous | 0.9387 |
76
+ | mmlu_moral_disputes | 0.8584 |
77
+ | mmlu_moral_scenarios | 0.7374 |
78
+ | mmlu_nutrition | 0.8987 |
79
+ | mmlu_other | 0.8745 |
80
+ | mmlu_philosophy | 0.8746 |
81
+ | mmlu_prehistory | 0.9043 |
82
+ | mmlu_professional_accounting | 0.7731 |
83
+ | mmlu_professional_law | 0.7086 |
84
+ | mmlu_professional_medicine | 0.9485 |
85
+ | mmlu_professional_psychology | 0.8873 |
86
+ | mmlu_public_relations | 0.7545 |
87
+ | mmlu_security_studies | 0.8327 |
88
+ | mmlu_social_sciences | 0.9155 |
89
+ | mmlu_sociology | 0.9403 |
90
+ | mmlu_stem | 0.8557 |
91
+ | mmlu_us_foreign_policy | 0.9500 |
92
+ | mmlu_virology | 0.5843 |
93
+ | mmlu_world_religions | 0.8889 |
94
+ | piqa | 0.8030 |
95
+
96
+ ## How to Use
97
+
98
+ ### HF Usage
99
+
100
+ **Step 1: Install [AutoRound](https://github.com/intel/auto-round)**
101
+
102
+ ```bash
103
+ pip install auto-round
104
+ ```
105
+
106
+ **Step 2: Load and run the quantized model**
107
+
108
+ ```python
109
+ from transformers import AutoModelForCausalLM, AutoTokenizer
110
+
111
+ model_name = "Qwen3.5-27B-Writer-V2-AutoRound-W4A16-Tuning"
112
+
113
+ # load the tokenizer and the model
114
+ tokenizer = AutoTokenizer.from_pretrained(model_name)
115
+ model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", device_map="auto")
116
+
117
+ # prepare the model input
118
+ prompt = "Write a quick sort algorithm."
119
+ messages = [{"role": "user", "content": prompt}]
120
+ text = tokenizer.apply_chat_template(
121
+ messages,
122
+ tokenize=False,
123
+ add_generation_prompt=True,
124
+ )
125
+ model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
126
+
127
+ # conduct text completion
128
+ generated_ids = model.generate(**model_inputs, max_new_tokens=512)
129
+ output_ids = generated_ids[0][len(model_inputs.input_ids[0]) :].tolist()
130
+
131
+ content = tokenizer.decode(output_ids, skip_special_tokens=True)
132
+ print("content:", content)
133
+ ```
134
+
135
+ ### VLLM Usage
136
+
137
+ ```bash
138
+ vllm serve Qwen3.5-27B-Writer-V2-AutoRound-W4A16-Tuning \
139
+ --trust-remote-code \
140
+ --dtype bfloat16 \
141
+ --tensor_parallel_size 1
142
+ ```
143
+
144
+ If you encounter any issues, feel free to open an issue on the [AutoRound GitHub repo](https://github.com/intel/auto-round/issues) or provide feedback on the [Low-Bit Open LLM Leaderboard](https://huggingface.co/spaces/Intel/low_bit_open_llm_leaderboard).
145
+
146
+ ## Ethical Considerations and Limitations
147
+
148
+ The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs.
149
+ Therefore, before deploying any applications of the model, developers should perform safety testing.
150
+
151
+ ## Caveats and Recommendations
152
+
153
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.
154
+ Here are a couple of useful links to learn more about Intel's AI software:
155
+
156
+ - [Intel Neural Compressor](https://github.com/intel/neural-compressor)
157
+ - [AutoRound](https://github.com/intel/auto-round)
158
+
159
+ ## Disclaimer
160
+
161
+ The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.
162
+
163
+ ## Cite
164
+
165
+ ```
166
+ @article{cheng2023optimize,
167
+ title={Optimize weight rounding via signed gradient descent for the quantization of llms},
168
+ author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi},
169
+ journal={arXiv preprint arXiv:2309.05516},
170
+ year={2023}
171
+ }
172
+ ```
173
+
174
+ [arxiv](https://arxiv.org/abs/2309.05516) [github](https://github.com/intel/auto-round)
175
+
176
+ ---
177
+
178
+ *This model is part of the [Intel Low-Bit Open LLM Leaderboard](https://huggingface.co/spaces/Intel/low_bit_open_llm_leaderboard) initiative.*
chat_template.jinja ADDED
@@ -0,0 +1,154 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- set image_count = namespace(value=0) %}
2
+ {%- set video_count = namespace(value=0) %}
3
+ {%- macro render_content(content, do_vision_count, is_system_content=false) %}
4
+ {%- if content is string %}
5
+ {{- content }}
6
+ {%- elif content is iterable and content is not mapping %}
7
+ {%- for item in content %}
8
+ {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
9
+ {%- if is_system_content %}
10
+ {{- raise_exception('System message cannot contain images.') }}
11
+ {%- endif %}
12
+ {%- if do_vision_count %}
13
+ {%- set image_count.value = image_count.value + 1 %}
14
+ {%- endif %}
15
+ {%- if add_vision_id %}
16
+ {{- 'Picture ' ~ image_count.value ~ ': ' }}
17
+ {%- endif %}
18
+ {{- '<|vision_start|><|image_pad|><|vision_end|>' }}
19
+ {%- elif 'video' in item or item.type == 'video' %}
20
+ {%- if is_system_content %}
21
+ {{- raise_exception('System message cannot contain videos.') }}
22
+ {%- endif %}
23
+ {%- if do_vision_count %}
24
+ {%- set video_count.value = video_count.value + 1 %}
25
+ {%- endif %}
26
+ {%- if add_vision_id %}
27
+ {{- 'Video ' ~ video_count.value ~ ': ' }}
28
+ {%- endif %}
29
+ {{- '<|vision_start|><|video_pad|><|vision_end|>' }}
30
+ {%- elif 'text' in item %}
31
+ {{- item.text }}
32
+ {%- else %}
33
