Image-Text-to-Text
Transformers
Safetensors
qwen3_5
qwen
multimodal
mtp
speculative-decoding
compressed-tensors
llm-compressor
nvfp4
w4a16
conversational
8-bit precision
Instructions to use huginnfork/Qwen3.8-27B-NVFP4A16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use huginnfork/Qwen3.8-27B-NVFP4A16 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="huginnfork/Qwen3.8-27B-NVFP4A16") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] pipe(text=messages)# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("huginnfork/Qwen3.8-27B-NVFP4A16") model = AutoModelForMultimodalLM.from_pretrained("huginnfork/Qwen3.8-27B-NVFP4A16", device_map="auto") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] inputs = processor.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=256) print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use huginnfork/Qwen3.8-27B-NVFP4A16 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "huginnfork/Qwen3.8-27B-NVFP4A16" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "huginnfork/Qwen3.8-27B-NVFP4A16", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'Use Docker
docker model run hf.co/huginnfork/Qwen3.8-27B-NVFP4A16
- SGLang
How to use huginnfork/Qwen3.8-27B-NVFP4A16 with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "huginnfork/Qwen3.8-27B-NVFP4A16" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "huginnfork/Qwen3.8-27B-NVFP4A16", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "huginnfork/Qwen3.8-27B-NVFP4A16" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "huginnfork/Qwen3.8-27B-NVFP4A16", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }' - Docker Model Runner
How to use huginnfork/Qwen3.8-27B-NVFP4A16 with Docker Model Runner:
docker model run hf.co/huginnfork/Qwen3.8-27B-NVFP4A16
Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- chat_template.jinja +170 -0
- config.json +697 -0
- generation_config.json +13 -0
- kld_qwen3.8-27b-nvfp4a16.json +28 -0
- merges.txt +0 -0
- model-00001-of-00006.safetensors +3 -0
- model-00002-of-00006.safetensors +3 -0
- model-00003-of-00006.safetensors +3 -0
- model-00004-of-00006.safetensors +3 -0
- model-00005-of-00006.safetensors +3 -0
- model-00006-of-00006.safetensors +3 -0
- model.safetensors.index.json +0 -0
- ppl_qwen3.8-27b-base.json +9 -0
- ppl_qwen3.8-27b-nvfp4a16.json +9 -0
- preprocessor_config.json +21 -0
- recipe.yaml +41 -0
- smoke_qwen3.8-27b-nvfp4a16.json +8 -0
- tokenizer.json +3 -0
- tokenizer_config.json +32 -0
- video_preprocessor_config.json +21 -0
- 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
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,170 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
{%- set reasoning_instructions = '' %}
|
| 46 |
+
{%- if enable_thinking is undefined or enable_thinking is true %}
|
| 47 |
+
{%- set resolved_reasoning_effort = reasoning_effort|default('xhigh') %}
|
| 48 |
+
{%- if resolved_reasoning_effort not in ('xhigh', 'medium', 'low') %}
|
| 49 |
+
{{- raise_exception('Unexpected reasoning effort ' ~ reasoning_effort ~ '. Supported types are xhigh (default), medium, and low.') }}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{%- if resolved_reasoning_effort == 'xhigh' %}
|
| 52 |
+
{%- set reasoning_instructions = 'Reasoning effort is set to xhigh. Please think carefully through the task, validate key assumptions, consider plausible alternatives, and prioritize correctness, consistency, and clarity in the final answer.' %}
|
| 53 |
+
{%- elif resolved_reasoning_effort == 'low' %}
|
| 54 |
+
{%- set reasoning_instructions = 'Reasoning effort is set to low. Keep your thinking brief and focused, moving directly to the conclusion without unnecessary elaboration.' %}
|
| 55 |
+
{%- endif %}
|
| 56 |
+
{%- endif %}
|
| 57 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 58 |
+
{{- '<|im_start|>system\n' }}
|
| 59 |
+
{%- if reasoning_instructions %}
|
| 60 |
+
{{- reasoning_instructions + '\n\n' }}
|
| 61 |
+
{%- endif %}
|
| 62 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 63 |
+
{%- for tool in tools %}
|
| 64 |
+
{{- "\n" }}
|
| 65 |
+
{{- tool | tojson }}
|
| 66 |
+
{%- endfor %}
|
| 67 |
+
{{- "\n</tools>" }}
|
| 68 |
+
{{- '\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>' }}
|
| 69 |
+
{%- if messages[0].role == 'system' %}
|
| 70 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 71 |
+
{%- if content %}
|
| 72 |
+
{{- '\n\n' + content }}
|
| 73 |
+
{%- endif %}
|
| 74 |
+
{%- endif %}
|
| 75 |
+
{{- '<|im_end|>\n' }}
|
| 76 |
+
{%- else %}
|
| 77 |
+
{%- if messages[0].role == 'system' %}
|
| 78 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 79 |
+
{%- if content %}
|
| 80 |
+
{{- '<|im_start|>system\n' + (reasoning_instructions + '\n\n' if reasoning_instructions else '') + content + '<|im_end|>\n' }}
|
| 81 |
+
{%- elif reasoning_instructions %}
|
| 82 |
+
{{- '<|im_start|>system\n' + reasoning_instructions + '<|im_end|>\n' }}
|
| 83 |
+
{%- endif %}
|
| 84 |
+
{%- elif reasoning_instructions %}
|
| 85 |
+
{{- '<|im_start|>system\n' + reasoning_instructions + '<|im_end|>\n' }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- endif %}
|
| 88 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 89 |
+
{%- for message in messages[::-1] %}
|
| 90 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 91 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 92 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 93 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 94 |
+
{%- set ns.multi_step_tool = false %}
|
| 95 |
+
{%- set ns.last_query_index = index %}
|
| 96 |
+
{%- endif %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endfor %}
|
| 99 |
+
{%- if ns.multi_step_tool %}
|
| 100 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 101 |
+
{%- endif %}
|
| 102 |
+
{%- for message in messages %}
|
| 103 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 104 |
+
{%- if message.role == "system" %}
|
| 105 |
+
{%- if not loop.first %}
|
| 106 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 107 |
+
{%- endif %}
|
| 108 |
+
{%- elif message.role == "user" %}
|
| 109 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 110 |
+
{%- elif message.role == "assistant" %}
|
| 111 |
+
{%- set reasoning_content = '' %}
|
| 112 |
+
{%- if message.reasoning_content is string %}
|
| 113 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 114 |
+
{%- endif %}
|
| 115 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 116 |
+
{%- if preserve_thinking is undefined or preserve_thinking is true or loop.index0 > ns.last_query_index %}
|
| 117 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 118 |
+
{%- else %}
|
| 119 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 120 |
+
{%- endif %}
|
| 121 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 122 |
