Upload folder using huggingface_hub
Browse files- .gitattributes +1 -0
- README.md +18 -0
- chat_template.jinja +154 -0
- config.json +1628 -0
- generation_config.json +10 -0
- model.safetensors +3 -0
- model.safetensors.index.json +999 -0
- optiq/metadata.json +756 -0
- optiq/mtp.safetensors +3 -0
- optiq/optiq_vision.safetensors +3 -0
- optiq/sensitivity.json +2246 -0
- tokenizer.json +3 -0
- tokenizer_config.json +33 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: empero-ai/Qwen3.8-2B-Distill
|
| 4 |
+
language:
|
| 5 |
+
- en
|
| 6 |
+
library_name: mlx
|
| 7 |
+
pipeline_tag: text-generation
|
| 8 |
+
tags:
|
| 9 |
+
- empero-ai
|
| 10 |
+
- qwen3.5
|
| 11 |
+
- qwen3.8
|
| 12 |
+
- distillation
|
| 13 |
+
- reasoning
|
| 14 |
+
- function-calling
|
| 15 |
+
- sft
|
| 16 |
+
- edge
|
| 17 |
+
- mlx
|
| 18 |
+
---
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set image_count = namespace(value=0) %}
|
| 2 |
+
{%- set video_count = namespace(value=0) %}
|
| 3 |
+
{%- macro render_content(content, do_vision_count, is_system_content=false) %}
|
| 4 |
+
{%- if content is string %}
|
| 5 |
+
{{- content }}
|
| 6 |
+
{%- elif content is iterable and content is not mapping %}
|
| 7 |
+
{%- for item in content %}
|
| 8 |
+
{%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
|
| 9 |
+
{%- if is_system_content %}
|
| 10 |
+
{{- raise_exception('System message cannot contain images.') }}
|
| 11 |
+
{%- endif %}
|
| 12 |
+
{%- if do_vision_count %}
|
| 13 |
+
{%- set image_count.value = image_count.value + 1 %}
|
| 14 |
+
{%- endif %}
|
| 15 |
+
{%- if add_vision_id %}
|
| 16 |
+
{{- 'Picture ' ~ image_count.value ~ ': ' }}
|
| 17 |
+
{%- endif %}
|
| 18 |
+
{{- '<|vision_start|><|image_pad|><|vision_end|>' }}
|
| 19 |
+
{%- elif 'video' in item or item.type == 'video' %}
|
| 20 |
+
{%- if is_system_content %}
|
| 21 |
+
{{- raise_exception('System message cannot contain videos.') }}
|
| 22 |
+
{%- endif %}
|
| 23 |
+
{%- if do_vision_count %}
|
| 24 |
+
{%- set video_count.value = video_count.value + 1 %}
|
| 25 |
+
{%- endif %}
|
| 26 |
+
{%- if add_vision_id %}
|
| 27 |
+
{{- 'Video ' ~ video_count.value ~ ': ' }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{{- '<|vision_start|><|video_pad|><|vision_end|>' }}
|
| 30 |
+
{%- elif 'text' in item %}
|
| 31 |
+
{{- item.text }}
|
| 32 |
+
{%- else %}
|
| 33 |
+
{{- raise_exception('Unexpected item type in content.') }}
|
| 34 |
+
{%- endif %}
|
| 35 |
+
{%- endfor %}
|
| 36 |
+
{%- elif content is none or content is undefined %}
|
| 37 |
+
{{- '' }}
|
| 38 |
+
{%- else %}
|
| 39 |
+
{{- raise_exception('Unexpected content type.') }}
|
| 40 |
+
{%- endif %}
|
| 41 |
+
{%- endmacro %}
|
| 42 |
+
{%- if not messages %}
|
| 43 |
+
{{- raise_exception('No messages provided.') }}
|
| 44 |
+
{%- endif %}
|
| 45 |
+
{%- if tools and tools is iterable and tools is not mapping %}
|
| 46 |
+
{{- '<|im_start|>system\n' }}
|
| 47 |
+
{{- "# Tools\n\nYou have access to the following functions:\n\n<tools>" }}
|
| 48 |
+
{%- for tool in tools %}
|
| 49 |
+
{{- "\n" }}
|
| 50 |
+
{{- tool | tojson }}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{{- "\n</tools>" }}
|
| 53 |
+
{{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
|
| 54 |
+
{%- if messages[0].role == 'system' %}
|
| 55 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 56 |
+
{%- if content %}
|
| 57 |
+
{{- '\n\n' + content }}
|
| 58 |
+
{%- endif %}
|
| 59 |
+
{%- endif %}
|
| 60 |
+
{{- '<|im_end|>\n' }}
|
| 61 |
+
{%- else %}
|
| 62 |
+
{%- if messages[0].role == 'system' %}
|
| 63 |
+
{%- set content = render_content(messages[0].content, false, true)|trim %}
|
| 64 |
+
{{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
|
| 65 |
+
{%- endif %}
|
| 66 |
+
{%- endif %}
|
| 67 |
+
{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
|
| 68 |
+
{%- for message in messages[::-1] %}
|
| 69 |
+
{%- set index = (messages|length - 1) - loop.index0 %}
|
| 70 |
+
{%- if ns.multi_step_tool and message.role == "user" %}
|
| 71 |
+
{%- set content = render_content(message.content, false)|trim %}
|
| 72 |
+
{%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
|
| 73 |
+
{%- set ns.multi_step_tool = false %}
|
| 74 |
+
{%- set ns.last_query_index = index %}
|
| 75 |
+
{%- endif %}
|
| 76 |
+
{%- endif %}
|
| 77 |
+
{%- endfor %}
|
| 78 |
+
{%- if ns.multi_step_tool %}
|
| 79 |
+
{{- raise_exception('No user query found in messages.') }}
|
| 80 |
+
{%- endif %}
|
| 81 |
+
{%- for message in messages %}
|
| 82 |
+
{%- set content = render_content(message.content, true)|trim %}
|
| 83 |
+
{%- if message.role == "system" %}
|
| 84 |
+
{%- if not loop.first %}
|
| 85 |
+
{{- raise_exception('System message must be at the beginning.') }}
|
| 86 |
+
{%- endif %}
|
| 87 |
+
{%- elif message.role == "user" %}
|
| 88 |
+
{{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
|
| 89 |
+
{%- elif message.role == "assistant" %}
|
| 90 |
+
{%- set reasoning_content = '' %}
|
| 91 |
+
{%- if message.reasoning_content is string %}
|
| 92 |
+
{%- set reasoning_content = message.reasoning_content %}
|
| 93 |
+
{%- else %}
|
| 94 |
+
{%- if '</think>' in content %}
|
| 95 |
+
{%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
|
| 96 |
+
{%- set content = content.split('</think>')[-1].lstrip('\n') %}
|
| 97 |
+
{%- endif %}
|
| 98 |
+
{%- endif %}
|
| 99 |
+
{%- set reasoning_content = reasoning_content|trim %}
|
| 100 |
+
{%- if loop.index0 > ns.last_query_index %}
|
| 101 |
+
{{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content + '\n</think>\n\n' + content }}
|
| 102 |
+
{%- else %}
|
| 103 |
+
{{- '<|im_start|>' + message.role + '\n' + content }}
|
| 104 |
+
{%- endif %}
|
| 105 |
+
{%- if message.tool_calls and message.tool_calls is iterable and message.tool_calls is not mapping %}
|
| 106 |
+
{%- for tool_call in message.tool_calls %}
|
| 107 |
+
{%- if tool_call.function is defined %}
|
| 108 |
+
{%- set tool_call = tool_call.function %}
|
| 109 |
+
{%- endif %}
|
| 110 |
+
{%- if loop.first %}
|
| 111 |
+
{%- if content|trim %}
|
| 112 |
+
{{- '\n\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 113 |
+
{%- else %}
|
| 114 |
+
{{- '<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 115 |
+
{%- endif %}
|
| 116 |
+
{%- else %}
|
| 117 |
+
{{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
|
| 118 |
+
{%- endif %}
|
| 119 |
+
{%- if tool_call.arguments is defined %}
|
| 120 |
+
{%- for args_name, args_value in tool_call.arguments|items %}
|
| 121 |
+
{{- '<parameter=' + args_name + '>\n' }}
|
| 122 |
+
{%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
|
| 123 |
+
{{- args_value }}
|
| 124 |
+
{{- '\n</parameter>\n' }}
|
| 125 |
+
{%- endfor %}
|
| 126 |
+
{%- endif %}
|
| 127 |
+
{{- '</function>\n</tool_call>' }}
|
| 128 |
+
{%- endfor %}
|
| 129 |
+
{%- endif %}
|
| 130 |
+
{{- '<|im_end|>\n' }}
|
| 131 |
+
{%- elif message.role == "tool" %}
|
| 132 |
+
{%- if loop.previtem and loop.previtem.role != "tool" %}
|
| 133 |
+
{{- '<|im_start|>user' }}
|
| 134 |
+
{%- endif %}
|
| 135 |
+
{{- '\n<tool_response>\n' }}
|
| 136 |
+
{{- content }}
|
| 137 |
+
{{- '\n</tool_response>' }}
|
| 138 |
+
{%- if not loop.last and loop.nextitem.role != "tool" %}
|
| 139 |
+
{{- '<|im_end|>\n' }}
|
| 140 |
+
{%- elif loop.last %}
|
| 141 |
+
{{- '<|im_end|>\n' }}
|
| 142 |
+
{%- endif %}
|
| 143 |
+
{%- else %}
|
| 144 |
+
{{- raise_exception('Unexpected message role.') }}
|
| 145 |
+
{%- endif %}
|
| 146 |
+
{%- endfor %}
|
| 147 |
+
{%- if add_generation_prompt %}
|
| 148 |
+
{{- '<|im_start|>assistant\n' }}
|
| 149 |
+
{%- if enable_thinking is defined and enable_thinking is true %}
|
| 150 |
+
{{- '<think>\n' }}
|
| 151 |
+
{%- else %}
|
| 152 |
+
{{- '<think>\n\n</think>\n\n' }}
|
| 153 |
+
{%- endif %}
|
| 154 |
+
{%- endif %}
|
config.json
ADDED
|
@@ -0,0 +1,1628 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3_5ForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"eos_token_id": [
|
| 6 |
+
248046,
|
| 7 |
+
248044
|
| 8 |
+
],
|
| 9 |
+
"image_token_id": 248056,
|
| 10 |
+
"model_type": "qwen3_5",
|
| 11 |
+
"pad_token_id": 248044,
|
| 12 |
+
"quantization": {
|
| 13 |
+
"group_size": 64,
|
| 14 |
+
"bits": 4,
|
| 15 |
+
"mode": "affine",
|
| 16 |
+
"language_model.model.embed_tokens": {
|
| 17 |
+
"bits": 6,
|
| 18 |
+
"group_size": 64
|
| 19 |
+
},
|
| 20 |
+
"language_model.model.layers.0.linear_attn.in_proj_qkv": {
|
| 21 |
+
"bits": 6,
|
| 22 |
+
"group_size": 64
|
| 23 |
+
},
|
| 24 |
+
"language_model.model.layers.0.linear_attn.in_proj_z": {
|
| 25 |
+
"bits": 6,
|
| 26 |
+
"group_size": 64
|
| 27 |
+
},
|
| 28 |
+
"language_model.model.layers.0.linear_attn.in_proj_b": {
|
| 29 |
+
"bits": 6,
|
| 30 |
+
"group_size": 64
|
| 31 |
+
},
|
| 32 |
+
"language_model.model.layers.0.linear_attn.in_proj_a": {
|
| 33 |
+
"bits": 6,
|
| 34 |
+
"group_size": 64
|
| 35 |
+
},
|
| 36 |
+
"language_model.model.layers.0.linear_attn.out_proj": {
|
| 37 |
+
"bits": 6,
|
| 38 |
+
"group_size": 64
|
| 39 |
+
},
|
| 40 |
+
"language_model.model.layers.0.mlp.gate_proj": {
|
| 41 |
+
"bits": 6,
|
| 42 |
+
"group_size": 64
|
| 43 |
+
},
|
| 44 |
+
"language_model.model.layers.0.mlp.down_proj": {
|
| 45 |
+
"bits": 6,
|
| 46 |
+
"group_size": 64
|
| 47 |
+
},
|
| 48 |
+
"language_model.model.layers.0.mlp.up_proj": {
|
| 49 |
+
"bits": 6,
|
| 50 |
+
"group_size": 64
|
| 51 |
+
},
|
| 52 |
+
"language_model.model.layers.1.linear_attn.in_proj_qkv": {
|
| 53 |
+
"bits": 3,
|
| 54 |
+
"group_size": 64
|
| 55 |
+
},
|
| 56 |
+
"language_model.model.layers.1.linear_attn.in_proj_z": {
|
| 57 |
+
"bits": 3,
|
| 58 |
+
"group_size": 64
|
| 59 |
+
},
|
| 60 |
+
"language_model.model.layers.1.linear_attn.in_proj_b": {
|
| 61 |
+
"bits": 4,
|
| 62 |
+
"group_size": 64
|
| 63 |
+
},
|
| 64 |
+
"language_model.model.layers.1.linear_attn.in_proj_a": {
|
| 65 |
+
"bits": 6,
|
| 66 |
+
"group_size": 64
|
| 67 |
+
},
|
| 68 |
+
"language_model.model.layers.1.linear_attn.out_proj": {
|
| 69 |
+
"bits": 4,
|
| 70 |
+
"group_size": 64
|
| 71 |
+
},
|
| 72 |
+
"language_model.model.layers.1.mlp.gate_proj": {
|
| 73 |
+
"bits": 3,
|
| 74 |
+
"group_size": 64
|
| 75 |
+
},
|
| 76 |
+
"language_model.model.layers.1.mlp.down_proj": {
|
| 77 |
+
"bits": 3,
|
| 78 |
+
"group_size": 64
|
| 79 |
+
},
|
| 80 |
+
"language_model.model.layers.1.mlp.up_proj": {
|
| 81 |
+
"bits": 3,
|
| 82 |
+
"group_size": 64
|
| 83 |
+
},
|
| 84 |
+
"language_model.model.layers.2.linear_attn.in_proj_qkv": {
|
| 85 |
+
"bits": 3,
|
| 86 |
+
"group_size": 64
|
| 87 |
+
},
|
| 88 |
+
"language_model.model.layers.2.linear_attn.in_proj_z": {
|
| 89 |
+
"bits": 3,
|
| 90 |
+
"group_size": 64
|
| 91 |
+
},
|
| 92 |
+
"language_model.model.layers.2.linear_attn.in_proj_b": {
|
| 93 |
+
"bits": 5,
|
| 94 |
+
"group_size": 64
|
| 95 |
+
},
|
| 96 |
+
"language_model.model.layers.2.linear_attn.in_proj_a": {
|
| 97 |
+
"bits": 4,
|
| 98 |
+
"group_size": 64
|
| 99 |
+
},
|
| 100 |
+
"language_model.model.layers.2.linear_attn.out_proj": {
|
| 101 |
+
"bits": 4,
|
| 102 |
+
"group_size": 64
|
| 103 |
+
},
|
| 104 |
+
"language_model.model.layers.2.mlp.gate_proj": {
|
| 105 |
+
"bits": 3,
|
| 106 |
+
"group_size": 64
|
| 107 |
+
},
|
| 108 |
+
"language_model.model.layers.2.mlp.down_proj": {
|
| 109 |
+
"bits": 3,
|
| 110 |
+
"group_size": 64
|
| 111 |
+
},
|
| 112 |
+
"language_model.model.layers.2.mlp.up_proj": {
|
| 113 |
+
"bits": 3,
|
| 114 |
+
"group_size": 64
|
| 115 |
+
},
|
| 116 |
+
"language_model.model.layers.3.self_attn.q_proj": {
|
| 117 |
+
"bits": 3,
|
| 118 |
+
"group_size": 64
|
| 119 |
+
},
|
| 120 |
+
"language_model.model.layers.3.self_attn.k_proj": {
|
| 121 |
+
"bits": 3,
|
| 122 |
+
"group_size": 64
|
| 123 |
+
},
|
| 124 |
+
"language_model.model.layers.3.self_attn.v_proj": {
|
| 125 |
+
"bits": 5,
|
| 126 |
+
"group_size": 64
|
| 127 |
+
},
|
| 128 |
+
"language_model.model.layers.3.self_attn.o_proj": {
|
| 129 |
+
"bits": 4,
|
| 130 |
+
"group_size": 64
|
| 131 |
+
},
|
| 132 |
+
"language_model.model.layers.3.mlp.gate_proj": {
|
| 133 |
+
"bits": 3,
|
| 134 |
+
"group_size": 64
|
| 135 |
+
},
|
| 136 |
+
"language_model.model.layers.3.mlp.down_proj": {
|
| 137 |
+
"bits": 3,
|
| 138 |
+
"group_size": 64
|
| 139 |
+
},
|
| 140 |
+
"language_model.model.layers.3.mlp.up_proj": {
|
| 141 |
+
"bits": 3,
|
| 142 |
+
"group_size": 64
|
| 143 |
+
},
|
| 144 |
+
"language_model.model.layers.4.linear_attn.in_proj_qkv": {
|
| 145 |
+
"bits": 3,
|
| 146 |
+
"group_size": 64
|
| 147 |
+
},
|
| 148 |
+
"language_model.model.layers.4.linear_attn.in_proj_z": {
|
| 149 |
+
"bits": 3,
|
| 150 |
+
"group_size": 64
|
| 151 |
+
},
|
| 152 |
+
"language_model.model.layers.4.linear_attn.in_proj_b": {
|
| 153 |
+
"bits": 5,
|
| 154 |
+
"group_size": 64
|
| 155 |
+
},
|
| 156 |
+
"language_model.model.layers.4.linear_attn.in_proj_a": {
|
| 157 |
+
"bits": 4,
|
| 158 |
+
"group_size": 64
|
| 159 |
+
},
|
| 160 |
+
"language_model.model.layers.4.linear_attn.out_proj": {
|
| 161 |
+
"bits": 3,
|
| 162 |
+
"group_size": 64
|
| 163 |
+
},
|
| 164 |
+
"language_model.model.layers.4.mlp.gate_proj": {
|
| 165 |
+
"bits": 3,
|
| 166 |
+
"group_size": 64
|
| 167 |
+
},
|
| 168 |
+
"language_model.model.layers.4.mlp.down_proj": {
|
| 169 |
+
"bits": 3,
|
| 170 |
+
"group_size": 64
|
| 171 |
+
},
|
| 172 |
+
"language_model.model.layers.4.mlp.up_proj": {
|
| 173 |
+
"bits": 3,
|
| 174 |
+
"group_size": 64
|
| 175 |
+
},
|
| 176 |
+
"language_model.model.layers.5.linear_attn.in_proj_qkv": {
|
| 177 |
+
"bits": 3,
|
| 178 |
+
"group_size": 64
|
| 179 |
+
},
|
| 180 |
+
"language_model.model.layers.5.linear_attn.in_proj_z": {
|
| 181 |
+
"bits": 3,
|
| 182 |
+
"group_size": 64
|
| 183 |
+
},
|
| 184 |
+
"language_model.model.layers.5.linear_attn.in_proj_b": {
|
| 185 |
+
"bits": 5,
|
| 186 |
+
"group_size": 64
|
| 187 |
+
},
|
| 188 |
+
"language_model.model.layers.5.linear_attn.in_proj_a": {
|
| 189 |
+
"bits": 4,
|
| 190 |
+
"group_size": 64
|
| 191 |
+
},
|
| 192 |
+
"language_model.model.layers.5.linear_attn.out_proj": {
|
| 193 |
+
"bits": 3,
|
| 194 |
+
"group_size": 64
|
| 195 |
+
},
|
| 196 |
+
"language_model.model.layers.5.mlp.gate_proj": {
|
| 197 |
+
"bits": 3,
|
| 198 |
+
"group_size": 64
|
| 199 |
+
},
|
| 200 |
+
"language_model.model.layers.5.mlp.down_proj": {
|
| 201 |
+
"bits": 3,
|
| 202 |
+
"group_size": 64
|
| 203 |
+
},
|
| 204 |
+
"language_model.model.layers.5.mlp.up_proj": {
|
| 205 |
+
"bits": 3,
|
| 206 |
+
"group_size": 64
|
| 207 |
+
},
|
| 208 |
+
"language_model.model.layers.6.linear_attn.in_proj_qkv": {
|
| 209 |
+
"bits": 3,
|
| 210 |
+
"group_size": 64
|
| 211 |
+
},
|
| 212 |
+
"language_model.model.layers.6.linear_attn.in_proj_z": {
|
| 213 |
+
"bits": 3,
|
| 214 |
+
"group_size": 64
|
| 215 |
+
},
|
| 216 |
+
"language_model.model.layers.6.linear_attn.in_proj_b": {
|
| 217 |
+
"bits": 5,
|
| 218 |
+
"group_size": 64
|
| 219 |
+
},
|
| 220 |
+
"language_model.model.layers.6.linear_attn.in_proj_a": {
|
| 221 |
+
"bits": 4,
|
| 222 |
+
"group_size": 64
|
| 223 |
+
},
|
| 224 |
+
"language_model.model.layers.6.linear_attn.out_proj": {
|
| 225 |
+
"bits": 4,
|
| 226 |
+
"group_size": 64
|
| 227 |
+
},
|
| 228 |
+
"language_model.model.layers.6.mlp.gate_proj": {
|
| 229 |
+
"bits": 3,
|
| 230 |
+
"group_size": 64
|
| 231 |
+
},
|
| 232 |
+
"language_model.model.layers.6.mlp.down_proj": {
|
| 233 |
+
"bits": 3,
|
| 234 |
+
"group_size": 64
|
| 235 |
+
},
|
| 236 |
+
"language_model.model.layers.6.mlp.up_proj": {
|
| 237 |
+
"bits": 3,
|
| 238 |
+
"group_size": 64
|
| 239 |
+
},
|
| 240 |
+
"language_model.model.layers.7.self_attn.q_proj": {
|
| 241 |
+
"bits": 3,
|
| 242 |
+
"group_size": 64
|
| 243 |
+
},
|
| 244 |
+
"language_model.model.layers.7.self_attn.k_proj": {
|
| 245 |
+
"bits": 3,
|
| 246 |
+
"group_size": 64
|
| 247 |
+
},
|
| 248 |
+
"language_model.model.layers.7.self_attn.v_proj": {
|
| 249 |
+
"bits": 5,
|
| 250 |
+
"group_size": 64
|
| 251 |
+
},
|
| 252 |
+
"language_model.model.layers.7.self_attn.o_proj": {
|
| 253 |
+
"bits": 4,
|
| 254 |
+
"group_size": 64
|
| 255 |
+
},
|
| 256 |
+
"language_model.model.layers.7.mlp.gate_proj": {
|
| 257 |
+
"bits": 2,
|
| 258 |
+
"group_size": 64
|
| 259 |
+
},
|
| 260 |
+
"language_model.model.layers.7.mlp.down_proj": {
|
| 261 |
+
"bits": 3,
|
| 262 |
+
"group_size": 64
|
| 263 |
+
},
|
| 264 |
+
"language_model.model.layers.7.mlp.up_proj": {
|
| 265 |
+
"bits": 3,
|
| 266 |
+
"group_size": 64
|
| 267 |
+
},
|
| 268 |
+
"language_model.model.layers.8.linear_attn.in_proj_qkv": {
|
| 269 |
+
"bits": 3,
|
| 270 |
+
"group_size": 64
|
| 271 |
+
},
|
| 272 |
+
"language_model.model.layers.8.linear_attn.in_proj_z": {
|
| 273 |
+
"bits": 3,
|
| 274 |
+
"group_size": 64
|
| 275 |
+
},
|
| 276 |
+
"language_model.model.layers.8.linear_attn.in_proj_b": {
|
| 277 |
+
"bits": 4,
|
| 278 |
+
"group_size": 64
|
| 279 |
+
},
|
| 280 |
+
"language_model.model.layers.8.linear_attn.in_proj_a": {
|
| 281 |
+
"bits": 4,
|
| 282 |
+
"group_size": 64
|
| 283 |
+
},
|
| 284 |
+
"language_model.model.layers.8.linear_attn.out_proj": {
|
| 285 |
+
"bits": 3,
|
| 286 |
+
"group_size": 64
|
| 287 |
+
},
|
| 288 |
+
"language_model.model.layers.8.mlp.gate_proj": {
|
| 289 |
+
"bits": 2,
|
| 290 |
+
"group_size": 64
|
| 291 |
+
},
|
| 292 |
+
"language_model.model.layers.8.mlp.down_proj": {
|
| 293 |
+
"bits": 3,
|
| 294 |
+
"group_size": 64
|
| 295 |
+
},
|
| 296 |
+
"language_model.model.layers.8.mlp.up_proj": {
|
| 297 |
+
"bits": 3,
|
| 298 |
+
"group_size": 64
|
| 299 |
+
},
|
| 300 |
+
"language_model.model.layers.9.linear_attn.in_proj_qkv": {
|
| 301 |
+
"bits": 3,
|
| 302 |
+
"group_size": 64
|
| 303 |
+
},
|
| 304 |
+
"language_model.model.layers.9.linear_attn.in_proj_z": {
|
| 305 |
+
"bits": 3,
|
| 306 |
+
"group_size": 64
|
| 307 |
+
},
|
| 308 |
+
"language_model.model.layers.9.linear_attn.in_proj_b": {
|
| 309 |
+
"bits": 5,
|
| 310 |
+
"group_size": 64
|
| 311 |
+
},
|
| 312 |
+
"language_model.model.layers.9.linear_attn.in_proj_a": {
|
| 313 |
+
"bits": 4,
|
| 314 |
+
"group_size": 64
|
| 315 |
+
},
|
| 316 |
+
"language_model.model.layers.9.linear_attn.out_proj": {
|
| 317 |
+
"bits": 3,
|
| 318 |
+
"group_size": 64
|
| 319 |
+
},
|
| 320 |
+
"language_model.model.layers.9.mlp.gate_proj": {
|
| 321 |
+
"bits": 2,
|
| 322 |
+
"group_size": 64
|
| 323 |
+
},
|
| 324 |
+
"language_model.model.layers.9.mlp.down_proj": {
|
| 325 |
+
"bits": 2,
|
| 326 |
+
"group_size": 64
|
| 327 |
+
},
|
| 328 |
+
"language_model.model.layers.9.mlp.up_proj": {
|
| 329 |
+
"bits": 2,
|
| 330 |
+
"group_size": 64
|
| 331 |
+
},
|
| 332 |
+
"language_model.model.layers.10.linear_attn.in_proj_qkv": {
|
| 333 |
+
"bits": 3,
|
| 334 |
+
"group_size": 64
|
| 335 |
+
},
|
| 336 |
+
"language_model.model.layers.10.linear_attn.in_proj_z": {
|
| 337 |
+
"bits": 2,
|
| 338 |
+
"group_size": 64
|
| 339 |
+
},
|
| 340 |
+
"language_model.model.layers.10.linear_attn.in_proj_b": {
|
| 341 |
+
"bits": 4,
|
| 342 |
+
"group_size": 64
|
| 343 |
+
},
|
| 344 |
+
"language_model.model.layers.10.linear_attn.in_proj_a": {
|
| 345 |
+
"bits": 4,
|
| 346 |
+
"group_size": 64
|
| 347 |
+
},
|
| 348 |
+
"language_model.model.layers.10.linear_attn.out_proj": {
|
| 349 |
+
"bits": 3,
|
| 350 |
+
"group_size": 64
|
| 351 |
+
},
|
| 352 |
+
"language_model.model.layers.10.mlp.gate_proj": {
|
| 353 |
+
"bits": 2,
|
| 354 |
+
"group_size": 64
|
| 355 |
+
},
|
| 356 |
+
"language_model.model.layers.10.mlp.down_proj": {
|
| 357 |
+
"bits": 3,
|
| 358 |
+
"group_size": 64
|
| 359 |
+
},
|
| 360 |
+
"language_model.model.layers.10.mlp.up_proj": {
|
| 361 |
+
"bits": 3,
|
| 362 |
+
"group_size": 64
|
| 363 |
+
},
|
| 364 |
+