+ {{- raise_exception('Unexpected item type in content.') }}
34
+ {%- endif %}
35
+ {%- endfor %}
36
+ {%- elif content is none or content is undefined %}
37
+ {{- '' }}
38
+ {%- else %}
39
+ {{- raise_exception('Unexpected content type.') }}
40
+ {%- endif %}
41
+ {%- endmacro %}
42
+ {%- if not messages %}
43
+ {{- raise_exception('No messages provided.') }}
44
+ {%- endif %}
45
+ {%- if tools and tools is iterable and tools is not mapping %}
46
+ {{- '<|im_start|>system\n' }}
47
+ {{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
48
+ {%- for tool in tools %}
49
+ {{- "\n" }}
50
+ {{- tool | tojson }}
51
+ {%- endfor %}
52
+ {{- "\n</tools>" }}
53
+ {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
54
+ {%- if messages[0].role == 'system' %}
55
+ {%- set content = render_content(messages[0].content, false, true)|trim %}
56
+ {%- if content %}
57
+ {{- '\n\n' + content }}
58
+ {%- endif %}
59
+ {%- endif %}
60
+ {{- '<|im_end|>\n' }}
61
+ {%- else %}
62
+ {%- if messages[0].role == 'system' %}
63
+ {%- set content = render_content(messages[0].content, false, true)|trim %}
64
+ {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
65
+ {%- endif %}
66
+ {%- endif %}
67
+ {%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
68
+ {%- for message in messages[::-1] %}
69
+ {%- set index = (messages|length - 1) - loop.index0 %}
70
+ {%- if ns.multi_step_tool and message.role == "user" %}
71
+ {%- set content = render_content(message.content, false)|trim %}
72
+ {%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
73
+ {%- set ns.multi_step_tool = false %}
74
+ {%- set ns.last_query_index = index %}
75
+ {%- endif %}
76
+ {%- endif %}
77
+ {%- endfor %}
78
+ {%- if ns.multi_step_tool %}
79
+ {{- raise_exception('No user query found in messages.') }}
80
+ {%- endif %}
81
+ {%- for message in messages %}
82
+ {%- set content = render_content(message.content, true)|trim %}
83
+ {%- if message.role == "system" %}
84
+ {%- if not loop.first %}
85
+ {{- raise_exception('System message must be at the beginning.') }}
86
+ {%- endif %}
87
+ {%- elif message.role == "user" %}
88
+ {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
89
+ {%- elif message.role == "assistant" %}
90
+ {%- set reasoning_content = '' %}
91
+ {%- if message.reasoning_content is string %}
92
+ {%- set reasoning_content = message.reasoning_content %}
93
+ {%- else %}
94
+ {%- if '</think>' in content %}
95
+ {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
96
+ {%- set content = content.split('</think>')[-1].lstrip('\n') %}
97
+ {%- endif %}
98
+ {%- endif %}
99
+ {%- set reasoning_content = reasoning_content|trim %}
100
+ {%- if loop.index0 > ns.last_query_index %}
101
+ {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
102
+ {%- else %}
103
+ {{- '<|im_start|>' + message.role + '\n' + content }}
104
+ {%- endif %}
105
+ {%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
106
+ {%- for tool_call in message.tool_calls %}
107
+ {%- if tool_call.function is defined %}
108
+ {%- set tool_call = tool_call.function %}
109
+ {%- endif %}
110
+ {%- if loop.first %}
111
+ {%- if content|trim %}
112
+ {{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
113
+ {%- else %}
114
+ {{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
115
+ {%- endif %}
116
+ {%- else %}
117
+ {{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
118
+ {%- endif %}
119
+ {%- if tool_call.arguments is defined %}
120
+ {%- for args_name, args_value in tool_call.arguments|items %}
121
+ {{- '<parameter=' + args_name + '>\n' }}
122
+ {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
123
+ {{- args_value }}
124
+ {{- '\n</parameter>\n' }}
125
+ {%- endfor %}
126
+ {%- endif %}
127
+ {{- '</function>\n</tool_call>' }}
128
+ {%- endfor %}
129
+ {%- endif %}
130
+ {{- '<|im_end|>\n' }}
131
+ {%- elif message.role == "tool" %}
132
+ {%- if loop.previtem and loop.previtem.role != "tool" %}
133
+ {{- '<|im_start|>user' }}
134
+ {%- endif %}
135
+ {{- '\n<tool_response>\n' }}
136
+ {{- content }}
137
+ {{- '\n</tool_response>' }}
138
+ {%- if not loop.last and loop.nextitem.role != "tool" %}
139
+ {{- '<|im_end|>\n' }}
140
+ {%- elif loop.last %}
141
+ {{- '<|im_end|>\n' }}
142
+ {%- endif %}
143
+ {%- else %}
144
+ {{- raise_exception('Unexpected message role.') }}
145
+ {%- endif %}
146
+ {%- endfor %}
147
+ {%- if add_generation_prompt %}
148
+ {{- '<|im_start|>assistant\n' }}
149
+ {%- if enable_thinking is defined and enable_thinking is false %}
150
+ {{- '<think>\n\n</think>\n\n' }}
151
+ {%- else %}
152
+ {{- '<think>\n' }}
153
+ {%- endif %}
154
+ {%- endif %}
config.json ADDED
@@ -0,0 +1,538 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3_5ForConditionalGeneration"
4
+ ],
5
+ "dtype": "bfloat16",
6
+ "image_token_id": 248056,
7
+ "model_type": "qwen3_5",
8
+ "quantization_config": {
9
+ "autoround_version": "0.13.0",
10
+ "bits": 4,
11
+ "block_name_to_quantize": "model.language_model.layers",
12
+ "data_type": "int",
13
+ "extra_config": {
14
+ "model.language_model.layers.0.linear_attn.in_proj_a": {
15
