Instructions to use Xananthium/Inference-Model-Archive-2026-10 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use Xananthium/Inference-Model-Archive-2026-10 with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf Xananthium/Inference-Model-Archive-2026-10:Q8_0 # Run inference directly in the terminal: llama cli -hf Xananthium/Inference-Model-Archive-2026-10:Q8_0
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf Xananthium/Inference-Model-Archive-2026-10:Q8_0 # Run inference directly in the terminal: llama cli -hf Xananthium/Inference-Model-Archive-2026-10:Q8_0
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf Xananthium/Inference-Model-Archive-2026-10:Q8_0 # Run inference directly in the terminal: ./build/bin/llama cli -hf Xananthium/Inference-Model-Archive-2026-10:Q8_0
Use Docker
docker model run hf.co/Xananthium/Inference-Model-Archive-2026-10:Q8_0
- Llama new
- LM Studio
- Jan
- Ollama
How to use Xananthium/Inference-Model-Archive-2026-10 with Ollama:
ollama run hf.co/Xananthium/Inference-Model-Archive-2026-10:Q8_0
- Unsloth Desktop
- Pi
How to use Xananthium/Inference-Model-Archive-2026-10 with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Xananthium/Inference-Model-Archive-2026-10:Q8_0
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:9931/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "Xananthium/Inference-Model-Archive-2026-10:Q8_0" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use Xananthium/Inference-Model-Archive-2026-10 with Docker Model Runner:
docker model run hf.co/Xananthium/Inference-Model-Archive-2026-10:Q8_0
- Lemonade
How to use Xananthium/Inference-Model-Archive-2026-10 with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Xananthium/Inference-Model-Archive-2026-10:Q8_0
Run and chat with the model
lemonade run user.Inference-Model-Archive-2026-10-Q8_0
List all available models
lemonade list
- Hermes Agent
How to use Xananthium/Inference-Model-Archive-2026-10 with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Xananthium/Inference-Model-Archive-2026-10:Q8_0
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:9931/v1 hermes config set model.default Xananthium/Inference-Model-Archive-2026-10:Q8_0
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use Xananthium/Inference-Model-Archive-2026-10 with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf Xananthium/Inference-Model-Archive-2026-10:Q8_0
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:9931/v1 \ --custom-model-id "Xananthium/Inference-Model-Archive-2026-10:Q8_0" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
Publish qwen3.8-27b-hauhau-w8a8-vision-mtp with measured results and source credits
Browse files- .gitattributes +1 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/README.md +17 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/chat_template.jinja +173 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/config.json +195 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/conversion-receipt.json +410 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/generation_config.json +13 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00000.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00001.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00002.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00003.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00004.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00005.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00006.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00007.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00008.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00009.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00010.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00011.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00012.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00013.safetensors +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/model.safetensors.index.json +0 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/processor_config.json +60 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/results/precision-comparison.json +382 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/tokenizer.json +3 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/tokenizer_config.json +33 -0
- qwen3.8-27b-hauhau-w8a8-vision-mtp/vision.safetensors +3 -0
.gitattributes
CHANGED
|
@@ -46,3 +46,4 @@ qwen3.6-35b-hauhau-nvfp4-trt/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
|
| 46 |
qwen3.6-35b-hauhau-reconstructed-bf16/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 47 |
qwen3.6-35b-hauhau-w8a16-trt/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 48 |
qwen3.6-35b-hauhau-w8a8-tq4-runtime/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 46 |
qwen3.6-35b-hauhau-reconstructed-bf16/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 47 |
qwen3.6-35b-hauhau-w8a16-trt/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 48 |
qwen3.6-35b-hauhau-w8a8-tq4-runtime/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 49 |
+
qwen3.8-27b-hauhau-w8a8-vision-mtp/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/README.md
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: HauhauCS/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUF
|
| 4 |
+
tags:
|
| 5 |
+
- local-inference
|
| 6 |
+
- quantization
|
| 7 |
+
- evaluation
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
# qwen3.8-27b-hauhau-w8a8-vision-mtp
|
| 11 |
+
|
| 12 |
+
Complete local checkpoint published by Xananthium, 2026-10-06. Not evaluated in the 2026-10-06 matched precision comparison. The attached report describes other checkpoints, not scores for this artifact.
|
| 13 |
+
|
| 14 |
+
Source credit: [HauhauCS/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUF](https://huggingface.co/HauhauCS/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUF). Original base: [Qwen/Qwen3.8-27B](https://huggingface.co/Qwen/Qwen3.8-27B). Original model card is preserved in `README_ORIGINAL.md`; upstream licenses apply.
|
| 15 |
+
|
| 16 |
+
Measured results and limitations are in `results/precision-comparison.json`. These are small synthetic tests, not published cybersecurity benchmark scores. Untested artifacts have no measured scores.
|
| 17 |
+
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/chat_template.jinja
ADDED
|
@@ -0,0 +1,173 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 in ('high', 'max') %}
|
| 49 |
+
{%- set resolved_reasoning_effort = 'xhigh' %}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{%- if resolved_reasoning_effort not in ('xhigh', 'medium', 'low') %}
|
| 52 |
+
{{- raise_exception('Unexpected reasoning effort ' ~ reasoning_effort ~ '. Supported types are xhigh (default), medium, and low.') }}
|
| 53 |
+
{%- endif %}
|
| 54 |
+
{%- if resolved_reasoning_effort == 'xhigh' %}
|
| 55 |
+
{%- 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.' %}
|
| 56 |
+
{%- elif resolved_reasoning_effort == 'low' %}
|
| 57 |
+
{%- set reasoning_instructions = 'Reasoning effort is set to low. Keep your thinking brief and focused, moving directly to the conclusion without unnecessary elaboration.' %}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 61 |
+
{{- '<|im_start|>system\n' }}
|
| 62 |
+
{%- if reasoning_instructions %}
|
| 63 |
+
{{- reasoning_instructions + '\n\n' }}
|
| 64 |
+
{%- endif %}
|
| 65 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 66 |
+
{%- for tool in tools %}
|
| 67 |
+
{{- "\n" }}
|
| 68 |
+
{{- tool | tojson }}
|
| 69 |
+
{%- endfor %}
|
| 70 |
+
{{- "\n</tools>" }}
|
| 71 |
+
{{- '\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>' }}
|
| 72 |
+
{%- if messages[0].role == 'system' %}
|
| 73 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 74 |
+
{%- if content %}
|
| 75 |
+
{{- '\n\n' + content }}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endif %}
|
| 78 |
+
{{- '<|im_end|>\n' }}
|
| 79 |
+
{%- else %}
|
| 80 |
+
{%- if messages[0].role == 'system' %}
|
| 81 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 82 |
+
{%- if content %}
|
| 83 |
+
{{- '<|im_start|>system\n' + (reasoning_instructions + '\n\n' if reasoning_instructions else '') + content + '<|im_end|>\n' }}
|
| 84 |
+
{%- elif reasoning_instructions %}
|
| 85 |
+
{{- '<|im_start|>system\n' + reasoning_instructions + '<|im_end|>\n' }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif reasoning_instructions %}
|
| 88 |
+
{{- '<|im_start|>system\n' + reasoning_instructions + '<|im_end|>\n' }}
|
| 89 |
+
{%- endif %}
|
| 90 |
+
{%- endif %}
|
| 91 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 92 |
+