"language_model.model.layers.11.self_attn.q_proj": {
|
| 365 |
+
"bits": 3,
|
| 366 |
+
"group_size": 64
|
| 367 |
+
},
|
| 368 |
+
"language_model.model.layers.11.self_attn.k_proj": {
|
| 369 |
+
"bits": 3,
|
| 370 |
+
"group_size": 64
|
| 371 |
+
},
|
| 372 |
+
"language_model.model.layers.11.self_attn.v_proj": {
|
| 373 |
+
"bits": 4,
|
| 374 |
+
"group_size": 64
|
| 375 |
+
},
|
| 376 |
+
"language_model.model.layers.11.self_attn.o_proj": {
|
| 377 |
+
"bits": 4,
|
| 378 |
+
"group_size": 64
|
| 379 |
+
},
|
| 380 |
+
"language_model.model.layers.11.mlp.gate_proj": {
|
| 381 |
+
"bits": 2,
|
| 382 |
+
"group_size": 64
|
| 383 |
+
},
|
| 384 |
+
"language_model.model.layers.11.mlp.down_proj": {
|
| 385 |
+
"bits": 2,
|
| 386 |
+
"group_size": 64
|
| 387 |
+
},
|
| 388 |
+
"language_model.model.layers.11.mlp.up_proj": {
|
| 389 |
+
"bits": 3,
|
| 390 |
+
"group_size": 64
|
| 391 |
+
},
|
| 392 |
+
"language_model.model.layers.12.linear_attn.in_proj_qkv": {
|
| 393 |
+
"bits": 3,
|
| 394 |
+
"group_size": 64
|
| 395 |
+
},
|
| 396 |
+
"language_model.model.layers.12.linear_attn.in_proj_z": {
|
| 397 |
+
"bits": 3,
|
| 398 |
+
"group_size": 64
|
| 399 |
+
},
|
| 400 |
+
"language_model.model.layers.12.linear_attn.in_proj_b": {
|
| 401 |
+
"bits": 5,
|
| 402 |
+
"group_size": 64
|
| 403 |
+
},
|
| 404 |
+
"language_model.model.layers.12.linear_attn.in_proj_a": {
|
| 405 |
+
"bits": 4,
|
| 406 |
+
"group_size": 64
|
| 407 |
+
},
|
| 408 |
+
"language_model.model.layers.12.linear_attn.out_proj": {
|
| 409 |
+
"bits": 3,
|
| 410 |
+
"group_size": 64
|
| 411 |
+
},
|
| 412 |
+
"language_model.model.layers.12.mlp.gate_proj": {
|
| 413 |
+
"bits": 2,
|
| 414 |
+
"group_size": 64
|
| 415 |
+
},
|
| 416 |
+
"language_model.model.layers.12.mlp.down_proj": {
|
| 417 |
+
"bits": 2,
|
| 418 |
+
"group_size": 64
|
| 419 |
+
},
|
| 420 |
+
"language_model.model.layers.12.mlp.up_proj": {
|
| 421 |
+
"bits": 2,
|
| 422 |
+
"group_size": 64
|
| 423 |
+
},
|
| 424 |
+
"language_model.model.layers.13.linear_attn.in_proj_qkv": {
|
| 425 |
+
"bits": 3,
|
| 426 |
+
"group_size": 64
|
| 427 |
+
},
|
| 428 |
+
"language_model.model.layers.13.linear_attn.in_proj_z": {
|
| 429 |
+
"bits": 2,
|
| 430 |
+
"group_size": 64
|
| 431 |
+
},
|
| 432 |
+
"language_model.model.layers.13.linear_attn.in_proj_b": {
|
| 433 |
+
"bits": 5,
|
| 434 |
+
"group_size": 64
|
| 435 |
+
},
|
| 436 |
+
"language_model.model.layers.13.linear_attn.in_proj_a": {
|
| 437 |
+
"bits": 4,
|
| 438 |
+
"group_size": 64
|
| 439 |
+
},
|
| 440 |
+
"language_model.model.layers.13.linear_attn.out_proj": {
|
| 441 |
+
"bits": 3,
|
| 442 |
+
"group_size": 64
|
| 443 |
+
},
|
| 444 |
+
"language_model.model.layers.13.mlp.gate_proj": {
|
| 445 |
+
"bits": 2,
|
| 446 |
+
"group_size": 64
|
| 447 |
+
},
|
| 448 |
+
"language_model.model.layers.13.mlp.down_proj": {
|
| 449 |
+
"bits": 2,
|
| 450 |
+
"group_size": 64
|
| 451 |
+
},
|
| 452 |
+
"language_model.model.layers.13.mlp.up_proj": {
|
| 453 |
+
"bits": 2,
|
| 454 |
+
"group_size": 64
|
| 455 |
+
},
|
| 456 |
+
"language_model.model.layers.14.linear_attn.in_proj_qkv": {
|
| 457 |
+
"bits": 3,
|
| 458 |
+
"group_size": 64
|
| 459 |
+
},
|
| 460 |
+
"language_model.model.layers.14.linear_attn.in_proj_z": {
|
| 461 |
+
"bits": 2,
|
| 462 |
+
"group_size": 64
|
| 463 |
+
},
|
| 464 |
+
"language_model.model.layers.14.linear_attn.in_proj_b": {
|
| 465 |
+
"bits": 4,
|
| 466 |
+
"group_size": 64
|
| 467 |
+
},
|
| 468 |
+
"language_model.model.layers.14.linear_attn.in_proj_a": {
|
| 469 |
+
"bits": 4,
|
| 470 |
+
"group_size": 64
|
| 471 |
+
},
|
| 472 |
+
"language_model.model.layers.14.linear_attn.out_proj": {
|
| 473 |
+
"bits": 3,
|
| 474 |
+
"group_size": 64
|
| 475 |
+
},
|
| 476 |
+
"language_model.model.layers.14.mlp.gate_proj": {
|
| 477 |
+
"bits": 2,
|
| 478 |
+
"group_size": 64
|
| 479 |
+
},
|
| 480 |
+
"language_model.model.layers.14.mlp.down_proj": {
|
| 481 |
+
"bits": 3,
|
| 482 |
+
"group_size": 64
|
| 483 |
+
},
|
| 484 |
+
"language_model.model.layers.14.mlp.up_proj": {
|
| 485 |
+
"bits": 3,
|
| 486 |
+
"group_size": 64
|
| 487 |
+
},
|
| 488 |
+
"language_model.model.layers.15.self_attn.q_proj": {
|
| 489 |
+
"bits": 2,
|
| 490 |
+
"group_size": 64
|
| 491 |
+
},
|
| 492 |
+
"language_model.model.layers.15.self_attn.k_proj": {
|
| 493 |
+
"bits": 3,
|
| 494 |
+
"group_size": 64
|
| 495 |
+
},
|
| 496 |
+
"language_model.model.layers.15.self_attn.v_proj": {
|
| 497 |
+
"bits": 4,
|
| 498 |
+
"group_size": 64
|
| 499 |
+
},
|
| 500 |
+
"language_model.model.layers.15.self_attn.o_proj": {
|
| 501 |
+
"bits": 4,
|
| 502 |
+
"group_size": 64
|
| 503 |
+
},
|
| 504 |
+
"language_model.model.layers.15.mlp.gate_proj": {
|
| 505 |
+
"bits": 2,
|
| 506 |
+
"group_size": 64
|
| 507 |
+
},
|
| 508 |
+
"language_model.model.layers.15.mlp.down_proj": {
|
| 509 |
+
"bits": 3,
|
| 510 |
+
"group_size": 64
|
| 511 |
+
},
|
| 512 |
+
"language_model.model.layers.15.mlp.up_proj": {
|
| 513 |
+
"bits": 3,
|
| 514 |
+
"group_size": 64
|
| 515 |
+
},
|
| 516 |
+
"language_model.model.layers.16.linear_attn.in_proj_qkv": {
|
| 517 |
+
"bits": 3,
|
| 518 |
+
"group_size": 64
|
| 519 |
+
},
|
| 520 |
+
"language_model.model.layers.16.linear_attn.in_proj_z": {
|
| 521 |
+
"bits": 2,
|
| 522 |
+
"group_size": 64
|
| 523 |
+
},
|
| 524 |
+
"language_model.model.layers.16.linear_attn.in_proj_b": {
|
| 525 |
+
"bits": 5,
|
| 526 |
+
"group_size": 64
|
| 527 |
+
},
|
| 528 |
+
"language_model.model.layers.16.linear_attn.in_proj_a": {
|
| 529 |
+
"bits": 4,
|
| 530 |
+
"group_size": 64
|
| 531 |
+
},
|
| 532 |
+
"language_model.model.layers.16.linear_attn.out_proj": {
|
| 533 |
+
"bits": 3,
|
| 534 |
+
"group_size": 64
|
| 535 |
+
},
|
| 536 |
+
"language_model.model.layers.16.mlp.gate_proj": {
|
| 537 |
+
"bits": 2,
|
| 538 |
+
"group_size": 64
|
| 539 |
+
},
|
| 540 |
+
"language_model.model.layers.16.mlp.down_proj": {
|
| 541 |
+
"bits": 3,
|
| 542 |
+
"group_size": 64
|
| 543 |
+
},
|
| 544 |
+
"language_model.model.layers.16.mlp.up_proj": {
|
| 545 |
+
"bits": 2,
|
| 546 |
+
"group_size": 64
|
| 547 |
+
},
|
| 548 |
+
"language_model.model.layers.17.linear_attn.in_proj_qkv": {
|
| 549 |
+
"bits": 3,
|
| 550 |
+
"group_size": 64
|
| 551 |
+
},
|
| 552 |
+
"language_model.model.layers.17.linear_attn.in_proj_z": {
|
| 553 |
+
"bits": 2,
|
| 554 |
+
"group_size": 64
|
| 555 |
+
},
|
| 556 |
+
"language_model.model.layers.17.linear_attn.in_proj_b": {
|
| 557 |
+
"bits": 5,
|
| 558 |
+
"group_size": 64
|
| 559 |
+
},
|
| 560 |
+
"language_model.model.layers.17.linear_attn.in_proj_a": {
|
| 561 |
+
"bits": 4,
|
| 562 |
+
"group_size": 64
|
| 563 |
+
},
|
| 564 |
+
"language_model.model.layers.17.linear_attn.out_proj": {
|
| 565 |
+
"bits": 3,
|
| 566 |
+
"group_size": 64
|
| 567 |
+
},
|
| 568 |
+
"language_model.model.layers.17.mlp.gate_proj": {
|
| 569 |
+
"bits": 2,
|
| 570 |
+
"group_size": 64
|
| 571 |
+
},
|
| 572 |
+
"language_model.model.layers.17.mlp.down_proj": {
|
| 573 |
+
"bits": 2,
|
| 574 |
+
"group_size": 64
|
| 575 |
+
},
|
| 576 |
+
"language_model.model.layers.17.mlp.up_proj": {
|
| 577 |
+
"bits": 2,
|
| 578 |
+
"group_size": 64
|
| 579 |
+
},
|
| 580 |
+
"language_model.model.layers.18.linear_attn.in_proj_qkv": {
|
| 581 |
+
"bits": 3,
|
| 582 |
+
"group_size": 64
|
| 583 |
+
},
|
| 584 |
+
"language_model.model.layers.18.linear_attn.in_proj_z": {
|
| 585 |
+
"bits": 2,
|
| 586 |
+
"group_size": 64
|
| 587 |
+
},
|
| 588 |
+
"language_model.model.layers.18.linear_attn.in_proj_b": {
|
| 589 |
+
"bits": 5,
|
| 590 |
+
"group_size": 64
|
| 591 |
+
},
|
| 592 |
+
"language_model.model.layers.18.linear_attn.in_proj_a": {
|
| 593 |
+
"bits": 3,
|
| 594 |
+
"group_size": 64
|
| 595 |
+
},
|
| 596 |
+
"language_model.model.layers.18.linear_attn.out_proj": {
|
| 597 |
+
"bits": 3,
|
| 598 |
+
"group_size": 64
|
| 599 |
+
},
|
| 600 |
+
"language_model.model.layers.18.mlp.gate_proj": {
|
| 601 |
+
"bits": 2,
|
| 602 |
+
"group_size": 64
|
| 603 |
+
},
|
| 604 |
+
"language_model.model.layers.18.mlp.down_proj": {
|
| 605 |
+
"bits": 2,
|
| 606 |
+
"group_size": 64
|
| 607 |
+
},
|
| 608 |
+
"language_model.model.layers.18.mlp.up_proj": {
|
| 609 |
+
"bits": 2,
|
| 610 |
+
"group_size": 64
|
| 611 |
+
},
|
| 612 |
+
"language_model.model.layers.19.self_attn.q_proj": {
|
| 613 |
+
"bits": 2,
|
| 614 |
+
"group_size": 64
|
| 615 |
+
},
|
| 616 |
+
"language_model.model.layers.19.self_attn.k_proj": {
|
| 617 |
+
"bits": 3,
|
| 618 |
+
"group_size": 64
|
| 619 |
+
},
|
| 620 |
+
"language_model.model.layers.19.self_attn.v_proj": {
|
| 621 |
+
"bits": 4,
|
| 622 |
+
"group_size": 64
|
| 623 |
+
},
|
| 624 |
+
"language_model.model.layers.19.self_attn.o_proj": {
|
| 625 |
+
"bits": 3,
|
| 626 |
+
"group_size": 64
|
| 627 |
+
},
|
| 628 |
+
"language_model.model.layers.19.mlp.gate_proj": {
|
| 629 |
+
"bits": 2,
|
| 630 |
+
"group_size": 64
|
| 631 |
+
},
|
| 632 |
+
"language_model.model.layers.19.mlp.down_proj": {
|
| 633 |
+
"bits": 2,
|
| 634 |
+
"group_size": 64
|
| 635 |
+
},
|
| 636 |
+
"language_model.model.layers.19.mlp.up_proj": {
|
| 637 |
+
"bits": 2,
|
| 638 |
+
"group_size": 64
|
| 639 |
+
},
|
| 640 |
+
"language_model.model.layers.20.linear_attn.in_proj_qkv": {
|
| 641 |
+
"bits": 3,
|
| 642 |
+
"group_size": 64
|
| 643 |
+
},
|
| 644 |
+
"language_model.model.layers.20.linear_attn.in_proj_z": {
|
| 645 |
+
"bits": 2,
|
| 646 |
+
"group_size": 64
|
| 647 |
+
},
|
| 648 |
+
"language_model.model.layers.20.linear_attn.in_proj_b": {
|
| 649 |
+
"bits": 4,
|
| 650 |
+
"group_size": 64
|
| 651 |
+
},
|
| 652 |
+
"language_model.model.layers.20.linear_attn.in_proj_a": {
|
| 653 |
+
"bits": 4,
|
| 654 |
+
"group_size": 64
|
| 655 |
+
},
|
| 656 |
+
"language_model.model.layers.20.linear_attn.out_proj": {
|
| 657 |
+
"bits": 3,
|
| 658 |
+
"group_size": 64
|
| 659 |
+
},
|
| 660 |
+
"language_model.model.layers.20.mlp.gate_proj": {
|
| 661 |
+
"bits": 2,
|
| 662 |
+
"group_size": 64
|
| 663 |
+
},
|
| 664 |
+
"language_model.model.layers.20.mlp.down_proj": {
|
| 665 |
+
"bits": 2,
|
| 666 |
+
"group_size": 64
|
| 667 |
+
},
|
| 668 |
+
"language_model.model.layers.20.mlp.up_proj": {
|
| 669 |
+
"bits": 2,
|
| 670 |
+
"group_size": 64
|
| 671 |
+
},
|
| 672 |
+
"language_model.model.layers.21.linear_attn.in_proj_qkv": {
|
| 673 |
+
"bits": 3,
|
| 674 |
+
"group_size": 64
|
| 675 |
+
},
|
| 676 |
+
"language_model.model.layers.21.linear_attn.in_proj_z": {
|
| 677 |
+
"bits": 2,
|
| 678 |
+
"group_size": 64
|
| 679 |
+
},
|
| 680 |
+
"language_model.model.layers.21.linear_attn.in_proj_b": {
|
| 681 |
+
"bits": 5,
|
| 682 |
+
"group_size": 64
|
| 683 |
+
},
|
| 684 |
+
"language_model.model.layers.21.linear_attn.in_proj_a": {
|
| 685 |
+
"bits": 3,
|
| 686 |
+
"group_size": 64
|
| 687 |
+
},
|
| 688 |
+
"language_model.model.layers.21.linear_attn.out_proj": {
|
| 689 |
+
"bits": 3,
|
| 690 |
+
"group_size": 64
|
| 691 |
+
},
|
| 692 |
+
"language_model.model.layers.21.mlp.gate_proj": {
|
| 693 |
+
"bits": 2,
|
| 694 |
+
"group_size": 64
|
| 695 |
+
},
|
| 696 |
+
"language_model.model.layers.21.mlp.down_proj": {
|
| 697 |
+
"bits": 2,
|
| 698 |
+
"group_size": 64
|
| 699 |
+
},
|
| 700 |
+
"language_model.model.layers.21.mlp.up_proj": {
|
| 701 |
+
"bits": 2,
|
| 702 |
+
"group_size": 64
|
| 703 |
+
},
|
| 704 |
+
"language_model.model.layers.22.linear_attn.in_proj_qkv": {
|
| 705 |
+
"bits": 3,
|
| 706 |
+
"group_size": 64
|
| 707 |
+
},
|
| 708 |
+
"language_model.model.layers.22.linear_attn.in_proj_z": {
|
| 709 |
+
"bits": 2,
|
| 710 |
+
"group_size": 64
|
| 711 |
+
},
|
| 712 |
+
"language_model.model.layers.22.linear_attn.in_proj_b": {
|
| 713 |
+
"bits": 5,
|
| 714 |
+
"group_size": 64
|
| 715 |
+
},
|
| 716 |
+
"language_model.model.layers.22.linear_attn.in_proj_a": {
|
| 717 |
+
"bits": 4,
|
| 718 |
+
"group_size": 64
|
| 719 |
+
},
|
| 720 |
+
"language_model.model.layers.22.linear_attn.out_proj": {
|
| 721 |
+
"bits": 3,
|
| 722 |
+
"group_size": 64
|
| 723 |
+
},
|
| 724 |
+
"language_model.model.layers.22.mlp.gate_proj": {
|
| 725 |
+
"bits": 2,
|
| 726 |
+
"group_size": 64
|
| 727 |
+
},
|
| 728 |
+
"language_model.model.layers.22.mlp.down_proj": {
|
| 729 |
+
"bits": 2,
|
| 730 |
+
"group_size": 64
|
| 731 |
+
},
|
| 732 |
+
"language_model.model.layers.22.mlp.up_proj": {
|
| 733 |
+
"bits": 2,
|
| 734 |
+
"group_size": 64
|
| 735 |
+
},
|
| 736 |
+
"language_model.model.layers.23.self_attn.q_proj": {
|
| 737 |
+
"bits": 6,
|
| 738 |
+
"group_size": 64
|
| 739 |
+
},
|
| 740 |
+
"language_model.model.layers.23.self_attn.k_proj": {
|
| 741 |
+
"bits": 6,
|
| 742 |
+
"group_size": 64
|
| 743 |
+
},
|
| 744 |
+
"language_model.model.layers.23.self_attn.v_proj": {
|
| 745 |
+
"bits": 6,
|
| 746 |
+
"group_size": 64
|
| 747 |
+
},
|
| 748 |
+
"language_model.model.layers.23.self_attn.o_proj": {
|
| 749 |
+
"bits": 6,
|
| 750 |
+
"group_size": 64
|
| 751 |
+
},
|
| 752 |
+
"language_model.model.layers.23.mlp.gate_proj": {
|
| 753 |
+
"bits": 6,
|
| 754 |
+
"group_size": 64
|
| 755 |
+
},
|
| 756 |
+
"language_model.model.layers.23.mlp.down_proj": {
|
| 757 |
+
"bits": 6,
|
| 758 |
+
"group_size": 64
|
| 759 |
+
},
|
| 760 |
+
"language_model.model.layers.23.mlp.up_proj": {
|
| 761 |
+
"bits": 6,
|
| 762 |
+
"group_size": 64
|
| 763 |
+
}
|
| 764 |
+
},
|
| 765 |
+
"quantization_config": {
|
| 766 |
+
"group_size": 64,
|
| 767 |
+
"bits": 4,
|
| 768 |
+
"mode": "affine",
|
| 769 |
+
"language_model.model.embed_tokens": {
|
| 770 |
+
"bits": 6,
|
| 771 |
+
"group_size": 64
|
| 772 |
+
},
|
| 773 |
+
"language_model.model.layers.0.linear_attn.in_proj_qkv": {
|
| 774 |
+
"bits": 6,
|
| 775 |
+
"group_size": 64
|
| 776 |
+
},
|
| 777 |
+
"language_model.model.layers.0.linear_attn.in_proj_z": {
|
| 778 |
+
"bits": 6,
|
| 779 |
+
"group_size": 64
|
| 780 |
+
},
|
| 781 |
+
"language_model.model.layers.0.linear_attn.in_proj_b": {
|
| 782 |
+
"bits": 6,
|
| 783 |
+
"group_size": 64
|
| 784 |
+
},
|
| 785 |
+
"language_model.model.layers.0.linear_attn.in_proj_a": {
|
| 786 |
+
"bits": 6,
|
| 787 |
+
"group_size": 64
|
| 788 |
+
},
|
| 789 |
+
"language_model.model.layers.0.linear_attn.out_proj": {
|
| 790 |
+
"bits": 6,
|
| 791 |
+
"group_size": 64
|
| 792 |
+
},
|
| 793 |
+
"language_model.model.layers.0.mlp.gate_proj": {
|
| 794 |
+
"bits": 6,
|
| 795 |
+
"group_size": 64
|
| 796 |
+
},
|
| 797 |
+
"language_model.model.layers.0.mlp.down_proj": {
|
| 798 |
+
"bits": 6,
|
| 799 |
+
"group_size": 64
|
| 800 |
+
},
|
| 801 |
+
"language_model.model.layers.0.mlp.up_proj": {
|
| 802 |
+
"bits": 6,
|
| 803 |
+
"group_size": 64
|
| 804 |
+
},
|
| 805 |
+
"language_model.model.layers.1.linear_attn.in_proj_qkv": {
|
| 806 |
+
"bits": 3,
|
| 807 |
+
"group_size": 64
|
| 808 |
+
},
|
| 809 |
+
"language_model.model.layers.1.linear_attn.in_proj_z": {
|
| 810 |
+
"bits": 3,
|
| 811 |
+
"group_size": 64
|
| 812 |
+
},
|
| 813 |
+
"language_model.model.layers.1.linear_attn.in_proj_b": {
|
| 814 |
+
"bits": 4,
|
| 815 |
+
"group_size": 64
|
| 816 |
+
},
|
| 817 |
+
"language_model.model.layers.1.linear_attn.in_proj_a": {
|
| 818 |
+
"bits": 6,
|
| 819 |
+
"group_size": 64
|
| 820 |
+
},
|
| 821 |
+
"language_model.model.layers.1.linear_attn.out_proj": {
|
| 822 |
+
"bits": 4,
|
| 823 |
+
"group_size": 64
|
| 824 |
+
},
|
| 825 |
+
"language_model.model.layers.1.mlp.gate_proj": {
|
| 826 |
+
"bits": 3,
|
| 827 |
+
"group_size": 64
|
| 828 |
+
},
|
| 829 |
+
"language_model.model.layers.1.mlp.down_proj": {
|
| 830 |
+
"bits": 3,
|
| 831 |
+
"group_size": 64
|
| 832 |
+
},
|
| 833 |
+
"language_model.model.layers.1.mlp.up_proj": {
|
| 834 |
+
"bits": 3,
|
| 835 |
+
"group_size": 64
|
| 836 |
+
},
|
| 837 |
+
"language_model.model.layers.2.linear_attn.in_proj_qkv": {
|
| 838 |
+
"bits": 3,
|
| 839 |
+
"group_size": 64
|
| 840 |
+
},
|
| 841 |
+
"language_model.model.layers.2.linear_attn.in_proj_z": {
|
| 842 |
+
"bits": 3,
|
| 843 |
+
"group_size": 64
|
| 844 |
+
},
|
| 845 |
+
"language_model.model.layers.2.linear_attn.in_proj_b": {
|
| 846 |
+
"bits": 5,
|
| 847 |
+
"group_size": 64
|
| 848 |
+
},
|
| 849 |
+
"language_model.model.layers.2.linear_attn.in_proj_a": {
|
| 850 |
+
"bits": 4,
|
| 851 |
+
"group_size": 64
|
| 852 |
+
},
|
| 853 |
+
"language_model.model.layers.2.linear_attn.out_proj": {
|
| 854 |
+
"bits": 4,
|
| 855 |
+
"group_size": 64
|
| 856 |
+
},
|
| 857 |
+
"language_model.model.layers.2.mlp.gate_proj": {
|
| 858 |
+
"bits": 3,
|
| 859 |
+
"group_size": 64
|
| 860 |
+
},
|
| 861 |
+
"language_model.model.layers.2.mlp.down_proj": {
|
| 862 |
+
"bits": 3,
|
| 863 |
+
"group_size": 64
|
| 864 |
+
},
|
| 865 |
+
"language_model.model.layers.2.mlp.up_proj": {
|
| 866 |
+
"bits": 3,
|
| 867 |
+
"group_size": 64
|
| 868 |
+
},
|
| 869 |
+
"language_model.model.layers.3.self_attn.q_proj": {
|
| 870 |
+
"bits": 3,
|
| 871 |
+
"group_size": 64
|
| 872 |
+
},
|
| 873 |
+
"language_model.model.layers.3.self_attn.k_proj": {
|
| 874 |
+
"bits": 3,
|
| 875 |
+
"group_size": 64
|
| 876 |
+
},
|
| 877 |
+
"language_model.model.layers.3.self_attn.v_proj": {
|
| 878 |
+
"bits": 5,
|
| 879 |
+
"group_size": 64
|
| 880 |
+
},
|
| 881 |
+
"language_model.model.layers.3.self_attn.o_proj": {
|
| 882 |
+
"bits": 4,
|
| 883 |
+
"group_size": 64
|
| 884 |
+
},
|
| 885 |
+
"language_model.model.layers.3.mlp.gate_proj": {
|
| 886 |
+
"bits": 3,
|
| 887 |
+
"group_size": 64
|
| 888 |
+
},
|
| 889 |
+
"language_model.model.layers.3.mlp.down_proj": {
|
| 890 |
+
"bits": 3,
|
| 891 |
+
"group_size": 64
|
| 892 |
+
},
|
| 893 |
+
"language_model.model.layers.3.mlp.up_proj": {
|
| 894 |
+
"bits": 3,
|
| 895 |
+
"group_size": 64
|
| 896 |
+
},
|
| 897 |
+
"language_model.model.layers.4.linear_attn.in_proj_qkv": {
|
| 898 |
+
"bits": 3,
|
| 899 |
+
"group_size": 64
|
| 900 |
+
},
|
| 901 |
+
"language_model.model.layers.4.linear_attn.in_proj_z": {
|
| 902 |
+
"bits": 3,
|
| 903 |
+
"group_size": 64
|
| 904 |
+
},
|
| 905 |
+
"language_model.model.layers.4.linear_attn.in_proj_b": {
|
| 906 |
+
"bits": 5,
|
| 907 |
+
"group_size": 64
|
| 908 |
+
},
|
| 909 |
+
"language_model.model.layers.4.linear_attn.in_proj_a": {
|
| 910 |
+
"bits": 4,
|
| 911 |
+
"group_size": 64
|
| 912 |
+
},
|
| 913 |
+
"language_model.model.layers.4.linear_attn.out_proj": {
|
| 914 |
+
"bits": 3,
|
| 915 |
+
"group_size": 64
|
| 916 |
+
},
|
| 917 |
+
"language_model.model.layers.4.mlp.gate_proj": {
|
| 918 |
+
"bits": 3,
|
| 919 |
+
"group_size": 64
|
| 920 |
+
},
|
| 921 |
+
"language_model.model.layers.4.mlp.down_proj": {
|
| 922 |
+
"bits": 3,
|
| 923 |
+
"group_size": 64
|
| 924 |
+
},
|
| 925 |
+
"language_model.model.layers.4.mlp.up_proj": {
|
| 926 |
+
"bits": 3,
|
| 927 |
+
"group_size": 64
|
| 928 |
+
},
|
| 929 |
+
"language_model.model.layers.5.linear_attn.in_proj_qkv": {
|
| 930 |
+
"bits": 3,
|
| 931 |
+
"group_size": 64
|
| 932 |
+
},
|
| 933 |
+
"language_model.model.layers.5.linear_attn.in_proj_z": {
|
| 934 |
+
"bits": 3,
|
| 935 |
+
"group_size": 64
|
| 936 |
+
},
|
| 937 |
+
"language_model.model.layers.5.linear_attn.in_proj_b": {
|
| 938 |
+
"bits": 5,
|
| 939 |
+
"group_size": 64
|
| 940 |
+
},
|
| 941 |
+
"language_model.model.layers.5.linear_attn.in_proj_a": {
|
| 942 |
+
"bits": 4,
|
| 943 |
+
"group_size": 64
|
| 944 |
+
},
|
| 945 |
+
"language_model.model.layers.5.linear_attn.out_proj": {
|
| 946 |
+
"bits": 3,
|
| 947 |
+
"group_size": 64
|
| 948 |
+
},
|
| 949 |
+
"language_model.model.layers.5.mlp.gate_proj": {
|
| 950 |
+
"bits": 3,
|
| 951 |
+
"group_size": 64
|
| 952 |
+
},
|
| 953 |
+
"language_model.model.layers.5.mlp.down_proj": {
|
| 954 |
+
"bits": 3,
|
| 955 |
+
"group_size": 64
|
| 956 |
+
},
|
| 957 |
+
"language_model.model.layers.5.mlp.up_proj": {
|
| 958 |
+
"bits": 3,
|