+
{%- for tool_call in message.tool_calls %}
|
| 123 |
+
{%- if tool_call.function is defined %}
|
| 124 |
+
{%- set tool_call = tool_call.function %}
|
| 125 |
+
{%- endif %}
|
| 126 |
+
{%- if loop.first %}
|
| 127 |
+
{%- if content|trim %}
|
| 128 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 129 |
+
{%- else %}
|
| 130 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 131 |
+
{%- endif %}
|
| 132 |
+
{%- else %}
|
| 133 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{%- if tool_call.arguments is defined and tool_call.arguments != '' %}
|
| 136 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 137 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 138 |
+
{%- set args_value = args_value | string if args_value is string else args_value | tojson | safe %}
|
| 139 |
+
{{- args_value }}
|
| 140 |
+
{{- '\n</parameter>\n' }}
|
| 141 |
+
{%- endfor %}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{{- '</function>\n</tool_call>' }}
|
| 144 |
+
{%- endfor %}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{{- '<|im_end|>\n' }}
|
| 147 |
+
{%- elif message.role == "tool" %}
|
| 148 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 149 |
+
{{- '<|im_start|>user' }}
|
| 150 |
+
{%- endif %}
|
| 151 |
+
{{- '\n<tool_response>\n' }}
|
| 152 |
+
{{- content }}
|
| 153 |
+
{{- '\n</tool_response>' }}
|
| 154 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 155 |
+
{{- '<|im_end|>\n' }}
|
| 156 |
+
{%- elif loop.last %}
|
| 157 |
+
{{- '<|im_end|>\n' }}
|
| 158 |
+
{%- endif %}
|
| 159 |
+
{%- else %}
|
| 160 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 161 |
+
{%- endif %}
|
| 162 |
+
{%- endfor %}
|
| 163 |
+
{%- if add_generation_prompt %}
|
| 164 |
+
{{- '<|im_start|>assistant\n' }}
|
| 165 |
+
{%- if enable_thinking is defined and enable_thinking is false %}
|
| 166 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 167 |
+
{%- else %}
|
| 168 |
+
{{- '<think>\n' }}
|
| 169 |
+
{%- endif %}
|
| 170 |
+
{%- endif %}
|
config.json
ADDED
|
@@ -0,0 +1,697 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3_5ForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"dtype": "bfloat16",
|
| 6 |
+
"image_token_id": 248056,
|
| 7 |
+
"language_model_only": false,
|
| 8 |
+
"model_type": "qwen3_5",
|
| 9 |
+
"quantization_config": {
|
| 10 |
+
"config_groups": {
|
| 11 |
+
"group_0": {
|
| 12 |
+
"format": "nvfp4-pack-quantized",
|
| 13 |
+
"input_activations": null,
|
| 14 |
+
"output_activations": null,
|
| 15 |
+
"targets": [
|
| 16 |
+
"Linear"
|
| 17 |
+
],
|
| 18 |
+
"weights": {
|
| 19 |
+
"actorder": null,
|
| 20 |
+
"block_structure": null,
|
| 21 |
+
"dynamic": false,
|
| 22 |
+
"group_size": 16,
|
| 23 |
+
"num_bits": 4,
|
| 24 |
+
"observer": "memoryless_minmax",
|
| 25 |
+
"observer_kwargs": {},
|
| 26 |
+
"scale_dtype": "torch.float8_e4m3fn",
|
| 27 |
+
"strategy": "tensor_group",
|
| 28 |
+
"symmetric": true,
|
| 29 |
+
"type": "float",
|
| 30 |
+
"zp_dtype": null
|
| 31 |
+
}
|
| 32 |
+
}
|
| 33 |
+
},
|
| 34 |
+
"format": "nvfp4-pack-quantized",
|
| 35 |
+
"global_compression_ratio": null,
|
| 36 |
+
"ignore": [
|
| 37 |
+
"model.visual.blocks.0.attn.qkv",
|
| 38 |
+
"model.visual.blocks.0.attn.proj",
|
| 39 |
+
"model.visual.blocks.0.mlp.linear_fc1",
|
| 40 |
+
"model.visual.blocks.0.mlp.linear_fc2",
|
| 41 |
+
"model.visual.blocks.1.attn.qkv",
|
| 42 |
+
"model.visual.blocks.1.attn.proj",
|
| 43 |
+
"model.visual.blocks.1.mlp.linear_fc1",
|
| 44 |
+
"model.visual.blocks.1.mlp.linear_fc2",
|
| 45 |
+
"model.visual.blocks.2.attn.qkv",
|
| 46 |
+
"model.visual.blocks.2.attn.proj",
|
| 47 |
+
"model.visual.blocks.2.mlp.linear_fc1",
|
| 48 |
+
"model.visual.blocks.2.mlp.linear_fc2",
|
| 49 |
+
"model.visual.blocks.3.attn.qkv",
|
| 50 |
+
"model.visual.blocks.3.attn.proj",
|
| 51 |
+
"model.visual.blocks.3.mlp.linear_fc1",
|
| 52 |
+
"model.visual.blocks.3.mlp.linear_fc2",
|
| 53 |
+
"model.visual.blocks.4.attn.qkv",
|
| 54 |
+
"model.visual.blocks.4.attn.proj",
|
| 55 |
+
"model.visual.blocks.4.mlp.linear_fc1",
|
| 56 |
+
"model.visual.blocks.4.mlp.linear_fc2",
|
| 57 |
+
"model.visual.blocks.5.attn.qkv",
|
| 58 |
+
"model.visual.blocks.5.attn.proj",
|
| 59 |
+
"model.visual.blocks.5.mlp.linear_fc1",
|
| 60 |
+
"model.visual.blocks.5.mlp.linear_fc2",
|
| 61 |
+
"model.visual.blocks.6.attn.qkv",
|
| 62 |
+
"model.visual.blocks.6.attn.proj",
|
| 63 |
+
"model.visual.blocks.6.mlp.linear_fc1",
|
| 64 |
+
"model.visual.blocks.6.mlp.linear_fc2",
|
| 65 |
+
"model.visual.blocks.7.attn.qkv",
|
| 66 |
+
"model.visual.blocks.7.attn.proj",
|
| 67 |
+
"model.visual.blocks.7.mlp.linear_fc1",
|
| 68 |
+
"model.visual.blocks.7.mlp.linear_fc2",
|
| 69 |
+
"model.visual.blocks.8.attn.qkv",
|
| 70 |
+
"model.visual.blocks.8.attn.proj",
|
| 71 |
+
"model.visual.blocks.8.mlp.linear_fc1",
|
| 72 |
+
"model.visual.blocks.8.mlp.linear_fc2",
|
| 73 |
+
"model.visual.blocks.9.attn.qkv",
|
| 74 |
+
"model.visual.blocks.9.attn.proj",
|
| 75 |
+
"model.visual.blocks.9.mlp.linear_fc1",
|
| 76 |
+
"model.visual.blocks.9.mlp.linear_fc2",
|
| 77 |
+
"model.visual.blocks.10.attn.qkv",
|
| 78 |
+
"model.visual.blocks.10.attn.proj",
|
| 79 |
+
"model.visual.blocks.10.mlp.linear_fc1",
|
| 80 |
+
"model.visual.blocks.10.mlp.linear_fc2",
|
| 81 |
+
"model.visual.blocks.11.attn.qkv",
|
| 82 |
+
"model.visual.blocks.11.attn.proj",
|
| 83 |
+
"model.visual.blocks.11.mlp.linear_fc1",
|
| 84 |
+
"model.visual.blocks.11.mlp.linear_fc2",
|
| 85 |
+
"model.visual.blocks.12.attn.qkv",
|
| 86 |
+
"model.visual.blocks.12.attn.proj",
|
| 87 |
+
"model.visual.blocks.12.mlp.linear_fc1",
|
| 88 |
+
"model.visual.blocks.12.mlp.linear_fc2",
|
| 89 |
+
"model.visual.blocks.13.attn.qkv",
|
| 90 |
+
"model.visual.blocks.13.attn.proj",
|
| 91 |
+
"model.visual.blocks.13.mlp.linear_fc1",
|
| 92 |
+
"model.visual.blocks.13.mlp.linear_fc2",
|
| 93 |
+
"model.visual.blocks.14.attn.qkv",
|
| 94 |
+
"model.visual.blocks.14.attn.proj",
|
| 95 |
+
"model.visual.blocks.14.mlp.linear_fc1",
|
| 96 |
+
"model.visual.blocks.14.mlp.linear_fc2",
|
| 97 |
+
"model.visual.blocks.15.attn.qkv",
|
| 98 |
+
"model.visual.blocks.15.attn.proj",
|
| 99 |
+
"model.visual.blocks.15.mlp.linear_fc1",
|
| 100 |
+
"model.visual.blocks.15.mlp.linear_fc2",
|
| 101 |
+
"model.visual.blocks.16.attn.qkv",
|
| 102 |
+
"model.visual.blocks.16.attn.proj",
|
| 103 |
+
"model.visual.blocks.16.mlp.linear_fc1",
|
| 104 |
+
"model.visual.blocks.16.mlp.linear_fc2",
|
| 105 |
+
"model.visual.blocks.17.attn.qkv",
|
| 106 |
+
"model.visual.blocks.17.attn.proj",
|
| 107 |
+
"model.visual.blocks.17.mlp.linear_fc1",
|
| 108 |
+
"model.visual.blocks.17.mlp.linear_fc2",
|
| 109 |
+
"model.visual.blocks.18.attn.qkv",
|
| 110 |
+
"model.visual.blocks.18.attn.proj",
|
| 111 |
+
"model.visual.blocks.18.mlp.linear_fc1",