{%- for message in messages[::-1] %}
|
| 93 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 94 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 95 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 96 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 97 |
+
{%- set ns.multi_step_tool = false %}
|
| 98 |
+
{%- set ns.last_query_index = index %}
|
| 99 |
+
{%- endif %}
|
| 100 |
+
{%- endif %}
|
| 101 |
+
{%- endfor %}
|
| 102 |
+
{%- if ns.multi_step_tool %}
|
| 103 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- for message in messages %}
|
| 106 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 107 |
+
{%- if message.role == "system" %}
|
| 108 |
+
{%- if not loop.first %}
|
| 109 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 110 |
+
{%- endif %}
|
| 111 |
+
{%- elif message.role == "user" %}
|
| 112 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 113 |
+
{%- elif message.role == "assistant" %}
|
| 114 |
+
{%- set reasoning_content = '' %}
|
| 115 |
+
{%- if message.reasoning_content is string %}
|
| 116 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 117 |
+
{%- endif %}
|
| 118 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 119 |
+
{%- if preserve_thinking is undefined or preserve_thinking is true or loop.index0 > ns.last_query_index %}
|
| 120 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 121 |
+
{%- else %}
|
| 122 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 123 |
+
{%- endif %}
|
| 124 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 125 |
+
{%- for tool_call in message.tool_calls %}
|
| 126 |
+
{%- if tool_call.function is defined %}
|
| 127 |
+
{%- set tool_call = tool_call.function %}
|
| 128 |
+
{%- endif %}
|
| 129 |
+
{%- if loop.first %}
|
| 130 |
+
{%- if content|trim %}
|
| 131 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 132 |
+
{%- else %}
|
| 133 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{%- else %}
|
| 136 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 137 |
+
{%- endif %}
|
| 138 |
+
{%- if tool_call.arguments is defined and tool_call.arguments != '' %}
|
| 139 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 140 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 141 |
+
{%- set args_value = args_value | string if args_value is string else args_value | tojson | safe %}
|
| 142 |
+
{{- args_value }}
|
| 143 |
+
{{- '\n</parameter>\n' }}
|
| 144 |
+
{%- endfor %}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{{- '</function>\n</tool_call>' }}
|
| 147 |
+
{%- endfor %}
|
| 148 |
+
{%- endif %}
|
| 149 |
+
{{- '<|im_end|>\n' }}
|
| 150 |
+
{%- elif message.role == "tool" %}
|
| 151 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 152 |
+
{{- '<|im_start|>user' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{{- '\n<tool_response>\n' }}
|
| 155 |
+
{{- content }}
|
| 156 |
+
{{- '\n</tool_response>' }}
|
| 157 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 158 |
+
{{- '<|im_end|>\n' }}
|
| 159 |
+
{%- elif loop.last %}
|
| 160 |
+
{{- '<|im_end|>\n' }}
|
| 161 |
+
{%- endif %}
|
| 162 |
+
{%- else %}
|
| 163 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 164 |
+
{%- endif %}
|
| 165 |
+
{%- endfor %}
|
| 166 |
+
{%- if add_generation_prompt %}
|
| 167 |
+
{{- '<|im_start|>assistant\n' }}
|
| 168 |
+
{%- if enable_thinking is defined and enable_thinking is false %}
|
| 169 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 170 |
+
{%- else %}
|
| 171 |
+
{{- '<think>\n' }}
|
| 172 |
+
{%- endif %}
|
| 173 |
+
{%- endif %}
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/config.json
ADDED
|
@@ -0,0 +1,195 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"config_group_0": {
|
| 12 |
+
"format": "int-quantized",
|
| 13 |
+
"input_activations": {
|
| 14 |
+
"actorder": null,
|
| 15 |
+
"block_structure": null,
|
| 16 |
+
"dynamic": true,
|
| 17 |
+
"group_size": null,
|
| 18 |
+
"num_bits": 8,
|
| 19 |
+
"observer": null,
|
| 20 |
+
"observer_kwargs": {},
|
| 21 |
+
"scale_dtype": null,
|
| 22 |
+
"strategy": "token",
|
| 23 |
+
"symmetric": true,
|
| 24 |
+
"type": "int",
|
| 25 |
+
"zp_dtype": null
|
| 26 |
+
},
|
| 27 |
+
"output_activations": null,
|
| 28 |
+
"targets": [
|
| 29 |
+
"Linear"
|
| 30 |
+
],
|
| 31 |
+
"weights": {
|
| 32 |
+
"actorder": null,
|
| 33 |
+
"block_structure": null,
|
| 34 |
+
"dynamic": false,
|
| 35 |
+
"group_size": null,
|
| 36 |
+
"num_bits": 8,
|
| 37 |
+
"observer": "memoryless_minmax",
|
| 38 |
+
"observer_kwargs": {},
|
| 39 |
+
"scale_dtype": null,
|
| 40 |
+
"strategy": "channel",
|
| 41 |
+
"symmetric": true,
|
| 42 |
+
"type": "int",
|
| 43 |
+
"zp_dtype": null
|
| 44 |
+
}
|
| 45 |
+
}
|
| 46 |
+
},
|
| 47 |
+
"format": "int-quantized",
|
| 48 |
+
"global_compression_ratio": null,
|
| 49 |
+
"ignore": [
|
| 50 |
+
"lm_head",
|
| 51 |
+
"model.language_model.embed_tokens",
|
| 52 |
+
"re:.*visual.*",
|
| 53 |
+
"re:.*mtp.*",
|
| 54 |
+
"re:.*linear_attn.in_proj_[ab]$"
|
| 55 |
+
],
|
| 56 |
+
"kv_cache_scheme": null,
|
| 57 |
+
"quant_method": "compressed-tensors",
|
| 58 |
+
"quantization_status": "compressed",
|
| 59 |
+
"transform_config": {},
|
| 60 |
+
"version": "0.17.1"
|
| 61 |
+
},
|
| 62 |
+
"text_config": {
|
| 63 |
+
"attention_bias": false,
|
| 64 |
+
"attention_dropout": 0.0,
|
| 65 |
+
"attn_output_gate": true,
|
| 66 |
+
"bos_token_id": 248044,
|
| 67 |
+
"dtype": "bfloat16",
|
| 68 |
+
"eos_token_id": 248044,
|
| 69 |
+
"full_attention_interval": 4,
|
| 70 |
+
"head_dim": 256,
|
| 71 |
+
"hidden_act": "silu",
|
| 72 |
+
"hidden_size": 5120,
|
| 73 |
+
"initializer_range": 0.02,
|
| 74 |
+
"intermediate_size": 17408,
|
| 75 |
+
"layer_types": [
|
| 76 |
+
"linear_attention",
|
| 77 |
+
"linear_attention",
|
| 78 |
+
"linear_attention",
|
| 79 |
+
"full_attention",
|
| 80 |
+
"linear_attention",
|
| 81 |
+
"linear_attention",
|
| 82 |
+
"linear_attention",
|
| 83 |
+
"full_attention",
|
| 84 |
+
"linear_attention",
|
| 85 |
+
"linear_attention",
|
| 86 |
+
"linear_attention",
|
| 87 |
+
"full_attention",
|
| 88 |
+
"linear_attention",
|
| 89 |
+
"linear_attention",
|
| 90 |
+
"linear_attention",
|
| 91 |
+
"full_attention",
|
| 92 |
+
"linear_attention",
|
| 93 |
+
"linear_attention",
|
| 94 |
+
"linear_attention",
|
| 95 |
+
"full_attention",
|
| 96 |
+
"linear_attention",
|
| 97 |
+
"linear_attention",
|
| 98 |
+
"linear_attention",
|
| 99 |
+
"full_attention",
|
| 100 |
+
"linear_attention",
|
| 101 |
+
"linear_attention",
|
| 102 |
+
"linear_attention",
|
| 103 |
+
"full_attention",
|
| 104 |
+
"linear_attention",
|
| 105 |
+
"linear_attention",
|
| 106 |
+
"linear_attention",
|
| 107 |
+
"full_attention",
|
| 108 |
+
"linear_attention",
|
| 109 |
+
"linear_attention",
|
| 110 |
+
"linear_attention",
|
| 111 |
+
"full_attention",
|
| 112 |
+
"linear_attention",
|
| 113 |
+
"linear_attention",
|
| 114 |
+
"linear_attention",
|
| 115 |
+
"full_attention",
|
| 116 |
+
"linear_attention",
|
| 117 |
+
"linear_attention",
|
| 118 |
+
"linear_attention",
|
| 119 |
+
"full_attention",
|
| 120 |
+
"linear_attention",
|
| 121 |
+
"linear_attention",
|
| 122 |
+
"linear_attention",
|
| 123 |
+
"full_attention",
|
| 124 |
+
"linear_attention",
|
| 125 |
+
"linear_attention",
|
| 126 |
+
"linear_attention",
|
| 127 |
+
"full_attention",
|
| 128 |
+
"linear_attention",
|
| 129 |
+
"linear_attention",
|
| 130 |
+
"linear_attention",
|
| 131 |
+
"full_attention",
|
| 132 |
+
"linear_attention",
|
| 133 |
+
"linear_attention",
|
| 134 |
+
"linear_attention",
|
| 135 |
+
"full_attention",
|
| 136 |
+
"linear_attention",
|
| 137 |
+
"linear_attention",
|
| 138 |
+
"linear_attention",
|
| 139 |
+
"full_attention"
|
| 140 |
+
],
|
| 141 |
+
"linear_conv_kernel_dim": 4,
|
| 142 |
+
"linear_key_head_dim": 128,
|
| 143 |
+
"linear_num_key_heads": 16,
|
| 144 |
+
"linear_num_value_heads": 48,
|
| 145 |
+
"linear_value_head_dim": 128,
|
| 146 |
+
"mamba_ssm_dtype": "float32",
|
| 147 |
+
"max_position_embeddings": 262144,
|
| 148 |
+
"model_type": "qwen3_5_text",
|
| 149 |
+
"mtp_num_hidden_layers": 1,
|
| 150 |
+
"mtp_use_dedicated_embeddings": false,
|
| 151 |
+
"num_attention_heads": 24,
|
| 152 |
+
"num_hidden_layers": 64,
|
| 153 |
+
"num_key_value_heads": 4,
|
| 154 |
+
"output_gate_type": "swish",
|
| 155 |
+
"pad_token_id": null,
|
| 156 |
+
"partial_rotary_factor": 0.25,
|
| 157 |
+
"rms_norm_eps": 1e-06,
|
| 158 |
+
"rope_parameters": {
|
| 159 |
+
"mrope_interleaved": true,
|
| 160 |
+
"mrope_section": [
|
| 161 |
+
11,
|
| 162 |
+
11,
|
| 163 |
+
10
|
| 164 |
+
],
|
| 165 |
+
"partial_rotary_factor": 0.25,
|
| 166 |
+
"rope_theta": 10000000,
|
| 167 |
+
"rope_type": "default"
|
| 168 |
+
},
|
| 169 |
+