| 959 |
+
"group_size": 64
|
| 960 |
+
},
|
| 961 |
+
"language_model.model.layers.6.linear_attn.in_proj_qkv": {
|
| 962 |
+
"bits": 3,
|
| 963 |
+
"group_size": 64
|
| 964 |
+
},
|
| 965 |
+
"language_model.model.layers.6.linear_attn.in_proj_z": {
|
| 966 |
+
"bits": 3,
|
| 967 |
+
"group_size": 64
|
| 968 |
+
},
|
| 969 |
+
"language_model.model.layers.6.linear_attn.in_proj_b": {
|
| 970 |
+
"bits": 5,
|
| 971 |
+
"group_size": 64
|
| 972 |
+
},
|
| 973 |
+
"language_model.model.layers.6.linear_attn.in_proj_a": {
|
| 974 |
+
"bits": 4,
|
| 975 |
+
"group_size": 64
|
| 976 |
+
},
|
| 977 |
+
"language_model.model.layers.6.linear_attn.out_proj": {
|
| 978 |
+
"bits": 4,
|
| 979 |
+
"group_size": 64
|
| 980 |
+
},
|
| 981 |
+
"language_model.model.layers.6.mlp.gate_proj": {
|
| 982 |
+
"bits": 3,
|
| 983 |
+
"group_size": 64
|
| 984 |
+
},
|
| 985 |
+
"language_model.model.layers.6.mlp.down_proj": {
|
| 986 |
+
"bits": 3,
|
| 987 |
+
"group_size": 64
|
| 988 |
+
},
|
| 989 |
+
"language_model.model.layers.6.mlp.up_proj": {
|
| 990 |
+
"bits": 3,
|
| 991 |
+
"group_size": 64
|
| 992 |
+
},
|
| 993 |
+
"language_model.model.layers.7.self_attn.q_proj": {
|
| 994 |
+
"bits": 3,
|
| 995 |
+
"group_size": 64
|
| 996 |
+
},
|
| 997 |
+
"language_model.model.layers.7.self_attn.k_proj": {
|
| 998 |
+
"bits": 3,
|
| 999 |
+
"group_size": 64
|
| 1000 |
+
},
|
| 1001 |
+
"language_model.model.layers.7.self_attn.v_proj": {
|
| 1002 |
+
"bits": 5,
|
| 1003 |
+
"group_size": 64
|
| 1004 |
+
},
|
| 1005 |
+
"language_model.model.layers.7.self_attn.o_proj": {
|
| 1006 |
+
"bits": 4,
|
| 1007 |
+
"group_size": 64
|
| 1008 |
+
},
|
| 1009 |
+
"language_model.model.layers.7.mlp.gate_proj": {
|
| 1010 |
+
"bits": 2,
|
| 1011 |
+
"group_size": 64
|
| 1012 |
+
},
|
| 1013 |
+
"language_model.model.layers.7.mlp.down_proj": {
|
| 1014 |
+
"bits": 3,
|
| 1015 |
+
"group_size": 64
|
| 1016 |
+
},
|
| 1017 |
+
"language_model.model.layers.7.mlp.up_proj": {
|
| 1018 |
+
"bits": 3,
|
| 1019 |
+
"group_size": 64
|
| 1020 |
+
},
|
| 1021 |
+
"language_model.model.layers.8.linear_attn.in_proj_qkv": {
|
| 1022 |
+
"bits": 3,
|
| 1023 |
+
"group_size": 64
|
| 1024 |
+
},
|
| 1025 |
+
"language_model.model.layers.8.linear_attn.in_proj_z": {
|
| 1026 |
+
"bits": 3,
|
| 1027 |
+
"group_size": 64
|
| 1028 |
+
},
|
| 1029 |
+
"language_model.model.layers.8.linear_attn.in_proj_b": {
|
| 1030 |
+
"bits": 4,
|
| 1031 |
+
"group_size": 64
|
| 1032 |
+
},
|
| 1033 |
+
"language_model.model.layers.8.linear_attn.in_proj_a": {
|
| 1034 |
+
"bits": 4,
|
| 1035 |
+
"group_size": 64
|
| 1036 |
+
},
|
| 1037 |
+
"language_model.model.layers.8.linear_attn.out_proj": {
|
| 1038 |
+
"bits": 3,
|
| 1039 |
+
"group_size": 64
|
| 1040 |
+
},
|
| 1041 |
+
"language_model.model.layers.8.mlp.gate_proj": {
|
| 1042 |
+
"bits": 2,
|
| 1043 |
+
"group_size": 64
|
| 1044 |
+
},
|
| 1045 |
+
"language_model.model.layers.8.mlp.down_proj": {
|
| 1046 |
+
"bits": 3,
|
| 1047 |
+
"group_size": 64
|
| 1048 |
+
},
|
| 1049 |
+
"language_model.model.layers.8.mlp.up_proj": {
|
| 1050 |
+
"bits": 3,
|
| 1051 |
+
"group_size": 64
|
| 1052 |
+
},
|
| 1053 |
+
"language_model.model.layers.9.linear_attn.in_proj_qkv": {
|
| 1054 |
+
"bits": 3,
|
| 1055 |
+
"group_size": 64
|
| 1056 |
+
},
|
| 1057 |
+
"language_model.model.layers.9.linear_attn.in_proj_z": {
|
| 1058 |
+
"bits": 3,
|
| 1059 |
+
"group_size": 64
|
| 1060 |
+
},
|
| 1061 |
+
"language_model.model.layers.9.linear_attn.in_proj_b": {
|
| 1062 |
+
"bits": 5,
|
| 1063 |
+
"group_size": 64
|
| 1064 |
+
},
|
| 1065 |
+
"language_model.model.layers.9.linear_attn.in_proj_a": {
|
| 1066 |
+
"bits": 4,
|
| 1067 |
+
"group_size": 64
|
| 1068 |
+
},
|
| 1069 |
+
"language_model.model.layers.9.linear_attn.out_proj": {
|
| 1070 |
+
"bits": 3,
|
| 1071 |
+
"group_size": 64
|
| 1072 |
+
},
|
| 1073 |
+
"language_model.model.layers.9.mlp.gate_proj": {
|
| 1074 |
+
"bits": 2,
|
| 1075 |
+
"group_size": 64
|
| 1076 |
+
},
|
| 1077 |
+
"language_model.model.layers.9.mlp.down_proj": {
|
| 1078 |
+
"bits": 2,
|
| 1079 |
+
"group_size": 64
|
| 1080 |
+
},
|
| 1081 |
+
"language_model.model.layers.9.mlp.up_proj": {
|
| 1082 |
+
"bits": 2,
|
| 1083 |
+
"group_size": 64
|
| 1084 |
+
},
|
| 1085 |
+
"language_model.model.layers.10.linear_attn.in_proj_qkv": {
|
| 1086 |
+
"bits": 3,
|
| 1087 |
+
"group_size": 64
|
| 1088 |
+
},
|
| 1089 |
+
"language_model.model.layers.10.linear_attn.in_proj_z": {
|
| 1090 |
+
"bits": 2,
|
| 1091 |
+
"group_size": 64
|
| 1092 |
+
},
|
| 1093 |
+
"language_model.model.layers.10.linear_attn.in_proj_b": {
|
| 1094 |
+
"bits": 4,
|
| 1095 |
+
"group_size": 64
|
| 1096 |
+
},
|
| 1097 |
+
"language_model.model.layers.10.linear_attn.in_proj_a": {
|
| 1098 |
+
"bits": 4,
|
| 1099 |
+
"group_size": 64
|
| 1100 |
+
},
|
| 1101 |
+
"language_model.model.layers.10.linear_attn.out_proj": {
|
| 1102 |
+
"bits": 3,
|
| 1103 |
+
"group_size": 64
|
| 1104 |
+
},
|
| 1105 |
+
"language_model.model.layers.10.mlp.gate_proj": {
|
| 1106 |
+
"bits": 2,
|
| 1107 |
+
"group_size": 64
|
| 1108 |
+
},
|
| 1109 |
+
"language_model.model.layers.10.mlp.down_proj": {
|
| 1110 |
+
"bits": 3,
|
| 1111 |
+
"group_size": 64
|
| 1112 |
+
},
|
| 1113 |
+
"language_model.model.layers.10.mlp.up_proj": {
|
| 1114 |
+
"bits": 3,
|
| 1115 |
+
"group_size": 64
|
| 1116 |
+
},
|
| 1117 |
+
"language_model.model.layers.11.self_attn.q_proj": {
|
| 1118 |
+
"bits": 3,
|
| 1119 |
+
"group_size": 64
|
| 1120 |
+
},
|
| 1121 |
+
"language_model.model.layers.11.self_attn.k_proj": {
|
| 1122 |
+
"bits": 3,
|
| 1123 |
+
"group_size": 64
|
| 1124 |
+
},
|
| 1125 |
+
"language_model.model.layers.11.self_attn.v_proj": {
|
| 1126 |
+
"bits": 4,
|
| 1127 |
+
"group_size": 64
|
| 1128 |
+
},
|
| 1129 |
+
"language_model.model.layers.11.self_attn.o_proj": {
|
| 1130 |
+
"bits": 4,
|
| 1131 |
+
"group_size": 64
|
| 1132 |
+
},
|
| 1133 |
+
"language_model.model.layers.11.mlp.gate_proj": {
|
| 1134 |
+
"bits": 2,
|
| 1135 |
+
"group_size": 64
|
| 1136 |
+
},
|
| 1137 |
+
"language_model.model.layers.11.mlp.down_proj": {
|
| 1138 |
+
"bits": 2,
|
| 1139 |
+
"group_size": 64
|
| 1140 |
+
},
|
| 1141 |
+
"language_model.model.layers.11.mlp.up_proj": {
|
| 1142 |
+
"bits": 3,
|
| 1143 |
+
"group_size": 64
|
| 1144 |
+
},
|
| 1145 |
+
"language_model.model.layers.12.linear_attn.in_proj_qkv": {
|
| 1146 |
+
"bits": 3,
|
| 1147 |
+
"group_size": 64
|
| 1148 |
+
},
|
| 1149 |
+
"language_model.model.layers.12.linear_attn.in_proj_z": {
|
| 1150 |
+
"bits": 3,
|
| 1151 |
+
"group_size": 64
|
| 1152 |
+
},
|
| 1153 |
+
"language_model.model.layers.12.linear_attn.in_proj_b": {
|
| 1154 |
+
"bits": 5,
|
| 1155 |
+
"group_size": 64
|
| 1156 |
+
},
|
| 1157 |
+
"language_model.model.layers.12.linear_attn.in_proj_a": {
|
| 1158 |
+
"bits": 4,
|
| 1159 |
+
"group_size": 64
|
| 1160 |
+
},
|
| 1161 |
+
"language_model.model.layers.12.linear_attn.out_proj": {
|
| 1162 |
+
"bits": 3,
|
| 1163 |
+
"group_size": 64
|
| 1164 |
+
},
|
| 1165 |
+
"language_model.model.layers.12.mlp.gate_proj": {
|
| 1166 |
+
"bits": 2,
|
| 1167 |
+
"group_size": 64
|
| 1168 |
+
},
|
| 1169 |
+
"language_model.model.layers.12.mlp.down_proj": {
|
| 1170 |
+
"bits": 2,
|
| 1171 |
+
"group_size": 64
|
| 1172 |
+
},
|
| 1173 |
+
"language_model.model.layers.12.mlp.up_proj": {
|
| 1174 |
+
"bits": 2,
|
| 1175 |
+
"group_size": 64
|
| 1176 |
+
},
|
| 1177 |
+
"language_model.model.layers.13.linear_attn.in_proj_qkv": {
|
| 1178 |
+
"bits": 3,
|
| 1179 |
+
"group_size": 64
|
| 1180 |
+
},
|
| 1181 |
+
"language_model.model.layers.13.linear_attn.in_proj_z": {
|
| 1182 |
+
"bits": 2,
|
| 1183 |
+
"group_size": 64
|
| 1184 |
+
},
|
| 1185 |
+
"language_model.model.layers.13.linear_attn.in_proj_b": {
|
| 1186 |
+
"bits": 5,
|
| 1187 |
+
"group_size": 64
|
| 1188 |
+
},
|
| 1189 |
+
"language_model.model.layers.13.linear_attn.in_proj_a": {
|
| 1190 |
+
"bits": 4,
|
| 1191 |
+
"group_size": 64
|
| 1192 |
+
},
|
| 1193 |
+
"language_model.model.layers.13.linear_attn.out_proj": {
|
| 1194 |
+
"bits": 3,
|
| 1195 |
+
"group_size": 64
|
| 1196 |
+
},
|
| 1197 |
+
"language_model.model.layers.13.mlp.gate_proj": {
|
| 1198 |
+
"bits": 2,
|
| 1199 |
+
"group_size": 64
|
| 1200 |
+
},
|
| 1201 |
+
"language_model.model.layers.13.mlp.down_proj": {
|
| 1202 |
+
"bits": 2,
|
| 1203 |
+
"group_size": 64
|
| 1204 |
+
},
|
| 1205 |
+
"language_model.model.layers.13.mlp.up_proj": {
|
| 1206 |
+
"bits": 2,
|
| 1207 |
+
"group_size": 64
|
| 1208 |
+
},
|
| 1209 |
+
"language_model.model.layers.14.linear_attn.in_proj_qkv": {
|
| 1210 |
+
"bits": 3,
|
| 1211 |
+
"group_size": 64
|
| 1212 |
+
},
|
| 1213 |
+
"language_model.model.layers.14.linear_attn.in_proj_z": {
|
| 1214 |
+
"bits": 2,
|
| 1215 |
+
"group_size": 64
|
| 1216 |
+
},
|
| 1217 |
+
"language_model.model.layers.14.linear_attn.in_proj_b": {
|
| 1218 |
+
"bits": 4,
|
| 1219 |
+
"group_size": 64
|
| 1220 |
+
},
|
| 1221 |
+
"language_model.model.layers.14.linear_attn.in_proj_a": {
|
| 1222 |
+
"bits": 4,
|
| 1223 |
+
"group_size": 64
|
| 1224 |
+
},
|
| 1225 |
+
"language_model.model.layers.14.linear_attn.out_proj": {
|
| 1226 |
+
"bits": 3,
|
| 1227 |
+
"group_size": 64
|
| 1228 |
+
},
|
| 1229 |
+
"language_model.model.layers.14.mlp.gate_proj": {
|
| 1230 |
+
"bits": 2,
|
| 1231 |
+
"group_size": 64
|
| 1232 |
+
},
|
| 1233 |
+
"language_model.model.layers.14.mlp.down_proj": {
|
| 1234 |
+
"bits": 3,
|
| 1235 |
+
"group_size": 64
|
| 1236 |
+
},
|
| 1237 |
+
"language_model.model.layers.14.mlp.up_proj": {
|
| 1238 |
+
"bits": 3,
|
| 1239 |
+
"group_size": 64
|
| 1240 |
+
},
|
| 1241 |
+
"language_model.model.layers.15.self_attn.q_proj": {
|
| 1242 |
+
"bits": 2,
|
| 1243 |
+
"group_size": 64
|
| 1244 |
+
},
|
| 1245 |
+
"language_model.model.layers.15.self_attn.k_proj": {
|
| 1246 |
+
"bits": 3,
|
| 1247 |
+
"group_size": 64
|
| 1248 |
+
},
|
| 1249 |
+
"language_model.model.layers.15.self_attn.v_proj": {
|
| 1250 |
+
"bits": 4,
|
| 1251 |
+
"group_size": 64
|
| 1252 |
+
},
|
| 1253 |
+
"language_model.model.layers.15.self_attn.o_proj": {
|
| 1254 |
+
"bits": 4,
|
| 1255 |
+
"group_size": 64
|
| 1256 |
+
},
|
| 1257 |
+
"language_model.model.layers.15.mlp.gate_proj": {
|
| 1258 |
+
"bits": 2,
|
| 1259 |
+
"group_size": 64
|
| 1260 |
+
},
|
| 1261 |
+
"language_model.model.layers.15.mlp.down_proj": {
|
| 1262 |
+
"bits": 3,
|
| 1263 |
+
"group_size": 64
|
| 1264 |
+
},
|
| 1265 |
+
"language_model.model.layers.15.mlp.up_proj": {
|
| 1266 |
+
"bits": 3,
|
| 1267 |
+
"group_size": 64
|
| 1268 |
+
},
|
| 1269 |
+
"language_model.model.layers.16.linear_attn.in_proj_qkv": {
|
| 1270 |
+
"bits": 3,
|
| 1271 |
+
"group_size": 64
|
| 1272 |
+
},
|
| 1273 |
+
"language_model.model.layers.16.linear_attn.in_proj_z": {
|
| 1274 |
+
"bits": 2,
|
| 1275 |
+
"group_size": 64
|
| 1276 |
+
},
|
| 1277 |
+
"language_model.model.layers.16.linear_attn.in_proj_b": {
|
| 1278 |
+
"bits": 5,
|
| 1279 |
+
"group_size": 64
|
| 1280 |
+
},
|
| 1281 |
+
"language_model.model.layers.16.linear_attn.in_proj_a": {
|
| 1282 |
+
"bits": 4,
|
| 1283 |
+
"group_size": 64
|
| 1284 |
+
},
|
| 1285 |
+
"language_model.model.layers.16.linear_attn.out_proj": {
|
| 1286 |
+
"bits": 3,
|
| 1287 |
+
"group_size": 64
|
| 1288 |
+
},
|
| 1289 |
+
"language_model.model.layers.16.mlp.gate_proj": {
|
| 1290 |
+
"bits": 2,
|
| 1291 |
+
"group_size": 64
|
| 1292 |
+
},
|
| 1293 |
+
"language_model.model.layers.16.mlp.down_proj": {
|
| 1294 |
+
"bits": 3,
|
| 1295 |
+
"group_size": 64
|
| 1296 |
+
},
|
| 1297 |
+
"language_model.model.layers.16.mlp.up_proj": {
|
| 1298 |
+
"bits": 2,
|
| 1299 |
+
"group_size": 64
|
| 1300 |
+
},
|
| 1301 |
+
"language_model.model.layers.17.linear_attn.in_proj_qkv": {
|
| 1302 |
+
"bits": 3,
|
| 1303 |
+
"group_size": 64
|
| 1304 |
+
},
|
| 1305 |
+
"language_model.model.layers.17.linear_attn.in_proj_z": {
|
| 1306 |
+
"bits": 2,
|
| 1307 |
+
"group_size": 64
|
| 1308 |
+
},
|
| 1309 |
+
"language_model.model.layers.17.linear_attn.in_proj_b": {
|
| 1310 |
+
"bits": 5,
|
| 1311 |
+
"group_size": 64
|
| 1312 |
+
},
|
| 1313 |
+
"language_model.model.layers.17.linear_attn.in_proj_a": {
|
| 1314 |
+
"bits": 4,
|
| 1315 |
+
"group_size": 64
|
| 1316 |
+
},
|
| 1317 |
+
"language_model.model.layers.17.linear_attn.out_proj": {
|
| 1318 |
+
"bits": 3,
|
| 1319 |
+
"group_size": 64
|
| 1320 |
+
},
|
| 1321 |
+
"language_model.model.layers.17.mlp.gate_proj": {
|
| 1322 |
+
"bits": 2,
|
| 1323 |
+
"group_size": 64
|
| 1324 |
+
},
|
| 1325 |
+
"language_model.model.layers.17.mlp.down_proj": {
|
| 1326 |
+
"bits": 2,
|
| 1327 |
+
"group_size": 64
|
| 1328 |
+
},
|
| 1329 |
+
"language_model.model.layers.17.mlp.up_proj": {
|
| 1330 |
+
"bits": 2,
|
| 1331 |
+
"group_size": 64
|
| 1332 |
+
},
|
| 1333 |
+
"language_model.model.layers.18.linear_attn.in_proj_qkv": {
|
| 1334 |
+
"bits": 3,
|
| 1335 |
+
"group_size": 64
|
| 1336 |
+
},
|
| 1337 |
+
"language_model.model.layers.18.linear_attn.in_proj_z": {
|
| 1338 |
+
"bits": 2,
|
| 1339 |
+
"group_size": 64
|
| 1340 |
+
},
|
| 1341 |
+
"language_model.model.layers.18.linear_attn.in_proj_b": {
|
| 1342 |
+
"bits": 5,
|
| 1343 |
+
"group_size": 64
|
| 1344 |
+
},
|
| 1345 |
+
"language_model.model.layers.18.linear_attn.in_proj_a": {
|
| 1346 |
+
"bits": 3,
|
| 1347 |
+
"group_size": 64
|
| 1348 |
+
},
|
| 1349 |
+
"language_model.model.layers.18.linear_attn.out_proj": {
|
| 1350 |
+
"bits": 3,
|
| 1351 |
+
"group_size": 64
|
| 1352 |
+
},
|
| 1353 |
+
"language_model.model.layers.18.mlp.gate_proj": {
|
| 1354 |
+
"bits": 2,
|
| 1355 |
+
"group_size": 64
|
| 1356 |
+
},
|
| 1357 |
+
"language_model.model.layers.18.mlp.down_proj": {
|
| 1358 |
+
"bits": 2,
|
| 1359 |
+
"group_size": 64
|
| 1360 |
+
},
|
| 1361 |
+
"language_model.model.layers.18.mlp.up_proj": {
|
| 1362 |
+
"bits": 2,
|
| 1363 |
+
"group_size": 64
|
| 1364 |
+
},
|
| 1365 |
+
"language_model.model.layers.19.self_attn.q_proj": {
|
| 1366 |
+
"bits": 2,
|
| 1367 |
+
"group_size": 64
|
| 1368 |
+
},
|
| 1369 |
+
"language_model.model.layers.19.self_attn.k_proj": {
|
| 1370 |
+
"bits": 3,
|
| 1371 |
+
"group_size": 64
|
| 1372 |
+
},
|
| 1373 |
+
"language_model.model.layers.19.self_attn.v_proj": {
|
| 1374 |
+
"bits": 4,
|
| 1375 |
+
"group_size": 64
|
| 1376 |
+
},
|
| 1377 |
+
"language_model.model.layers.19.self_attn.o_proj": {
|
| 1378 |
+
"bits": 3,
|
| 1379 |
+
"group_size": 64
|
| 1380 |
+
},
|
| 1381 |
+
"language_model.model.layers.19.mlp.gate_proj": {
|
| 1382 |
+
"bits": 2,
|
| 1383 |
+
"group_size": 64
|
| 1384 |
+
},
|
| 1385 |
+
"language_model.model.layers.19.mlp.down_proj": {
|
| 1386 |
+
"bits": 2,
|
| 1387 |
+
"group_size": 64
|
| 1388 |
+
},
|
| 1389 |
+
"language_model.model.layers.19.mlp.up_proj": {
|
| 1390 |
+
"bits": 2,
|
| 1391 |
+
"group_size": 64
|
| 1392 |
+
},
|
| 1393 |
+
"language_model.model.layers.20.linear_attn.in_proj_qkv": {
|
| 1394 |
+
"bits": 3,
|
| 1395 |
+
"group_size": 64
|
| 1396 |
+
},
|
| 1397 |
+
"language_model.model.layers.20.linear_attn.in_proj_z": {
|
| 1398 |
+
"bits": 2,
|
| 1399 |
+
"group_size": 64
|
| 1400 |
+
},
|
| 1401 |
+
"language_model.model.layers.20.linear_attn.in_proj_b": {
|
| 1402 |
+
"bits": 4,
|
| 1403 |
+
"group_size": 64
|
| 1404 |
+
},
|
| 1405 |
+
"language_model.model.layers.20.linear_attn.in_proj_a": {
|
| 1406 |
+
"bits": 4,
|
| 1407 |
+
"group_size": 64
|
| 1408 |
+
},
|
| 1409 |
+
"language_model.model.layers.20.linear_attn.out_proj": {
|
| 1410 |
+
"bits": 3,
|
| 1411 |
+
"group_size": 64
|
| 1412 |
+
},
|
| 1413 |
+
"language_model.model.layers.20.mlp.gate_proj": {
|
| 1414 |
+
"bits": 2,
|
| 1415 |
+
"group_size": 64
|
| 1416 |
+
},
|
| 1417 |
+
"language_model.model.layers.20.mlp.down_proj": {
|
| 1418 |
+
"bits": 2,
|
| 1419 |
+
"group_size": 64
|
| 1420 |
+
},
|
| 1421 |
+
"language_model.model.layers.20.mlp.up_proj": {
|
| 1422 |
+
"bits": 2,
|
| 1423 |
+
"group_size": 64
|
| 1424 |
+
},
|
| 1425 |
+
"language_model.model.layers.21.linear_attn.in_proj_qkv": {
|
| 1426 |
+
"bits": 3,
|
| 1427 |
+
"group_size": 64
|
| 1428 |
+
},
|
| 1429 |
+
"language_model.model.layers.21.linear_attn.in_proj_z": {
|
| 1430 |
+
"bits": 2,
|
| 1431 |
+
"group_size": 64
|
| 1432 |
+
},
|
| 1433 |
+
"language_model.model.layers.21.linear_attn.in_proj_b": {
|
| 1434 |
+
"bits": 5,
|
| 1435 |
+
"group_size": 64
|
| 1436 |
+
},
|
| 1437 |
+
"language_model.model.layers.21.linear_attn.in_proj_a": {
|
| 1438 |
+
"bits": 3,
|
| 1439 |
+
"group_size": 64
|
| 1440 |
+
},
|
| 1441 |
+
"language_model.model.layers.21.linear_attn.out_proj": {
|
| 1442 |
+
"bits": 3,
|
| 1443 |
+
"group_size": 64
|
| 1444 |
+
},
|
| 1445 |
+
"language_model.model.layers.21.mlp.gate_proj": {
|
| 1446 |
+
"bits": 2,
|
| 1447 |
+
"group_size": 64
|
| 1448 |
+
},
|
| 1449 |
+
"language_model.model.layers.21.mlp.down_proj": {
|
| 1450 |
+
"bits": 2,
|
| 1451 |
+
"group_size": 64
|
| 1452 |
+
},
|
| 1453 |
+
"language_model.model.layers.21.mlp.up_proj": {
|
| 1454 |
+
"bits": 2,
|
| 1455 |
+
"group_size": 64
|
| 1456 |
+
},
|
| 1457 |
+
"language_model.model.layers.22.linear_attn.in_proj_qkv": {
|
| 1458 |
+
"bits": 3,
|
| 1459 |
+
"group_size": 64
|
| 1460 |
+
},
|
| 1461 |
+
"language_model.model.layers.22.linear_attn.in_proj_z": {
|
| 1462 |
+
"bits": 2,
|
| 1463 |
+
"group_size": 64
|
| 1464 |
+
},
|
| 1465 |
+
"language_model.model.layers.22.linear_attn.in_proj_b": {
|
| 1466 |
+
"bits": 5,
|
| 1467 |
+
"group_size": 64
|
| 1468 |
+
},
|
| 1469 |
+
"language_model.model.layers.22.linear_attn.in_proj_a": {
|
| 1470 |
+
"bits": 4,
|
| 1471 |
+
"group_size": 64
|
| 1472 |
+
},
|
| 1473 |
+
"language_model.model.layers.22.linear_attn.out_proj": {
|
| 1474 |
+
"bits": 3,
|
| 1475 |
+
"group_size": 64
|
| 1476 |
+
},
|
| 1477 |
+
"language_model.model.layers.22.mlp.gate_proj": {
|
| 1478 |
+
"bits": 2,
|
| 1479 |
+
"group_size": 64
|
| 1480 |
+
},
|
| 1481 |
+
"language_model.model.layers.22.mlp.down_proj": {
|
| 1482 |
+
"bits": 2,
|
| 1483 |
+
"group_size": 64
|
| 1484 |
+
},
|
| 1485 |
+
"language_model.model.layers.22.mlp.up_proj": {
|
| 1486 |
+
"bits": 2,
|
| 1487 |
+
"group_size": 64
|
| 1488 |
+
},
|
| 1489 |
+
"language_model.model.layers.23.self_attn.q_proj": {
|
| 1490 |
+
"bits": 6,
|
| 1491 |
+
"group_size": 64
|
| 1492 |
+
},
|
| 1493 |
+
"language_model.model.layers.23.self_attn.k_proj": {
|
| 1494 |
+
"bits": 6,
|
| 1495 |
+
"group_size": 64
|
| 1496 |
+
},
|
| 1497 |
+
"language_model.model.layers.23.self_attn.v_proj": {
|
| 1498 |
+
"bits": 6,
|
| 1499 |
+
"group_size": 64
|
| 1500 |
+
},
|
| 1501 |
+
"language_model.model.layers.23.self_attn.o_proj": {
|
| 1502 |
+
"bits": 6,
|
| 1503 |
+
"group_size": 64
|
| 1504 |
+
},
|
| 1505 |
+
"language_model.model.layers.23.mlp.gate_proj": {
|
| 1506 |
+
"bits": 6,
|
| 1507 |
+
"group_size": 64
|
| 1508 |
+
},
|
| 1509 |
+
"language_model.model.layers.23.mlp.down_proj": {
|
| 1510 |
+
"bits": 6,
|
| 1511 |
+
"group_size": 64
|
| 1512 |
+
},
|
| 1513 |
+
"language_model.model.layers.23.mlp.up_proj": {
|
| 1514 |
+
"bits": 6,
|
| 1515 |
+
"group_size": 64
|
| 1516 |
+
}
|
| 1517 |
+
},
|
| 1518 |
+
"text_config": {
|
| 1519 |
+
"attention_bias": false,
|
| 1520 |
+
"attention_dropout": 0.0,
|
| 1521 |
+
"attn_output_gate": true,
|
| 1522 |
+
"dtype": "bfloat16",
|
| 1523 |
+
"eos_token_id": 248046,
|
| 1524 |
+
"full_attention_interval": 4,
|
| 1525 |
+
"head_dim": 256,
|
| 1526 |
+
"hidden_act": "silu",
|
| 1527 |
+
"hidden_size": 2048,
|
| 1528 |
+
"initializer_range": 0.02,
|
| 1529 |
+
"intermediate_size": 6144,
|
| 1530 |
+
"layer_types": [
|
| 1531 |
+
"linear_attention",
|
| 1532 |
+
"linear_attention",
|
| 1533 |
+
"linear_attention",
|
| 1534 |
+
"full_attention",
|
| 1535 |
+
"linear_attention",
|
| 1536 |
+
"linear_attention",
|