+ "bits": 16,
16
+ "data_type": "fp"
17
+ },
18
+ "model.language_model.layers.0.linear_attn.in_proj_b": {
19
+ "bits": 16,
20
+ "data_type": "fp"
21
+ },
22
+ "model.language_model.layers.1.linear_attn.in_proj_a": {
23
+ "bits": 16,
24
+ "data_type": "fp"
25
+ },
26
+ "model.language_model.layers.1.linear_attn.in_proj_b": {
27
+ "bits": 16,
28
+ "data_type": "fp"
29
+ },
30
+ "model.language_model.layers.10.linear_attn.in_proj_a": {
31
+ "bits": 16,
32
+ "data_type": "fp"
33
+ },
34
+ "model.language_model.layers.10.linear_attn.in_proj_b": {
35
+ "bits": 16,
36
+ "data_type": "fp"
37
+ },
38
+ "model.language_model.layers.12.linear_attn.in_proj_a": {
39
+ "bits": 16,
40
+ "data_type": "fp"
41
+ },
42
+ "model.language_model.layers.12.linear_attn.in_proj_b": {
43
+ "bits": 16,
44
+ "data_type": "fp"
45
+ },
46
+ "model.language_model.layers.13.linear_attn.in_proj_a": {
47
+ "bits": 16,
48
+ "data_type": "fp"
49
+ },
50
+ "model.language_model.layers.13.linear_attn.in_proj_b": {
51
+ "bits": 16,
52
+ "data_type": "fp"
53
+ },
54
+ "model.language_model.layers.14.linear_attn.in_proj_a": {
55
+ "bits": 16,
56
+ "data_type": "fp"
57
+ },
58
+ "model.language_model.layers.14.linear_attn.in_proj_b": {
59
+ "bits": 16,
60
+ "data_type": "fp"
61
+ },
62
+ "model.language_model.layers.16.linear_attn.in_proj_a": {
63
+ "bits": 16,
64
+ "data_type": "fp"
65
+ },
66
+ "model.language_model.layers.16.linear_attn.in_proj_b": {
67
+ "bits": 16,
68
+ "data_type": "fp"
69
+ },
70
+ "model.language_model.layers.17.linear_attn.in_proj_a": {
71
+ "bits": 16,
72
+ "data_type": "fp"
73
+ },
74
+ "model.language_model.layers.17.linear_attn.in_proj_b": {
75
+ "bits": 16,
76
+ "data_type": "fp"
77
+ },
78
+ "model.language_model.layers.18.linear_attn.in_proj_a": {
79
+ "bits": 16,
80
+ "data_type": "fp"
81
+ },
82
+ "model.language_model.layers.18.linear_attn.in_proj_b": {
83
+ "bits": 16,
84
+ "data_type": "fp"
85
+ },
86
+ "model.language_model.layers.2.linear_attn.in_proj_a": {
87
+ "bits": 16,
88
+ "data_type": "fp"
89
+ },
90
+ "model.language_model.layers.2.linear_attn.in_proj_b": {
91
+ "bits": 16,
92
+ "data_type": "fp"
93
+ },
94
+ "model.language_model.layers.20.linear_attn.in_proj_a": {
95
+ "bits": 16,
96
+ "data_type": "fp"
97
+ },
98
+ "model.language_model.layers.20.linear_attn.in_proj_b": {
99
+ "bits": 16,
100
+ "data_type": "fp"
101
+ },
102
+ "model.language_model.layers.21.linear_attn.in_proj_a": {
103
+ "bits": 16,
104
+ "data_type": "fp"
105
+ },
106
+ "model.language_model.layers.21.linear_attn.in_proj_b": {
107
+ "bits": 16,
108
+ "data_type": "fp"
109
+ },
110
+ "model.language_model.layers.22.linear_attn.in_proj_a": {
111
+ "bits": 16,
112
+ "data_type": "fp"
113
+ },
114
+ "model.language_model.layers.22.linear_attn.in_proj_b": {
115
+ "bits": 16,
116
+ "data_type": "fp"
117
+ },
118
+ "model.language_model.layers.24.linear_attn.in_proj_a": {
119
+ "bits": 16,
120
+ "data_type": "fp"
121
+ },
122
+ "model.language_model.layers.24.linear_attn.in_proj_b": {
123
+ "bits": 16,
124
+ "data_type": "fp"
125
+ },
126
+ "model.language_model.layers.25.linear_attn.in_proj_a": {
127
+ "bits": 16,
128
+ "data_type": "fp"
129
+ },
130
+ "model.language_model.layers.25.linear_attn.in_proj_b": {
131
+ "bits": 16,
132
+ "data_type": "fp"
133
+ },
134
+ "model.language_model.layers.26.linear_attn.in_proj_a": {
135
+ "bits": 16,
136
+ "data_type": "fp"
137
+ },
138
+ "model.language_model.layers.26.linear_attn.in_proj_b": {
139
+ "bits": 16,
140
+ "data_type": "fp"
141
+ },
142
+ "model.language_model.layers.28.linear_attn.in_proj_a": {
143
+ "bits": 16,
144
+ "data_type": "fp"
145
+ },
146
+ "model.language_model.layers.28.linear_attn.in_proj_b": {
147
+ "bits": 16,
148
+ "data_type": "fp"
149
+ },
150
+ "model.language_model.layers.29.linear_attn.in_proj_a": {
151
+ "bits": 16,
152
+ "data_type": "fp"
153
+ },
154
+ "model.language_model.layers.29.linear_attn.in_proj_b": {
155
+ "bits": 16,
156
+ "data_type": "fp"
157
+ },
158
+ "model.language_model.layers.30.linear_attn.in_proj_a": {
159
+ "bits": 16,
160
+ "data_type": "fp"
161
+ },
162
+ "model.language_model.layers.30.linear_attn.in_proj_b": {
163
+ "bits": 16,
164
+ "data_type": "fp"
165
+ },
166
+ "model.language_model.layers.32.linear_attn.in_proj_a": {
167
+ "bits": 16,
168
+ "data_type": "fp"
169
+ },
170
+ "model.language_model.layers.32.linear_attn.in_proj_b": {
171
+ "bits": 16,
172
+ "data_type": "fp"
173
+ },
174
+ "model.language_model.layers.33.linear_attn.in_proj_a": {
175
+ "bits": 16,
176
+ "data_type": "fp"
177
+ },
178
+ "model.language_model.layers.33.linear_attn.in_proj_b": {
179
+ "bits": 16,
180
+ "data_type": "fp"
181
+ },
182
+ "model.language_model.layers.34.linear_attn.in_proj_a": {
183
+ "bits": 16,
184
+ "data_type": "fp"
185
+ },
186
+ "model.language_model.layers.34.linear_attn.in_proj_b": {
187
+ "bits": 16,
188
+ "data_type": "fp"
189
+ },
190
+ "model.language_model.layers.36.linear_attn.in_proj_a": {
191
+ "bits": 16,
192
+ "data_type": "fp"
193
+ },
194
+ "model.language_model.layers.36.linear_attn.in_proj_b": {