"tie_word_embeddings": false,
|
| 170 |
+
"use_cache": true,
|
| 171 |
+
"vocab_size": 248320
|
| 172 |
+
},
|
| 173 |
+
"tie_word_embeddings": false,
|
| 174 |
+
"transformers_version": "5.15.0",
|
| 175 |
+
"video_token_id": 248057,
|
| 176 |
+
"vision_config": {
|
| 177 |
+
"deepstack_visual_indexes": [],
|
| 178 |
+
"depth": 27,
|
| 179 |
+
"dtype": "bfloat16",
|
| 180 |
+
"hidden_act": "gelu_pytorch_tanh",
|
| 181 |
+
"hidden_size": 1152,
|
| 182 |
+
"in_channels": 3,
|
| 183 |
+
"initializer_range": 0.02,
|
| 184 |
+
"intermediate_size": 4304,
|
| 185 |
+
"model_type": "qwen3_5_vision",
|
| 186 |
+
"num_heads": 16,
|
| 187 |
+
"num_position_embeddings": 2304,
|
| 188 |
+
"out_hidden_size": 5120,
|
| 189 |
+
"patch_size": 16,
|
| 190 |
+
"spatial_merge_size": 2,
|
| 191 |
+
"temporal_patch_size": 2
|
| 192 |
+
},
|
| 193 |
+
"vision_end_token_id": 248054,
|
| 194 |
+
"vision_start_token_id": 248053
|
| 195 |
+
}
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/conversion-receipt.json
ADDED
|
@@ -0,0 +1,410 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"source_repo": "HauhauCS/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUF",
|
| 3 |
+
"source_file": "Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-Q8_K_P.gguf",
|
| 4 |
+
"precision": "INT8 weights, dynamic per-token INT8 activations; embeddings, lm_head, vision, MTP, small a/b projections BF16",
|
| 5 |
+
"quantized_linears": [
|
| 6 |
+
"model.language_model.layers.0.linear_attn.out_proj.weight",
|
| 7 |
+
"model.language_model.layers.0.linear_attn.in_proj_qkv.weight",
|
| 8 |
+
"model.language_model.layers.0.linear_attn.in_proj_z.weight",
|
| 9 |
+
"model.language_model.layers.0.mlp.gate_proj.weight",
|
| 10 |
+
"model.language_model.layers.0.mlp.up_proj.weight",
|
| 11 |
+
"model.language_model.layers.0.mlp.down_proj.weight",
|
| 12 |
+
"model.language_model.layers.1.linear_attn.out_proj.weight",
|
| 13 |
+
"model.language_model.layers.1.linear_attn.in_proj_qkv.weight",
|
| 14 |
+
"model.language_model.layers.1.linear_attn.in_proj_z.weight",
|
| 15 |
+
"model.language_model.layers.1.mlp.gate_proj.weight",
|
| 16 |
+
"model.language_model.layers.1.mlp.up_proj.weight",
|
| 17 |
+
"model.language_model.layers.1.mlp.down_proj.weight",
|
| 18 |
+
"model.language_model.layers.2.linear_attn.out_proj.weight",
|
| 19 |
+
"model.language_model.layers.2.linear_attn.in_proj_qkv.weight",
|
| 20 |
+
"model.language_model.layers.2.linear_attn.in_proj_z.weight",
|
| 21 |
+
"model.language_model.layers.2.mlp.gate_proj.weight",
|
| 22 |
+
"model.language_model.layers.2.mlp.up_proj.weight",
|
| 23 |
+
"model.language_model.layers.2.mlp.down_proj.weight",
|
| 24 |
+
"model.language_model.layers.3.self_attn.q_proj.weight",
|
| 25 |
+
"model.language_model.layers.3.self_attn.k_proj.weight",
|
| 26 |
+
"model.language_model.layers.3.self_attn.v_proj.weight",
|
| 27 |
+
"model.language_model.layers.3.self_attn.o_proj.weight",
|
| 28 |
+
"model.language_model.layers.3.mlp.gate_proj.weight",
|
| 29 |
+
"model.language_model.layers.3.mlp.up_proj.weight",
|
| 30 |
+
"model.language_model.layers.3.mlp.down_proj.weight",
|
| 31 |
+
"model.language_model.layers.4.linear_attn.out_proj.weight",
|
| 32 |
+
"model.language_model.layers.4.linear_attn.in_proj_qkv.weight",
|
| 33 |
+
"model.language_model.layers.4.linear_attn.in_proj_z.weight",
|
| 34 |
+
"model.language_model.layers.4.mlp.gate_proj.weight",
|
| 35 |
+
"model.language_model.layers.4.mlp.up_proj.weight",
|
| 36 |
+
"model.language_model.layers.4.mlp.down_proj.weight",
|
| 37 |
+
"model.language_model.layers.5.linear_attn.out_proj.weight",
|
| 38 |
+
"model.language_model.layers.5.linear_attn.in_proj_qkv.weight",
|
| 39 |
+
"model.language_model.layers.5.linear_attn.in_proj_z.weight",
|
| 40 |
+
"model.language_model.layers.5.mlp.gate_proj.weight",
|
| 41 |
+
"model.language_model.layers.5.mlp.up_proj.weight",
|
| 42 |
+
"model.language_model.layers.5.mlp.down_proj.weight",
|
| 43 |
+
"model.language_model.layers.6.linear_attn.out_proj.weight",
|
| 44 |
+
"model.language_model.layers.6.linear_attn.in_proj_qkv.weight",
|
| 45 |
+
"model.language_model.layers.6.linear_attn.in_proj_z.weight",
|
| 46 |
+
"model.language_model.layers.6.mlp.gate_proj.weight",
|
| 47 |
+
"model.language_model.layers.6.mlp.up_proj.weight",
|
| 48 |
+
"model.language_model.layers.6.mlp.down_proj.weight",
|
| 49 |
+
"model.language_model.layers.7.self_attn.q_proj.weight",
|
| 50 |
+
"model.language_model.layers.7.self_attn.k_proj.weight",
|
| 51 |
+
"model.language_model.layers.7.self_attn.v_proj.weight",
|
| 52 |
+
"model.language_model.layers.7.self_attn.o_proj.weight",
|
| 53 |
+
"model.language_model.layers.7.mlp.gate_proj.weight",
|
| 54 |
+
"model.language_model.layers.7.mlp.up_proj.weight",
|
| 55 |
+
"model.language_model.layers.7.mlp.down_proj.weight",
|
| 56 |
+
"model.language_model.layers.8.linear_attn.out_proj.weight",
|
| 57 |
+
"model.language_model.layers.8.linear_attn.in_proj_qkv.weight",
|
| 58 |
+
"model.language_model.layers.8.linear_attn.in_proj_z.weight",
|
| 59 |
+
"model.language_model.layers.8.mlp.gate_proj.weight",
|
| 60 |
+
"model.language_model.layers.8.mlp.up_proj.weight",
|
| 61 |
+
"model.language_model.layers.8.mlp.down_proj.weight",
|
| 62 |
+
"model.language_model.layers.9.linear_attn.out_proj.weight",
|
| 63 |
+
"model.language_model.layers.9.linear_attn.in_proj_qkv.weight",
|
| 64 |
+
"model.language_model.layers.9.linear_attn.in_proj_z.weight",
|
| 65 |
+
"model.language_model.layers.9.mlp.gate_proj.weight",
|
| 66 |
+
"model.language_model.layers.9.mlp.up_proj.weight",
|
| 67 |
+
"model.language_model.layers.9.mlp.down_proj.weight",
|
| 68 |
+
"model.language_model.layers.10.linear_attn.out_proj.weight",
|
| 69 |
+
"model.language_model.layers.10.linear_attn.in_proj_qkv.weight",
|
| 70 |
+
"model.language_model.layers.10.linear_attn.in_proj_z.weight",
|
| 71 |
+
"model.language_model.layers.10.mlp.gate_proj.weight",
|
| 72 |
+
"model.language_model.layers.10.mlp.up_proj.weight",
|
| 73 |
+
"model.language_model.layers.10.mlp.down_proj.weight",
|
| 74 |
+
"model.language_model.layers.11.self_attn.q_proj.weight",
|
| 75 |
+
"model.language_model.layers.11.self_attn.k_proj.weight",
|
| 76 |
+
"model.language_model.layers.11.self_attn.v_proj.weight",
|
| 77 |
+
"model.language_model.layers.11.self_attn.o_proj.weight",
|
| 78 |
+
"model.language_model.layers.11.mlp.gate_proj.weight",
|
| 79 |
+
"model.language_model.layers.11.mlp.up_proj.weight",
|
| 80 |
+
"model.language_model.layers.11.mlp.down_proj.weight",
|
| 81 |
+
"model.language_model.layers.12.linear_attn.out_proj.weight",
|
| 82 |
+
"model.language_model.layers.12.linear_attn.in_proj_qkv.weight",
|
| 83 |
+
"model.language_model.layers.12.linear_attn.in_proj_z.weight",
|
| 84 |
+
"model.language_model.layers.12.mlp.gate_proj.weight",
|
| 85 |
+
"model.language_model.layers.12.mlp.up_proj.weight",
|
| 86 |
+
"model.language_model.layers.12.mlp.down_proj.weight",
|
| 87 |
+
"model.language_model.layers.13.linear_attn.out_proj.weight",
|
| 88 |
+
"model.language_model.layers.13.linear_attn.in_proj_qkv.weight",
|
| 89 |
+
"model.language_model.layers.13.linear_attn.in_proj_z.weight",
|
| 90 |
+
"model.language_model.layers.13.mlp.gate_proj.weight",
|
| 91 |
+
"model.language_model.layers.13.mlp.up_proj.weight",
|
| 92 |
+
"model.language_model.layers.13.mlp.down_proj.weight",
|
| 93 |
+
"model.language_model.layers.14.linear_attn.out_proj.weight",
|
| 94 |
+
"model.language_model.layers.14.linear_attn.in_proj_qkv.weight",
|
| 95 |
+
"model.language_model.layers.14.linear_attn.in_proj_z.weight",
|
| 96 |
+
"model.language_model.layers.14.mlp.gate_proj.weight",
|
| 97 |
+
"model.language_model.layers.14.mlp.up_proj.weight",
|
| 98 |
+
"model.language_model.layers.14.mlp.down_proj.weight",
|
| 99 |
+
"model.language_model.layers.15.self_attn.q_proj.weight",
|
| 100 |
+
"model.language_model.layers.15.self_attn.k_proj.weight",
|
| 101 |
+
"model.language_model.layers.15.self_attn.v_proj.weight",
|
| 102 |
+
"model.language_model.layers.15.self_attn.o_proj.weight",
|
| 103 |
+
"model.language_model.layers.15.mlp.gate_proj.weight",
|
| 104 |
+
"model.language_model.layers.15.mlp.up_proj.weight",
|
| 105 |
+
"model.language_model.layers.15.mlp.down_proj.weight",
|
| 106 |
+
"model.language_model.layers.16.linear_attn.out_proj.weight",
|
| 107 |
+
"model.language_model.layers.16.linear_attn.in_proj_qkv.weight",
|
| 108 |
+
"model.language_model.layers.16.linear_attn.in_proj_z.weight",
|
| 109 |
+
"model.language_model.layers.16.mlp.gate_proj.weight",
|
| 110 |
+
"model.language_model.layers.16.mlp.up_proj.weight",
|
| 111 |
+
"model.language_model.layers.16.mlp.down_proj.weight",