|
| 112 |
+
"model.visual.blocks.18.mlp.linear_fc2",
|
| 113 |
+
"model.visual.blocks.19.attn.qkv",
|
| 114 |
+
"model.visual.blocks.19.attn.proj",
|
| 115 |
+
"model.visual.blocks.19.mlp.linear_fc1",
|
| 116 |
+
"model.visual.blocks.19.mlp.linear_fc2",
|
| 117 |
+
"model.visual.blocks.20.attn.qkv",
|
| 118 |
+
"model.visual.blocks.20.attn.proj",
|
| 119 |
+
"model.visual.blocks.20.mlp.linear_fc1",
|
| 120 |
+
"model.visual.blocks.20.mlp.linear_fc2",
|
| 121 |
+
"model.visual.blocks.21.attn.qkv",
|
| 122 |
+
"model.visual.blocks.21.attn.proj",
|
| 123 |
+
"model.visual.blocks.21.mlp.linear_fc1",
|
| 124 |
+
"model.visual.blocks.21.mlp.linear_fc2",
|
| 125 |
+
"model.visual.blocks.22.attn.qkv",
|
| 126 |
+
"model.visual.blocks.22.attn.proj",
|
| 127 |
+
"model.visual.blocks.22.mlp.linear_fc1",
|
| 128 |
+
"model.visual.blocks.22.mlp.linear_fc2",
|
| 129 |
+
"model.visual.blocks.23.attn.qkv",
|
| 130 |
+
"model.visual.blocks.23.attn.proj",
|
| 131 |
+
"model.visual.blocks.23.mlp.linear_fc1",
|
| 132 |
+
"model.visual.blocks.23.mlp.linear_fc2",
|
| 133 |
+
"model.visual.blocks.24.attn.qkv",
|
| 134 |
+
"model.visual.blocks.24.attn.proj",
|
| 135 |
+
"model.visual.blocks.24.mlp.linear_fc1",
|
| 136 |
+
"model.visual.blocks.24.mlp.linear_fc2",
|
| 137 |
+
"model.visual.blocks.25.attn.qkv",
|
| 138 |
+
"model.visual.blocks.25.attn.proj",
|
| 139 |
+
"model.visual.blocks.25.mlp.linear_fc1",
|
| 140 |
+
"model.visual.blocks.25.mlp.linear_fc2",
|
| 141 |
+
"model.visual.blocks.26.attn.qkv",
|
| 142 |
+
"model.visual.blocks.26.attn.proj",
|
| 143 |
+
"model.visual.blocks.26.mlp.linear_fc1",
|
| 144 |
+
"model.visual.blocks.26.mlp.linear_fc2",
|
| 145 |
+
"model.visual.merger.linear_fc1",
|
| 146 |
+
"model.visual.merger.linear_fc2",
|
| 147 |
+
"model.language_model.layers.0.linear_attn",
|
| 148 |
+
"model.language_model.layers.0.linear_attn.norm",
|
| 149 |
+
"model.language_model.layers.0.linear_attn.out_proj",
|
| 150 |
+
"model.language_model.layers.0.linear_attn.in_proj_qkv",
|
| 151 |
+
"model.language_model.layers.0.linear_attn.in_proj_z",
|
| 152 |
+
"model.language_model.layers.0.linear_attn.in_proj_b",
|
| 153 |
+
"model.language_model.layers.0.linear_attn.in_proj_a",
|
| 154 |
+
"model.language_model.layers.1.linear_attn",
|
| 155 |
+
"model.language_model.layers.1.linear_attn.norm",
|
| 156 |
+
"model.language_model.layers.1.linear_attn.out_proj",
|
| 157 |
+
"model.language_model.layers.1.linear_attn.in_proj_qkv",
|
| 158 |
+
"model.language_model.layers.1.linear_attn.in_proj_z",
|
| 159 |
+
"model.language_model.layers.1.linear_attn.in_proj_b",
|
| 160 |
+
"model.language_model.layers.1.linear_attn.in_proj_a",
|
| 161 |
+
"model.language_model.layers.2.linear_attn",
|
| 162 |
+
"model.language_model.layers.2.linear_attn.norm",
|
| 163 |
+
"model.language_model.layers.2.linear_attn.out_proj",
|
| 164 |
+
"model.language_model.layers.2.linear_attn.in_proj_qkv",
|
| 165 |
+
"model.language_model.layers.2.linear_attn.in_proj_z",
|
| 166 |
+
"model.language_model.layers.2.linear_attn.in_proj_b",
|
| 167 |
+
"model.language_model.layers.2.linear_attn.in_proj_a",
|
| 168 |
+
"model.language_model.layers.3.self_attn.q_proj",
|
| 169 |
+
"model.language_model.layers.3.self_attn.k_proj",
|
| 170 |
+
"model.language_model.layers.3.self_attn.v_proj",
|
| 171 |
+
"model.language_model.layers.3.self_attn.o_proj",
|
| 172 |
+
"model.language_model.layers.4.linear_attn",
|
| 173 |
+
"model.language_model.layers.4.linear_attn.norm",
|
| 174 |
+
"model.language_model.layers.4.linear_attn.out_proj",
|
| 175 |
+
"model.language_model.layers.4.linear_attn.in_proj_qkv",
|
| 176 |
+
"model.language_model.layers.4.linear_attn.in_proj_z",
|
| 177 |
+
"model.language_model.layers.4.linear_attn.in_proj_b",
|
| 178 |
+
"model.language_model.layers.4.linear_attn.in_proj_a",
|
| 179 |
+
"model.language_model.layers.5.linear_attn",
|
| 180 |
+
"model.language_model.layers.5.linear_attn.norm",
|
| 181 |
+
"model.language_model.layers.5.linear_attn.out_proj",
|
| 182 |
+
"model.language_model.layers.5.linear_attn.in_proj_qkv",
|
| 183 |
+
"model.language_model.layers.5.linear_attn.in_proj_z",
|
| 184 |
+
"model.language_model.layers.5.linear_attn.in_proj_b",
|
| 185 |
+
"model.language_model.layers.5.linear_attn.in_proj_a",
|
| 186 |
+
"model.language_model.layers.6.linear_attn",
|
| 187 |
+
"model.language_model.layers.6.linear_attn.norm",
|
| 188 |
+
"model.language_model.layers.6.linear_attn.out_proj",
|
| 189 |
+
"model.language_model.layers.6.linear_attn.in_proj_qkv",
|
| 190 |
+
"model.language_model.layers.6.linear_attn.in_proj_z",
|
| 191 |
+
"model.language_model.layers.6.linear_attn.in_proj_b",
|
| 192 |
+
"model.language_model.layers.6.linear_attn.in_proj_a",
|
| 193 |
+
"model.language_model.layers.7.self_attn.q_proj",
|
| 194 |
+
"model.language_model.layers.7.self_attn.k_proj",
|
| 195 |
+
"model.language_model.layers.7.self_attn.v_proj",
|
| 196 |
+
"model.language_model.layers.7.self_attn.o_proj",
|
| 197 |
+
"model.language_model.layers.8.linear_attn",
|
| 198 |
+
"model.language_model.layers.8.linear_attn.norm",
|
| 199 |
+
"model.language_model.layers.8.linear_attn.out_proj",
|
| 200 |
+
"model.language_model.layers.8.linear_attn.in_proj_qkv",
|
| 201 |
+
"model.language_model.layers.8.linear_attn.in_proj_z",
|
| 202 |
+
"model.language_model.layers.8.linear_attn.in_proj_b",
|
| 203 |
+
"model.language_model.layers.8.linear_attn.in_proj_a",
|
| 204 |
+
"model.language_model.layers.9.linear_attn",
|
| 205 |
+
"model.language_model.layers.9.linear_attn.norm",
|
| 206 |
+
"model.language_model.layers.9.linear_attn.out_proj",
|
| 207 |
+
"model.language_model.layers.9.linear_attn.in_proj_qkv",
|
| 208 |
+
"model.language_model.layers.9.linear_attn.in_proj_z",
|
| 209 |
+
"model.language_model.layers.9.linear_attn.in_proj_b",
|
| 210 |
+
"model.language_model.layers.9.linear_attn.in_proj_a",
|
| 211 |
+
"model.language_model.layers.10.linear_attn",
|
| 212 |
+
"model.language_model.layers.10.linear_attn.norm",
|
| 213 |
+
"model.language_model.layers.10.linear_attn.out_proj",
|
| 214 |
+
"model.language_model.layers.10.linear_attn.in_proj_qkv",
|
| 215 |
+
"model.language_model.layers.10.linear_attn.in_proj_z",
|
| 216 |
+
"model.language_model.layers.10.linear_attn.in_proj_b",
|
| 217 |
+
"model.language_model.layers.10.linear_attn.in_proj_a",
|
| 218 |
+
"model.language_model.layers.11.self_attn.q_proj",
|
| 219 |
+
"model.language_model.layers.11.self_attn.k_proj",
|
| 220 |
+
"model.language_model.layers.11.self_attn.v_proj",
|
| 221 |
+
"model.language_model.layers.11.self_attn.o_proj",
|
| 222 |
+
"model.language_model.layers.12.linear_attn",
|
| 223 |
+
"model.language_model.layers.12.linear_attn.norm",
|
| 224 |
+
"model.language_model.layers.12.linear_attn.out_proj",