195
+ "bits": 16,
196
+ "data_type": "fp"
197
+ },
198
+ "model.language_model.layers.37.linear_attn.in_proj_a": {
199
+ "bits": 16,
200
+ "data_type": "fp"
201
+ },
202
+ "model.language_model.layers.37.linear_attn.in_proj_b": {
203
+ "bits": 16,
204
+ "data_type": "fp"
205
+ },
206
+ "model.language_model.layers.38.linear_attn.in_proj_a": {
207
+ "bits": 16,
208
+ "data_type": "fp"
209
+ },
210
+ "model.language_model.layers.38.linear_attn.in_proj_b": {
211
+ "bits": 16,
212
+ "data_type": "fp"
213
+ },
214
+ "model.language_model.layers.4.linear_attn.in_proj_a": {
215
+ "bits": 16,
216
+ "data_type": "fp"
217
+ },
218
+ "model.language_model.layers.4.linear_attn.in_proj_b": {
219
+ "bits": 16,
220
+ "data_type": "fp"
221
+ },
222
+ "model.language_model.layers.40.linear_attn.in_proj_a": {
223
+ "bits": 16,
224
+ "data_type": "fp"
225
+ },
226
+ "model.language_model.layers.40.linear_attn.in_proj_b": {
227
+ "bits": 16,
228
+ "data_type": "fp"
229
+ },
230
+ "model.language_model.layers.41.linear_attn.in_proj_a": {
231
+ "bits": 16,
232
+ "data_type": "fp"
233
+ },
234
+ "model.language_model.layers.41.linear_attn.in_proj_b": {
235
+ "bits": 16,
236
+ "data_type": "fp"
237
+ },
238
+ "model.language_model.layers.42.linear_attn.in_proj_a": {
239
+ "bits": 16,
240
+ "data_type": "fp"
241
+ },
242
+ "model.language_model.layers.42.linear_attn.in_proj_b": {
243
+ "bits": 16,
244
+ "data_type": "fp"
245
+ },
246
+ "model.language_model.layers.44.linear_attn.in_proj_a": {
247
+ "bits": 16,
248
+ "data_type": "fp"
249
+ },
250
+ "model.language_model.layers.44.linear_attn.in_proj_b": {
251
+ "bits": 16,
252
+ "data_type": "fp"
253
+ },
254
+ "model.language_model.layers.45.linear_attn.in_proj_a": {
255
+ "bits": 16,
256
+ "data_type": "fp"
257
+ },
258
+ "model.language_model.layers.45.linear_attn.in_proj_b": {
259
+ "bits": 16,
260
+ "data_type": "fp"
261
+ },
262
+ "model.language_model.layers.46.linear_attn.in_proj_a": {
263
+ "bits": 16,
264
+ "data_type": "fp"
265
+ },
266
+ "model.language_model.layers.46.linear_attn.in_proj_b": {
267
+ "bits": 16,
268
+ "data_type": "fp"
269
+ },
270
+ "model.language_model.layers.48.linear_attn.in_proj_a": {
271
+ "bits": 16,
272
+ "data_type": "fp"
273
+ },
274
+ "model.language_model.layers.48.linear_attn.in_proj_b": {
275
+ "bits": 16,
276
+ "data_type": "fp"
277
+ },
278
+ "model.language_model.layers.49.linear_attn.in_proj_a": {
279
+ "bits": 16,
280
+ "data_type": "fp"
281
+ },
282
+ "model.language_model.layers.49.linear_attn.in_proj_b": {
283
+ "bits": 16,
284
+ "data_type": "fp"
285
+ },
286
+ "model.language_model.layers.5.linear_attn.in_proj_a": {
287
+ "bits": 16,
288
+ "data_type": "fp"
289
+ },
290
+ "model.language_model.layers.5.linear_attn.in_proj_b": {
291
+ "bits": 16,
292
+ "data_type": "fp"
293
+ },
294
+ "model.language_model.layers.50.linear_attn.in_proj_a": {
295
+ "bits": 16,
296
+ "data_type": "fp"
297
+ },
298
+ "model.language_model.layers.50.linear_attn.in_proj_b": {
299
+ "bits": 16,
300
+ "data_type": "fp"
301
+ },
302
+ "model.language_model.layers.52.linear_attn.in_proj_a": {
303
+ "bits": 16,
304
+ "data_type": "fp"
305
+ },
306
+ "model.language_model.layers.52.linear_attn.in_proj_b": {
307
+ "bits": 16,
308
+ "data_type": "fp"
309
+ },
310
+ "model.language_model.layers.53.linear_attn.in_proj_a": {
311
+ "bits": 16,
312
+ "data_type": "fp"
313
+ },
314
+ "model.language_model.layers.53.linear_attn.in_proj_b": {
315
+ "bits": 16,
316
+ "data_type": "fp"
317
+ },
318
+ "model.language_model.layers.54.linear_attn.in_proj_a": {
319
+ "bits": 16,
320
+ "data_type": "fp"
321
+ },
322
+ "model.language_model.layers.54.linear_attn.in_proj_b": {
323
+ "bits": 16,
324
+ "data_type": "fp"
325
+ },
326
+ "model.language_model.layers.56.linear_attn.in_proj_a": {
327
+ "bits": 16,
328
+ "data_type": "fp"
329
+ },
330
+ "model.language_model.layers.56.linear_attn.in_proj_b": {
331
+ "bits": 16,
332
+ "data_type": "fp"
333
+ },
334
+ "model.language_model.layers.57.linear_attn.in_proj_a": {
335
+ "bits": 16,
336
+ "data_type": "fp"
337
+ },
338
+ "model.language_model.layers.57.linear_attn.in_proj_b": {
339
+ "bits": 16,
340
+ "data_type": "fp"
341
+ },
342
+ "model.language_model.layers.58.linear_attn.in_proj_a": {
343
+ "bits": 16,
344
+ "data_type": "fp"
345
+ },
346
+ "model.language_model.layers.58.linear_attn.in_proj_b": {
347
+ "bits": 16,
348
+ "data_type": "fp"
349
+ },
350
+ "model.language_model.layers.6.linear_attn.in_proj_a": {
351
+ "bits": 16,
352
+ "data_type": "fp"
353
+ },
354
+ "model.language_model.layers.6.linear_attn.in_proj_b": {
355
+ "bits": 16,
356
+ "data_type": "fp"
357
+ },
358
+ "model.language_model.layers.60.linear_attn.in_proj_a": {
359
+ "bits": 16,
360
+ "data_type": "fp"
361
+ },
362
+ "model.language_model.layers.60.linear_attn.in_proj_b": {
363
+ "bits": 16,
364
+ "data_type": "fp"
365
+ },
366
+ "model.language_model.layers.61.linear_attn.in_proj_a": {
367
+ "bits": 16,
368
+ "data_type": "fp"
369
+ },
370
+ "model.language_model.layers.61.linear_attn.in_proj_b": {