|
| 112 |
+
"model.language_model.layers.17.linear_attn.out_proj.weight",
|
| 113 |
+
"model.language_model.layers.17.linear_attn.in_proj_qkv.weight",
|
| 114 |
+
"model.language_model.layers.17.linear_attn.in_proj_z.weight",
|
| 115 |
+
"model.language_model.layers.17.mlp.gate_proj.weight",
|
| 116 |
+
"model.language_model.layers.17.mlp.up_proj.weight",
|
| 117 |
+
"model.language_model.layers.17.mlp.down_proj.weight",
|
| 118 |
+
"model.language_model.layers.18.linear_attn.out_proj.weight",
|
| 119 |
+
"model.language_model.layers.18.linear_attn.in_proj_qkv.weight",
|
| 120 |
+
"model.language_model.layers.18.linear_attn.in_proj_z.weight",
|
| 121 |
+
"model.language_model.layers.18.mlp.gate_proj.weight",
|
| 122 |
+
"model.language_model.layers.18.mlp.up_proj.weight",
|
| 123 |
+
"model.language_model.layers.18.mlp.down_proj.weight",
|
| 124 |
+
"model.language_model.layers.19.self_attn.q_proj.weight",
|
| 125 |
+
"model.language_model.layers.19.self_attn.k_proj.weight",
|
| 126 |
+
"model.language_model.layers.19.self_attn.v_proj.weight",
|
| 127 |
+
"model.language_model.layers.19.self_attn.o_proj.weight",
|
| 128 |
+
"model.language_model.layers.19.mlp.gate_proj.weight",
|
| 129 |
+
"model.language_model.layers.19.mlp.up_proj.weight",
|
| 130 |
+
"model.language_model.layers.19.mlp.down_proj.weight",
|
| 131 |
+
"model.language_model.layers.20.linear_attn.out_proj.weight",
|
| 132 |
+
"model.language_model.layers.20.linear_attn.in_proj_qkv.weight",
|
| 133 |
+
"model.language_model.layers.20.linear_attn.in_proj_z.weight",
|
| 134 |
+
"model.language_model.layers.20.mlp.gate_proj.weight",
|
| 135 |
+
"model.language_model.layers.20.mlp.up_proj.weight",
|
| 136 |
+
"model.language_model.layers.20.mlp.down_proj.weight",
|
| 137 |
+
"model.language_model.layers.21.linear_attn.out_proj.weight",
|
| 138 |
+
"model.language_model.layers.21.linear_attn.in_proj_qkv.weight",
|
| 139 |
+
"model.language_model.layers.21.linear_attn.in_proj_z.weight",
|
| 140 |
+
"model.language_model.layers.21.mlp.gate_proj.weight",
|
| 141 |
+
"model.language_model.layers.21.mlp.up_proj.weight",
|
| 142 |
+
"model.language_model.layers.21.mlp.down_proj.weight",
|
| 143 |
+
"model.language_model.layers.22.linear_attn.out_proj.weight",
|
| 144 |
+
"model.language_model.layers.22.linear_attn.in_proj_qkv.weight",
|
| 145 |
+
"model.language_model.layers.22.linear_attn.in_proj_z.weight",
|
| 146 |
+
"model.language_model.layers.22.mlp.gate_proj.weight",
|
| 147 |
+
"model.language_model.layers.22.mlp.up_proj.weight",
|
| 148 |
+
"model.language_model.layers.22.mlp.down_proj.weight",
|
| 149 |
+
"model.language_model.layers.23.self_attn.q_proj.weight",
|
| 150 |
+
"model.language_model.layers.23.self_attn.k_proj.weight",
|
| 151 |
+
"model.language_model.layers.23.self_attn.v_proj.weight",
|
| 152 |
+
"model.language_model.layers.23.self_attn.o_proj.weight",
|
| 153 |
+
"model.language_model.layers.23.mlp.gate_proj.weight",
|
| 154 |
+
"model.language_model.layers.23.mlp.up_proj.weight",
|
| 155 |
+
"model.language_model.layers.23.mlp.down_proj.weight",
|
| 156 |
+
"model.language_model.layers.24.linear_attn.out_proj.weight",
|
| 157 |
+
"model.language_model.layers.24.linear_attn.in_proj_qkv.weight",
|
| 158 |
+
"model.language_model.layers.24.linear_attn.in_proj_z.weight",
|
| 159 |
+
"model.language_model.layers.24.mlp.gate_proj.weight",
|
| 160 |
+
"model.language_model.layers.24.mlp.up_proj.weight",
|
| 161 |
+
"model.language_model.layers.24.mlp.down_proj.weight",
|
| 162 |
+
"model.language_model.layers.25.linear_attn.out_proj.weight",
|
| 163 |
+
"model.language_model.layers.25.linear_attn.in_proj_qkv.weight",
|
| 164 |
+
"model.language_model.layers.25.linear_attn.in_proj_z.weight",
|
| 165 |
+
"model.language_model.layers.25.mlp.gate_proj.weight",
|
| 166 |
+
"model.language_model.layers.25.mlp.up_proj.weight",
|
| 167 |
+
"model.language_model.layers.25.mlp.down_proj.weight",
|
| 168 |
+
"model.language_model.layers.26.linear_attn.out_proj.weight",
|
| 169 |
+
"model.language_model.layers.26.linear_attn.in_proj_qkv.weight",
|
| 170 |
+
"model.language_model.layers.26.linear_attn.in_proj_z.weight",
|
| 171 |
+
"model.language_model.layers.26.mlp.gate_proj.weight",
|
| 172 |
+
"model.language_model.layers.26.mlp.up_proj.weight",
|
| 173 |
+
"model.language_model.layers.26.mlp.down_proj.weight",
|
| 174 |
+
"model.language_model.layers.27.self_attn.q_proj.weight",
|
| 175 |
+
"model.language_model.layers.27.self_attn.k_proj.weight",
|
| 176 |
+
"model.language_model.layers.27.self_attn.v_proj.weight",
|
| 177 |
+
"model.language_model.layers.27.self_attn.o_proj.weight",
|
| 178 |
+
"model.language_model.layers.27.mlp.gate_proj.weight",
|
| 179 |
+
"model.language_model.layers.27.mlp.up_proj.weight",
|
| 180 |
+
"model.language_model.layers.27.mlp.down_proj.weight",
|
| 181 |
+
"model.language_model.layers.28.linear_attn.out_proj.weight",
|
| 182 |
+
"model.language_model.layers.28.linear_attn.in_proj_qkv.weight",
|
| 183 |
+
"model.language_model.layers.28.linear_attn.in_proj_z.weight",
|
| 184 |
+
"model.language_model.layers.28.mlp.gate_proj.weight",
|
| 185 |
+
"model.language_model.layers.28.mlp.up_proj.weight",
|
| 186 |
+
"model.language_model.layers.28.mlp.down_proj.weight",
|
| 187 |
+
"model.language_model.layers.29.linear_attn.out_proj.weight",
|
| 188 |
+
"model.language_model.layers.29.linear_attn.in_proj_qkv.weight",
|
| 189 |
+
"model.language_model.layers.29.linear_attn.in_proj_z.weight",
|
| 190 |
+
"model.language_model.layers.29.mlp.gate_proj.weight",
|
| 191 |
+
"model.language_model.layers.29.mlp.up_proj.weight",
|
| 192 |
+
"model.language_model.layers.29.mlp.down_proj.weight",
|
| 193 |
+
"model.language_model.layers.30.linear_attn.out_proj.weight",
|
| 194 |
+
"model.language_model.layers.30.linear_attn.in_proj_qkv.weight",
|
| 195 |
+
"model.language_model.layers.30.linear_attn.in_proj_z.weight",
|
| 196 |
+
"model.language_model.layers.30.mlp.gate_proj.weight",
|
| 197 |
+
"model.language_model.layers.30.mlp.up_proj.weight",
|
| 198 |
+
"model.language_model.layers.30.mlp.down_proj.weight",
|
| 199 |
+
"model.language_model.layers.31.self_attn.q_proj.weight",
|
| 200 |
+
"model.language_model.layers.31.self_attn.k_proj.weight",
|
| 201 |
+
"model.language_model.layers.31.self_attn.v_proj.weight",
|
| 202 |
+
"model.language_model.layers.31.self_attn.o_proj.weight",
|
| 203 |
+
"model.language_model.layers.31.mlp.gate_proj.weight",
|
| 204 |
+
"model.language_model.layers.31.mlp.up_proj.weight",
|
| 205 |
+
"model.language_model.layers.31.mlp.down_proj.weight",
|
| 206 |
+
"model.language_model.layers.32.linear_attn.out_proj.weight",
|
| 207 |
+
"model.language_model.layers.32.linear_attn.in_proj_qkv.weight",
|
| 208 |
+
"model.language_model.layers.32.linear_attn.in_proj_z.weight",
|
| 209 |
+
"model.language_model.layers.32.mlp.gate_proj.weight",
|
| 210 |
+
"model.language_model.layers.32.mlp.up_proj.weight",
|
| 211 |
+
"model.language_model.layers.32.mlp.down_proj.weight",
|
| 212 |
+
"model.language_model.layers.33.linear_attn.out_proj.weight",
|
| 213 |
+
"model.language_model.layers.33.linear_attn.in_proj_qkv.weight",
|
| 214 |
+
"model.language_model.layers.33.linear_attn.in_proj_z.weight",
|
| 215 |
+
"model.language_model.layers.33.mlp.gate_proj.weight",
|
| 216 |
+
"model.language_model.layers.33.mlp.up_proj.weight",
|
| 217 |
+
"model.language_model.layers.33.mlp.down_proj.weight",
|
| 218 |
+
"model.language_model.layers.34.linear_attn.out_proj.weight",
|
| 219 |
+
"model.language_model.layers.34.linear_attn.in_proj_qkv.weight",
|
| 220 |
+
"model.language_model.layers.34.linear_attn.in_proj_z.weight",
|
| 221 |
+
"model.language_model.layers.34.mlp.gate_proj.weight",
|
| 222 |
+
"model.language_model.layers.34.mlp.up_proj.weight",
|
| 223 |
+
"model.language_model.layers.34.mlp.down_proj.weight",
|
| 224 |
+
"model.language_model.layers.35.self_attn.q_proj.weight",
|
| 225 |
+
"model.language_model.layers.35.self_attn.k_proj.weight",
|
| 226 |
+
"model.language_model.layers.35.self_attn.v_proj.weight",
|
| 227 |
+
"model.language_model.layers.35.self_attn.o_proj.weight",
|
| 228 |
+
"model.language_model.layers.35.mlp.gate_proj.weight",
|
| 229 |
+
"model.language_model.layers.35.mlp.up_proj.weight",
|
| 230 |
+
"model.language_model.layers.35.mlp.down_proj.weight",
|
| 231 |
+
"model.language_model.layers.36.linear_attn.out_proj.weight",
|
| 232 |
+
"model.language_model.layers.36.linear_attn.in_proj_qkv.weight",
|
| 233 |
+
"model.language_model.layers.36.linear_attn.in_proj_z.weight",
|
| 234 |
+
"model.language_model.layers.36.mlp.gate_proj.weight",
|
| 235 |
+