|
| 225 |
+
"model.language_model.layers.12.linear_attn.in_proj_qkv",
|
| 226 |
+
"model.language_model.layers.12.linear_attn.in_proj_z",
|
| 227 |
+
"model.language_model.layers.12.linear_attn.in_proj_b",
|
| 228 |
+
"model.language_model.layers.12.linear_attn.in_proj_a",
|
| 229 |
+
"model.language_model.layers.13.linear_attn",
|
| 230 |
+
"model.language_model.layers.13.linear_attn.norm",
|
| 231 |
+
"model.language_model.layers.13.linear_attn.out_proj",
|
| 232 |
+
"model.language_model.layers.13.linear_attn.in_proj_qkv",
|
| 233 |
+
"model.language_model.layers.13.linear_attn.in_proj_z",
|
| 234 |
+
"model.language_model.layers.13.linear_attn.in_proj_b",
|
| 235 |
+
"model.language_model.layers.13.linear_attn.in_proj_a",
|
| 236 |
+
"model.language_model.layers.14.linear_attn",
|
| 237 |
+
"model.language_model.layers.14.linear_attn.norm",
|
| 238 |
+
"model.language_model.layers.14.linear_attn.out_proj",
|
| 239 |
+
"model.language_model.layers.14.linear_attn.in_proj_qkv",
|
| 240 |
+
"model.language_model.layers.14.linear_attn.in_proj_z",
|
| 241 |
+
"model.language_model.layers.14.linear_attn.in_proj_b",
|
| 242 |
+
"model.language_model.layers.14.linear_attn.in_proj_a",
|
| 243 |
+
"model.language_model.layers.15.self_attn.q_proj",
|
| 244 |
+
"model.language_model.layers.15.self_attn.k_proj",
|
| 245 |
+
"model.language_model.layers.15.self_attn.v_proj",
|
| 246 |
+
"model.language_model.layers.15.self_attn.o_proj",
|
| 247 |
+
"model.language_model.layers.16.linear_attn",
|
| 248 |
+
"model.language_model.layers.16.linear_attn.norm",
|
| 249 |
+
"model.language_model.layers.16.linear_attn.out_proj",
|
| 250 |
+
"model.language_model.layers.16.linear_attn.in_proj_qkv",
|
| 251 |
+
"model.language_model.layers.16.linear_attn.in_proj_z",
|
| 252 |
+
"model.language_model.layers.16.linear_attn.in_proj_b",
|
| 253 |
+
"model.language_model.layers.16.linear_attn.in_proj_a",
|
| 254 |
+
"model.language_model.layers.17.linear_attn",
|
| 255 |
+
"model.language_model.layers.17.linear_attn.norm",
|
| 256 |
+
"model.language_model.layers.17.linear_attn.out_proj",
|
| 257 |
+
"model.language_model.layers.17.linear_attn.in_proj_qkv",
|
| 258 |
+
"model.language_model.layers.17.linear_attn.in_proj_z",
|
| 259 |
+
"model.language_model.layers.17.linear_attn.in_proj_b",
|
| 260 |
+
"model.language_model.layers.17.linear_attn.in_proj_a",
|
| 261 |
+
"model.language_model.layers.18.linear_attn",
|
| 262 |
+
"model.language_model.layers.18.linear_attn.norm",
|
| 263 |
+
"model.language_model.layers.18.linear_attn.out_proj",
|
| 264 |
+
"model.language_model.layers.18.linear_attn.in_proj_qkv",
|
| 265 |
+
"model.language_model.layers.18.linear_attn.in_proj_z",
|
| 266 |
+
"model.language_model.layers.18.linear_attn.in_proj_b",
|
| 267 |
+
"model.language_model.layers.18.linear_attn.in_proj_a",
|
| 268 |
+
"model.language_model.layers.19.self_attn.q_proj",
|
| 269 |
+
"model.language_model.layers.19.self_attn.k_proj",
|
| 270 |
+
"model.language_model.layers.19.self_attn.v_proj",
|
| 271 |
+
"model.language_model.layers.19.self_attn.o_proj",
|
| 272 |
+
"model.language_model.layers.20.linear_attn",
|
| 273 |
+
"model.language_model.layers.20.linear_attn.norm",
|
| 274 |
+
"model.language_model.layers.20.linear_attn.out_proj",
|
| 275 |
+
"model.language_model.layers.20.linear_attn.in_proj_qkv",
|
| 276 |
+
"model.language_model.layers.20.linear_attn.in_proj_z",
|
| 277 |
+
"model.language_model.layers.20.linear_attn.in_proj_b",
|
| 278 |
+
"model.language_model.layers.20.linear_attn.in_proj_a",
|
| 279 |
+
"model.language_model.layers.21.linear_attn",
|
| 280 |
+
"model.language_model.layers.21.linear_attn.norm",
|
| 281 |
+
"model.language_model.layers.21.linear_attn.out_proj",
|
| 282 |
+
"model.language_model.layers.21.linear_attn.in_proj_qkv",
|
| 283 |
+
"model.language_model.layers.21.linear_attn.in_proj_z",
|
| 284 |
+
"model.language_model.layers.21.linear_attn.in_proj_b",
|
| 285 |
+
"model.language_model.layers.21.linear_attn.in_proj_a",
|
| 286 |
+
"model.language_model.layers.22.linear_attn",
|
| 287 |
+
"model.language_model.layers.22.linear_attn.norm",
|
| 288 |
+
"model.language_model.layers.22.linear_attn.out_proj",
|
| 289 |
+
"model.language_model.layers.22.linear_attn.in_proj_qkv",
|
| 290 |
+
"model.language_model.layers.22.linear_attn.in_proj_z",
|
| 291 |
+
"model.language_model.layers.22.linear_attn.in_proj_b",
|
| 292 |
+
"model.language_model.layers.22.linear_attn.in_proj_a",
|
| 293 |
+
"model.language_model.layers.23.self_attn.q_proj",
|
| 294 |
+
"model.language_model.layers.23.self_attn.k_proj",
|
| 295 |
+
"model.language_model.layers.23.self_attn.v_proj",
|
| 296 |
+
"model.language_model.layers.23.self_attn.o_proj",
|
| 297 |
+
"model.language_model.layers.24.linear_attn",
|
| 298 |
+
"model.language_model.layers.24.linear_attn.norm",
|
| 299 |
+
"model.language_model.layers.24.linear_attn.out_proj",
|
| 300 |
+
"model.language_model.layers.24.linear_attn.in_proj_qkv",
|
| 301 |
+
"model.language_model.layers.24.linear_attn.in_proj_z",
|
| 302 |
+
"model.language_model.layers.24.linear_attn.in_proj_b",
|
| 303 |
+
"model.language_model.layers.24.linear_attn.in_proj_a",
|
| 304 |
+
"model.language_model.layers.25.linear_attn",
|
| 305 |
+
"model.language_model.layers.25.linear_attn.norm",
|
| 306 |
+
"model.language_model.layers.25.linear_attn.out_proj",
|
| 307 |
+
"model.language_model.layers.25.linear_attn.in_proj_qkv",
|
| 308 |
+
"model.language_model.layers.25.linear_attn.in_proj_z",
|
| 309 |
+
"model.language_model.layers.25.linear_attn.in_proj_b",
|
| 310 |
+
"model.language_model.layers.25.linear_attn.in_proj_a",
|
| 311 |
+
"model.language_model.layers.26.linear_attn",
|
| 312 |
+
"model.language_model.layers.26.linear_attn.norm",
|
| 313 |
+
"model.language_model.layers.26.linear_attn.out_proj",
|
| 314 |
+
"model.language_model.layers.26.linear_attn.in_proj_qkv",
|
| 315 |
+
"model.language_model.layers.26.linear_attn.in_proj_z",
|
| 316 |
+
"model.language_model.layers.26.linear_attn.in_proj_b",
|
| 317 |
+
"model.language_model.layers.26.linear_attn.in_proj_a",
|
| 318 |
+
"model.language_model.layers.27.self_attn.q_proj",
|
| 319 |
+
"model.language_model.layers.27.self_attn.k_proj",
|
| 320 |
+
"model.language_model.layers.27.self_attn.v_proj",
|
| 321 |
+
"model.language_model.layers.27.self_attn.o_proj",
|
| 322 |
+
"model.language_model.layers.28.linear_attn",
|
| 323 |
+
"model.language_model.layers.28.linear_attn.norm",
|
| 324 |
+
"model.language_model.layers.28.linear_attn.out_proj",
|
| 325 |
+
"model.language_model.layers.28.linear_attn.in_proj_qkv",
|
| 326 |
+
"model.language_model.layers.28.linear_attn.in_proj_z",
|
| 327 |
+
"model.language_model.layers.28.linear_attn.in_proj_b",
|
| 328 |
+