| 1537 |
+
"linear_attention",
|
| 1538 |
+
"full_attention",
|
| 1539 |
+
"linear_attention",
|
| 1540 |
+
"linear_attention",
|
| 1541 |
+
"linear_attention",
|
| 1542 |
+
"full_attention",
|
| 1543 |
+
"linear_attention",
|
| 1544 |
+
"linear_attention",
|
| 1545 |
+
"linear_attention",
|
| 1546 |
+
"full_attention",
|
| 1547 |
+
"linear_attention",
|
| 1548 |
+
"linear_attention",
|
| 1549 |
+
"linear_attention",
|
| 1550 |
+
"full_attention",
|
| 1551 |
+
"linear_attention",
|
| 1552 |
+
"linear_attention",
|
| 1553 |
+
"linear_attention",
|
| 1554 |
+
"full_attention"
|
| 1555 |
+
],
|
| 1556 |
+
"linear_conv_kernel_dim": 4,
|
| 1557 |
+
"linear_key_head_dim": 128,
|
| 1558 |
+
"linear_num_key_heads": 16,
|
| 1559 |
+
"linear_num_value_heads": 16,
|
| 1560 |
+
"linear_value_head_dim": 128,
|
| 1561 |
+
"max_position_embeddings": 262144,
|
| 1562 |
+
"mlp_only_layers": [],
|
| 1563 |
+
"model_type": "qwen3_5_text",
|
| 1564 |
+
"mtp_num_hidden_layers": 1,
|
| 1565 |
+
"mtp_use_dedicated_embeddings": false,
|
| 1566 |
+
"num_attention_heads": 8,
|
| 1567 |
+
"num_hidden_layers": 24,
|
| 1568 |
+
"num_key_value_heads": 2,
|
| 1569 |
+
"rms_norm_eps": 1e-06,
|
| 1570 |
+
"tie_word_embeddings": true,
|
| 1571 |
+
"use_cache": true,
|
| 1572 |
+
"vocab_size": 248320,
|
| 1573 |
+
"mamba_ssm_dtype": "float32",
|
| 1574 |
+
"rope_parameters": {
|
| 1575 |
+
"mrope_interleaved": true,
|
| 1576 |
+
"mrope_section": [
|
| 1577 |
+
11,
|
| 1578 |
+
11,
|
| 1579 |
+
10
|
| 1580 |
+
],
|
| 1581 |
+
"rope_theta": 10000000,
|
| 1582 |
+
"partial_rotary_factor": 0.25,
|
| 1583 |
+
"type": "default"
|
| 1584 |
+
},
|
| 1585 |
+
"pad_token_id": 248044
|
| 1586 |
+
},
|
| 1587 |
+
"tie_word_embeddings": true,
|
| 1588 |
+
"transformers_version": "4.57.0.dev0",
|
| 1589 |
+
"video_token_id": 248057,
|
| 1590 |
+
"vision_end_token_id": 248054,
|
| 1591 |
+
"vision_start_token_id": 248053,
|
| 1592 |
+
"mtp_file": "optiq/mtp.safetensors",
|
| 1593 |
+
"mtp_tensor_count": 29,
|
| 1594 |
+
"mtp_policy": "optiq-int4-prequantized-gs64",
|
| 1595 |
+
"mtplx_mtp_quantization": {
|
| 1596 |
+
"bits": 4,
|
| 1597 |
+
"group_size": 64,
|
| 1598 |
+
"mode": "affine",
|
| 1599 |
+
"policy": "cyankiwi",
|
| 1600 |
+
"prequantized": true
|
| 1601 |
+
},
|
| 1602 |
+
"mlx_lm_extra_tensors": {
|
| 1603 |
+
"mtp_file": "optiq/mtp.safetensors"
|
| 1604 |
+
},
|
| 1605 |
+
"mtp_num_hidden_layers": 1,
|
| 1606 |
+
"vision_config": {
|
| 1607 |
+
"deepstack_visual_indexes": [],
|
| 1608 |
+
"depth": 24,
|
| 1609 |
+
"hidden_act": "gelu_pytorch_tanh",
|
| 1610 |
+
"hidden_size": 1024,
|
| 1611 |
+
"in_channels": 3,
|
| 1612 |
+
"initializer_range": 0.02,
|
| 1613 |
+
"intermediate_size": 4096,
|
| 1614 |
+
"model_type": "qwen3_5",
|
| 1615 |
+
"num_heads": 16,
|
| 1616 |
+
"num_position_embeddings": 2304,
|
| 1617 |
+
"out_hidden_size": 2048,
|
| 1618 |
+
"patch_size": 16,
|
| 1619 |
+
"spatial_merge_size": 2,
|
| 1620 |
+
"temporal_patch_size": 2
|
| 1621 |
+
},
|
| 1622 |
+
"optiq_vision": {
|
| 1623 |
+
"sidecar": "optiq/optiq_vision.safetensors",
|
| 1624 |
+
"dtype": "bfloat16",
|
| 1625 |
+
"n_tensors": 297,
|
| 1626 |
+
"base_model": "empero-ai/Qwen3.8-2B-Distill"
|
| 1627 |
+
}
|
| 1628 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"eos_token_id": [
|
| 4 |
+
248046,
|
| 5 |
+
248044
|
| 6 |
+
],
|
| 7 |
+
"pad_token_id": 248044,
|
| 8 |
+
"transformers_version": "5.15.0",
|
| 9 |
+
"use_cache": true
|
| 10 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a5a2b5113340ed457ef6be22f9161f9ac862f9447dad9494fdb7900501ea67f3
|
| 3 |
+
size 997996964
|
model.safetensors.index.json
ADDED
|
@@ -0,0 +1,999 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 997908096,
|
| 4 |
+
"total_parameters": 1881824512
|
| 5 |
+
},
|
| 6 |
+
"weight_map": {
|
| 7 |
+
"language_model.model.embed_tokens.biases": "model.safetensors",
|
| 8 |
+
"language_model.model.embed_tokens.scales": "model.safetensors",
|
| 9 |
+
"language_model.model.embed_tokens.weight": "model.safetensors",
|
| 10 |
+
"language_model.model.layers.0.input_layernorm.weight": "model.safetensors",
|
| 11 |
+
"language_model.model.layers.0.linear_attn.A_log": "model.safetensors",
|
| 12 |
+
"language_model.model.layers.0.linear_attn.conv1d.weight": "model.safetensors",
|
| 13 |
+
"language_model.model.layers.0.linear_attn.dt_bias": "model.safetensors",
|
| 14 |
+
"language_model.model.layers.0.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 15 |
+
"language_model.model.layers.0.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 16 |
+
"language_model.model.layers.0.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 17 |
+
"language_model.model.layers.0.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 18 |
+
"language_model.model.layers.0.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 19 |
+
"language_model.model.layers.0.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 20 |
+
"language_model.model.layers.0.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 21 |
+
"language_model.model.layers.0.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 22 |
+
"language_model.model.layers.0.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 23 |
+
"language_model.model.layers.0.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 24 |
+
"language_model.model.layers.0.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 25 |
+
"language_model.model.layers.0.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 26 |
+
"language_model.model.layers.0.linear_attn.norm.weight": "model.safetensors",
|
| 27 |
+
"language_model.model.layers.0.linear_attn.out_proj.biases": "model.safetensors",
|
| 28 |
+
"language_model.model.layers.0.linear_attn.out_proj.scales": "model.safetensors",
|
| 29 |
+
"language_model.model.layers.0.linear_attn.out_proj.weight": "model.safetensors",
|
| 30 |
+
"language_model.model.layers.0.mlp.down_proj.biases": "model.safetensors",
|
| 31 |
+
"language_model.model.layers.0.mlp.down_proj.scales": "model.safetensors",
|
| 32 |
+
"language_model.model.layers.0.mlp.down_proj.weight": "model.safetensors",
|
| 33 |
+
"language_model.model.layers.0.mlp.gate_proj.biases": "model.safetensors",
|
| 34 |
+
"language_model.model.layers.0.mlp.gate_proj.scales": "model.safetensors",
|
| 35 |
+
"language_model.model.layers.0.mlp.gate_proj.weight": "model.safetensors",
|
| 36 |
+
"language_model.model.layers.0.mlp.up_proj.biases": "model.safetensors",
|
| 37 |
+
"language_model.model.layers.0.mlp.up_proj.scales": "model.safetensors",
|
| 38 |
+
"language_model.model.layers.0.mlp.up_proj.weight": "model.safetensors",
|
| 39 |
+
"language_model.model.layers.0.post_attention_layernorm.weight": "model.safetensors",
|
| 40 |
+
"language_model.model.layers.1.input_layernorm.weight": "model.safetensors",
|
| 41 |
+
"language_model.model.layers.1.linear_attn.A_log": "model.safetensors",
|
| 42 |
+
"language_model.model.layers.1.linear_attn.conv1d.weight": "model.safetensors",
|
| 43 |
+
"language_model.model.layers.1.linear_attn.dt_bias": "model.safetensors",
|
| 44 |
+
"language_model.model.layers.1.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 45 |
+
"language_model.model.layers.1.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 46 |
+
"language_model.model.layers.1.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 47 |
+
"language_model.model.layers.1.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 48 |
+
"language_model.model.layers.1.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 49 |
+
"language_model.model.layers.1.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 50 |
+
"language_model.model.layers.1.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 51 |
+
"language_model.model.layers.1.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 52 |
+
"language_model.model.layers.1.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 53 |
+
"language_model.model.layers.1.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 54 |
+
"language_model.model.layers.1.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 55 |
+
"language_model.model.layers.1.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 56 |
+
"language_model.model.layers.1.linear_attn.norm.weight": "model.safetensors",
|
| 57 |
+
"language_model.model.layers.1.linear_attn.out_proj.biases": "model.safetensors",
|
| 58 |
+
"language_model.model.layers.1.linear_attn.out_proj.scales": "model.safetensors",
|
| 59 |
+
"language_model.model.layers.1.linear_attn.out_proj.weight": "model.safetensors",
|
| 60 |
+
"language_model.model.layers.1.mlp.down_proj.biases": "model.safetensors",
|
| 61 |
+
"language_model.model.layers.1.mlp.down_proj.scales": "model.safetensors",
|
| 62 |
+
"language_model.model.layers.1.mlp.down_proj.weight": "model.safetensors",
|
| 63 |
+
"language_model.model.layers.1.mlp.gate_proj.biases": "model.safetensors",
|
| 64 |
+
"language_model.model.layers.1.mlp.gate_proj.scales": "model.safetensors",
|
| 65 |
+
"language_model.model.layers.1.mlp.gate_proj.weight": "model.safetensors",
|
| 66 |
+
"language_model.model.layers.1.mlp.up_proj.biases": "model.safetensors",
|
| 67 |
+
"language_model.model.layers.1.mlp.up_proj.scales": "model.safetensors",
|
| 68 |
+
"language_model.model.layers.1.mlp.up_proj.weight": "model.safetensors",
|
| 69 |
+
"language_model.model.layers.1.post_attention_layernorm.weight": "model.safetensors",
|
| 70 |
+
"language_model.model.layers.10.input_layernorm.weight": "model.safetensors",
|
| 71 |
+
"language_model.model.layers.10.linear_attn.A_log": "model.safetensors",
|
| 72 |
+
"language_model.model.layers.10.linear_attn.conv1d.weight": "model.safetensors",
|
| 73 |
+
"language_model.model.layers.10.linear_attn.dt_bias": "model.safetensors",
|
| 74 |
+
"language_model.model.layers.10.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 75 |
+
"language_model.model.layers.10.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 76 |
+
"language_model.model.layers.10.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 77 |
+
"language_model.model.layers.10.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 78 |
+
"language_model.model.layers.10.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 79 |
+
"language_model.model.layers.10.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 80 |
+
"language_model.model.layers.10.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 81 |
+
"language_model.model.layers.10.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 82 |
+
"language_model.model.layers.10.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 83 |
+
"language_model.model.layers.10.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 84 |
+
"language_model.model.layers.10.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 85 |
+
"language_model.model.layers.10.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 86 |
+
"language_model.model.layers.10.linear_attn.norm.weight": "model.safetensors",
|
| 87 |
+
"language_model.model.layers.10.linear_attn.out_proj.biases": "model.safetensors",
|
| 88 |
+
"language_model.model.layers.10.linear_attn.out_proj.scales": "model.safetensors",
|
| 89 |
+
"language_model.model.layers.10.linear_attn.out_proj.weight": "model.safetensors",
|
| 90 |
+
"language_model.model.layers.10.mlp.down_proj.biases": "model.safetensors",
|
| 91 |
+
"language_model.model.layers.10.mlp.down_proj.scales": "model.safetensors",
|
| 92 |
+
"language_model.model.layers.10.mlp.down_proj.weight": "model.safetensors",
|
| 93 |
+
"language_model.model.layers.10.mlp.gate_proj.biases": "model.safetensors",
|
| 94 |
+
"language_model.model.layers.10.mlp.gate_proj.scales": "model.safetensors",
|
| 95 |
+
"language_model.model.layers.10.mlp.gate_proj.weight": "model.safetensors",
|
| 96 |
+
"language_model.model.layers.10.mlp.up_proj.biases": "model.safetensors",
|
| 97 |
+
"language_model.model.layers.10.mlp.up_proj.scales": "model.safetensors",
|
| 98 |
+
"language_model.model.layers.10.mlp.up_proj.weight": "model.safetensors",
|
| 99 |
+
"language_model.model.layers.10.post_attention_layernorm.weight": "model.safetensors",
|
| 100 |
+
"language_model.model.layers.11.input_layernorm.weight": "model.safetensors",
|
| 101 |
+
"language_model.model.layers.11.mlp.down_proj.biases": "model.safetensors",
|
| 102 |
+
"language_model.model.layers.11.mlp.down_proj.scales": "model.safetensors",
|
| 103 |
+
"language_model.model.layers.11.mlp.down_proj.weight": "model.safetensors",
|
| 104 |
+
"language_model.model.layers.11.mlp.gate_proj.biases": "model.safetensors",
|
| 105 |
+
"language_model.model.layers.11.mlp.gate_proj.scales": "model.safetensors",
|
| 106 |
+
"language_model.model.layers.11.mlp.gate_proj.weight": "model.safetensors",
|
| 107 |
+
"language_model.model.layers.11.mlp.up_proj.biases": "model.safetensors",
|
| 108 |
+
"language_model.model.layers.11.mlp.up_proj.scales": "model.safetensors",
|
| 109 |
+
"language_model.model.layers.11.mlp.up_proj.weight": "model.safetensors",
|
| 110 |
+
"language_model.model.layers.11.post_attention_layernorm.weight": "model.safetensors",
|
| 111 |
+
"language_model.model.layers.11.self_attn.k_norm.weight": "model.safetensors",
|
| 112 |
+
"language_model.model.layers.11.self_attn.k_proj.biases": "model.safetensors",
|
| 113 |
+
"language_model.model.layers.11.self_attn.k_proj.scales": "model.safetensors",
|
| 114 |
+
"language_model.model.layers.11.self_attn.k_proj.weight": "model.safetensors",
|
| 115 |
+
"language_model.model.layers.11.self_attn.o_proj.biases": "model.safetensors",
|
| 116 |
+
"language_model.model.layers.11.self_attn.o_proj.scales": "model.safetensors",
|
| 117 |
+
"language_model.model.layers.11.self_attn.o_proj.weight": "model.safetensors",
|
| 118 |
+
"language_model.model.layers.11.self_attn.q_norm.weight": "model.safetensors",
|
| 119 |
+
"language_model.model.layers.11.self_attn.q_proj.biases": "model.safetensors",
|
| 120 |
+
"language_model.model.layers.11.self_attn.q_proj.scales": "model.safetensors",
|
| 121 |
+
"language_model.model.layers.11.self_attn.q_proj.weight": "model.safetensors",
|
| 122 |
+
"language_model.model.layers.11.self_attn.v_proj.biases": "model.safetensors",
|
| 123 |
+
"language_model.model.layers.11.self_attn.v_proj.scales": "model.safetensors",
|
| 124 |
+
"language_model.model.layers.11.self_attn.v_proj.weight": "model.safetensors",
|
| 125 |
+
"language_model.model.layers.12.input_layernorm.weight": "model.safetensors",
|
| 126 |
+
"language_model.model.layers.12.linear_attn.A_log": "model.safetensors",
|
| 127 |
+
"language_model.model.layers.12.linear_attn.conv1d.weight": "model.safetensors",
|
| 128 |
+
"language_model.model.layers.12.linear_attn.dt_bias": "model.safetensors",
|
| 129 |
+
"language_model.model.layers.12.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 130 |
+
"language_model.model.layers.12.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 131 |
+
"language_model.model.layers.12.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 132 |
+
"language_model.model.layers.12.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 133 |
+
"language_model.model.layers.12.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 134 |
+
"language_model.model.layers.12.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 135 |
+
"language_model.model.layers.12.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 136 |
+
"language_model.model.layers.12.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 137 |
+
"language_model.model.layers.12.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 138 |
+
"language_model.model.layers.12.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 139 |
+
"language_model.model.layers.12.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 140 |
+
"language_model.model.layers.12.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 141 |
+
"language_model.model.layers.12.linear_attn.norm.weight": "model.safetensors",
|
| 142 |
+
"language_model.model.layers.12.linear_attn.out_proj.biases": "model.safetensors",
|
| 143 |
+
"language_model.model.layers.12.linear_attn.out_proj.scales": "model.safetensors",
|
| 144 |
+
"language_model.model.layers.12.linear_attn.out_proj.weight": "model.safetensors",
|
| 145 |
+
"language_model.model.layers.12.mlp.down_proj.biases": "model.safetensors",
|
| 146 |
+
"language_model.model.layers.12.mlp.down_proj.scales": "model.safetensors",
|
| 147 |
+
"language_model.model.layers.12.mlp.down_proj.weight": "model.safetensors",
|
| 148 |
+
"language_model.model.layers.12.mlp.gate_proj.biases": "model.safetensors",
|
| 149 |
+
"language_model.model.layers.12.mlp.gate_proj.scales": "model.safetensors",
|
| 150 |
+
"language_model.model.layers.12.mlp.gate_proj.weight": "model.safetensors",
|
| 151 |
+
"language_model.model.layers.12.mlp.up_proj.biases": "model.safetensors",
|
| 152 |
+
"language_model.model.layers.12.mlp.up_proj.scales": "model.safetensors",
|
| 153 |
+
"language_model.model.layers.12.mlp.up_proj.weight": "model.safetensors",
|
| 154 |
+
"language_model.model.layers.12.post_attention_layernorm.weight": "model.safetensors",
|
| 155 |
+
"language_model.model.layers.13.input_layernorm.weight": "model.safetensors",
|
| 156 |
+
"language_model.model.layers.13.linear_attn.A_log": "model.safetensors",
|
| 157 |
+
"language_model.model.layers.13.linear_attn.conv1d.weight": "model.safetensors",
|
| 158 |
+
"language_model.model.layers.13.linear_attn.dt_bias": "model.safetensors",
|
| 159 |
+
"language_model.model.layers.13.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 160 |
+
"language_model.model.layers.13.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 161 |
+
"language_model.model.layers.13.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 162 |
+
"language_model.model.layers.13.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 163 |
+
"language_model.model.layers.13.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 164 |
+
"language_model.model.layers.13.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 165 |
+
"language_model.model.layers.13.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 166 |
+
"language_model.model.layers.13.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 167 |
+
"language_model.model.layers.13.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 168 |
+
"language_model.model.layers.13.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 169 |
+
"language_model.model.layers.13.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 170 |
+
"language_model.model.layers.13.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 171 |
+
"language_model.model.layers.13.linear_attn.norm.weight": "model.safetensors",
|
| 172 |
+
"language_model.model.layers.13.linear_attn.out_proj.biases": "model.safetensors",
|
| 173 |
+
"language_model.model.layers.13.linear_attn.out_proj.scales": "model.safetensors",
|
| 174 |
+
"language_model.model.layers.13.linear_attn.out_proj.weight": "model.safetensors",
|
| 175 |
+
"language_model.model.layers.13.mlp.down_proj.biases": "model.safetensors",
|
| 176 |
+
"language_model.model.layers.13.mlp.down_proj.scales": "model.safetensors",
|
| 177 |
+
"language_model.model.layers.13.mlp.down_proj.weight": "model.safetensors",
|
| 178 |
+
"language_model.model.layers.13.mlp.gate_proj.biases": "model.safetensors",
|
| 179 |
+
"language_model.model.layers.13.mlp.gate_proj.scales": "model.safetensors",
|
| 180 |
+
"language_model.model.layers.13.mlp.gate_proj.weight": "model.safetensors",
|
| 181 |
+
"language_model.model.layers.13.mlp.up_proj.biases": "model.safetensors",
|
| 182 |
+
"language_model.model.layers.13.mlp.up_proj.scales": "model.safetensors",
|
| 183 |
+