371
+ "bits": 16,
372
+ "data_type": "fp"
373
+ },
374
+ "model.language_model.layers.62.linear_attn.in_proj_a": {
375
+ "bits": 16,
376
+ "data_type": "fp"
377
+ },
378
+ "model.language_model.layers.62.linear_attn.in_proj_b": {
379
+ "bits": 16,
380
+ "data_type": "fp"
381
+ },
382
+ "model.language_model.layers.8.linear_attn.in_proj_a": {
383
+ "bits": 16,
384
+ "data_type": "fp"
385
+ },
386
+ "model.language_model.layers.8.linear_attn.in_proj_b": {
387
+ "bits": 16,
388
+ "data_type": "fp"
389
+ },
390
+ "model.language_model.layers.9.linear_attn.in_proj_a": {
391
+ "bits": 16,
392
+ "data_type": "fp"
393
+ },
394
+ "model.language_model.layers.9.linear_attn.in_proj_b": {
395
+ "bits": 16,
396
+ "data_type": "fp"
397
+ }
398
+ },
399
+ "group_size": 128,
400
+ "low_gpu_mem_usage": true,
401
+ "packing_format": "auto_round:auto_gptq",
402
+ "quant_method": "auto-round",
403
+ "sym": true
404
+ },
405
+ "text_config": {
406
+ "attention_bias": false,
407
+ "attention_dropout": 0.0,
408
+ "attn_output_gate": true,
409
+ "bos_token_id": null,
410
+ "dtype": "bfloat16",
411
+ "eos_token_id": 248044,
412
+ "full_attention_interval": 4,
413
+ "head_dim": 256,
414
+ "hidden_act": "silu",
415
+ "hidden_size": 5120,
416
+ "initializer_range": 0.02,
417
+ "intermediate_size": 17408,
418
+ "layer_types": [
419
+ "linear_attention",
420
+ "linear_attention",
421
+ "linear_attention",
422
+ "full_attention",
423
+ "linear_attention",
424
+ "linear_attention",
425
+ "linear_attention",
426
+ "full_attention",
427
+ "linear_attention",
428
+ "linear_attention",
429
+ "linear_attention",
430
+ "full_attention",
431
+ "linear_attention",
432
+ "linear_attention",
433
+ "linear_attention",
434
+ "full_attention",
435
+ "linear_attention",
436
+ "linear_attention",
437
+ "linear_attention",
438
+ "full_attention",
439
+ "linear_attention",
440
+ "linear_attention",
441
+ "linear_attention",
442
+ "full_attention",
443
+ "linear_attention",
444
+ "linear_attention",
445
+ "linear_attention",
446
+ "full_attention",
447
+ "linear_attention",
448
+ "linear_attention",
449
+ "linear_attention",
450
+ "full_attention",
451
+ "linear_attention",
452
+ "linear_attention",
453
+ "linear_attention",
454
+ "full_attention",
455
+ "linear_attention",
456
+ "linear_attention",
457
+ "linear_attention",
458
+ "full_attention",
459
+ "linear_attention",
460
+ "linear_attention",
461
+ "linear_attention",
462
+ "full_attention",
463
+ "linear_attention",
464
+ "linear_attention",
465
+ "linear_attention",
466
+ "full_attention",
467
+ "linear_attention",
468
+ "linear_attention",
469
+ "linear_attention",
470
+ "full_attention",
471
+ "linear_attention",
472
+ "linear_attention",
473
+ "linear_attention",
474
+ "full_attention",
475
+ "linear_attention",
476
+ "linear_attention",
477
+ "linear_attention",
478
+ "full_attention",
479
+ "linear_attention",
480
+ "linear_attention",
481
+ "linear_attention",
482
+ "full_attention"
483
+ ],
484
+ "linear_conv_kernel_dim": 4,
485
+ "linear_key_head_dim": 128,
486
+ "linear_num_key_heads": 16,
487
+ "linear_num_value_heads": 48,
488
+ "linear_value_head_dim": 128,
489
+ "mamba_ssm_dtype": "float32",
490
+ "max_position_embeddings": 262144,
491
+ "mlp_only_layers": [],
492
+ "model_type": "qwen3_5_text",
493
+ "mtp_num_hidden_layers": 1,
494
+ "mtp_use_dedicated_embeddings": false,
495
+ "num_attention_heads": 24,
496
+ "num_hidden_layers": 64,
497
+ "num_key_value_heads": 4,
498
+ "pad_token_id": null,
499
+ "partial_rotary_factor": 0.25,
500
+ "rms_norm_eps": 1e-06,
501
+ "rope_parameters": {
502
+ "mrope_interleaved": true,
503
+ "mrope_section": [
504
+ 11,
505
+ 11,
506
+ 10
507
+ ],
508
+ "partial_rotary_factor": 0.25,
509
+ "rope_theta": 10000000,
510
+ "rope_type": "default"
511
+ },
512
+ "tie_word_embeddings": false,
513
+ "use_cache": false,
514
+ "vocab_size": 248320
515
+ },
516
+ "tie_word_embeddings": false,
517
+ "transformers_version": "5.10.2",
518
+ "video_token_id": 248057,
519
+ "vision_config": {
520
+ "deepstack_visual_indexes": [],
521
+ "depth": 27,
522
+ "dtype": "bfloat16",
523
+ "hidden_act": "gelu_pytorch_tanh",
524
+ "hidden_size": 1152,
525
+ "in_channels": 3,
526
+ "initializer_range": 0.02,
527
+ "intermediate_size": 4304,
528
+ "model_type": "qwen3_5_vision",
529
+ "num_heads": 16,
530
+ "num_position_embeddings": 2304,
531
+ "out_hidden_size": 5120,
532
+ "patch_size": 16,
533
+ "spatial_merge_size": 2,
534
+ "temporal_patch_size": 2
535
+ },
536
+ "vision_end_token_id": 248054,
537
+ "vision_start_token_id": 248053
538
+ }
generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 248044,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 248046,
6
+ 248044
7
+ ],
8
+ "pad_token_id": 248044,
9
+ "temperature": 0.6,
10
+ "top_k": 20,
11
+ "top_p": 0.95,
12
+ "transformers_version": "5.10.2"
13
+ }
model-00001-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c1ebaabbaca559789304b32d41530fb890cb406d30b4113f2e0e754a05dc93d
3
+ size 2542796928