"model.language_model.layers.36.mlp.up_proj.weight",
|
| 236 |
+
"model.language_model.layers.36.mlp.down_proj.weight",
|
| 237 |
+
"model.language_model.layers.37.linear_attn.out_proj.weight",
|
| 238 |
+
"model.language_model.layers.37.linear_attn.in_proj_qkv.weight",
|
| 239 |
+
"model.language_model.layers.37.linear_attn.in_proj_z.weight",
|
| 240 |
+
"model.language_model.layers.37.mlp.gate_proj.weight",
|
| 241 |
+
"model.language_model.layers.37.mlp.up_proj.weight",
|
| 242 |
+
"model.language_model.layers.37.mlp.down_proj.weight",
|
| 243 |
+
"model.language_model.layers.38.linear_attn.out_proj.weight",
|
| 244 |
+
"model.language_model.layers.38.linear_attn.in_proj_qkv.weight",
|
| 245 |
+
"model.language_model.layers.38.linear_attn.in_proj_z.weight",
|
| 246 |
+
"model.language_model.layers.38.mlp.gate_proj.weight",
|
| 247 |
+
"model.language_model.layers.38.mlp.up_proj.weight",
|
| 248 |
+
"model.language_model.layers.38.mlp.down_proj.weight",
|
| 249 |
+
"model.language_model.layers.39.self_attn.q_proj.weight",
|
| 250 |
+
"model.language_model.layers.39.self_attn.k_proj.weight",
|
| 251 |
+
"model.language_model.layers.39.self_attn.v_proj.weight",
|
| 252 |
+
"model.language_model.layers.39.self_attn.o_proj.weight",
|
| 253 |
+
"model.language_model.layers.39.mlp.gate_proj.weight",
|
| 254 |
+
"model.language_model.layers.39.mlp.up_proj.weight",
|
| 255 |
+
"model.language_model.layers.39.mlp.down_proj.weight",
|
| 256 |
+
"model.language_model.layers.40.linear_attn.out_proj.weight",
|
| 257 |
+
"model.language_model.layers.40.linear_attn.in_proj_qkv.weight",
|
| 258 |
+
"model.language_model.layers.40.linear_attn.in_proj_z.weight",
|
| 259 |
+
"model.language_model.layers.40.mlp.gate_proj.weight",
|
| 260 |
+
"model.language_model.layers.40.mlp.up_proj.weight",
|
| 261 |
+
"model.language_model.layers.40.mlp.down_proj.weight",
|
| 262 |
+
"model.language_model.layers.41.linear_attn.out_proj.weight",
|
| 263 |
+
"model.language_model.layers.41.linear_attn.in_proj_qkv.weight",
|
| 264 |
+
"model.language_model.layers.41.linear_attn.in_proj_z.weight",
|
| 265 |
+
"model.language_model.layers.41.mlp.gate_proj.weight",
|
| 266 |
+
"model.language_model.layers.41.mlp.up_proj.weight",
|
| 267 |
+
"model.language_model.layers.41.mlp.down_proj.weight",
|
| 268 |
+
"model.language_model.layers.42.linear_attn.out_proj.weight",
|
| 269 |
+
"model.language_model.layers.42.linear_attn.in_proj_qkv.weight",
|
| 270 |
+
"model.language_model.layers.42.linear_attn.in_proj_z.weight",
|
| 271 |
+
"model.language_model.layers.42.mlp.gate_proj.weight",
|
| 272 |
+
"model.language_model.layers.42.mlp.up_proj.weight",
|
| 273 |
+
"model.language_model.layers.42.mlp.down_proj.weight",
|
| 274 |
+
"model.language_model.layers.43.self_attn.q_proj.weight",
|
| 275 |
+
"model.language_model.layers.43.self_attn.k_proj.weight",
|
| 276 |
+
"model.language_model.layers.43.self_attn.v_proj.weight",
|
| 277 |
+
"model.language_model.layers.43.self_attn.o_proj.weight",
|
| 278 |
+
"model.language_model.layers.43.mlp.gate_proj.weight",
|
| 279 |
+
"model.language_model.layers.43.mlp.up_proj.weight",
|
| 280 |
+
"model.language_model.layers.43.mlp.down_proj.weight",
|
| 281 |
+
"model.language_model.layers.44.linear_attn.out_proj.weight",
|
| 282 |
+
"model.language_model.layers.44.linear_attn.in_proj_qkv.weight",
|
| 283 |
+
"model.language_model.layers.44.linear_attn.in_proj_z.weight",
|
| 284 |
+
"model.language_model.layers.44.mlp.gate_proj.weight",
|
| 285 |
+
"model.language_model.layers.44.mlp.up_proj.weight",
|
| 286 |
+
"model.language_model.layers.44.mlp.down_proj.weight",
|
| 287 |
+
"model.language_model.layers.45.linear_attn.out_proj.weight",
|
| 288 |
+
"model.language_model.layers.45.linear_attn.in_proj_qkv.weight",
|
| 289 |
+
"model.language_model.layers.45.linear_attn.in_proj_z.weight",
|
| 290 |
+
"model.language_model.layers.45.mlp.gate_proj.weight",
|
| 291 |
+
"model.language_model.layers.45.mlp.up_proj.weight",
|
| 292 |
+
"model.language_model.layers.45.mlp.down_proj.weight",
|
| 293 |
+
"model.language_model.layers.46.linear_attn.out_proj.weight",
|
| 294 |
+
"model.language_model.layers.46.linear_attn.in_proj_qkv.weight",
|
| 295 |
+
"model.language_model.layers.46.linear_attn.in_proj_z.weight",
|
| 296 |
+
"model.language_model.layers.46.mlp.gate_proj.weight",
|
| 297 |
+
"model.language_model.layers.46.mlp.up_proj.weight",
|
| 298 |
+
"model.language_model.layers.46.mlp.down_proj.weight",
|
| 299 |
+
"model.language_model.layers.47.self_attn.q_proj.weight",
|
| 300 |
+
"model.language_model.layers.47.self_attn.k_proj.weight",
|
| 301 |
+
"model.language_model.layers.47.self_attn.v_proj.weight",
|
| 302 |
+
"model.language_model.layers.47.self_attn.o_proj.weight",
|
| 303 |
+
"model.language_model.layers.47.mlp.gate_proj.weight",
|
| 304 |
+
"model.language_model.layers.47.mlp.up_proj.weight",
|
| 305 |
+
"model.language_model.layers.47.mlp.down_proj.weight",
|
| 306 |
+
"model.language_model.layers.48.linear_attn.out_proj.weight",
|
| 307 |
+
"model.language_model.layers.48.linear_attn.in_proj_qkv.weight",
|
| 308 |
+
"model.language_model.layers.48.linear_attn.in_proj_z.weight",
|
| 309 |
+
"model.language_model.layers.48.mlp.gate_proj.weight",
|
| 310 |
+
"model.language_model.layers.48.mlp.up_proj.weight",
|
| 311 |
+
"model.language_model.layers.48.mlp.down_proj.weight",
|
| 312 |
+
"model.language_model.layers.49.linear_attn.out_proj.weight",
|
| 313 |
+
"model.language_model.layers.49.linear_attn.in_proj_qkv.weight",
|
| 314 |
+
"model.language_model.layers.49.linear_attn.in_proj_z.weight",
|
| 315 |
+
"model.language_model.layers.49.mlp.gate_proj.weight",
|
| 316 |
+
"model.language_model.layers.49.mlp.up_proj.weight",
|
| 317 |
+
"model.language_model.layers.49.mlp.down_proj.weight",
|
| 318 |
+
"model.language_model.layers.50.linear_attn.out_proj.weight",
|
| 319 |
+
"model.language_model.layers.50.linear_attn.in_proj_qkv.weight",
|
| 320 |
+
"model.language_model.layers.50.linear_attn.in_proj_z.weight",
|
| 321 |
+
"model.language_model.layers.50.mlp.gate_proj.weight",
|
| 322 |
+
"model.language_model.layers.50.mlp.up_proj.weight",
|
| 323 |
+
"model.language_model.layers.50.mlp.down_proj.weight",
|
| 324 |
+
"model.language_model.layers.51.self_attn.q_proj.weight",
|
| 325 |
+
"model.language_model.layers.51.self_attn.k_proj.weight",
|
| 326 |
+
"model.language_model.layers.51.self_attn.v_proj.weight",
|
| 327 |
+
"model.language_model.layers.51.self_attn.o_proj.weight",
|
| 328 |
+
"model.language_model.layers.51.mlp.gate_proj.weight",
|
| 329 |
+
"model.language_model.layers.51.mlp.up_proj.weight",
|
| 330 |
+
"model.language_model.layers.51.mlp.down_proj.weight",
|
| 331 |
+
"model.language_model.layers.52.linear_attn.out_proj.weight",
|
| 332 |
+
"model.language_model.layers.52.linear_attn.in_proj_qkv.weight",
|
| 333 |
+
"model.language_model.layers.52.linear_attn.in_proj_z.weight",
|
| 334 |
+
"model.language_model.layers.52.mlp.gate_proj.weight",
|
| 335 |
+
"model.language_model.layers.52.mlp.up_proj.weight",
|
| 336 |
+
"model.language_model.layers.52.mlp.down_proj.weight",
|
| 337 |
+
"model.language_model.layers.53.linear_attn.out_proj.weight",
|
| 338 |
+
"model.language_model.layers.53.linear_attn.in_proj_qkv.weight",
|
| 339 |
+
"model.language_model.layers.53.linear_attn.in_proj_z.weight",
|
| 340 |
+
"model.language_model.layers.53.mlp.gate_proj.weight",
|
| 341 |
+
"model.language_model.layers.53.mlp.up_proj.weight",
|
| 342 |
+
"model.language_model.layers.53.mlp.down_proj.weight",
|
| 343 |
+
"model.language_model.layers.54.linear_attn.out_proj.weight",
|
| 344 |
+
"model.language_model.layers.54.linear_attn.in_proj_qkv.weight",
|
| 345 |
+
"model.language_model.layers.54.linear_attn.in_proj_z.weight",
|
| 346 |
+
"model.language_model.layers.54.mlp.gate_proj.weight",
|
| 347 |
+
"model.language_model.layers.54.mlp.up_proj.weight",
|
| 348 |
+
"model.language_model.layers.54.mlp.down_proj.weight",
|
| 349 |
+
"model.language_model.layers.55.self_attn.q_proj.weight",
|
| 350 |
+
"model.language_model.layers.55.self_attn.k_proj.weight",
|
| 351 |
+
"model.language_model.layers.55.self_attn.v_proj.weight",
|
| 352 |
+
"model.language_model.layers.55.self_attn.o_proj.weight",
|
| 353 |
+
"model.language_model.layers.55.mlp.gate_proj.weight",
|
| 354 |
+
"model.language_model.layers.55.mlp.up_proj.weight",
|
| 355 |
+
"model.language_model.layers.55.mlp.down_proj.weight",
|
| 356 |
+
"model.language_model.layers.56.linear_attn.out_proj.weight",
|
| 357 |
+
"model.language_model.layers.56.linear_attn.in_proj_qkv.weight",
|
| 358 |
+
"model.language_model.layers.56.linear_attn.in_proj_z.weight",
|