"model.language_model.layers.28.linear_attn.in_proj_a",
|
| 329 |
+
"model.language_model.layers.29.linear_attn",
|
| 330 |
+
"model.language_model.layers.29.linear_attn.norm",
|
| 331 |
+
"model.language_model.layers.29.linear_attn.out_proj",
|
| 332 |
+
"model.language_model.layers.29.linear_attn.in_proj_qkv",
|
| 333 |
+
"model.language_model.layers.29.linear_attn.in_proj_z",
|
| 334 |
+
"model.language_model.layers.29.linear_attn.in_proj_b",
|
| 335 |
+
"model.language_model.layers.29.linear_attn.in_proj_a",
|
| 336 |
+
"model.language_model.layers.30.linear_attn",
|
| 337 |
+
"model.language_model.layers.30.linear_attn.norm",
|
| 338 |
+
"model.language_model.layers.30.linear_attn.out_proj",
|
| 339 |
+
"model.language_model.layers.30.linear_attn.in_proj_qkv",
|
| 340 |
+
"model.language_model.layers.30.linear_attn.in_proj_z",
|
| 341 |
+
"model.language_model.layers.30.linear_attn.in_proj_b",
|
| 342 |
+
"model.language_model.layers.30.linear_attn.in_proj_a",
|
| 343 |
+
"model.language_model.layers.31.self_attn.q_proj",
|
| 344 |
+
"model.language_model.layers.31.self_attn.k_proj",
|
| 345 |
+
"model.language_model.layers.31.self_attn.v_proj",
|
| 346 |
+
"model.language_model.layers.31.self_attn.o_proj",
|
| 347 |
+
"model.language_model.layers.32.linear_attn",
|
| 348 |
+
"model.language_model.layers.32.linear_attn.norm",
|
| 349 |
+
"model.language_model.layers.32.linear_attn.out_proj",
|
| 350 |
+
"model.language_model.layers.32.linear_attn.in_proj_qkv",
|
| 351 |
+
"model.language_model.layers.32.linear_attn.in_proj_z",
|
| 352 |
+
"model.language_model.layers.32.linear_attn.in_proj_b",
|
| 353 |
+
"model.language_model.layers.32.linear_attn.in_proj_a",
|
| 354 |
+
"model.language_model.layers.33.linear_attn",
|
| 355 |
+
"model.language_model.layers.33.linear_attn.norm",
|
| 356 |
+
"model.language_model.layers.33.linear_attn.out_proj",
|
| 357 |
+
"model.language_model.layers.33.linear_attn.in_proj_qkv",
|
| 358 |
+
"model.language_model.layers.33.linear_attn.in_proj_z",
|
| 359 |
+
"model.language_model.layers.33.linear_attn.in_proj_b",
|
| 360 |
+
"model.language_model.layers.33.linear_attn.in_proj_a",
|
| 361 |
+
"model.language_model.layers.34.linear_attn",
|
| 362 |
+
"model.language_model.layers.34.linear_attn.norm",
|
| 363 |
+
"model.language_model.layers.34.linear_attn.out_proj",
|
| 364 |
+
"model.language_model.layers.34.linear_attn.in_proj_qkv",
|
| 365 |
+
"model.language_model.layers.34.linear_attn.in_proj_z",
|
| 366 |
+
"model.language_model.layers.34.linear_attn.in_proj_b",
|
| 367 |
+
"model.language_model.layers.34.linear_attn.in_proj_a",
|
| 368 |
+
"model.language_model.layers.35.self_attn.q_proj",
|
| 369 |
+
"model.language_model.layers.35.self_attn.k_proj",
|
| 370 |
+
"model.language_model.layers.35.self_attn.v_proj",
|
| 371 |
+
"model.language_model.layers.35.self_attn.o_proj",
|
| 372 |
+
"model.language_model.layers.36.linear_attn",
|
| 373 |
+
"model.language_model.layers.36.linear_attn.norm",
|
| 374 |
+
"model.language_model.layers.36.linear_attn.out_proj",
|
| 375 |
+
"model.language_model.layers.36.linear_attn.in_proj_qkv",
|
| 376 |
+
"model.language_model.layers.36.linear_attn.in_proj_z",
|
| 377 |
+
"model.language_model.layers.36.linear_attn.in_proj_b",
|
| 378 |
+
"model.language_model.layers.36.linear_attn.in_proj_a",
|
| 379 |
+
"model.language_model.layers.37.linear_attn",
|
| 380 |
+
"model.language_model.layers.37.linear_attn.norm",
|
| 381 |
+
"model.language_model.layers.37.linear_attn.out_proj",
|
| 382 |
+
"model.language_model.layers.37.linear_attn.in_proj_qkv",
|
| 383 |
+
"model.language_model.layers.37.linear_attn.in_proj_z",
|
| 384 |
+
"model.language_model.layers.37.linear_attn.in_proj_b",
|
| 385 |
+
"model.language_model.layers.37.linear_attn.in_proj_a",
|
| 386 |
+
"model.language_model.layers.38.linear_attn",
|
| 387 |
+
"model.language_model.layers.38.linear_attn.norm",
|
| 388 |
+
"model.language_model.layers.38.linear_attn.out_proj",
|
| 389 |
+
"model.language_model.layers.38.linear_attn.in_proj_qkv",
|
| 390 |
+
"model.language_model.layers.38.linear_attn.in_proj_z",
|
| 391 |
+
"model.language_model.layers.38.linear_attn.in_proj_b",
|
| 392 |
+
"model.language_model.layers.38.linear_attn.in_proj_a",
|
| 393 |
+
"model.language_model.layers.39.self_attn.q_proj",
|
| 394 |
+
"model.language_model.layers.39.self_attn.k_proj",
|
| 395 |
+
"model.language_model.layers.39.self_attn.v_proj",
|
| 396 |
+
"model.language_model.layers.39.self_attn.o_proj",
|
| 397 |
+
"model.language_model.layers.40.linear_attn",
|
| 398 |
+
"model.language_model.layers.40.linear_attn.norm",
|
| 399 |
+
"model.language_model.layers.40.linear_attn.out_proj",
|
| 400 |
+
"model.language_model.layers.40.linear_attn.in_proj_qkv",
|
| 401 |
+
"model.language_model.layers.40.linear_attn.in_proj_z",
|
| 402 |
+
"model.language_model.layers.40.linear_attn.in_proj_b",
|
| 403 |
+
"model.language_model.layers.40.linear_attn.in_proj_a",
|
| 404 |
+
"model.language_model.layers.41.linear_attn",
|
| 405 |
+
"model.language_model.layers.41.linear_attn.norm",
|
| 406 |
+
"model.language_model.layers.41.linear_attn.out_proj",
|
| 407 |
+
"model.language_model.layers.41.linear_attn.in_proj_qkv",
|
| 408 |
+
"model.language_model.layers.41.linear_attn.in_proj_z",
|
| 409 |
+
"model.language_model.layers.41.linear_attn.in_proj_b",
|
| 410 |
+
"model.language_model.layers.41.linear_attn.in_proj_a",
|
| 411 |
+
"model.language_model.layers.42.linear_attn",
|
| 412 |
+
"model.language_model.layers.42.linear_attn.norm",
|
| 413 |
+
"model.language_model.layers.42.linear_attn.out_proj",
|
| 414 |
+
"model.language_model.layers.42.linear_attn.in_proj_qkv",
|
| 415 |
+
"model.language_model.layers.42.linear_attn.in_proj_z",
|
| 416 |
+
"model.language_model.layers.42.linear_attn.in_proj_b",
|
| 417 |
+
"model.language_model.layers.42.linear_attn.in_proj_a",
|
| 418 |
+
"model.language_model.layers.43.self_attn.q_proj",
|
| 419 |
+
"model.language_model.layers.43.self_attn.k_proj",
|
| 420 |
+
"model.language_model.layers.43.self_attn.v_proj",
|
| 421 |
+
"model.language_model.layers.43.self_attn.o_proj",
|
| 422 |
+
"model.language_model.layers.44.linear_attn",
|
| 423 |
+
"model.language_model.layers.44.linear_attn.norm",
|
| 424 |
+
"model.language_model.layers.44.linear_attn.out_proj",
|
| 425 |
+
"model.language_model.layers.44.linear_attn.in_proj_qkv",
|
| 426 |
+
"model.language_model.layers.44.linear_attn.in_proj_z",
|
| 427 |
+
"model.language_model.layers.44.linear_attn.in_proj_b",
|
| 428 |
+
"model.language_model.layers.44.linear_attn.in_proj_a",
|
| 429 |
+
"model.language_model.layers.45.linear_attn",
|
| 430 |
+
"model.language_model.layers.45.linear_attn.norm",
|
| 431 |
+
"model.language_model.layers.45.linear_attn.out_proj",
|
| 432 |
+