model-00002-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6bfb319ac4496d2acf14007ce7f7098da87c0abe5f4837b4905c1df6c4fb24c2
3
+ size 4983417520
model-00003-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f6af48edbfa824f0c162dbc718d3edfb2c9d8eede1eab13a95e136954c139b33
3
+ size 4959166944
model-00004-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:93c4728ede4b48757d7bd7c4a49db826e3c0968d4b55177a445379cd9905edf6
3
+ size 4959166944
model-00005-of-00005.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:81bfbc6c76450a4bcf4fd7a180a241fa139894b8e79af269edd131d4c66784b6
3
+ size 1253852480
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
preprocessor_config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "data_format": "channels_first",
3
+ "do_convert_rgb": true,
4
+ "do_normalize": true,
5
+ "do_rescale": true,
6
+ "do_resize": true,
7
+ "image_mean": [
8
+ 0.5,
9
+ 0.5,
10
+ 0.5
11
+ ],
12
+ "image_processor_type": "Qwen2VLImageProcessor",
13
+ "image_std": [
14
+ 0.5,
15
+ 0.5,
16
+ 0.5
17
+ ],
18
+ "merge_size": 2,
19
+ "patch_size": 16,
20
+ "resample": 3,
21
+ "rescale_factor": 0.00392156862745098,
22
+ "size": {
23
+ "longest_edge": 16777216,
24
+ "shortest_edge": 65536
25
+ },
26
+ "temporal_patch_size": 2
27
+ }
processor_config.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "data_format": "channels_first",
4
+ "do_convert_rgb": true,
5
+ "do_normalize": true,
6
+ "do_rescale": true,
7
+ "do_resize": true,
8
+ "image_mean": [
9
+ 0.5,
10
+ 0.5,
11
+ 0.5
12
+ ],
13
+ "image_processor_type": "Qwen2VLImageProcessor",
14
+ "image_std": [
15
+ 0.5,
16
+ 0.5,
17
+ 0.5
18
+ ],
19
+ "merge_size": 2,
20
+ "patch_size": 16,
21
+ "resample": 3,
22
+ "rescale_factor": 0.00392156862745098,
23
+ "size": {
24
+ "longest_edge": 16777216,
25
+ "shortest_edge": 65536
26
+ },
27
+ "temporal_patch_size": 2
28
+ },
29
+ "processor_class": "Qwen3VLProcessor",
30
+ "video_processor": {
31
+ "data_format": "channels_first",
32
+ "default_to_square": true,
33
+ "do_convert_rgb": true,
34
+ "do_normalize": true,
35
+ "do_rescale": true,
36
+ "do_resize": true,
37
+ "do_sample_frames": true,
38
+ "fps": 2,
39
+ "image_mean": [
40
+ 0.5,
41
+ 0.5,
42
+ 0.5
43
+ ],
44
+ "image_std": [
45
+ 0.5,
46
+ 0.5,
47
+ 0.5
48
+ ],
49
+ "max_frames": 768,
50
+ "merge_size": 2,
51
+ "min_frames": 4,
52
+ "patch_size": 16,
53
+ "resample": 3,
54
+ "rescale_factor": 0.00392156862745098,
55
+ "return_metadata": false,
56
+ "size": {
57
+ "longest_edge": 25165824,
58
+ "shortest_edge": 4096
59
+ },
60
+ "temporal_patch_size": 2,
61
+ "video_processor_type": "Qwen3VLVideoProcessor"
62
+ }
63
+ }
quantization_config.json ADDED
@@ -0,0 +1,397 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bits": 4,
3
+ "data_type": "int",
4
+ "group_size": 128,
5
+ "sym": true,
6
+ "low_gpu_mem_usage": true,
7
+ "autoround_version": "0.13.0",
8
+ "block_name_to_quantize": "model.language_model.layers",
9
+ "quant_method": "auto-round",
10
+ "packing_format": "auto_round:auto_gptq",
11
+ "extra_config": {
12
+ "model.language_model.layers.0.linear_attn.in_proj_b": {
13
+ "bits": 16,
14
+ "data_type": "fp"
15
+ },
16
+ "model.language_model.layers.0.linear_attn.in_proj_a": {
17
+ "bits": 16,
18
+ "data_type": "fp"
19
+ },
20
+ "model.language_model.layers.1.linear_attn.in_proj_b": {
21
+ "bits": 16,
22
+ "data_type": "fp"
23
+ },
24
+ "model.language_model.layers.1.linear_attn.in_proj_a": {
25
+ "bits": 16,
26
+ "data_type": "fp"
27
+ },
28
+ "model.language_model.layers.2.linear_attn.in_proj_b": {
29
+ "bits": 16,
30
+ "data_type": "fp"
31
+ },
32
+ "model.language_model.layers.2.linear_attn.in_proj_a": {
33
+ "bits": 16,
34
+ "data_type": "fp"
35
+ },
36
+ "model.language_model.layers.4.linear_attn.in_proj_b": {
37
+ "bits": 16,
38
+ "data_type": "fp"
39
+ },
40
+ "model.language_model.layers.4.linear_attn.in_proj_a": {
41
+ "bits": 16,
42
+ "data_type": "fp"
43
+ },
44
+ "model.language_model.layers.5.linear_attn.in_proj_b": {
45
+ "bits": 16,
46
+ "data_type": "fp"
47
+ },
48
+ "model.language_model.layers.5.linear_attn.in_proj_a": {
49
+ "bits": 16,
50
+ "data_type": "fp"
51
+ },
52
+ "model.language_model.layers.6.linear_attn.in_proj_b": {
53
+ "bits": 16,
54
+ "data_type": "fp"
55
+ },
56
+ "model.language_model.layers.6.linear_attn.in_proj_a": {
57
+ "bits": 16,
58
+ "data_type": "fp"
59
+ },
60
+ "model.language_model.layers.8.linear_attn.in_proj_b": {
61
+ "bits": 16,
62
+ "data_type": "fp"
63
+ },
64
+ "model.language_model.layers.8.linear_attn.in_proj_a": {
65
+ "bits": 16,
66
+ "data_type": "fp"
67
+ },
68
+ "model.language_model.layers.9.linear_attn.in_proj_b": {
69
+ "bits": 16,
70
+ "data_type": "fp"
71
+ },
72
+ "model.language_model.layers.9.linear_attn.in_proj_a": {
73
+ "bits": 16,
74
+ "data_type": "fp"
75
+ },
76
+ "model.language_model.layers.10.linear_attn.in_proj_b": {
77
+ "bits": 16,
78
+ "data_type": "fp"
79
+ },
80
+ "model.language_model.layers.10.linear_attn.in_proj_a": {
81
+ "bits": 16,
82
+ "data_type": "fp"
83
+ },
84
+ "model.language_model.layers.12.linear_attn.in_proj_b": {
85
+ "bits": 16,
86