| 359 |
+
"model.language_model.layers.56.mlp.gate_proj.weight",
|
| 360 |
+
"model.language_model.layers.56.mlp.up_proj.weight",
|
| 361 |
+
"model.language_model.layers.56.mlp.down_proj.weight",
|
| 362 |
+
"model.language_model.layers.57.linear_attn.out_proj.weight",
|
| 363 |
+
"model.language_model.layers.57.linear_attn.in_proj_qkv.weight",
|
| 364 |
+
"model.language_model.layers.57.linear_attn.in_proj_z.weight",
|
| 365 |
+
"model.language_model.layers.57.mlp.gate_proj.weight",
|
| 366 |
+
"model.language_model.layers.57.mlp.up_proj.weight",
|
| 367 |
+
"model.language_model.layers.57.mlp.down_proj.weight",
|
| 368 |
+
"model.language_model.layers.58.linear_attn.out_proj.weight",
|
| 369 |
+
"model.language_model.layers.58.linear_attn.in_proj_qkv.weight",
|
| 370 |
+
"model.language_model.layers.58.linear_attn.in_proj_z.weight",
|
| 371 |
+
"model.language_model.layers.58.mlp.gate_proj.weight",
|
| 372 |
+
"model.language_model.layers.58.mlp.up_proj.weight",
|
| 373 |
+
"model.language_model.layers.58.mlp.down_proj.weight",
|
| 374 |
+
"model.language_model.layers.59.self_attn.q_proj.weight",
|
| 375 |
+
"model.language_model.layers.59.self_attn.k_proj.weight",
|
| 376 |
+
"model.language_model.layers.59.self_attn.v_proj.weight",
|
| 377 |
+
"model.language_model.layers.59.self_attn.o_proj.weight",
|
| 378 |
+
"model.language_model.layers.59.mlp.gate_proj.weight",
|
| 379 |
+
"model.language_model.layers.59.mlp.up_proj.weight",
|
| 380 |
+
"model.language_model.layers.59.mlp.down_proj.weight",
|
| 381 |
+
"model.language_model.layers.60.linear_attn.out_proj.weight",
|
| 382 |
+
"model.language_model.layers.60.linear_attn.in_proj_qkv.weight",
|
| 383 |
+
"model.language_model.layers.60.linear_attn.in_proj_z.weight",
|
| 384 |
+
"model.language_model.layers.60.mlp.gate_proj.weight",
|
| 385 |
+
"model.language_model.layers.60.mlp.up_proj.weight",
|
| 386 |
+
"model.language_model.layers.60.mlp.down_proj.weight",
|
| 387 |
+
"model.language_model.layers.61.linear_attn.out_proj.weight",
|
| 388 |
+
"model.language_model.layers.61.linear_attn.in_proj_qkv.weight",
|
| 389 |
+
"model.language_model.layers.61.linear_attn.in_proj_z.weight",
|
| 390 |
+
"model.language_model.layers.61.mlp.gate_proj.weight",
|
| 391 |
+
"model.language_model.layers.61.mlp.up_proj.weight",
|
| 392 |
+
"model.language_model.layers.61.mlp.down_proj.weight",
|
| 393 |
+
"model.language_model.layers.62.linear_attn.out_proj.weight",
|
| 394 |
+
"model.language_model.layers.62.linear_attn.in_proj_qkv.weight",
|
| 395 |
+
"model.language_model.layers.62.linear_attn.in_proj_z.weight",
|
| 396 |
+
"model.language_model.layers.62.mlp.gate_proj.weight",
|
| 397 |
+
"model.language_model.layers.62.mlp.up_proj.weight",
|
| 398 |
+
"model.language_model.layers.62.mlp.down_proj.weight",
|
| 399 |
+
"model.language_model.layers.63.self_attn.q_proj.weight",
|
| 400 |
+
"model.language_model.layers.63.self_attn.k_proj.weight",
|
| 401 |
+
"model.language_model.layers.63.self_attn.v_proj.weight",
|
| 402 |
+
"model.language_model.layers.63.self_attn.o_proj.weight",
|
| 403 |
+
"model.language_model.layers.63.mlp.gate_proj.weight",
|
| 404 |
+
"model.language_model.layers.63.mlp.up_proj.weight",
|
| 405 |
+
"model.language_model.layers.63.mlp.down_proj.weight"
|
| 406 |
+
],
|
| 407 |
+
"mtp": "matching Hauhau head retained in BF16",
|
| 408 |
+
"elapsed_seconds": 124.82400365800004,
|
| 409 |
+
"complete": true
|
| 410 |
+
}
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/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": "5.15.0"
|
| 13 |
+
}
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00000.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9371b3c01e6b8fc586c9da33d0ae1b29d206cb930e45847f02c3626caf121ecb
|
| 3 |
+
size 2542796928
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00001.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:65df23e620a8371d8a3b791947d1a119dc3355f006a115f3d3deb1ff46972159
|
| 3 |
+
size 2023623272
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00002.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9cb14086fd0460295b1f7e4d48e98232eb5fbc3268a0662fd6562564367b7963
|
| 3 |
+
size 2087464624
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00003.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d3186cfca81f36c644da237725556d0a2bd3d9b535baef2080a5658cce290cfa
|
| 3 |
+
size 2017309632
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00004.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1418e7e4176e3ef22fc8673123c20ac8dc84686ecc863bfa8468f49064cca453
|
| 3 |
+
size 2082220088
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00005.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:935660b5cce2276b74fb40dda107f1bcd8fc0abe87dccb60f22d0e43814c6ab4
|
| 3 |
+
size 2086456824
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00006.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d51d3c24a2327815aa58214a7a45d95821717f237f556477fe74ac50d224f5e8
|
| 3 |
+
size 2012089576
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00007.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d51ce7e051480e36218f8dcce7db3199643b85c79c6e977d1fbc90230fc56b69
|
| 3 |
+
size 2087464728
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00008.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b6d56d06507d1d0e9c77e37c3c7036cd8d30b5abeababe6b70f660084b7dc16f
|
| 3 |
+
size 2060252928
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00009.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c9e95379f567007f312a5e24033bc59935b1ee4c2df7f6356a2f3e5ca434c204
|
| 3 |
+
size 2039276792
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00010.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3191b39e99871d97be9e1b6bc5485c3caa7e77ce54a5a4cde8269766461cf932
|
| 3 |
+
size 2086456832
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00011.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cea09579aa996755bc365ab2ab0ae501998b48672b0b12fd607930e2fcf741e0
|
| 3 |
+
size 2013071632
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00012.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bc531dde00bd7609180aa18b282e391a48de1b3c3032016d647a5505cf1cb5d2
|
| 3 |
+
size 4334508936
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model-00013.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5a3db9777591c229af34aeb30198f99c7653b8d10beff38ffc752b6dae60b576
|
| 3 |
+
size 849400392
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/processor_config.json
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"image_processor": {
|
| 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 |
+
},
|
| 28 |
+
"processor_class": "Qwen3VLProcessor",
|
| 29 |
+
"video_processor": {
|
| 30 |
+
"do_convert_rgb": true,
|
| 31 |
+
"do_normalize": true,
|
| 32 |
+
"do_rescale": true,
|
| 33 |
+
"do_resize": true,
|
| 34 |
+
"do_sample_frames": true,
|
| 35 |
+
"fps": 2,
|
| 36 |
+
"image_mean": [
|
| 37 |
+
0.5,
|
| 38 |
+
0.5,
|
| 39 |
+
0.5
|
| 40 |
+
],
|
| 41 |
+
"image_std": [
|
| 42 |
+
0.5,
|
| 43 |
+
0.5,
|
| 44 |
+
0.5
|
| 45 |
+
],
|
| 46 |
+
"max_frames": 768,
|
| 47 |
+
"merge_size": 2,
|
| 48 |
+
"min_frames": 4,
|
| 49 |
+
"patch_size": 16,
|
| 50 |
+
"resample": 3,
|
| 51 |
+
"rescale_factor": 0.00392156862745098,
|
| 52 |
+
"return_metadata": false,
|
| 53 |
+
"size": {
|
| 54 |
+
"longest_edge": 25165824,
|
| 55 |
+
"shortest_edge": 4096
|
| 56 |
+
},
|
| 57 |
+
"temporal_patch_size": 2,
|
| 58 |
+
"video_processor_type": "Qwen3VLVideoProcessor"
|
| 59 |
+
}
|
| 60 |
+
}
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/results/precision-comparison.json
ADDED
|
@@ -0,0 +1,382 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"date_utc": "2026-10-06",
|
| 3 |
+
"hardware": "2 x NVIDIA RTX 3090 24 GiB, NVLink, 125 GiB host RAM",
|
| 4 |
+
"vllm": "0.31.0",
|
| 5 |
+
"context_limit": 200000,
|
| 6 |
+
"kv_cache": "turboquant_4bit_nc",
|
| 7 |
+
"host_kv_offload_gib": 90,
|
| 8 |
+
"tensor_parallel_size": 2,
|
| 9 |
+
"sampling": {
|
| 10 |
+
"temperature": 0.6,
|
| 11 |
+
"top_p": 0.95,
|
| 12 |
+
"top_k": 20,
|
| 13 |
+
"presence_penalty": 0.0
|
| 14 |
+
},
|
| 15 |
+
"thinking": true,
|
| 16 |
+
"variants": {
|
| 17 |
+
"hauhau-w8a8": {
|
| 18 |
+
"long_context": {
|
| 19 |
+
"variant": "hauhau-w8a8",
|
| 20 |
+
"baseline": "model-tools/endy-long-context-20261005T172133Z",
|
| 21 |
+
"turns": 8,
|
| 22 |
+
"checks_passed": 53,
|
| 23 |
+
"checks_total": 54,
|
| 24 |
+
"strict_json_turns": 7,
|
| 25 |
+
"tool_calls": 2,
|
| 26 |
+
"median_decode_tokens_per_second": 46.885000000000005,
|
| 27 |
+
"grading_note": "Original checks retained; inspect semantic cyber findings separately."