"model.language_model.layers.45.linear_attn.in_proj_qkv",
|
| 433 |
+
"model.language_model.layers.45.linear_attn.in_proj_z",
|
| 434 |
+
"model.language_model.layers.45.linear_attn.in_proj_b",
|
| 435 |
+
"model.language_model.layers.45.linear_attn.in_proj_a",
|
| 436 |
+
"model.language_model.layers.46.linear_attn",
|
| 437 |
+
"model.language_model.layers.46.linear_attn.norm",
|
| 438 |
+
"model.language_model.layers.46.linear_attn.out_proj",
|
| 439 |
+
"model.language_model.layers.46.linear_attn.in_proj_qkv",
|
| 440 |
+
"model.language_model.layers.46.linear_attn.in_proj_z",
|
| 441 |
+
"model.language_model.layers.46.linear_attn.in_proj_b",
|
| 442 |
+
"model.language_model.layers.46.linear_attn.in_proj_a",
|
| 443 |
+
"model.language_model.layers.47.self_attn.q_proj",
|
| 444 |
+
"model.language_model.layers.47.self_attn.k_proj",
|
| 445 |
+
"model.language_model.layers.47.self_attn.v_proj",
|
| 446 |
+
"model.language_model.layers.47.self_attn.o_proj",
|
| 447 |
+
"model.language_model.layers.48.linear_attn",
|
| 448 |
+
"model.language_model.layers.48.linear_attn.norm",
|
| 449 |
+
"model.language_model.layers.48.linear_attn.out_proj",
|
| 450 |
+
"model.language_model.layers.48.linear_attn.in_proj_qkv",
|
| 451 |
+
"model.language_model.layers.48.linear_attn.in_proj_z",
|
| 452 |
+
"model.language_model.layers.48.linear_attn.in_proj_b",
|
| 453 |
+
"model.language_model.layers.48.linear_attn.in_proj_a",
|
| 454 |
+
"model.language_model.layers.49.linear_attn",
|
| 455 |
+
"model.language_model.layers.49.linear_attn.norm",
|
| 456 |
+
"model.language_model.layers.49.linear_attn.out_proj",
|
| 457 |
+
"model.language_model.layers.49.linear_attn.in_proj_qkv",
|
| 458 |
+
"model.language_model.layers.49.linear_attn.in_proj_z",
|
| 459 |
+
"model.language_model.layers.49.linear_attn.in_proj_b",
|
| 460 |
+
"model.language_model.layers.49.linear_attn.in_proj_a",
|
| 461 |
+
"model.language_model.layers.50.linear_attn",
|
| 462 |
+
"model.language_model.layers.50.linear_attn.norm",
|
| 463 |
+
"model.language_model.layers.50.linear_attn.out_proj",
|
| 464 |
+
"model.language_model.layers.50.linear_attn.in_proj_qkv",
|
| 465 |
+
"model.language_model.layers.50.linear_attn.in_proj_z",
|
| 466 |
+
"model.language_model.layers.50.linear_attn.in_proj_b",
|
| 467 |
+
"model.language_model.layers.50.linear_attn.in_proj_a",
|
| 468 |
+
"model.language_model.layers.51.self_attn.q_proj",
|
| 469 |
+
"model.language_model.layers.51.self_attn.k_proj",
|
| 470 |
+
"model.language_model.layers.51.self_attn.v_proj",
|
| 471 |
+
"model.language_model.layers.51.self_attn.o_proj",
|
| 472 |
+
"model.language_model.layers.52.linear_attn",
|
| 473 |
+
"model.language_model.layers.52.linear_attn.norm",
|
| 474 |
+
"model.language_model.layers.52.linear_attn.out_proj",
|
| 475 |
+
"model.language_model.layers.52.linear_attn.in_proj_qkv",
|
| 476 |
+
"model.language_model.layers.52.linear_attn.in_proj_z",
|
| 477 |
+
"model.language_model.layers.52.linear_attn.in_proj_b",
|
| 478 |
+
"model.language_model.layers.52.linear_attn.in_proj_a",
|
| 479 |
+
"model.language_model.layers.53.linear_attn",
|
| 480 |
+
"model.language_model.layers.53.linear_attn.norm",
|
| 481 |
+
"model.language_model.layers.53.linear_attn.out_proj",
|
| 482 |
+
"model.language_model.layers.53.linear_attn.in_proj_qkv",
|
| 483 |
+
"model.language_model.layers.53.linear_attn.in_proj_z",
|
| 484 |
+
"model.language_model.layers.53.linear_attn.in_proj_b",
|
| 485 |
+
"model.language_model.layers.53.linear_attn.in_proj_a",
|
| 486 |
+
"model.language_model.layers.54.linear_attn",
|
| 487 |
+
"model.language_model.layers.54.linear_attn.norm",
|
| 488 |
+
"model.language_model.layers.54.linear_attn.out_proj",
|
| 489 |
+
"model.language_model.layers.54.linear_attn.in_proj_qkv",
|
| 490 |
+
"model.language_model.layers.54.linear_attn.in_proj_z",
|
| 491 |
+
"model.language_model.layers.54.linear_attn.in_proj_b",
|
| 492 |
+
"model.language_model.layers.54.linear_attn.in_proj_a",
|
| 493 |
+
"model.language_model.layers.55.self_attn.q_proj",
|
| 494 |
+
"model.language_model.layers.55.self_attn.k_proj",
|
| 495 |
+
"model.language_model.layers.55.self_attn.v_proj",
|
| 496 |
+
"model.language_model.layers.55.self_attn.o_proj",
|
| 497 |
+
"model.language_model.layers.56.linear_attn",
|
| 498 |
+
"model.language_model.layers.56.linear_attn.norm",
|
| 499 |
+
"model.language_model.layers.56.linear_attn.out_proj",
|
| 500 |
+
"model.language_model.layers.56.linear_attn.in_proj_qkv",
|
| 501 |
+
"model.language_model.layers.56.linear_attn.in_proj_z",
|
| 502 |
+
"model.language_model.layers.56.linear_attn.in_proj_b",
|
| 503 |
+
"model.language_model.layers.56.linear_attn.in_proj_a",
|
| 504 |
+
"model.language_model.layers.57.linear_attn",
|
| 505 |
+
"model.language_model.layers.57.linear_attn.norm",
|
| 506 |
+
"model.language_model.layers.57.linear_attn.out_proj",
|
| 507 |
+
"model.language_model.layers.57.linear_attn.in_proj_qkv",
|
| 508 |
+
"model.language_model.layers.57.linear_attn.in_proj_z",
|
| 509 |
+
"model.language_model.layers.57.linear_attn.in_proj_b",
|
| 510 |
+
"model.language_model.layers.57.linear_attn.in_proj_a",
|
| 511 |
+
"model.language_model.layers.58.linear_attn",
|
| 512 |
+
"model.language_model.layers.58.linear_attn.norm",
|
| 513 |
+
"model.language_model.layers.58.linear_attn.out_proj",
|
| 514 |
+
"model.language_model.layers.58.linear_attn.in_proj_qkv",
|
| 515 |
+
"model.language_model.layers.58.linear_attn.in_proj_z",
|
| 516 |
+
"model.language_model.layers.58.linear_attn.in_proj_b",
|
| 517 |
+
"model.language_model.layers.58.linear_attn.in_proj_a",
|
| 518 |
+
"model.language_model.layers.59.self_attn.q_proj",
|
| 519 |
+
"model.language_model.layers.59.self_attn.k_proj",
|
| 520 |
+
"model.language_model.layers.59.self_attn.v_proj",
|
| 521 |
+
"model.language_model.layers.59.self_attn.o_proj",
|
| 522 |
+
"model.language_model.layers.60.linear_attn",
|
| 523 |
+
"model.language_model.layers.60.linear_attn.norm",
|
| 524 |
+
"model.language_model.layers.60.linear_attn.out_proj",
|
| 525 |
+
"model.language_model.layers.60.linear_attn.in_proj_qkv",
|
| 526 |
+
"model.language_model.layers.60.linear_attn.in_proj_z",
|
| 527 |
+
"model.language_model.layers.60.linear_attn.in_proj_b",
|
| 528 |
+
"model.language_model.layers.60.linear_attn.in_proj_a",
|
| 529 |
+
"model.language_model.layers.61.linear_attn",
|
| 530 |
+
"model.language_model.layers.61.linear_attn.norm",
|
| 531 |
+
"model.language_model.layers.61.linear_attn.out_proj",
|
| 532 |
+
"model.language_model.layers.61.linear_attn.in_proj_qkv",
|
| 533 |
+
"model.language_model.layers.61.linear_attn.in_proj_z",
|
| 534 |
+
"model.language_model.layers.61.linear_attn.in_proj_b",
|
| 535 |
+
"model.language_model.layers.61.linear_attn.in_proj_a",