"language_model.model.layers.13.mlp.up_proj.weight": "model.safetensors",
|
| 184 |
+
"language_model.model.layers.13.post_attention_layernorm.weight": "model.safetensors",
|
| 185 |
+
"language_model.model.layers.14.input_layernorm.weight": "model.safetensors",
|
| 186 |
+
"language_model.model.layers.14.linear_attn.A_log": "model.safetensors",
|
| 187 |
+
"language_model.model.layers.14.linear_attn.conv1d.weight": "model.safetensors",
|
| 188 |
+
"language_model.model.layers.14.linear_attn.dt_bias": "model.safetensors",
|
| 189 |
+
"language_model.model.layers.14.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 190 |
+
"language_model.model.layers.14.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 191 |
+
"language_model.model.layers.14.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 192 |
+
"language_model.model.layers.14.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 193 |
+
"language_model.model.layers.14.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 194 |
+
"language_model.model.layers.14.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 195 |
+
"language_model.model.layers.14.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 196 |
+
"language_model.model.layers.14.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 197 |
+
"language_model.model.layers.14.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 198 |
+
"language_model.model.layers.14.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 199 |
+
"language_model.model.layers.14.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 200 |
+
"language_model.model.layers.14.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 201 |
+
"language_model.model.layers.14.linear_attn.norm.weight": "model.safetensors",
|
| 202 |
+
"language_model.model.layers.14.linear_attn.out_proj.biases": "model.safetensors",
|
| 203 |
+
"language_model.model.layers.14.linear_attn.out_proj.scales": "model.safetensors",
|
| 204 |
+
"language_model.model.layers.14.linear_attn.out_proj.weight": "model.safetensors",
|
| 205 |
+
"language_model.model.layers.14.mlp.down_proj.biases": "model.safetensors",
|
| 206 |
+
"language_model.model.layers.14.mlp.down_proj.scales": "model.safetensors",
|
| 207 |
+
"language_model.model.layers.14.mlp.down_proj.weight": "model.safetensors",
|
| 208 |
+
"language_model.model.layers.14.mlp.gate_proj.biases": "model.safetensors",
|
| 209 |
+
"language_model.model.layers.14.mlp.gate_proj.scales": "model.safetensors",
|
| 210 |
+
"language_model.model.layers.14.mlp.gate_proj.weight": "model.safetensors",
|
| 211 |
+
"language_model.model.layers.14.mlp.up_proj.biases": "model.safetensors",
|
| 212 |
+
"language_model.model.layers.14.mlp.up_proj.scales": "model.safetensors",
|
| 213 |
+
"language_model.model.layers.14.mlp.up_proj.weight": "model.safetensors",
|
| 214 |
+
"language_model.model.layers.14.post_attention_layernorm.weight": "model.safetensors",
|
| 215 |
+
"language_model.model.layers.15.input_layernorm.weight": "model.safetensors",
|
| 216 |
+
"language_model.model.layers.15.mlp.down_proj.biases": "model.safetensors",
|
| 217 |
+
"language_model.model.layers.15.mlp.down_proj.scales": "model.safetensors",
|
| 218 |
+
"language_model.model.layers.15.mlp.down_proj.weight": "model.safetensors",
|
| 219 |
+
"language_model.model.layers.15.mlp.gate_proj.biases": "model.safetensors",
|
| 220 |
+
"language_model.model.layers.15.mlp.gate_proj.scales": "model.safetensors",
|
| 221 |
+
"language_model.model.layers.15.mlp.gate_proj.weight": "model.safetensors",
|
| 222 |
+
"language_model.model.layers.15.mlp.up_proj.biases": "model.safetensors",
|
| 223 |
+
"language_model.model.layers.15.mlp.up_proj.scales": "model.safetensors",
|
| 224 |
+
"language_model.model.layers.15.mlp.up_proj.weight": "model.safetensors",
|
| 225 |
+
"language_model.model.layers.15.post_attention_layernorm.weight": "model.safetensors",
|
| 226 |
+
"language_model.model.layers.15.self_attn.k_norm.weight": "model.safetensors",
|
| 227 |
+
"language_model.model.layers.15.self_attn.k_proj.biases": "model.safetensors",
|
| 228 |
+
"language_model.model.layers.15.self_attn.k_proj.scales": "model.safetensors",
|
| 229 |
+
"language_model.model.layers.15.self_attn.k_proj.weight": "model.safetensors",
|
| 230 |
+
"language_model.model.layers.15.self_attn.o_proj.biases": "model.safetensors",
|
| 231 |
+
"language_model.model.layers.15.self_attn.o_proj.scales": "model.safetensors",
|
| 232 |
+
"language_model.model.layers.15.self_attn.o_proj.weight": "model.safetensors",
|
| 233 |
+
"language_model.model.layers.15.self_attn.q_norm.weight": "model.safetensors",
|
| 234 |
+
"language_model.model.layers.15.self_attn.q_proj.biases": "model.safetensors",
|
| 235 |
+
"language_model.model.layers.15.self_attn.q_proj.scales": "model.safetensors",
|
| 236 |
+
"language_model.model.layers.15.self_attn.q_proj.weight": "model.safetensors",
|
| 237 |
+
"language_model.model.layers.15.self_attn.v_proj.biases": "model.safetensors",
|
| 238 |
+
"language_model.model.layers.15.self_attn.v_proj.scales": "model.safetensors",
|
| 239 |
+
"language_model.model.layers.15.self_attn.v_proj.weight": "model.safetensors",
|
| 240 |
+
"language_model.model.layers.16.input_layernorm.weight": "model.safetensors",
|
| 241 |
+
"language_model.model.layers.16.linear_attn.A_log": "model.safetensors",
|
| 242 |
+
"language_model.model.layers.16.linear_attn.conv1d.weight": "model.safetensors",
|
| 243 |
+
"language_model.model.layers.16.linear_attn.dt_bias": "model.safetensors",
|
| 244 |
+
"language_model.model.layers.16.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 245 |
+
"language_model.model.layers.16.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 246 |
+
"language_model.model.layers.16.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 247 |
+
"language_model.model.layers.16.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 248 |
+
"language_model.model.layers.16.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 249 |
+
"language_model.model.layers.16.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 250 |
+
"language_model.model.layers.16.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 251 |
+
"language_model.model.layers.16.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 252 |
+
"language_model.model.layers.16.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 253 |
+
"language_model.model.layers.16.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 254 |
+
"language_model.model.layers.16.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 255 |
+
"language_model.model.layers.16.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 256 |
+
"language_model.model.layers.16.linear_attn.norm.weight": "model.safetensors",
|
| 257 |
+
"language_model.model.layers.16.linear_attn.out_proj.biases": "model.safetensors",
|
| 258 |
+
"language_model.model.layers.16.linear_attn.out_proj.scales": "model.safetensors",
|
| 259 |
+
"language_model.model.layers.16.linear_attn.out_proj.weight": "model.safetensors",
|
| 260 |
+
"language_model.model.layers.16.mlp.down_proj.biases": "model.safetensors",
|
| 261 |
+
"language_model.model.layers.16.mlp.down_proj.scales": "model.safetensors",
|
| 262 |
+
"language_model.model.layers.16.mlp.down_proj.weight": "model.safetensors",
|
| 263 |
+
"language_model.model.layers.16.mlp.gate_proj.biases": "model.safetensors",
|
| 264 |
+
"language_model.model.layers.16.mlp.gate_proj.scales": "model.safetensors",
|
| 265 |
+
"language_model.model.layers.16.mlp.gate_proj.weight": "model.safetensors",
|
| 266 |
+
"language_model.model.layers.16.mlp.up_proj.biases": "model.safetensors",
|
| 267 |
+
"language_model.model.layers.16.mlp.up_proj.scales": "model.safetensors",
|
| 268 |
+
"language_model.model.layers.16.mlp.up_proj.weight": "model.safetensors",
|
| 269 |
+
"language_model.model.layers.16.post_attention_layernorm.weight": "model.safetensors",
|
| 270 |
+
"language_model.model.layers.17.input_layernorm.weight": "model.safetensors",
|
| 271 |
+
"language_model.model.layers.17.linear_attn.A_log": "model.safetensors",
|
| 272 |
+
"language_model.model.layers.17.linear_attn.conv1d.weight": "model.safetensors",
|
| 273 |
+
"language_model.model.layers.17.linear_attn.dt_bias": "model.safetensors",
|
| 274 |
+
"language_model.model.layers.17.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 275 |
+
"language_model.model.layers.17.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 276 |
+
"language_model.model.layers.17.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 277 |
+
"language_model.model.layers.17.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 278 |
+
"language_model.model.layers.17.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 279 |
+
"language_model.model.layers.17.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 280 |
+
"language_model.model.layers.17.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 281 |
+
"language_model.model.layers.17.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 282 |
+
"language_model.model.layers.17.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 283 |
+
"language_model.model.layers.17.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 284 |
+
"language_model.model.layers.17.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 285 |
+
"language_model.model.layers.17.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 286 |
+
"language_model.model.layers.17.linear_attn.norm.weight": "model.safetensors",
|
| 287 |
+
"language_model.model.layers.17.linear_attn.out_proj.biases": "model.safetensors",
|
| 288 |
+
"language_model.model.layers.17.linear_attn.out_proj.scales": "model.safetensors",
|
| 289 |
+
"language_model.model.layers.17.linear_attn.out_proj.weight": "model.safetensors",
|
| 290 |
+
"language_model.model.layers.17.mlp.down_proj.biases": "model.safetensors",
|
| 291 |
+
"language_model.model.layers.17.mlp.down_proj.scales": "model.safetensors",
|
| 292 |
+
"language_model.model.layers.17.mlp.down_proj.weight": "model.safetensors",
|
| 293 |
+
"language_model.model.layers.17.mlp.gate_proj.biases": "model.safetensors",
|
| 294 |
+
"language_model.model.layers.17.mlp.gate_proj.scales": "model.safetensors",
|
| 295 |
+
"language_model.model.layers.17.mlp.gate_proj.weight": "model.safetensors",
|
| 296 |
+
"language_model.model.layers.17.mlp.up_proj.biases": "model.safetensors",
|
| 297 |
+
"language_model.model.layers.17.mlp.up_proj.scales": "model.safetensors",
|
| 298 |
+
"language_model.model.layers.17.mlp.up_proj.weight": "model.safetensors",
|
| 299 |
+
"language_model.model.layers.17.post_attention_layernorm.weight": "model.safetensors",
|
| 300 |
+
"language_model.model.layers.18.input_layernorm.weight": "model.safetensors",
|
| 301 |
+
"language_model.model.layers.18.linear_attn.A_log": "model.safetensors",
|
| 302 |
+
"language_model.model.layers.18.linear_attn.conv1d.weight": "model.safetensors",
|
| 303 |
+
"language_model.model.layers.18.linear_attn.dt_bias": "model.safetensors",
|
| 304 |
+
"language_model.model.layers.18.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 305 |
+
"language_model.model.layers.18.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 306 |
+
"language_model.model.layers.18.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 307 |
+
"language_model.model.layers.18.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 308 |
+
"language_model.model.layers.18.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 309 |
+
"language_model.model.layers.18.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 310 |
+
"language_model.model.layers.18.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 311 |
+
"language_model.model.layers.18.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 312 |
+
"language_model.model.layers.18.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 313 |
+
"language_model.model.layers.18.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 314 |
+
"language_model.model.layers.18.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 315 |
+
"language_model.model.layers.18.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 316 |
+
"language_model.model.layers.18.linear_attn.norm.weight": "model.safetensors",
|
| 317 |
+
"language_model.model.layers.18.linear_attn.out_proj.biases": "model.safetensors",
|
| 318 |
+
"language_model.model.layers.18.linear_attn.out_proj.scales": "model.safetensors",
|
| 319 |
+
"language_model.model.layers.18.linear_attn.out_proj.weight": "model.safetensors",
|
| 320 |
+
"language_model.model.layers.18.mlp.down_proj.biases": "model.safetensors",
|
| 321 |
+
"language_model.model.layers.18.mlp.down_proj.scales": "model.safetensors",
|
| 322 |
+
"language_model.model.layers.18.mlp.down_proj.weight": "model.safetensors",
|
| 323 |
+
"language_model.model.layers.18.mlp.gate_proj.biases": "model.safetensors",
|
| 324 |
+
"language_model.model.layers.18.mlp.gate_proj.scales": "model.safetensors",
|
| 325 |
+
"language_model.model.layers.18.mlp.gate_proj.weight": "model.safetensors",
|
| 326 |
+
"language_model.model.layers.18.mlp.up_proj.biases": "model.safetensors",
|
| 327 |
+
"language_model.model.layers.18.mlp.up_proj.scales": "model.safetensors",
|
| 328 |
+
"language_model.model.layers.18.mlp.up_proj.weight": "model.safetensors",
|
| 329 |
+
"language_model.model.layers.18.post_attention_layernorm.weight": "model.safetensors",
|
| 330 |
+
"language_model.model.layers.19.input_layernorm.weight": "model.safetensors",
|
| 331 |
+
"language_model.model.layers.19.mlp.down_proj.biases": "model.safetensors",
|
| 332 |
+
"language_model.model.layers.19.mlp.down_proj.scales": "model.safetensors",
|
| 333 |
+
"language_model.model.layers.19.mlp.down_proj.weight": "model.safetensors",
|
| 334 |
+
"language_model.model.layers.19.mlp.gate_proj.biases": "model.safetensors",
|
| 335 |
+
"language_model.model.layers.19.mlp.gate_proj.scales": "model.safetensors",
|
| 336 |
+
"language_model.model.layers.19.mlp.gate_proj.weight": "model.safetensors",
|
| 337 |
+
"language_model.model.layers.19.mlp.up_proj.biases": "model.safetensors",
|
| 338 |
+
"language_model.model.layers.19.mlp.up_proj.scales": "model.safetensors",
|
| 339 |
+
"language_model.model.layers.19.mlp.up_proj.weight": "model.safetensors",
|
| 340 |
+
"language_model.model.layers.19.post_attention_layernorm.weight": "model.safetensors",
|
| 341 |
+
"language_model.model.layers.19.self_attn.k_norm.weight": "model.safetensors",
|
| 342 |
+
"language_model.model.layers.19.self_attn.k_proj.biases": "model.safetensors",
|
| 343 |
+
"language_model.model.layers.19.self_attn.k_proj.scales": "model.safetensors",
|
| 344 |
+
"language_model.model.layers.19.self_attn.k_proj.weight": "model.safetensors",
|
| 345 |
+
"language_model.model.layers.19.self_attn.o_proj.biases": "model.safetensors",
|
| 346 |
+
"language_model.model.layers.19.self_attn.o_proj.scales": "model.safetensors",
|
| 347 |
+
"language_model.model.layers.19.self_attn.o_proj.weight": "model.safetensors",
|
| 348 |
+
"language_model.model.layers.19.self_attn.q_norm.weight": "model.safetensors",
|
| 349 |
+
"language_model.model.layers.19.self_attn.q_proj.biases": "model.safetensors",
|
| 350 |
+
"language_model.model.layers.19.self_attn.q_proj.scales": "model.safetensors",
|
| 351 |
+
"language_model.model.layers.19.self_attn.q_proj.weight": "model.safetensors",
|
| 352 |
+
"language_model.model.layers.19.self_attn.v_proj.biases": "model.safetensors",
|
| 353 |
+
"language_model.model.layers.19.self_attn.v_proj.scales": "model.safetensors",
|
| 354 |
+
"language_model.model.layers.19.self_attn.v_proj.weight": "model.safetensors",
|
| 355 |
+
"language_model.model.layers.2.input_layernorm.weight": "model.safetensors",
|
| 356 |
+
"language_model.model.layers.2.linear_attn.A_log": "model.safetensors",
|
| 357 |
+
"language_model.model.layers.2.linear_attn.conv1d.weight": "model.safetensors",
|
| 358 |
+
"language_model.model.layers.2.linear_attn.dt_bias": "model.safetensors",
|
| 359 |
+
"language_model.model.layers.2.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 360 |
+
"language_model.model.layers.2.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 361 |
+
"language_model.model.layers.2.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 362 |
+
"language_model.model.layers.2.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 363 |
+
"language_model.model.layers.2.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 364 |
+
"language_model.model.layers.2.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 365 |
+
"language_model.model.layers.2.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 366 |
+
"language_model.model.layers.2.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 367 |
+
"language_model.model.layers.2.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 368 |
+
"language_model.model.layers.2.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 369 |
+
"language_model.model.layers.2.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 370 |
+
"language_model.model.layers.2.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 371 |
+
"language_model.model.layers.2.linear_attn.norm.weight": "model.safetensors",
|
| 372 |
+
"language_model.model.layers.2.linear_attn.out_proj.biases": "model.safetensors",
|
| 373 |
+
"language_model.model.layers.2.linear_attn.out_proj.scales": "model.safetensors",
|
| 374 |
+
"language_model.model.layers.2.linear_attn.out_proj.weight": "model.safetensors",
|
| 375 |
+
"language_model.model.layers.2.mlp.down_proj.biases": "model.safetensors",
|
| 376 |
+
"language_model.model.layers.2.mlp.down_proj.scales": "model.safetensors",
|
| 377 |
+
"language_model.model.layers.2.mlp.down_proj.weight": "model.safetensors",
|
| 378 |
+
"language_model.model.layers.2.mlp.gate_proj.biases": "model.safetensors",
|
| 379 |
+
"language_model.model.layers.2.mlp.gate_proj.scales": "model.safetensors",
|
| 380 |
+
"language_model.model.layers.2.mlp.gate_proj.weight": "model.safetensors",
|
| 381 |
+
"language_model.model.layers.2.mlp.up_proj.biases": "model.safetensors",
|
| 382 |
+
"language_model.model.layers.2.mlp.up_proj.scales": "model.safetensors",
|
| 383 |
+
"language_model.model.layers.2.mlp.up_proj.weight": "model.safetensors",
|
| 384 |
+
"language_model.model.layers.2.post_attention_layernorm.weight": "model.safetensors",
|
| 385 |
+
"language_model.model.layers.20.input_layernorm.weight": "model.safetensors",
|
| 386 |
+
"language_model.model.layers.20.linear_attn.A_log": "model.safetensors",
|
| 387 |
+
"language_model.model.layers.20.linear_attn.conv1d.weight": "model.safetensors",
|
| 388 |
+
"language_model.model.layers.20.linear_attn.dt_bias": "model.safetensors",
|
| 389 |
+
"language_model.model.layers.20.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 390 |
+
"language_model.model.layers.20.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 391 |
+
"language_model.model.layers.20.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 392 |
+
"language_model.model.layers.20.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 393 |
+
"language_model.model.layers.20.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 394 |
+
"language_model.model.layers.20.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 395 |
+
"language_model.model.layers.20.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 396 |
+
"language_model.model.layers.20.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 397 |
+
"language_model.model.layers.20.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 398 |
+
"language_model.model.layers.20.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 399 |
+
"language_model.model.layers.20.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 400 |