|
| 28 |
+
},
|
| 29 |
+
"cybersecurity": {
|
| 30 |
+
"variant": "hauhau-w8a8",
|
| 31 |
+
"cases": 6,
|
| 32 |
+
"checks_passed": 30,
|
| 33 |
+
"checks_total": 43,
|
| 34 |
+
"strict_json_cases": 4,
|
| 35 |
+
"checks_passed_after_clarification": 34,
|
| 36 |
+
"note": "Small synthetic assessment, not a published benchmark or comprehensive security validation."
|
| 37 |
+
},
|
| 38 |
+
"speed": {
|
| 39 |
+
"variant": "hauhau-w8a8",
|
| 40 |
+
"output_tokens_per_request": 1024,
|
| 41 |
+
"sampling": {
|
| 42 |
+
"temperature": 0.6,
|
| 43 |
+
"top_p": 0.95,
|
| 44 |
+
"top_k": 20,
|
| 45 |
+
"presence_penalty": 0
|
| 46 |
+
},
|
| 47 |
+
"thinking_enabled": true,
|
| 48 |
+
"requests": [
|
| 49 |
+
{
|
| 50 |
+
"label": "short_0",
|
| 51 |
+
"input_tokens": 34,
|
| 52 |
+
"output_tokens": 1024,
|
| 53 |
+
"wall_seconds": 6.534,
|
| 54 |
+
"first_token_seconds": 0.753,
|
| 55 |
+
"decode_tokens_per_second": 177.144,
|
| 56 |
+
"warmup": true
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"label": "short_1",
|
| 60 |
+
"input_tokens": 34,
|
| 61 |
+
"output_tokens": 1024,
|
| 62 |
+
"wall_seconds": 6.022,
|
| 63 |
+
"first_token_seconds": 0.126,
|
| 64 |
+
"decode_tokens_per_second": 173.698,
|
| 65 |
+
"warmup": false
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"label": "short_2",
|
| 69 |
+
"input_tokens": 34,
|
| 70 |
+
"output_tokens": 1024,
|
| 71 |
+
"wall_seconds": 5.938,
|
| 72 |
+
"first_token_seconds": 0.116,
|
| 73 |
+
"decode_tokens_per_second": 175.89,
|
| 74 |
+
"warmup": false
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"label": "long_0",
|
| 78 |
+
"input_tokens": 121616,
|
| 79 |
+
"output_tokens": 1024,
|
| 80 |
+
"wall_seconds": 40.625,
|
| 81 |
+
"first_token_seconds": 18.331,
|
| 82 |
+
"decode_tokens_per_second": 45.932,
|
| 83 |
+
"warmup": true
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"label": "long_1",
|
| 87 |
+
"input_tokens": 121616,
|
| 88 |
+
"output_tokens": 1024,
|
| 89 |
+
"wall_seconds": 23.206,
|
| 90 |
+
"first_token_seconds": 1.08,
|
| 91 |
+
"decode_tokens_per_second": 46.28,
|
| 92 |
+
"warmup": false
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"label": "long_2",
|
| 96 |
+
"input_tokens": 121616,
|
| 97 |
+
"output_tokens": 1024,
|
| 98 |
+
"wall_seconds": 23.118,
|
| 99 |
+
"first_token_seconds": 1.079,
|
| 100 |
+
"decode_tokens_per_second": 46.462,
|
| 101 |
+
"warmup": false
|
| 102 |
+
}
|
| 103 |
+
],
|
| 104 |
+
"short_median_decode_tps": 174.79399999999998,
|
| 105 |
+
"long_median_decode_tps": 46.371,
|
| 106 |
+
"four_concurrent": {
|
| 107 |
+
"wall_seconds": 11.758,
|
| 108 |
+
"aggregate_output_tokens_per_second": 348.353,
|
| 109 |
+
"requests": [
|
| 110 |
+
{
|
| 111 |
+
"label": "concurrent_0",
|
| 112 |
+
"input_tokens": 34,
|
| 113 |
+
"output_tokens": 1024,
|
| 114 |
+
"wall_seconds": 11.74,
|
| 115 |
+
"first_token_seconds": 0.128,
|
| 116 |
+
"decode_tokens_per_second": 88.188
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"label": "concurrent_1",
|
| 120 |
+
"input_tokens": 34,
|
| 121 |
+
"output_tokens": 1024,
|
| 122 |
+
"wall_seconds": 11.754,
|
| 123 |
+
"first_token_seconds": 3.292,
|
| 124 |
+
"decode_tokens_per_second": 121.005
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"label": "concurrent_2",
|
| 128 |
+
"input_tokens": 34,
|
| 129 |
+
"output_tokens": 1024,
|
| 130 |
+
"wall_seconds": 11.753,
|
| 131 |
+
"first_token_seconds": 3.291,
|
| 132 |
+
"decode_tokens_per_second": 121.006
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"label": "concurrent_3",
|
| 136 |
+
"input_tokens": 34,
|
| 137 |
+
"output_tokens": 1024,
|
| 138 |
+
"wall_seconds": 11.752,
|
| 139 |
+
"first_token_seconds": 3.29,
|
| 140 |
+
"decode_tokens_per_second": 121.009
|
| 141 |
+
}
|
| 142 |
+
]
|
| 143 |
+
}
|
| 144 |
+
},
|
| 145 |
+
"custom_metadata": {
|
| 146 |
+
"valid_tool_cases": 6,
|
| 147 |
+
"total_cases": 6,
|
| 148 |
+
"correct_fields": 30,
|
| 149 |
+
"total_fields": 30,
|
| 150 |
+
"raw_json_cases": 5
|
| 151 |
+
},
|
| 152 |
+
"executed_sql_fixture": {
|
| 153 |
+
"passed": 2,
|
| 154 |
+
"total": 2
|
| 155 |
+
},
|
| 156 |
+
"native_api_integration": [
|
| 157 |
+
{
|
| 158 |
+
"test": "model_alias_and_underlying_checkpoint",
|
| 159 |
+
"passed": true
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"test": "anthropic_text_through_caddy",
|
| 163 |
+
"passed": true
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"test": "anthropic_streaming_through_caddy",
|
| 167 |
+
"passed": true
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"test": "anthropic_tool_call_and_result_roundtrip",
|
| 171 |
+
"passed": true
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"test": "anthropic_vision_two_distinct_images",
|
| 175 |
+
"passed": true
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"test": "eight_concurrent_anthropic_requests",
|
| 179 |
+
"passed": true,
|
| 180 |
+
"wall_seconds": 1.0,
|
| 181 |
+
"total_output_tokens": 578
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"test": "public_https_anthropic_endpoint",
|
| 185 |
+
"passed": true
|
| 186 |
+
}
|
| 187 |
+
],
|
| 188 |
+
"claude_code": {
|
| 189 |
+
"wall_seconds": 2.721,
|
| 190 |
+
"exit_code": 0
|
| 191 |
+
}
|
| 192 |
+
},
|
| 193 |
+
"hauhau-w8a16": {
|
| 194 |
+
"long_context": {
|
| 195 |
+
"variant": "hauhau-w8a16",
|
| 196 |
+
"baseline": "/home/isnotreal/model-tools/endy-long-context-20261005T172133Z",
|
| 197 |
+
"turns": 8,
|
| 198 |
+
"checks_passed": 53,
|
| 199 |
+
"checks_total": 54,
|
| 200 |
+
"strict_json_turns": 7,
|
| 201 |
+
"tool_calls": 2,
|
| 202 |
+
"median_decode_tokens_per_second": 47.269999999999996,
|
| 203 |
+
"grading_note": "Original checks retained; inspect semantic cyber findings separately."