|
| 536 |
+
"model.language_model.layers.62.linear_attn",
|
| 537 |
+
"model.language_model.layers.62.linear_attn.norm",
|
| 538 |
+
"model.language_model.layers.62.linear_attn.out_proj",
|
| 539 |
+
"model.language_model.layers.62.linear_attn.in_proj_qkv",
|
| 540 |
+
"model.language_model.layers.62.linear_attn.in_proj_z",
|
| 541 |
+
"model.language_model.layers.62.linear_attn.in_proj_b",
|
| 542 |
+
"model.language_model.layers.62.linear_attn.in_proj_a",
|
| 543 |
+
"model.language_model.layers.63.self_attn.q_proj",
|
| 544 |
+
"model.language_model.layers.63.self_attn.k_proj",
|
| 545 |
+
"model.language_model.layers.63.self_attn.v_proj",
|
| 546 |
+
"model.language_model.layers.63.self_attn.o_proj",
|
| 547 |
+
"lm_head",
|
| 548 |
+
"mtp.fc",
|
| 549 |
+
"mtp.layers.0.mlp.down_proj",
|
| 550 |
+
"mtp.layers.0.mlp.gate_proj",
|
| 551 |
+
"mtp.layers.0.mlp.up_proj",
|
| 552 |
+
"mtp.layers.0.self_attn.k_proj",
|
| 553 |
+
"mtp.layers.0.self_attn.o_proj",
|
| 554 |
+
"mtp.layers.0.self_attn.q_proj",
|
| 555 |
+
"mtp.layers.0.self_attn.v_proj"
|
| 556 |
+
],
|
| 557 |
+
"kv_cache_scheme": null,
|
| 558 |
+
"quant_method": "compressed-tensors",
|
| 559 |
+
"quantization_status": "compressed",
|
| 560 |
+
"sparsity_config": {},
|
| 561 |
+
"transform_config": {},
|
| 562 |
+
"version": "0.18.0"
|
| 563 |
+
},
|
| 564 |
+
"text_config": {
|
| 565 |
+
"attention_bias": false,
|
| 566 |
+
"attention_dropout": 0.0,
|
| 567 |
+
"attn_output_gate": true,
|
| 568 |
+
"bos_token_id": 248044,
|
| 569 |
+
"dtype": "bfloat16",
|
| 570 |
+
"eos_token_id": 248044,
|
| 571 |
+
"full_attention_interval": 4,
|
| 572 |
+
"head_dim": 256,
|
| 573 |
+
"hidden_act": "silu",
|
| 574 |
+
"hidden_size": 5120,
|
| 575 |
+
"initializer_range": 0.02,
|
| 576 |
+
"intermediate_size": 17408,
|
| 577 |
+
"layer_types": [
|
| 578 |
+
"linear_attention",
|
| 579 |
+
"linear_attention",
|
| 580 |
+
"linear_attention",
|
| 581 |
+
"full_attention",
|
| 582 |
+
"linear_attention",
|
| 583 |
+
"linear_attention",
|
| 584 |
+
"linear_attention",
|
| 585 |
+
"full_attention",
|
| 586 |
+
"linear_attention",
|
| 587 |
+
"linear_attention",
|
| 588 |
+
"linear_attention",
|
| 589 |
+
"full_attention",
|
| 590 |
+
"linear_attention",
|
| 591 |
+
"linear_attention",
|
| 592 |
+
"linear_attention",
|
| 593 |
+
"full_attention",
|
| 594 |
+
"linear_attention",
|
| 595 |
+
"linear_attention",
|
| 596 |
+
"linear_attention",
|
| 597 |
+
"full_attention",
|
| 598 |
+
"linear_attention",
|
| 599 |
+
"linear_attention",
|
| 600 |
+
"linear_attention",
|
| 601 |
+
"full_attention",
|
| 602 |
+
"linear_attention",
|
| 603 |
+
"linear_attention",
|
| 604 |
+
"linear_attention",
|
| 605 |
+
"full_attention",
|
| 606 |
+
"linear_attention",
|
| 607 |
+
"linear_attention",
|
| 608 |
+
"linear_attention",
|
| 609 |
+
"full_attention",
|
| 610 |
+
"linear_attention",
|
| 611 |
+
"linear_attention",
|
| 612 |
+
"linear_attention",
|
| 613 |
+
"full_attention",
|
| 614 |
+
"linear_attention",
|
| 615 |
+
"linear_attention",
|
| 616 |
+
"linear_attention",
|
| 617 |
+
"full_attention",
|
| 618 |
+
"linear_attention",
|
| 619 |
+
"linear_attention",
|
| 620 |
+
"linear_attention",
|
| 621 |
+
"full_attention",
|
| 622 |
+
"linear_attention",
|
| 623 |
+
"linear_attention",
|
| 624 |
+
"linear_attention",
|
| 625 |
+
"full_attention",
|
| 626 |
+
"linear_attention",
|
| 627 |
+
"linear_attention",
|
| 628 |
+
"linear_attention",
|
| 629 |
+
"full_attention",
|
| 630 |
+
"linear_attention",
|
| 631 |
+
"linear_attention",
|
| 632 |
+
"linear_attention",
|
| 633 |
+
"full_attention",
|
| 634 |
+
"linear_attention",
|
| 635 |
+
"linear_attention",
|
| 636 |
+
"linear_attention",
|
| 637 |
+
"full_attention",
|
| 638 |
+
"linear_attention",
|
| 639 |
+
"linear_attention",
|
| 640 |
+
"linear_attention",
|
| 641 |
+
"full_attention"
|
| 642 |
+
],
|
| 643 |
+
"linear_conv_kernel_dim": 4,
|
| 644 |
+
"linear_key_head_dim": 128,
|
| 645 |
+
"linear_num_key_heads": 16,
|
| 646 |
+
"linear_num_value_heads": 48,
|
| 647 |
+
"linear_value_head_dim": 128,
|
| 648 |
+
"mamba_ssm_dtype": "float32",
|
| 649 |
+
"max_position_embeddings": 262144,
|
| 650 |
+
"model_type": "qwen3_5_text",
|
| 651 |
+
"mtp_num_hidden_layers": 1,
|
| 652 |
+
"mtp_use_dedicated_embeddings": false,
|
| 653 |
+
"num_attention_heads": 24,
|
| 654 |
+
"num_hidden_layers": 64,
|
| 655 |
+
"num_key_value_heads": 4,
|
| 656 |
+
"output_gate_type": "swish",
|
| 657 |
+
"pad_token_id": null,
|
| 658 |
+
"partial_rotary_factor": 0.25,
|
| 659 |
+
"rms_norm_eps": 1e-06,
|
| 660 |
+
"rope_parameters": {
|
| 661 |
+
"mrope_interleaved": true,
|
| 662 |
+
"mrope_section": [
|
| 663 |
+
11,
|
| 664 |
+
11,
|
| 665 |
+
10
|
| 666 |
+
],
|
| 667 |
+
"partial_rotary_factor": 0.25,
|
| 668 |
+
"rope_theta": 10000000,
|
| 669 |
+
"rope_type": "default"
|
| 670 |
+
},
|
| 671 |
+
"tie_word_embeddings": false,
|
| 672 |
+
"use_cache": true,
|
| 673 |
+
"vocab_size": 248320
|
| 674 |
+
},
|
| 675 |
+
"tie_word_embeddings": false,
|
| 676 |
+
"transformers_version": "5.14.1",
|
| 677 |
+
"video_token_id": 248057,
|
| 678 |
+
"vision_config": {
|
| 679 |
+
"deepstack_visual_indexes": [],
|
| 680 |
+
"depth": 27,
|
| 681 |
+
"dtype": "bfloat16",
|
| 682 |
+
"hidden_act": "gelu_pytorch_tanh",
|
| 683 |
+
"hidden_size": 1152,
|
| 684 |
+
"in_channels": 3,
|
| 685 |
+
"initializer_range": 0.02,
|
| 686 |
+
"intermediate_size": 4304,
|
| 687 |
+
"model_type": "qwen3_5_vision",
|
| 688 |
+
"num_heads": 16,
|
| 689 |
+
"num_position_embeddings": 2304,
|
| 690 |
+
"out_hidden_size": 5120,
|
| 691 |
+
"patch_size": 16,
|
| 692 |
+
"spatial_merge_size": 2,
|
| 693 |
+
"temporal_patch_size": 2
|
| 694 |
+
},
|
| 695 |
+
"vision_end_token_id": 248054,
|
| 696 |
+
"vision_start_token_id": 248053
|
| 697 |
+
}
|
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": 1.0,
|
| 10 |
+
"top_k": 20,
|
| 11 |
+
"top_p": 0.95,
|
| 12 |
+
"transformers_version": "5.14.1"
|
| 13 |
+
}
|
kld_qwen3.8-27b-nvfp4a16.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"mean_kld_nats": 0.12671151501126587,
|
| 3 |
+
"max_sample_kld_nats": 0.2964491546154022,
|
| 4 |
+
"max_token_kld_nats": 23.491870880126953,
|
| 5 |
+
"per_sample": [
|
| 6 |
+
0.10766392201185226,
|
| 7 |
+
0.030203072354197502,
|
| 8 |
+
0.2964491546154022,
|
| 9 |
+
0.10052302479743958,
|
| 10 |
+
0.08012005686759949,
|
| 11 |
+
0.2676840126514435,
|
| 12 |
+
0.04444360360503197,
|
| 13 |
+
0.08660527318716049
|
| 14 |
+
],
|
| 15 |
+
"per_sample_max_token": [
|
| 16 |
+
3.9194209575653076,
|
| 17 |
+
1.233764410018921,
|
| 18 |
+
23.491870880126953,
|
| 19 |
+
2.2289748191833496,
|
| 20 |
+
3.3153929710388184,
|
| 21 |
+
19.507736206054688,
|
| 22 |
+
1.8779168128967285,
|
| 23 |
+
5.843461513519287
|
| 24 |
+
],
|
| 25 |
+
"num_samples": 8,