+
"language_model.model.layers.20.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 401 |
+
"language_model.model.layers.20.linear_attn.norm.weight": "model.safetensors",
|
| 402 |
+
"language_model.model.layers.20.linear_attn.out_proj.biases": "model.safetensors",
|
| 403 |
+
"language_model.model.layers.20.linear_attn.out_proj.scales": "model.safetensors",
|
| 404 |
+
"language_model.model.layers.20.linear_attn.out_proj.weight": "model.safetensors",
|
| 405 |
+
"language_model.model.layers.20.mlp.down_proj.biases": "model.safetensors",
|
| 406 |
+
"language_model.model.layers.20.mlp.down_proj.scales": "model.safetensors",
|
| 407 |
+
"language_model.model.layers.20.mlp.down_proj.weight": "model.safetensors",
|
| 408 |
+
"language_model.model.layers.20.mlp.gate_proj.biases": "model.safetensors",
|
| 409 |
+
"language_model.model.layers.20.mlp.gate_proj.scales": "model.safetensors",
|
| 410 |
+
"language_model.model.layers.20.mlp.gate_proj.weight": "model.safetensors",
|
| 411 |
+
"language_model.model.layers.20.mlp.up_proj.biases": "model.safetensors",
|
| 412 |
+
"language_model.model.layers.20.mlp.up_proj.scales": "model.safetensors",
|
| 413 |
+
"language_model.model.layers.20.mlp.up_proj.weight": "model.safetensors",
|
| 414 |
+
"language_model.model.layers.20.post_attention_layernorm.weight": "model.safetensors",
|
| 415 |
+
"language_model.model.layers.21.input_layernorm.weight": "model.safetensors",
|
| 416 |
+
"language_model.model.layers.21.linear_attn.A_log": "model.safetensors",
|
| 417 |
+
"language_model.model.layers.21.linear_attn.conv1d.weight": "model.safetensors",
|
| 418 |
+
"language_model.model.layers.21.linear_attn.dt_bias": "model.safetensors",
|
| 419 |
+
"language_model.model.layers.21.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 420 |
+
"language_model.model.layers.21.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 421 |
+
"language_model.model.layers.21.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 422 |
+
"language_model.model.layers.21.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 423 |
+
"language_model.model.layers.21.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 424 |
+
"language_model.model.layers.21.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 425 |
+
"language_model.model.layers.21.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 426 |
+
"language_model.model.layers.21.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 427 |
+
"language_model.model.layers.21.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 428 |
+
"language_model.model.layers.21.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 429 |
+
"language_model.model.layers.21.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 430 |
+
"language_model.model.layers.21.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 431 |
+
"language_model.model.layers.21.linear_attn.norm.weight": "model.safetensors",
|
| 432 |
+
"language_model.model.layers.21.linear_attn.out_proj.biases": "model.safetensors",
|
| 433 |
+
"language_model.model.layers.21.linear_attn.out_proj.scales": "model.safetensors",
|
| 434 |
+
"language_model.model.layers.21.linear_attn.out_proj.weight": "model.safetensors",
|
| 435 |
+
"language_model.model.layers.21.mlp.down_proj.biases": "model.safetensors",
|
| 436 |
+
"language_model.model.layers.21.mlp.down_proj.scales": "model.safetensors",
|
| 437 |
+
"language_model.model.layers.21.mlp.down_proj.weight": "model.safetensors",
|
| 438 |
+
"language_model.model.layers.21.mlp.gate_proj.biases": "model.safetensors",
|
| 439 |
+
"language_model.model.layers.21.mlp.gate_proj.scales": "model.safetensors",
|
| 440 |
+
"language_model.model.layers.21.mlp.gate_proj.weight": "model.safetensors",
|
| 441 |
+
"language_model.model.layers.21.mlp.up_proj.biases": "model.safetensors",
|
| 442 |
+
"language_model.model.layers.21.mlp.up_proj.scales": "model.safetensors",
|
| 443 |
+
"language_model.model.layers.21.mlp.up_proj.weight": "model.safetensors",
|
| 444 |
+
"language_model.model.layers.21.post_attention_layernorm.weight": "model.safetensors",
|
| 445 |
+
"language_model.model.layers.22.input_layernorm.weight": "model.safetensors",
|
| 446 |
+
"language_model.model.layers.22.linear_attn.A_log": "model.safetensors",
|
| 447 |
+
"language_model.model.layers.22.linear_attn.conv1d.weight": "model.safetensors",
|
| 448 |
+
"language_model.model.layers.22.linear_attn.dt_bias": "model.safetensors",
|
| 449 |
+
"language_model.model.layers.22.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 450 |
+
"language_model.model.layers.22.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 451 |
+
"language_model.model.layers.22.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 452 |
+
"language_model.model.layers.22.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 453 |
+
"language_model.model.layers.22.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 454 |
+
"language_model.model.layers.22.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 455 |
+
"language_model.model.layers.22.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 456 |
+
"language_model.model.layers.22.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 457 |
+
"language_model.model.layers.22.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 458 |
+
"language_model.model.layers.22.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 459 |
+
"language_model.model.layers.22.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 460 |
+
"language_model.model.layers.22.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 461 |
+
"language_model.model.layers.22.linear_attn.norm.weight": "model.safetensors",
|
| 462 |
+
"language_model.model.layers.22.linear_attn.out_proj.biases": "model.safetensors",
|
| 463 |
+
"language_model.model.layers.22.linear_attn.out_proj.scales": "model.safetensors",
|
| 464 |
+
"language_model.model.layers.22.linear_attn.out_proj.weight": "model.safetensors",
|
| 465 |
+
"language_model.model.layers.22.mlp.down_proj.biases": "model.safetensors",
|
| 466 |
+
"language_model.model.layers.22.mlp.down_proj.scales": "model.safetensors",
|
| 467 |
+
"language_model.model.layers.22.mlp.down_proj.weight": "model.safetensors",
|
| 468 |
+
"language_model.model.layers.22.mlp.gate_proj.biases": "model.safetensors",
|
| 469 |
+
"language_model.model.layers.22.mlp.gate_proj.scales": "model.safetensors",
|
| 470 |
+
"language_model.model.layers.22.mlp.gate_proj.weight": "model.safetensors",
|
| 471 |
+
"language_model.model.layers.22.mlp.up_proj.biases": "model.safetensors",
|
| 472 |
+
"language_model.model.layers.22.mlp.up_proj.scales": "model.safetensors",
|
| 473 |
+
"language_model.model.layers.22.mlp.up_proj.weight": "model.safetensors",
|
| 474 |
+
"language_model.model.layers.22.post_attention_layernorm.weight": "model.safetensors",
|
| 475 |
+
"language_model.model.layers.23.input_layernorm.weight": "model.safetensors",
|
| 476 |
+
"language_model.model.layers.23.mlp.down_proj.biases": "model.safetensors",
|
| 477 |
+
"language_model.model.layers.23.mlp.down_proj.scales": "model.safetensors",
|
| 478 |
+
"language_model.model.layers.23.mlp.down_proj.weight": "model.safetensors",
|
| 479 |
+
"language_model.model.layers.23.mlp.gate_proj.biases": "model.safetensors",
|
| 480 |
+
"language_model.model.layers.23.mlp.gate_proj.scales": "model.safetensors",
|
| 481 |
+
"language_model.model.layers.23.mlp.gate_proj.weight": "model.safetensors",
|
| 482 |
+
"language_model.model.layers.23.mlp.up_proj.biases": "model.safetensors",
|
| 483 |
+
"language_model.model.layers.23.mlp.up_proj.scales": "model.safetensors",
|
| 484 |
+
"language_model.model.layers.23.mlp.up_proj.weight": "model.safetensors",
|
| 485 |
+
"language_model.model.layers.23.post_attention_layernorm.weight": "model.safetensors",
|
| 486 |
+
"language_model.model.layers.23.self_attn.k_norm.weight": "model.safetensors",
|
| 487 |
+
"language_model.model.layers.23.self_attn.k_proj.biases": "model.safetensors",
|
| 488 |
+
"language_model.model.layers.23.self_attn.k_proj.scales": "model.safetensors",
|
| 489 |
+
"language_model.model.layers.23.self_attn.k_proj.weight": "model.safetensors",
|
| 490 |
+
"language_model.model.layers.23.self_attn.o_proj.biases": "model.safetensors",
|
| 491 |
+
"language_model.model.layers.23.self_attn.o_proj.scales": "model.safetensors",
|
| 492 |
+
"language_model.model.layers.23.self_attn.o_proj.weight": "model.safetensors",
|
| 493 |
+
"language_model.model.layers.23.self_attn.q_norm.weight": "model.safetensors",
|
| 494 |
+
"language_model.model.layers.23.self_attn.q_proj.biases": "model.safetensors",
|
| 495 |
+
"language_model.model.layers.23.self_attn.q_proj.scales": "model.safetensors",
|
| 496 |
+
"language_model.model.layers.23.self_attn.q_proj.weight": "model.safetensors",
|
| 497 |
+
"language_model.model.layers.23.self_attn.v_proj.biases": "model.safetensors",
|
| 498 |
+
"language_model.model.layers.23.self_attn.v_proj.scales": "model.safetensors",
|
| 499 |
+
"language_model.model.layers.23.self_attn.v_proj.weight": "model.safetensors",
|
| 500 |
+
"language_model.model.layers.3.input_layernorm.weight": "model.safetensors",
|
| 501 |
+
"language_model.model.layers.3.mlp.down_proj.biases": "model.safetensors",
|
| 502 |
+
"language_model.model.layers.3.mlp.down_proj.scales": "model.safetensors",
|
| 503 |
+
"language_model.model.layers.3.mlp.down_proj.weight": "model.safetensors",
|
| 504 |
+
"language_model.model.layers.3.mlp.gate_proj.biases": "model.safetensors",
|
| 505 |
+
"language_model.model.layers.3.mlp.gate_proj.scales": "model.safetensors",
|
| 506 |
+
"language_model.model.layers.3.mlp.gate_proj.weight": "model.safetensors",
|
| 507 |
+
"language_model.model.layers.3.mlp.up_proj.biases": "model.safetensors",
|
| 508 |
+
"language_model.model.layers.3.mlp.up_proj.scales": "model.safetensors",
|
| 509 |
+
"language_model.model.layers.3.mlp.up_proj.weight": "model.safetensors",
|
| 510 |
+
"language_model.model.layers.3.post_attention_layernorm.weight": "model.safetensors",
|
| 511 |
+
"language_model.model.layers.3.self_attn.k_norm.weight": "model.safetensors",
|
| 512 |
+
"language_model.model.layers.3.self_attn.k_proj.biases": "model.safetensors",
|
| 513 |
+
"language_model.model.layers.3.self_attn.k_proj.scales": "model.safetensors",
|
| 514 |
+
"language_model.model.layers.3.self_attn.k_proj.weight": "model.safetensors",
|
| 515 |
+
"language_model.model.layers.3.self_attn.o_proj.biases": "model.safetensors",
|
| 516 |
+
"language_model.model.layers.3.self_attn.o_proj.scales": "model.safetensors",
|
| 517 |
+
"language_model.model.layers.3.self_attn.o_proj.weight": "model.safetensors",
|
| 518 |
+
"language_model.model.layers.3.self_attn.q_norm.weight": "model.safetensors",
|
| 519 |
+
"language_model.model.layers.3.self_attn.q_proj.biases": "model.safetensors",
|
| 520 |
+
"language_model.model.layers.3.self_attn.q_proj.scales": "model.safetensors",
|
| 521 |
+
"language_model.model.layers.3.self_attn.q_proj.weight": "model.safetensors",
|
| 522 |
+
"language_model.model.layers.3.self_attn.v_proj.biases": "model.safetensors",
|
| 523 |
+
"language_model.model.layers.3.self_attn.v_proj.scales": "model.safetensors",
|
| 524 |
+
"language_model.model.layers.3.self_attn.v_proj.weight": "model.safetensors",
|
| 525 |
+
"language_model.model.layers.4.input_layernorm.weight": "model.safetensors",
|
| 526 |
+
"language_model.model.layers.4.linear_attn.A_log": "model.safetensors",
|
| 527 |
+
"language_model.model.layers.4.linear_attn.conv1d.weight": "model.safetensors",
|
| 528 |
+
"language_model.model.layers.4.linear_attn.dt_bias": "model.safetensors",
|
| 529 |
+
"language_model.model.layers.4.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 530 |
+
"language_model.model.layers.4.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 531 |
+
"language_model.model.layers.4.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 532 |
+
"language_model.model.layers.4.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 533 |
+
"language_model.model.layers.4.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 534 |
+
"language_model.model.layers.4.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 535 |
+
"language_model.model.layers.4.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 536 |
+
"language_model.model.layers.4.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 537 |
+
"language_model.model.layers.4.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 538 |
+
"language_model.model.layers.4.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 539 |
+
"language_model.model.layers.4.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 540 |
+
"language_model.model.layers.4.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 541 |
+
"language_model.model.layers.4.linear_attn.norm.weight": "model.safetensors",
|
| 542 |
+
"language_model.model.layers.4.linear_attn.out_proj.biases": "model.safetensors",
|
| 543 |
+
"language_model.model.layers.4.linear_attn.out_proj.scales": "model.safetensors",
|
| 544 |
+
"language_model.model.layers.4.linear_attn.out_proj.weight": "model.safetensors",
|
| 545 |
+
"language_model.model.layers.4.mlp.down_proj.biases": "model.safetensors",
|
| 546 |
+
"language_model.model.layers.4.mlp.down_proj.scales": "model.safetensors",
|
| 547 |
+
"language_model.model.layers.4.mlp.down_proj.weight": "model.safetensors",
|
| 548 |
+
"language_model.model.layers.4.mlp.gate_proj.biases": "model.safetensors",
|
| 549 |
+
"language_model.model.layers.4.mlp.gate_proj.scales": "model.safetensors",
|
| 550 |
+
"language_model.model.layers.4.mlp.gate_proj.weight": "model.safetensors",
|
| 551 |
+
"language_model.model.layers.4.mlp.up_proj.biases": "model.safetensors",
|
| 552 |
+
"language_model.model.layers.4.mlp.up_proj.scales": "model.safetensors",
|
| 553 |
+
"language_model.model.layers.4.mlp.up_proj.weight": "model.safetensors",
|
| 554 |
+
"language_model.model.layers.4.post_attention_layernorm.weight": "model.safetensors",
|
| 555 |
+
"language_model.model.layers.5.input_layernorm.weight": "model.safetensors",
|
| 556 |
+
"language_model.model.layers.5.linear_attn.A_log": "model.safetensors",
|
| 557 |
+
"language_model.model.layers.5.linear_attn.conv1d.weight": "model.safetensors",
|
| 558 |
+
"language_model.model.layers.5.linear_attn.dt_bias": "model.safetensors",
|
| 559 |
+
"language_model.model.layers.5.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 560 |
+
"language_model.model.layers.5.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 561 |
+
"language_model.model.layers.5.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 562 |
+
"language_model.model.layers.5.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 563 |
+
"language_model.model.layers.5.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 564 |
+
"language_model.model.layers.5.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 565 |
+
"language_model.model.layers.5.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 566 |
+
"language_model.model.layers.5.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 567 |
+
"language_model.model.layers.5.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 568 |
+
"language_model.model.layers.5.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 569 |
+
"language_model.model.layers.5.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 570 |
+
"language_model.model.layers.5.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 571 |
+
"language_model.model.layers.5.linear_attn.norm.weight": "model.safetensors",
|
| 572 |
+
"language_model.model.layers.5.linear_attn.out_proj.biases": "model.safetensors",
|
| 573 |
+
"language_model.model.layers.5.linear_attn.out_proj.scales": "model.safetensors",
|
| 574 |
+
"language_model.model.layers.5.linear_attn.out_proj.weight": "model.safetensors",
|
| 575 |
+
"language_model.model.layers.5.mlp.down_proj.biases": "model.safetensors",
|
| 576 |
+
"language_model.model.layers.5.mlp.down_proj.scales": "model.safetensors",
|
| 577 |
+
"language_model.model.layers.5.mlp.down_proj.weight": "model.safetensors",
|
| 578 |
+
"language_model.model.layers.5.mlp.gate_proj.biases": "model.safetensors",
|
| 579 |
+
"language_model.model.layers.5.mlp.gate_proj.scales": "model.safetensors",
|
| 580 |
+
"language_model.model.layers.5.mlp.gate_proj.weight": "model.safetensors",
|
| 581 |
+
"language_model.model.layers.5.mlp.up_proj.biases": "model.safetensors",
|
| 582 |
+
"language_model.model.layers.5.mlp.up_proj.scales": "model.safetensors",
|
| 583 |
+
"language_model.model.layers.5.mlp.up_proj.weight": "model.safetensors",
|
| 584 |
+
"language_model.model.layers.5.post_attention_layernorm.weight": "model.safetensors",
|
| 585 |
+
"language_model.model.layers.6.input_layernorm.weight": "model.safetensors",
|
| 586 |
+
"language_model.model.layers.6.linear_attn.A_log": "model.safetensors",
|
| 587 |
+
"language_model.model.layers.6.linear_attn.conv1d.weight": "model.safetensors",
|
| 588 |
+
"language_model.model.layers.6.linear_attn.dt_bias": "model.safetensors",
|
| 589 |
+
"language_model.model.layers.6.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 590 |
+
"language_model.model.layers.6.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 591 |
+
"language_model.model.layers.6.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 592 |
+
"language_model.model.layers.6.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 593 |
+
"language_model.model.layers.6.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 594 |
+
"language_model.model.layers.6.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 595 |
+
"language_model.model.layers.6.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 596 |
+
"language_model.model.layers.6.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 597 |
+
"language_model.model.layers.6.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 598 |
+
"language_model.model.layers.6.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 599 |
+
"language_model.model.layers.6.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 600 |
+
"language_model.model.layers.6.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 601 |
+
"language_model.model.layers.6.linear_attn.norm.weight": "model.safetensors",
|
| 602 |
+
"language_model.model.layers.6.linear_attn.out_proj.biases": "model.safetensors",
|
| 603 |
+
"language_model.model.layers.6.linear_attn.out_proj.scales": "model.safetensors",
|
| 604 |
+
"language_model.model.layers.6.linear_attn.out_proj.weight": "model.safetensors",
|
| 605 |
+
"language_model.model.layers.6.mlp.down_proj.biases": "model.safetensors",
|
| 606 |
+
"language_model.model.layers.6.mlp.down_proj.scales": "model.safetensors",
|
| 607 |
+
"language_model.model.layers.6.mlp.down_proj.weight": "model.safetensors",
|
| 608 |
+
"language_model.model.layers.6.mlp.gate_proj.biases": "model.safetensors",
|
| 609 |
+
"language_model.model.layers.6.mlp.gate_proj.scales": "model.safetensors",
|
| 610 |
+
"language_model.model.layers.6.mlp.gate_proj.weight": "model.safetensors",
|
| 611 |
+
"language_model.model.layers.6.mlp.up_proj.biases": "model.safetensors",
|
| 612 |
+
"language_model.model.layers.6.mlp.up_proj.scales": "model.safetensors",
|
| 613 |
+
"language_model.model.layers.6.mlp.up_proj.weight": "model.safetensors",
|
| 614 |
+
"language_model.model.layers.6.post_attention_layernorm.weight": "model.safetensors",
|
| 615 |
+
"language_model.model.layers.7.input_layernorm.weight": "model.safetensors",
|
| 616 |
+
"language_model.model.layers.7.mlp.down_proj.biases": "model.safetensors",
|
| 617 |
+