+ "data_type": "fp"
87
+ },
88
+ "model.language_model.layers.12.linear_attn.in_proj_a": {
89
+ "bits": 16,
90
+ "data_type": "fp"
91
+ },
92
+ "model.language_model.layers.13.linear_attn.in_proj_b": {
93
+ "bits": 16,
94
+ "data_type": "fp"
95
+ },
96
+ "model.language_model.layers.13.linear_attn.in_proj_a": {
97
+ "bits": 16,
98
+ "data_type": "fp"
99
+ },
100
+ "model.language_model.layers.14.linear_attn.in_proj_b": {
101
+ "bits": 16,
102
+ "data_type": "fp"
103
+ },
104
+ "model.language_model.layers.14.linear_attn.in_proj_a": {
105
+ "bits": 16,
106
+ "data_type": "fp"
107
+ },
108
+ "model.language_model.layers.16.linear_attn.in_proj_b": {
109
+ "bits": 16,
110
+ "data_type": "fp"
111
+ },
112
+ "model.language_model.layers.16.linear_attn.in_proj_a": {
113
+ "bits": 16,
114
+ "data_type": "fp"
115
+ },
116
+ "model.language_model.layers.17.linear_attn.in_proj_b": {
117
+ "bits": 16,
118
+ "data_type": "fp"
119
+ },
120
+ "model.language_model.layers.17.linear_attn.in_proj_a": {
121
+ "bits": 16,
122
+ "data_type": "fp"
123
+ },
124
+ "model.language_model.layers.18.linear_attn.in_proj_b": {
125
+ "bits": 16,
126
+ "data_type": "fp"
127
+ },
128
+ "model.language_model.layers.18.linear_attn.in_proj_a": {
129
+ "bits": 16,
130
+ "data_type": "fp"
131
+ },
132
+ "model.language_model.layers.20.linear_attn.in_proj_b": {
133
+ "bits": 16,
134
+ "data_type": "fp"
135
+ },
136
+ "model.language_model.layers.20.linear_attn.in_proj_a": {
137
+ "bits": 16,
138
+ "data_type": "fp"
139
+ },
140
+ "model.language_model.layers.21.linear_attn.in_proj_b": {
141
+ "bits": 16,
142
+ "data_type": "fp"
143
+ },
144
+ "model.language_model.layers.21.linear_attn.in_proj_a": {
145
+ "bits": 16,
146
+ "data_type": "fp"
147
+ },
148
+ "model.language_model.layers.22.linear_attn.in_proj_b": {
149
+ "bits": 16,
150
+ "data_type": "fp"
151
+ },
152
+ "model.language_model.layers.22.linear_attn.in_proj_a": {
153
+ "bits": 16,
154
+ "data_type": "fp"
155
+ },
156
+ "model.language_model.layers.24.linear_attn.in_proj_b": {
157
+ "bits": 16,
158
+ "data_type": "fp"
159
+ },
160
+ "model.language_model.layers.24.linear_attn.in_proj_a": {
161
+ "bits": 16,
162
+ "data_type": "fp"
163
+ },
164
+ "model.language_model.layers.25.linear_attn.in_proj_b": {
165
+ "bits": 16,
166
+ "data_type": "fp"
167
+ },
168
+ "model.language_model.layers.25.linear_attn.in_proj_a": {
169
+ "bits": 16,
170
+ "data_type": "fp"
171
+ },
172
+ "model.language_model.layers.26.linear_attn.in_proj_b": {
173
+ "bits": 16,
174
+ "data_type": "fp"
175
+ },
176
+ "model.language_model.layers.26.linear_attn.in_proj_a": {
177
+ "bits": 16,
178
+ "data_type": "fp"
179
+ },
180
+ "model.language_model.layers.28.linear_attn.in_proj_b": {
181
+ "bits": 16,
182
+ "data_type": "fp"
183
+ },
184
+ "model.language_model.layers.28.linear_attn.in_proj_a": {
185
+ "bits": 16,
186
+ "data_type": "fp"
187
+ },
188
+ "model.language_model.layers.29.linear_attn.in_proj_b": {
189
+ "bits": 16,
190
+ "data_type": "fp"
191
+ },
192
+ "model.language_model.layers.29.linear_attn.in_proj_a": {
193
+ "bits": 16,
194
+ "data_type": "fp"
195
+ },
196
+ "model.language_model.layers.30.linear_attn.in_proj_b": {
197
+ "bits": 16,
198
+ "data_type": "fp"
199
+ },
200
+ "model.language_model.layers.30.linear_attn.in_proj_a": {
201
+ "bits": 16,
202
+ "data_type": "fp"
203
+ },
204
+ "model.language_model.layers.32.linear_attn.in_proj_b": {
205
+ "bits": 16,
206
+ "data_type": "fp"
207
+ },
208
+ "model.language_model.layers.32.linear_attn.in_proj_a": {
209
+ "bits": 16,
210
+ "data_type": "fp"
211
+ },
212
+ "model.language_model.layers.33.linear_attn.in_proj_b": {
213
+ "bits": 16,
214
+ "data_type": "fp"
215
+ },
216
+ "model.language_model.layers.33.linear_attn.in_proj_a": {
217
+ "bits": 16,
218
+ "data_type": "fp"
219
+ },
220
+ "model.language_model.layers.34.linear_attn.in_proj_b": {
221
+ "bits": 16,
222
+ "data_type": "fp"
223
+ },
224
+ "model.language_model.layers.34.linear_attn.in_proj_a": {
225
+ "bits": 16,
226
+ "data_type": "fp"
227
+ },
228
+ "model.language_model.layers.36.linear_attn.in_proj_b": {
229
+ "bits": 16,
230
+ "data_type": "fp"
231
+ },
232
+ "model.language_model.layers.36.linear_attn.in_proj_a": {
233
+ "bits": 16,
234
+ "data_type": "fp"
235
+ },
236
+ "model.language_model.layers.37.linear_attn.in_proj_b": {
237
+ "bits": 16,
238
+ "data_type": "fp"
239
+ },
240
+ "model.language_model.layers.37.linear_attn.in_proj_a": {
241
+ "bits": 16,
242
+ "data_type": "fp"
243
+ },
244
+ "model.language_model.layers.38.linear_attn.in_proj_b": {
245
+ "bits": 16,
246
+ "data_type": "fp"
247
+ },
248
+ "model.language_model.layers.38.linear_attn.in_proj_a": {
249
+ "bits": 16,
250
+ "data_type": "fp"
251
+ },
252
+ "model.language_model.layers.40.linear_attn.in_proj_b": {
253
+ "bits": 16,
254
+ "data_type": "fp"
255
+ },
256
+ "model.language_model.layers.40.linear_attn.in_proj_a": {
257
+ "bits": 16,
258
+ "data_type": "fp"
259
+ },
260
+ "model.language_model.layers.41.linear_attn.in_proj_b": {
261
+ "bits": 16,
262
+ "data_type": "fp"