|
| 26 |
+
"max_seq": 1024,
|
| 27 |
+
"dataset": "neuralmagic/calibration"
|
| 28 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model-00001-of-00006.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5fb697e3d053622ed1b56aafb835914628bf4f57e9c139ce09158e0098bdc4d1
|
| 3 |
+
size 5346460600
|
model-00002-of-00006.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:852f2e7d354fe2f529d11dab02a8aba55cc53352328a9526904289621c452360
|
| 3 |
+
size 4779928464
|
model-00003-of-00006.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c4f76126f5e4a55fcb5e8bae1a35d6e6560a8ea96683e9f450b60adbf8690fd9
|
| 3 |
+
size 5356936752
|
model-00004-of-00006.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:604ea1479a282b34d64a52d02569423609c3f626a34cb206962ebc086435e2b0
|
| 3 |
+
size 5333230088
|
model-00005-of-00006.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7623117e49cfe93f1ed2c186117d86535e6d6e3ef4556a2b9c5fc41b43f2c74b
|
| 3 |
+
size 5362650672
|
model-00006-of-00006.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0fc27015f380cc2fc70988949609241675c1a268ded16ab6c0feac6cfc02756e
|
| 3 |
+
size 4784258112
|
model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ppl_qwen3.8-27b-base.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": "/workspace/models/qwen3.8-27b-base",
|
| 3 |
+
"dataset": "Salesforce/wikitext/wikitext-2-raw-v1/test",
|
| 4 |
+
"text_field": "text",
|
| 5 |
+
"max_seq": 2048,
|
| 6 |
+
"tokens_scored": 296907,
|
| 7 |
+
"mean_nll": 1.9375337600328988,
|
| 8 |
+
"perplexity": 6.941609859466553
|
| 9 |
+
}
|
ppl_qwen3.8-27b-nvfp4a16.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": "/workspace/models/qwen3.8-27b-nvfp4a16-5g",
|
| 3 |
+
"dataset": "Salesforce/wikitext/wikitext-2-raw-v1/test",
|
| 4 |
+
"text_field": "text",
|
| 5 |
+
"max_seq": 2048,
|
| 6 |
+
"tokens_scored": 296907,
|
| 7 |
+
"mean_nll": 1.9670747163509692,
|
| 8 |
+
"perplexity": 7.149731159210205
|
| 9 |
+
}
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"size": {
|
| 3 |
+
"longest_edge": 16777216,
|
| 4 |
+
"shortest_edge": 65536
|
| 5 |
+
},
|
| 6 |
+
"patch_size": 16,
|
| 7 |
+
"temporal_patch_size": 2,
|
| 8 |
+
"merge_size": 2,
|
| 9 |
+
"image_mean": [
|
| 10 |
+
0.5,
|
| 11 |
+
0.5,
|
| 12 |
+
0.5
|
| 13 |
+
],
|
| 14 |
+
"image_std": [
|
| 15 |
+
0.5,
|
| 16 |
+
0.5,
|
| 17 |
+
0.5
|
| 18 |
+
],
|
| 19 |
+
"processor_class": "Qwen3VLProcessor",
|
| 20 |
+
"image_processor_type": "Qwen2VLImageProcessorFast"
|
| 21 |
+
}
|
recipe.yaml
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
name: nvfp4a16_attnbf16
|
| 2 |
+
scheme: NVFP4A16 # FP4 weights, bf16 activations — skips activation quant for much lower KLD
|
| 3 |
+
engine: llmcompressor
|
| 4 |
+
|
| 5 |
+
# Variant of `nvfp4a16.yaml` that additionally keeps the ENTIRE self-attention
|
| 6 |
+
# block in bf16, leaving only the MLPs quantised — the same lever measured on
|
| 7 |
+
# ThinkingCap for FP8 (landmine 36), applied to the W4A16 path.
|
| 8 |
+
#
|
| 9 |
+
# The motivating fact is architectural, not scheme-specific: `attn_output_gate:
|
| 10 |
+
# true` fuses the attention output gate into `q_proj`, making it
|
| 11 |
+
# [2*heads*head_dim, hidden] instead of [heads*head_dim, hidden]. Half that
|
| 12 |
+
# tensor is a multiplicative per-head gate on what attention writes into the
|
| 13 |
+
# residual stream, and only the full_attention layers (a quarter of the stack)
|
| 14 |
+
# carry long-range retrieval. Quantisation error on a multiplicative gate
|
| 15 |
+
# behaves worse than on an additive projection.
|
| 16 |
+
#
|
| 17 |
+
# Cost/benefit differs from the FP8 case and should be re-measured, not assumed:
|
| 18 |
+
# at W4A16 the attention block would otherwise compress 4x rather than 2x, so
|
| 19 |
+
# holding it in bf16 costs proportionally more on-disk than the +4.6% measured
|
| 20 |
+
# for FP8. The KLD win is expected to be larger too, since NVFP4A16's error
|
| 21 |
+
# floor is 2-3x FP8's. Ship whichever the measured KLD/size tradeoff justifies —
|
| 22 |
+
# `nvfp4a16.yaml` remains the standard build.
|
| 23 |
+
|
| 24 |
+
calibration:
|
| 25 |
+
dataset: neuralmagic/calibration
|
| 26 |
+
config: LLM
|
| 27 |
+
split: train
|
| 28 |
+
num_samples: 128
|
| 29 |
+
max_seq_length: 2048
|
| 30 |
+
|
| 31 |
+
ignore:
|
| 32 |
+
- lm_head
|
| 33 |
+
- "re:.*visual.*"
|
| 34 |
+
- "re:.*linear_attn.*" # entire SSM block kept in bf16 — same rationale as the standard build
|
| 35 |
+
- "re:.*self_attn.*" # THE VARIANT: q/k/v/o_proj too, incl. the output gate fused into q_proj
|
| 36 |
+
- "re:.*mtp.*"
|
| 37 |
+
# Note: dense base. On MoE bases also add "re:.*mlp.gate$" and
|
| 38 |
+
# "re:.*mlp.shared_expert_gate$" — no-ops on dense models.
|
| 39 |
+
|
| 40 |
+
export:
|
| 41 |
+
save_compressed: true
|
smoke_qwen3.8-27b-nvfp4a16.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_dir": "/workspace/models/qwen3.8-27b-nvfp4a16-5g",
|
| 3 |
+
"mtp_present": true,
|
| 4 |
+
"text_ok": true,
|
| 5 |
+
"text_output": "We need to answer user: \"Briefly explain why 1+1 equals 2 in everyday terms. Then list three real-world examples.\" Simple. Need final concise. Explain addition as combining groups, one plus one makes two. Examples: one apple + one apple = two apples; one hand + one hand =",
|
| 6 |
+
"vision_ok": true,
|
| 7 |
+
"vision_output": "The user wants a single short sentence describing the image. I need to identify the key subject and setting. The image shows a small, fluffy wild cat with thick fur, likely a Pallas's cat, walking on snow with birch trees in the background. I should keep it concise and capture the essence: the cat"
|
| 8 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f399b3cd12fa270d51457bb749fb30863521e8359b8a27059c71b6c2f7d6dd6c
|
| 3 |
+
size 19989424
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": true,
|
| 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 |
+
"split_special_tokens": false,
|
| 27 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 28 |
+
"unk_token": null,
|
| 29 |
+
"video_token": "<|video_pad|>",
|
| 30 |
+
"vision_bos_token": "<|vision_start|>",
|
| 31 |
+
"vision_eos_token": "<|vision_end|>"
|
| 32 |
+
}
|
video_preprocessor_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"size": {
|
| 3 |
+
"longest_edge": 25165824,
|
| 4 |
+
"shortest_edge": 4096
|
| 5 |
+
},
|
| 6 |
+
"patch_size": 16,
|
| 7 |
+
"temporal_patch_size": 2,
|
| 8 |
+
"merge_size": 2,
|
| 9 |
+
"image_mean": [
|
| 10 |
+
0.5,
|
| 11 |
+
0.5,
|
| 12 |
+
0.5
|
| 13 |
+
],
|
| 14 |
+
"image_std": [
|
| 15 |
+
0.5,
|
| 16 |
+
0.5,
|
| 17 |
+
0.5
|
| 18 |
+
],
|
| 19 |
+
"processor_class": "Qwen3VLProcessor",
|
| 20 |
+
"video_processor_type": "Qwen3VLVideoProcessor"
|
| 21 |
+
}
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|