"language_model.model.layers.7.mlp.down_proj.scales": "model.safetensors",
|
| 618 |
+
"language_model.model.layers.7.mlp.down_proj.weight": "model.safetensors",
|
| 619 |
+
"language_model.model.layers.7.mlp.gate_proj.biases": "model.safetensors",
|
| 620 |
+
"language_model.model.layers.7.mlp.gate_proj.scales": "model.safetensors",
|
| 621 |
+
"language_model.model.layers.7.mlp.gate_proj.weight": "model.safetensors",
|
| 622 |
+
"language_model.model.layers.7.mlp.up_proj.biases": "model.safetensors",
|
| 623 |
+
"language_model.model.layers.7.mlp.up_proj.scales": "model.safetensors",
|
| 624 |
+
"language_model.model.layers.7.mlp.up_proj.weight": "model.safetensors",
|
| 625 |
+
"language_model.model.layers.7.post_attention_layernorm.weight": "model.safetensors",
|
| 626 |
+
"language_model.model.layers.7.self_attn.k_norm.weight": "model.safetensors",
|
| 627 |
+
"language_model.model.layers.7.self_attn.k_proj.biases": "model.safetensors",
|
| 628 |
+
"language_model.model.layers.7.self_attn.k_proj.scales": "model.safetensors",
|
| 629 |
+
"language_model.model.layers.7.self_attn.k_proj.weight": "model.safetensors",
|
| 630 |
+
"language_model.model.layers.7.self_attn.o_proj.biases": "model.safetensors",
|
| 631 |
+
"language_model.model.layers.7.self_attn.o_proj.scales": "model.safetensors",
|
| 632 |
+
"language_model.model.layers.7.self_attn.o_proj.weight": "model.safetensors",
|
| 633 |
+
"language_model.model.layers.7.self_attn.q_norm.weight": "model.safetensors",
|
| 634 |
+
"language_model.model.layers.7.self_attn.q_proj.biases": "model.safetensors",
|
| 635 |
+
"language_model.model.layers.7.self_attn.q_proj.scales": "model.safetensors",
|
| 636 |
+
"language_model.model.layers.7.self_attn.q_proj.weight": "model.safetensors",
|
| 637 |
+
"language_model.model.layers.7.self_attn.v_proj.biases": "model.safetensors",
|
| 638 |
+
"language_model.model.layers.7.self_attn.v_proj.scales": "model.safetensors",
|
| 639 |
+
"language_model.model.layers.7.self_attn.v_proj.weight": "model.safetensors",
|
| 640 |
+
"language_model.model.layers.8.input_layernorm.weight": "model.safetensors",
|
| 641 |
+
"language_model.model.layers.8.linear_attn.A_log": "model.safetensors",
|
| 642 |
+
"language_model.model.layers.8.linear_attn.conv1d.weight": "model.safetensors",
|
| 643 |
+
"language_model.model.layers.8.linear_attn.dt_bias": "model.safetensors",
|
| 644 |
+
"language_model.model.layers.8.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 645 |
+
"language_model.model.layers.8.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 646 |
+
"language_model.model.layers.8.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 647 |
+
"language_model.model.layers.8.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 648 |
+
"language_model.model.layers.8.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 649 |
+
"language_model.model.layers.8.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 650 |
+
"language_model.model.layers.8.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 651 |
+
"language_model.model.layers.8.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 652 |
+
"language_model.model.layers.8.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 653 |
+
"language_model.model.layers.8.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 654 |
+
"language_model.model.layers.8.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 655 |
+
"language_model.model.layers.8.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 656 |
+
"language_model.model.layers.8.linear_attn.norm.weight": "model.safetensors",
|
| 657 |
+
"language_model.model.layers.8.linear_attn.out_proj.biases": "model.safetensors",
|
| 658 |
+
"language_model.model.layers.8.linear_attn.out_proj.scales": "model.safetensors",
|
| 659 |
+
"language_model.model.layers.8.linear_attn.out_proj.weight": "model.safetensors",
|
| 660 |
+
"language_model.model.layers.8.mlp.down_proj.biases": "model.safetensors",
|
| 661 |
+
"language_model.model.layers.8.mlp.down_proj.scales": "model.safetensors",
|
| 662 |
+
"language_model.model.layers.8.mlp.down_proj.weight": "model.safetensors",
|
| 663 |
+
"language_model.model.layers.8.mlp.gate_proj.biases": "model.safetensors",
|
| 664 |
+
"language_model.model.layers.8.mlp.gate_proj.scales": "model.safetensors",
|
| 665 |
+
"language_model.model.layers.8.mlp.gate_proj.weight": "model.safetensors",
|
| 666 |
+
"language_model.model.layers.8.mlp.up_proj.biases": "model.safetensors",
|
| 667 |
+
"language_model.model.layers.8.mlp.up_proj.scales": "model.safetensors",
|
| 668 |
+
"language_model.model.layers.8.mlp.up_proj.weight": "model.safetensors",
|
| 669 |
+
"language_model.model.layers.8.post_attention_layernorm.weight": "model.safetensors",
|
| 670 |
+
"language_model.model.layers.9.input_layernorm.weight": "model.safetensors",
|
| 671 |
+
"language_model.model.layers.9.linear_attn.A_log": "model.safetensors",
|
| 672 |
+
"language_model.model.layers.9.linear_attn.conv1d.weight": "model.safetensors",
|
| 673 |
+
"language_model.model.layers.9.linear_attn.dt_bias": "model.safetensors",
|
| 674 |
+
"language_model.model.layers.9.linear_attn.in_proj_a.biases": "model.safetensors",
|
| 675 |
+
"language_model.model.layers.9.linear_attn.in_proj_a.scales": "model.safetensors",
|
| 676 |
+
"language_model.model.layers.9.linear_attn.in_proj_a.weight": "model.safetensors",
|
| 677 |
+
"language_model.model.layers.9.linear_attn.in_proj_b.biases": "model.safetensors",
|
| 678 |
+
"language_model.model.layers.9.linear_attn.in_proj_b.scales": "model.safetensors",
|
| 679 |
+
"language_model.model.layers.9.linear_attn.in_proj_b.weight": "model.safetensors",
|
| 680 |
+
"language_model.model.layers.9.linear_attn.in_proj_qkv.biases": "model.safetensors",
|
| 681 |
+
"language_model.model.layers.9.linear_attn.in_proj_qkv.scales": "model.safetensors",
|
| 682 |
+
"language_model.model.layers.9.linear_attn.in_proj_qkv.weight": "model.safetensors",
|
| 683 |
+
"language_model.model.layers.9.linear_attn.in_proj_z.biases": "model.safetensors",
|
| 684 |
+
"language_model.model.layers.9.linear_attn.in_proj_z.scales": "model.safetensors",
|
| 685 |
+
"language_model.model.layers.9.linear_attn.in_proj_z.weight": "model.safetensors",
|
| 686 |
+
"language_model.model.layers.9.linear_attn.norm.weight": "model.safetensors",
|
| 687 |
+
"language_model.model.layers.9.linear_attn.out_proj.biases": "model.safetensors",
|
| 688 |
+
"language_model.model.layers.9.linear_attn.out_proj.scales": "model.safetensors",
|
| 689 |
+
"language_model.model.layers.9.linear_attn.out_proj.weight": "model.safetensors",
|
| 690 |
+
"language_model.model.layers.9.mlp.down_proj.biases": "model.safetensors",
|
| 691 |
+
"language_model.model.layers.9.mlp.down_proj.scales": "model.safetensors",
|
| 692 |
+
"language_model.model.layers.9.mlp.down_proj.weight": "model.safetensors",
|
| 693 |
+
"language_model.model.layers.9.mlp.gate_proj.biases": "model.safetensors",
|
| 694 |
+
"language_model.model.layers.9.mlp.gate_proj.scales": "model.safetensors",
|
| 695 |
+
"language_model.model.layers.9.mlp.gate_proj.weight": "model.safetensors",
|
| 696 |
+
"language_model.model.layers.9.mlp.up_proj.biases": "model.safetensors",
|
| 697 |
+
"language_model.model.layers.9.mlp.up_proj.scales": "model.safetensors",
|
| 698 |
+
"language_model.model.layers.9.mlp.up_proj.weight": "model.safetensors",
|
| 699 |
+
"language_model.model.layers.9.post_attention_layernorm.weight": "model.safetensors",
|
| 700 |
+
"language_model.model.norm.weight": "model.safetensors",
|
| 701 |
+
"vision_tower.pos_embed.weight": "optiq/optiq_vision.safetensors",
|
| 702 |
+
"vision_tower.patch_embed.proj.weight": "optiq/optiq_vision.safetensors",
|
| 703 |
+
"vision_tower.patch_embed.proj.bias": "optiq/optiq_vision.safetensors",
|
| 704 |
+
"vision_tower.merger.norm.bias": "optiq/optiq_vision.safetensors",
|
| 705 |
+
"vision_tower.merger.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 706 |
+
"vision_tower.blocks.9.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 707 |
+
"vision_tower.blocks.9.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 708 |
+
"vision_tower.blocks.9.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 709 |
+
"vision_tower.blocks.9.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 710 |
+
"vision_tower.blocks.8.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 711 |
+
"vision_tower.blocks.8.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 712 |
+
"vision_tower.blocks.8.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 713 |
+
"vision_tower.blocks.8.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 714 |
+
"vision_tower.blocks.8.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 715 |
+
"vision_tower.blocks.8.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 716 |
+
"vision_tower.blocks.7.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 717 |
+
"vision_tower.blocks.7.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 718 |
+
"vision_tower.blocks.7.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 719 |
+
"vision_tower.blocks.7.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 720 |
+
"vision_tower.blocks.7.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 721 |
+
"vision_tower.blocks.7.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 722 |
+
"vision_tower.blocks.7.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 723 |
+
"vision_tower.blocks.6.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 724 |
+
"vision_tower.blocks.6.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 725 |
+
"vision_tower.blocks.6.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 726 |
+
"vision_tower.blocks.6.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 727 |
+
"vision_tower.blocks.6.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 728 |
+
"vision_tower.blocks.5.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 729 |
+
"vision_tower.blocks.6.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 730 |
+
"vision_tower.blocks.5.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 731 |
+
"vision_tower.blocks.5.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 732 |
+
"vision_tower.blocks.5.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 733 |
+
"vision_tower.blocks.5.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 734 |
+
"vision_tower.blocks.5.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 735 |
+
"vision_tower.blocks.5.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 736 |
+
"vision_tower.blocks.5.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 737 |
+
"vision_tower.blocks.5.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 738 |
+
"vision_tower.blocks.4.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 739 |
+
"vision_tower.blocks.4.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 740 |
+
"vision_tower.blocks.4.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 741 |
+
"vision_tower.blocks.4.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 742 |
+
"vision_tower.blocks.4.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 743 |
+
"vision_tower.blocks.4.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 744 |
+
"vision_tower.blocks.6.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 745 |
+
"vision_tower.blocks.3.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 746 |
+
"vision_tower.blocks.3.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 747 |
+
"vision_tower.blocks.3.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 748 |
+
"vision_tower.blocks.3.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 749 |
+
"vision_tower.blocks.3.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 750 |
+
"vision_tower.blocks.3.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 751 |
+
"vision_tower.blocks.3.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 752 |
+
"vision_tower.blocks.3.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 753 |
+
"vision_tower.blocks.23.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 754 |
+
"vision_tower.blocks.23.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 755 |
+
"vision_tower.blocks.23.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 756 |
+
"vision_tower.blocks.23.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 757 |
+
"vision_tower.blocks.23.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 758 |
+
"vision_tower.blocks.23.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 759 |
+
"vision_tower.blocks.22.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 760 |
+
"vision_tower.blocks.22.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 761 |
+
"vision_tower.blocks.22.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 762 |
+
"vision_tower.blocks.22.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 763 |
+
"vision_tower.blocks.22.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 764 |
+
"vision_tower.blocks.22.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 765 |
+
"vision_tower.blocks.22.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 766 |
+
"vision_tower.blocks.21.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 767 |
+
"vision_tower.blocks.21.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 768 |
+
"vision_tower.blocks.21.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 769 |
+
"vision_tower.blocks.21.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 770 |
+
"vision_tower.blocks.21.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 771 |
+
"vision_tower.blocks.8.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 772 |
+
"vision_tower.blocks.21.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 773 |
+
"vision_tower.blocks.20.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 774 |
+
"vision_tower.blocks.20.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 775 |
+
"vision_tower.blocks.20.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 776 |
+
"vision_tower.blocks.20.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 777 |
+
"vision_tower.blocks.20.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 778 |
+
"vision_tower.blocks.20.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 779 |
+
"vision_tower.blocks.7.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 780 |
+
"vision_tower.blocks.20.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 781 |
+
"vision_tower.blocks.2.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 782 |
+
"vision_tower.blocks.3.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 783 |
+
"vision_tower.blocks.2.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 784 |
+
"vision_tower.blocks.2.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 785 |
+
"vision_tower.blocks.2.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 786 |
+
"vision_tower.blocks.21.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 787 |
+
"vision_tower.blocks.2.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 788 |
+
"vision_tower.blocks.2.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 789 |
+
"vision_tower.blocks.2.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 790 |
+
"vision_tower.blocks.2.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 791 |
+
"vision_tower.blocks.19.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 792 |
+
"vision_tower.blocks.19.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 793 |
+
"vision_tower.blocks.19.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 794 |
+
"vision_tower.merger.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 795 |
+
"vision_tower.merger.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 796 |
+
"vision_tower.blocks.18.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 797 |
+
"vision_tower.blocks.18.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 798 |
+
"vision_tower.blocks.18.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 799 |
+
"vision_tower.blocks.18.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 800 |
+
"vision_tower.blocks.18.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 801 |
+
"vision_tower.blocks.18.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 802 |
+
"vision_tower.blocks.18.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 803 |
+
"vision_tower.blocks.18.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 804 |
+
"vision_tower.blocks.18.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 805 |
+
"vision_tower.blocks.17.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 806 |
+
"vision_tower.blocks.22.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 807 |
+
"vision_tower.blocks.17.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 808 |
+
"vision_tower.blocks.18.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 809 |
+
"vision_tower.blocks.17.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 810 |
+
"vision_tower.blocks.17.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 811 |
+
"vision_tower.blocks.17.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 812 |
+
"vision_tower.blocks.17.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 813 |
+
"vision_tower.blocks.17.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 814 |
+
"vision_tower.blocks.16.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 815 |
+
"vision_tower.blocks.16.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 816 |
+
"vision_tower.blocks.16.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 817 |
+
"vision_tower.blocks.19.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 818 |
+
"vision_tower.blocks.16.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 819 |
+
"vision_tower.blocks.9.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 820 |
+
"vision_tower.blocks.16.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 821 |
+
"vision_tower.blocks.16.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 822 |
+
"vision_tower.blocks.15.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 823 |
+
"vision_tower.blocks.15.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 824 |
+
"vision_tower.blocks.15.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 825 |
+
"vision_tower.blocks.15.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 826 |
+
"vision_tower.blocks.15.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 827 |
+
"vision_tower.blocks.15.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 828 |
+
"vision_tower.blocks.15.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 829 |
+
"vision_tower.blocks.14.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 830 |
+
"vision_tower.blocks.23.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 831 |
+
"vision_tower.blocks.14.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 832 |
+
"vision_tower.blocks.14.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 833 |
+
"vision_tower.blocks.14.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 834 |
+
"vision_tower.blocks.14.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 835 |
+
"vision_tower.blocks.14.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 836 |
+
"vision_tower.blocks.14.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 837 |
+
"vision_tower.blocks.11.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 838 |
+
"vision_tower.blocks.14.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 839 |
+
"vision_tower.blocks.19.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 840 |
+
"vision_tower.blocks.16.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 841 |
+
"vision_tower.blocks.14.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 842 |
+
"vision_tower.blocks.13.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 843 |
+