|
| 204 |
+
},
|
| 205 |
+
"cybersecurity": {
|
| 206 |
+
"variant": "hauhau-w8a16",
|
| 207 |
+
"cases": 6,
|
| 208 |
+
"checks_passed": 28,
|
| 209 |
+
"checks_total": 43,
|
| 210 |
+
"strict_json_cases": 6,
|
| 211 |
+
"checks_passed_after_clarification": 36,
|
| 212 |
+
"note": "Small synthetic assessment, not a published benchmark or comprehensive security validation."
|
| 213 |
+
},
|
| 214 |
+
"speed": {
|
| 215 |
+
"variant": "hauhau-w8a16",
|
| 216 |
+
"output_tokens_per_request": 1024,
|
| 217 |
+
"sampling": {
|
| 218 |
+
"temperature": 0.6,
|
| 219 |
+
"top_p": 0.95,
|
| 220 |
+
"top_k": 20,
|
| 221 |
+
"presence_penalty": 0
|
| 222 |
+
},
|
| 223 |
+
"thinking_enabled": true,
|
| 224 |
+
"requests": [
|
| 225 |
+
{
|
| 226 |
+
"label": "short_0",
|
| 227 |
+
"input_tokens": 34,
|
| 228 |
+
"output_tokens": 1024,
|
| 229 |
+
"wall_seconds": 5.891,
|
| 230 |
+
"first_token_seconds": 0.131,
|
| 231 |
+
"decode_tokens_per_second": 177.768,
|
| 232 |
+
"warmup": true
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"label": "short_1",
|
| 236 |
+
"input_tokens": 34,
|
| 237 |
+
"output_tokens": 1024,
|
| 238 |
+
"wall_seconds": 6.107,
|
| 239 |
+
"first_token_seconds": 0.113,
|
| 240 |
+
"decode_tokens_per_second": 170.834,
|
| 241 |
+
"warmup": false
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"label": "short_2",
|
| 245 |
+
"input_tokens": 34,
|
| 246 |
+
"output_tokens": 1024,
|
| 247 |
+
"wall_seconds": 6.021,
|
| 248 |
+
"first_token_seconds": 0.115,
|
| 249 |
+
"decode_tokens_per_second": 173.38,
|
| 250 |
+
"warmup": false
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"label": "long_0",
|
| 254 |
+
"input_tokens": 121616,
|
| 255 |
+
"output_tokens": 1024,
|
| 256 |
+
"wall_seconds": 44.096,
|
| 257 |
+
"first_token_seconds": 21.895,
|
| 258 |
+
"decode_tokens_per_second": 46.124,
|
| 259 |
+
"warmup": true
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"label": "long_1",
|
| 263 |
+
"input_tokens": 121616,
|
| 264 |
+
"output_tokens": 1024,
|
| 265 |
+
"wall_seconds": 23.291,
|
| 266 |
+
"first_token_seconds": 1.18,
|
| 267 |
+
"decode_tokens_per_second": 46.312,
|
| 268 |
+
"warmup": false
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"label": "long_2",
|
| 272 |
+
"input_tokens": 121616,
|
| 273 |
+
"output_tokens": 1024,
|
| 274 |
+
"wall_seconds": 23.12,
|
| 275 |
+
"first_token_seconds": 1.17,
|
| 276 |
+
"decode_tokens_per_second": 46.652,
|
| 277 |
+
"warmup": false
|
| 278 |
+
}
|
| 279 |
+
],
|
| 280 |
+
"short_median_decode_tps": 172.107,
|
| 281 |
+
"long_median_decode_tps": 46.482,
|
| 282 |
+
"four_concurrent": {
|
| 283 |
+
"wall_seconds": 10.924,
|
| 284 |
+
"aggregate_output_tokens_per_second": 374.959,
|
| 285 |
+
"requests": [
|
| 286 |
+
{
|
| 287 |
+
"label": "concurrent_0",
|
| 288 |
+
"input_tokens": 34,
|
| 289 |
+
"output_tokens": 1024,
|
| 290 |
+
"wall_seconds": 10.907,
|
| 291 |
+
"first_token_seconds": 0.118,
|
| 292 |
+
"decode_tokens_per_second": 94.911
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"label": "concurrent_1",
|
| 296 |
+
"input_tokens": 34,
|
| 297 |
+
"output_tokens": 1024,
|
| 298 |
+
"wall_seconds": 10.914,
|
| 299 |
+
"first_token_seconds": 0.221,
|
| 300 |
+
"decode_tokens_per_second": 95.758
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"label": "concurrent_2",
|
| 304 |
+
"input_tokens": 34,
|
| 305 |
+
"output_tokens": 1024,
|
| 306 |
+
"wall_seconds": 10.921,
|
| 307 |
+
"first_token_seconds": 2.561,
|
| 308 |
+
"decode_tokens_per_second": 122.487
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"label": "concurrent_3",
|
| 312 |
+
"input_tokens": 34,
|
| 313 |
+
"output_tokens": 1024,
|
| 314 |
+
"wall_seconds": 10.92,
|
| 315 |
+
"first_token_seconds": 2.56,
|
| 316 |
+
"decode_tokens_per_second": 122.494
|
| 317 |
+
}
|
| 318 |
+
]
|
| 319 |
+
}
|
| 320 |
+
},
|
| 321 |
+
"custom_metadata": {
|
| 322 |
+
"valid_tool_cases": 6,
|
| 323 |
+
"total_cases": 6,
|
| 324 |
+
"correct_fields": 30,
|
| 325 |
+
"total_fields": 30,
|
| 326 |
+
"raw_json_cases": 4
|
| 327 |
+
},
|
| 328 |
+
"executed_sql_fixture": {
|
| 329 |
+
"passed": 2,
|
| 330 |
+
"total": 2
|
| 331 |
+
}
|
| 332 |
+
},
|
| 333 |
+
"endy-w8a16": {
|
| 334 |
+
"long_context": {
|
| 335 |
+
"variant": "endy-w8a16-vllm-0.31-qwen-coding-preset",
|
| 336 |
+
"baseline": "/home/isnotreal/model-tools/endy-long-context-20261005T172133Z",
|
| 337 |
+
"turns": 8,
|
| 338 |
+
"checks_passed": 46,
|
| 339 |
+
"checks_total": 54,
|
| 340 |
+
"strict_json_turns": 0,
|
| 341 |
+
"tool_calls": 0,
|
| 342 |
+
"median_decode_tokens_per_second": 47.545,
|
| 343 |
+
"grading_note": "Original checks retained; inspect semantic cyber findings separately."
|
| 344 |
+
},
|
| 345 |
+
"custom_metadata": {
|
| 346 |
+
"fresh_valid_tool_cases": 2,
|
| 347 |
+
"fresh_cases": 2,
|
| 348 |
+
"compacted_valid_tool_cases": 0,
|
| 349 |
+
"compacted_cases": 2,
|
| 350 |
+
"long_valid_tool_cases": 0,
|
| 351 |
+
"long_cases": 2
|
| 352 |
+
},
|
| 353 |
+
"note": "Endy can call tools on fresh prompts. Reliability degraded on the compacted and long-history custom-metadata fixtures."
|
| 354 |
+
},
|
| 355 |
+
"endy-w8a8": {
|
| 356 |
+
"long_context": {
|
| 357 |
+
"variant": "endy-w8a8-vllm-0.31",
|
| 358 |
+
"baseline": "/home/isnotreal/model-tools/endy-long-context-20261005T172133Z",
|
| 359 |
+
"turns": 8,
|
| 360 |
+
"checks_passed": 27,
|
| 361 |
+
"checks_total": 54,
|
| 362 |
+
"strict_json_turns": 0,
|
| 363 |
+
"tool_calls": 0,
|
| 364 |
+
"median_decode_tokens_per_second": 47.515,
|
| 365 |
+
"grading_note": "Original checks retained; inspect semantic cyber findings separately."
|
| 366 |
+
},
|
| 367 |
+
"note": "Earlier W8A8 long-context run used different sampling; not a matched activation-precision comparison."
|
| 368 |
+
}
|
| 369 |
+
},
|
| 370 |
+
"limitations": [
|
| 371 |
+
"Small synthetic local assessment, not a published cybersecurity benchmark.",
|
| 372 |
+
"Long-context tests independently replayed eight original 122k-124k-token prefixes, not an eight-turn fresh conversation.",
|
| 373 |
+
"The one failed long-context strict check per Hauhau variant was a literal regex mismatch; the proposed SQL was executed successfully in a fixture.",
|
| 374 |
+
"The four-concurrent speed measurement is one batch and can include first-shape compilation; it is not a stable concurrency ranking.",
|
| 375 |
+
"Both variants made substantive security-policy mistakes on the broader quiz; manual review is necessary.",
|
| 376 |
+
"W8A8 SSRF clarification exhausted the 4096-output-token cap and truncated JSON; its secondary score includes that failure.",
|
| 377 |
+
"The native W8A8 export uses dynamic INT8 activations throughout eligible language linears; the older SmoothQuant export is a different mixed W8A8/W8A16 artifact.",
|
| 378 |
+
"Hauhau weights originate in Q8_K_P GGUF, then reconstruction and INT8 conversion. W8A16 and W8A8 use identical INT8 weight values, but native kernels and scale handling differ.",
|
| 379 |
+
"MTP and vision weights are retained; speculative decoding was not enabled.",
|
| 380 |
+
"Two serial repeats after warmup and one concurrency batch are too few to attribute small differences to precision alone."
|
| 381 |
+
]
|
| 382 |
+
}
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523
|
| 3 |
+
size 19989325
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/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": 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 |
+
"processor_class": "Qwen3VLProcessor",
|
| 27 |
+
"split_special_tokens": false,
|
| 28 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 29 |
+
"unk_token": null,
|
| 30 |
+
"video_token": "<|video_pad|>",
|
| 31 |
+
"vision_bos_token": "<|vision_start|>",
|
| 32 |
+
"vision_eos_token": "<|vision_end|>"
|
| 33 |
+
}
|
qwen3.8-27b-hauhau-w8a8-vision-mtp/vision.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7defc90994f7bbc58a8feb8d3c4144895babb434a9be9f47e88ae3646e1df33
|
| 3 |
+
size 921497224
|