263
+ },
264
+ "model.language_model.layers.41.linear_attn.in_proj_a": {
265
+ "bits": 16,
266
+ "data_type": "fp"
267
+ },
268
+ "model.language_model.layers.42.linear_attn.in_proj_b": {
269
+ "bits": 16,
270
+ "data_type": "fp"
271
+ },
272
+ "model.language_model.layers.42.linear_attn.in_proj_a": {
273
+ "bits": 16,
274
+ "data_type": "fp"
275
+ },
276
+ "model.language_model.layers.44.linear_attn.in_proj_b": {
277
+ "bits": 16,
278
+ "data_type": "fp"
279
+ },
280
+ "model.language_model.layers.44.linear_attn.in_proj_a": {
281
+ "bits": 16,
282
+ "data_type": "fp"
283
+ },
284
+ "model.language_model.layers.45.linear_attn.in_proj_b": {
285
+ "bits": 16,
286
+ "data_type": "fp"
287
+ },
288
+ "model.language_model.layers.45.linear_attn.in_proj_a": {
289
+ "bits": 16,
290
+ "data_type": "fp"
291
+ },
292
+ "model.language_model.layers.46.linear_attn.in_proj_b": {
293
+ "bits": 16,
294
+ "data_type": "fp"
295
+ },
296
+ "model.language_model.layers.46.linear_attn.in_proj_a": {
297
+ "bits": 16,
298
+ "data_type": "fp"
299
+ },
300
+ "model.language_model.layers.48.linear_attn.in_proj_b": {
301
+ "bits": 16,
302
+ "data_type": "fp"
303
+ },
304
+ "model.language_model.layers.48.linear_attn.in_proj_a": {
305
+ "bits": 16,
306
+ "data_type": "fp"
307
+ },
308
+ "model.language_model.layers.49.linear_attn.in_proj_b": {
309
+ "bits": 16,
310
+ "data_type": "fp"
311
+ },
312
+ "model.language_model.layers.49.linear_attn.in_proj_a": {
313
+ "bits": 16,
314
+ "data_type": "fp"
315
+ },
316
+ "model.language_model.layers.50.linear_attn.in_proj_b": {
317
+ "bits": 16,
318
+ "data_type": "fp"
319
+ },
320
+ "model.language_model.layers.50.linear_attn.in_proj_a": {
321
+ "bits": 16,
322
+ "data_type": "fp"
323
+ },
324
+ "model.language_model.layers.52.linear_attn.in_proj_b": {
325
+ "bits": 16,
326
+ "data_type": "fp"
327
+ },
328
+ "model.language_model.layers.52.linear_attn.in_proj_a": {
329
+ "bits": 16,
330
+ "data_type": "fp"
331
+ },
332
+ "model.language_model.layers.53.linear_attn.in_proj_b": {
333
+ "bits": 16,
334
+ "data_type": "fp"
335
+ },
336
+ "model.language_model.layers.53.linear_attn.in_proj_a": {
337
+ "bits": 16,
338
+ "data_type": "fp"
339
+ },
340
+ "model.language_model.layers.54.linear_attn.in_proj_b": {
341
+ "bits": 16,
342
+ "data_type": "fp"
343
+ },
344
+ "model.language_model.layers.54.linear_attn.in_proj_a": {
345
+ "bits": 16,
346
+ "data_type": "fp"
347
+ },
348
+ "model.language_model.layers.56.linear_attn.in_proj_b": {
349
+ "bits": 16,
350
+ "data_type": "fp"
351
+ },
352
+ "model.language_model.layers.56.linear_attn.in_proj_a": {
353
+ "bits": 16,
354
+ "data_type": "fp"
355
+ },
356
+ "model.language_model.layers.57.linear_attn.in_proj_b": {
357
+ "bits": 16,
358
+ "data_type": "fp"
359
+ },
360
+ "model.language_model.layers.57.linear_attn.in_proj_a": {
361
+ "bits": 16,
362
+ "data_type": "fp"
363
+ },
364
+ "model.language_model.layers.58.linear_attn.in_proj_b": {
365
+ "bits": 16,
366
+ "data_type": "fp"
367
+ },
368
+ "model.language_model.layers.58.linear_attn.in_proj_a": {
369
+ "bits": 16,
370
+ "data_type": "fp"
371
+ },
372
+ "model.language_model.layers.60.linear_attn.in_proj_b": {
373
+ "bits": 16,
374
+ "data_type": "fp"
375
+ },
376
+ "model.language_model.layers.60.linear_attn.in_proj_a": {
377
+ "bits": 16,
378
+ "data_type": "fp"
379
+ },
380
+ "model.language_model.layers.61.linear_attn.in_proj_b": {
381
+ "bits": 16,
382
+ "data_type": "fp"
383
+ },
384
+ "model.language_model.layers.61.linear_attn.in_proj_a": {
385
+ "bits": 16,
386
+ "data_type": "fp"
387
+ },
388
+ "model.language_model.layers.62.linear_attn.in_proj_b": {
389
+ "bits": 16,
390
+ "data_type": "fp"
391
+ },
392
+ "model.language_model.layers.62.linear_attn.in_proj_a": {
393
+ "bits": 16,
394
+ "data_type": "fp"
395
+ }
396
+ }
397
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4
3
+ size 19989343
tokenizer_config.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "audio_bos_token": "<|audio_start|>",
4
+ "audio_eos_token": "<|audio_end|>",
5
+ "audio_token": "<|audio_pad|>",
6
+ "backend": "tokenizers",
7
+ "bos_token": null,
8
+ "clean_up_tokenization_spaces": false,
9
+ "eos_token": "<|im_end|>",
10
+ "errors": "replace",
11
+ "image_token": "<|image_pad|>",
12
+ "is_local": false,
13
+ "local_files_only": false,
14
+ "model_max_length": 262144,
15
+ "model_specific_special_tokens": {
16
+ "audio_bos_token": "<|audio_start|>",
17
+ "audio_eos_token": "<|audio_end|>",
18
+ "audio_token": "<|audio_pad|>",
19
+ "image_token": "<|image_pad|>",
20
+ "video_token": "<|video_pad|>",
21
+ "vision_bos_token": "<|vision_start|>",
22
+ "vision_eos_token": "<|vision_end|>"
23
+ },
24
+ "pad_token": "<|endoftext|>",
25
+ "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
26
+ "processor_class": "Qwen3VLProcessor",
27
+ "split_special_tokens": false,
28
+ "tokenizer_class": "TokenizersBackend",
29
+ "unk_token": null,
30
+ "video_token": "<|video_pad|>",
31
+ "vision_bos_token": "<|vision_start|>",
32
+ "vision_eos_token": "<|vision_end|>"
33
+ }