"vision_tower.blocks.4.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 844 |
+
"vision_tower.blocks.13.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 845 |
+
"vision_tower.blocks.13.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 846 |
+
"vision_tower.blocks.12.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 847 |
+
"vision_tower.blocks.13.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 848 |
+
"vision_tower.blocks.13.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 849 |
+
"vision_tower.blocks.12.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 850 |
+
"vision_tower.blocks.12.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 851 |
+
"vision_tower.blocks.13.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 852 |
+
"vision_tower.blocks.22.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 853 |
+
"vision_tower.blocks.11.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 854 |
+
"vision_tower.blocks.11.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 855 |
+
"vision_tower.blocks.10.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 856 |
+
"vision_tower.blocks.4.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 857 |
+
"vision_tower.blocks.10.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 858 |
+
"vision_tower.blocks.10.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 859 |
+
"vision_tower.blocks.4.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 860 |
+
"vision_tower.blocks.10.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 861 |
+
"vision_tower.blocks.0.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 862 |
+
"vision_tower.blocks.10.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 863 |
+
"vision_tower.blocks.20.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 864 |
+
"vision_tower.blocks.13.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 865 |
+
"vision_tower.blocks.1.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 866 |
+
"vision_tower.blocks.1.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 867 |
+
"vision_tower.blocks.1.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 868 |
+
"vision_tower.blocks.0.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 869 |
+
"vision_tower.blocks.20.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 870 |
+
"vision_tower.blocks.0.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 871 |
+
"vision_tower.blocks.23.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 872 |
+
"vision_tower.blocks.0.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 873 |
+
"vision_tower.blocks.15.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 874 |
+
"vision_tower.blocks.17.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 875 |
+
"vision_tower.blocks.0.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 876 |
+
"vision_tower.blocks.11.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 877 |
+
"vision_tower.blocks.12.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 878 |
+
"vision_tower.blocks.20.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 879 |
+
"vision_tower.blocks.11.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 880 |
+
"vision_tower.blocks.8.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 881 |
+
"vision_tower.blocks.11.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 882 |
+
"vision_tower.blocks.1.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 883 |
+
"vision_tower.blocks.13.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 884 |
+
"vision_tower.blocks.10.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 885 |
+
"vision_tower.blocks.23.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 886 |
+
"vision_tower.blocks.2.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 887 |
+
"vision_tower.blocks.13.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 888 |
+
"vision_tower.blocks.0.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 889 |
+
"vision_tower.blocks.6.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 890 |
+
"vision_tower.blocks.15.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 891 |
+
"vision_tower.blocks.13.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 892 |
+
"vision_tower.blocks.11.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 893 |
+
"vision_tower.merger.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 894 |
+
"vision_tower.blocks.20.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 895 |
+
"vision_tower.blocks.10.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 896 |
+
"vision_tower.blocks.0.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 897 |
+
"vision_tower.blocks.8.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 898 |
+
"vision_tower.blocks.23.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 899 |
+
"vision_tower.blocks.19.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 900 |
+
"vision_tower.blocks.11.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 901 |
+
"vision_tower.blocks.1.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 902 |
+
"vision_tower.blocks.0.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 903 |
+
"vision_tower.blocks.22.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 904 |
+
"vision_tower.blocks.14.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 905 |
+
"vision_tower.blocks.9.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 906 |
+
"vision_tower.blocks.11.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 907 |
+
"vision_tower.blocks.14.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 908 |
+
"vision_tower.blocks.0.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 909 |
+
"vision_tower.blocks.16.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 910 |
+
"vision_tower.blocks.10.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 911 |
+
"vision_tower.blocks.2.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 912 |
+
"vision_tower.blocks.12.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 913 |
+
"vision_tower.blocks.3.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 914 |
+
"vision_tower.blocks.1.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 915 |
+
"vision_tower.blocks.6.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 916 |
+
"vision_tower.blocks.14.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 917 |
+
"vision_tower.blocks.19.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 918 |
+
"vision_tower.blocks.7.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 919 |
+
"vision_tower.blocks.2.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 920 |
+
"vision_tower.blocks.3.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 921 |
+
"vision_tower.blocks.11.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 922 |
+
"vision_tower.blocks.4.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 923 |
+
"vision_tower.blocks.11.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 924 |
+
"vision_tower.blocks.7.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 925 |
+
"vision_tower.blocks.20.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 926 |
+
"vision_tower.blocks.5.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 927 |
+
"vision_tower.blocks.17.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 928 |
+
"vision_tower.blocks.1.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 929 |
+
"vision_tower.blocks.18.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 930 |
+
"vision_tower.blocks.1.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 931 |
+
"vision_tower.blocks.3.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 932 |
+
"vision_tower.blocks.1.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 933 |
+
"vision_tower.blocks.2.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 934 |
+
"vision_tower.blocks.12.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 935 |
+
"vision_tower.blocks.9.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 936 |
+
"vision_tower.blocks.5.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 937 |
+
"vision_tower.blocks.16.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 938 |
+
"vision_tower.merger.norm.weight": "optiq/optiq_vision.safetensors",
|
| 939 |
+
"vision_tower.blocks.23.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 940 |
+
"vision_tower.blocks.19.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 941 |
+
"vision_tower.blocks.18.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 942 |
+
"vision_tower.blocks.12.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 943 |
+
"vision_tower.blocks.12.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 944 |
+
"vision_tower.blocks.13.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 945 |
+
"vision_tower.blocks.23.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 946 |
+
"vision_tower.blocks.7.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 947 |
+
"vision_tower.blocks.0.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 948 |
+
"vision_tower.blocks.8.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 949 |
+
"vision_tower.blocks.12.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 950 |
+
"vision_tower.blocks.19.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 951 |
+
"vision_tower.blocks.4.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 952 |
+
"vision_tower.blocks.10.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 953 |
+
"vision_tower.blocks.9.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 954 |
+
"vision_tower.blocks.7.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 955 |
+
"vision_tower.blocks.6.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 956 |
+
"vision_tower.blocks.12.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 957 |
+
"vision_tower.blocks.16.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 958 |
+
"vision_tower.blocks.9.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 959 |
+
"vision_tower.blocks.8.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 960 |
+
"vision_tower.blocks.17.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 961 |
+
"vision_tower.blocks.0.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 962 |
+
"vision_tower.blocks.21.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 963 |
+
"vision_tower.blocks.21.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 964 |
+
"vision_tower.blocks.15.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 965 |
+
"vision_tower.blocks.1.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 966 |
+
"vision_tower.blocks.9.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 967 |
+
"vision_tower.blocks.22.norm1.bias": "optiq/optiq_vision.safetensors",
|
| 968 |
+
"vision_tower.blocks.21.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 969 |
+
"vision_tower.blocks.15.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 970 |
+
"vision_tower.blocks.10.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 971 |
+
"vision_tower.blocks.1.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 972 |
+
"vision_tower.blocks.9.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 973 |
+
"vision_tower.blocks.6.attn.qkv.bias": "optiq/optiq_vision.safetensors",
|
| 974 |
+
"vision_tower.blocks.12.attn.proj.bias": "optiq/optiq_vision.safetensors",
|
| 975 |
+
"vision_tower.blocks.0.norm2.bias": "optiq/optiq_vision.safetensors",
|
| 976 |
+
"vision_tower.blocks.5.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 977 |
+
"vision_tower.blocks.16.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 978 |
+
"vision_tower.blocks.13.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 979 |
+
"vision_tower.blocks.17.norm1.weight": "optiq/optiq_vision.safetensors",
|
| 980 |
+
"vision_tower.blocks.1.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 981 |
+
"vision_tower.blocks.9.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 982 |
+
"vision_tower.blocks.17.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 983 |
+
"vision_tower.blocks.21.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 984 |
+
"vision_tower.blocks.19.mlp.linear_fc1.bias": "optiq/optiq_vision.safetensors",
|
| 985 |
+
"vision_tower.blocks.8.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 986 |
+
"vision_tower.blocks.22.mlp.linear_fc2.bias": "optiq/optiq_vision.safetensors",
|
| 987 |
+
"vision_tower.blocks.19.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 988 |
+
"vision_tower.blocks.6.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 989 |
+
"vision_tower.blocks.10.mlp.linear_fc1.weight": "optiq/optiq_vision.safetensors",
|
| 990 |
+
"vision_tower.blocks.21.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 991 |
+
"vision_tower.blocks.4.attn.proj.weight": "optiq/optiq_vision.safetensors",
|
| 992 |
+
"vision_tower.blocks.10.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 993 |
+
"vision_tower.blocks.15.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 994 |
+
"vision_tower.blocks.16.attn.qkv.weight": "optiq/optiq_vision.safetensors",
|
| 995 |
+
"vision_tower.blocks.12.norm2.weight": "optiq/optiq_vision.safetensors",
|
| 996 |
+
"vision_tower.blocks.11.mlp.linear_fc2.weight": "optiq/optiq_vision.safetensors",
|
| 997 |
+
"vision_tower.blocks.19.norm1.bias": "optiq/optiq_vision.safetensors"
|
| 998 |
+
}
|
| 999 |
+
}
|
optiq/metadata.json
ADDED
|
@@ -0,0 +1,756 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"method": "optiq_mixed_precision",
|
| 3 |
+
"base_model": "empero-ai/Qwen3.8-2B-Distill",
|
| 4 |
+
"reference": "bf16",
|
| 5 |
+
"target_bpw": 2.9,
|
| 6 |
+
"achieved_bpw": 2.901174448582068,
|
| 7 |
+
"n_high_bits": 16,
|
| 8 |
+
"n_low_bits": 47,
|
| 9 |
+
"threshold": 0.0,
|
| 10 |
+
"per_layer": {
|
| 11 |
+
"language_model.model.layers.23.mlp.up_proj": {
|
| 12 |
+
"bits": 6,
|
| 13 |
+
"group_size": 64
|
| 14 |
+
},
|
| 15 |
+
"language_model.model.layers.23.mlp.down_proj": {
|
| 16 |
+
"bits": 6,
|
| 17 |
+
"group_size": 64
|
| 18 |
+
},
|
| 19 |
+
"language_model.model.layers.23.mlp.gate_proj": {
|
| 20 |
+
"bits": 6,
|
| 21 |
+
"group_size": 64
|
| 22 |
+
},
|
| 23 |
+
"language_model.model.layers.23.self_attn.o_proj": {
|
| 24 |
+
"bits": 6,
|
| 25 |
+
"group_size": 64
|
| 26 |
+
},
|
| 27 |
+
"language_model.model.layers.23.self_attn.v_proj": {
|
| 28 |
+
"bits": 6,
|
| 29 |
+
"group_size": 64
|
| 30 |
+
},
|
| 31 |
+
"language_model.model.layers.23.self_attn.k_proj": {
|
| 32 |
+
"bits": 6,
|
| 33 |
+
"group_size": 64
|
| 34 |
+
},
|
| 35 |
+
"language_model.model.layers.23.self_attn.q_proj": {
|
| 36 |
+
"bits": 6,
|
| 37 |
+
"group_size": 64
|
| 38 |
+
},
|
| 39 |
+
"language_model.model.layers.22.mlp.up_proj": {
|
| 40 |
+
"bits": 2,
|
| 41 |
+
"group_size": 64
|
| 42 |
+
},
|
| 43 |
+
"language_model.model.layers.22.mlp.down_proj": {
|
| 44 |
+
"bits": 2,
|
| 45 |
+
"group_size": 64
|
| 46 |
+
},
|
| 47 |
+
"language_model.model.layers.22.mlp.gate_proj": {
|
| 48 |
+
"bits": 2,
|
| 49 |
+
"group_size": 64
|
| 50 |
+
},
|
| 51 |
+
"language_model.model.layers.22.linear_attn.out_proj": {
|
| 52 |
+
"bits": 3,
|
| 53 |
+
"group_size": 64
|
| 54 |
+
},
|
| 55 |
+
"language_model.model.layers.22.linear_attn.in_proj_a": {
|
| 56 |
+
"bits": 4,
|
| 57 |
+
"group_size": 64
|
| 58 |
+
},
|
| 59 |
+
"language_model.model.layers.22.linear_attn.in_proj_b": {
|
| 60 |
+
"bits": 5,
|
| 61 |
+
"group_size": 64
|
| 62 |
+
},
|
| 63 |
+
"language_model.model.layers.22.linear_attn.in_proj_z": {
|
| 64 |
+
"bits": 2,
|
| 65 |
+
"group_size": 64
|
| 66 |
+
},
|
| 67 |
+
"language_model.model.layers.22.linear_attn.in_proj_qkv": {
|
| 68 |
+
"bits": 3,
|
| 69 |
+
"group_size": 64
|
| 70 |
+
},
|
| 71 |
+
"language_model.model.layers.21.mlp.up_proj": {
|
| 72 |
+
"bits": 2,
|
| 73 |
+
"group_size": 64
|
| 74 |
+
},
|
| 75 |
+
"language_model.model.layers.21.mlp.down_proj": {
|
| 76 |
+
"bits": 2,
|
| 77 |
+
"group_size": 64
|
| 78 |
+
},
|
| 79 |
+
"language_model.model.layers.21.mlp.gate_proj": {
|
| 80 |
+
"bits": 2,
|
| 81 |
+
"group_size": 64
|
| 82 |
+
},
|
| 83 |
+
"language_model.model.layers.21.linear_attn.out_proj": {
|
| 84 |
+
"bits": 3,
|
| 85 |
+
"group_size": 64
|
| 86 |
+
},
|
| 87 |
+
"language_model.model.layers.21.linear_attn.in_proj_a": {
|
| 88 |
+
"bits": 3,
|
| 89 |
+
"group_size": 64
|
| 90 |
+
},
|
| 91 |
+
"language_model.model.layers.21.linear_attn.in_proj_b": {
|
| 92 |
+
"bits": 5,
|
| 93 |
+
"group_size": 64
|
| 94 |
+
},
|
| 95 |
+
"language_model.model.layers.21.linear_attn.in_proj_z": {
|
| 96 |
+
"bits": 2,
|
| 97 |
+
"group_size": 64
|
| 98 |
+
},
|
| 99 |
+
"language_model.model.layers.21.linear_attn.in_proj_qkv": {
|
| 100 |
+
"bits": 3,
|
| 101 |
+
"group_size": 64
|
| 102 |
+
},
|
| 103 |
+
"language_model.model.layers.20.mlp.up_proj": {
|
| 104 |
+
"bits": 2,
|
| 105 |
+
"group_size": 64
|
| 106 |
+
},
|
| 107 |
+
"language_model.model.layers.20.mlp.down_proj": {
|
| 108 |
+
"bits": 2,
|
| 109 |
+
"group_size": 64
|
| 110 |
+
},
|
| 111 |
+
"language_model.model.layers.20.mlp.gate_proj": {
|
| 112 |
+
"bits": 2,
|
| 113 |
+
"group_size": 64
|
| 114 |
+
},
|
| 115 |
+
"language_model.model.layers.20.linear_attn.out_proj": {
|
| 116 |
+
"bits": 3,
|
| 117 |
+
"group_size": 64
|
| 118 |
+
},
|
| 119 |
+
"language_model.model.layers.20.linear_attn.in_proj_a": {
|
| 120 |
+
"bits": 4,
|
| 121 |
+
"group_size": 64
|
| 122 |
+
},
|
| 123 |
+
"language_model.model.layers.20.linear_attn.in_proj_b": {
|
| 124 |
+
"bits": 4,
|
| 125 |
+
"group_size": 64
|
| 126 |
+
},
|
| 127 |
+
"language_model.model.layers.20.linear_attn.in_proj_z": {
|
| 128 |
+
"bits": 2,
|
| 129 |
+
"group_size": 64
|
| 130 |
+
},
|
| 131 |
+
"language_model.model.layers.20.linear_attn.in_proj_qkv": {
|
| 132 |
+
"bits": 3,
|
| 133 |
+
"group_size": 64
|
| 134 |
+
},
|
| 135 |
+
"language_model.model.layers.19.mlp.up_proj": {
|
| 136 |
+
"bits": 2,
|
| 137 |
+
"group_size": 64
|
| 138 |
+
},
|
| 139 |
+
"language_model.model.layers.19.mlp.down_proj": {
|
| 140 |
+
"bits": 2,
|
| 141 |
+
"group_size": 64
|
| 142 |
+
},
|
| 143 |
+
"language_model.model.layers.19.mlp.gate_proj": {
|
| 144 |
+
"bits": 2,
|
| 145 |
+
"group_size": 64
|
| 146 |
+
},
|
| 147 |
+
"language_model.model.layers.19.self_attn.o_proj": {
|
| 148 |
+
"bits": 3,
|
| 149 |
+
"group_size": 64
|
| 150 |
+
},
|
| 151 |
+
"language_model.model.layers.19.self_attn.v_proj": {
|
| 152 |
+
"bits": 4,
|
| 153 |
+
"group_size": 64
|
| 154 |
+
},
|
| 155 |
+
"language_model.model.layers.19.self_attn.k_proj": {
|
| 156 |
+
"bits": 3,
|
| 157 |
+
"group_size": 64
|
| 158 |
+
},
|
| 159 |
+
"language_model.model.layers.19.self_attn.q_proj": {
|
| 160 |
+
"bits": 2,
|
| 161 |
+
"group_size": 64
|
| 162 |
+
},
|
| 163 |
+
"language_model.model.layers.18.mlp.up_proj": {
|
| 164 |
+
"bits": 2,
|
| 165 |
+
"group_size": 64
|
| 166 |
+
},
|
| 167 |
+
"language_model.model.layers.18.mlp.down_proj": {
|
| 168 |
+
"bits": 2,
|
| 169 |
+
"group_size": 64
|
| 170 |
+
},
|
| 171 |
+
"language_model.model.layers.18.mlp.gate_proj": {
|
| 172 |
+
"bits": 2,
|
| 173 |
+
"group_size": 64
|
| 174 |
+
},
|
| 175 |
+
"language_model.model.layers.18.linear_attn.out_proj": {
|
| 176 |
+
"bits": 3,
|
| 177 |
+
"group_size": 64
|
| 178 |
+
},
|
| 179 |
+
"language_model.model.layers.18.linear_attn.in_proj_a": {
|
| 180 |
+
"bits": 3,
|
| 181 |
+
"group_size": 64
|
| 182 |
+
},
|
| 183 |
+
"language_model.model.layers.18.linear_attn.in_proj_b": {
|
| 184 |
+
"bits": 5,
|
| 185 |
+
"group_size": 64
|
| 186 |
+
},
|
| 187 |
+
"language_model.model.layers.18.linear_attn.in_proj_z": {
|
| 188 |
+
"bits": 2,
|
| 189 |
+
"group_size": 64
|
| 190 |
+
},
|
| 191 |
+
"language_model.model.layers.18.linear_attn.in_proj_qkv": {
|
| 192 |
+
"bits": 3,
|
| 193 |
+
"group_size": 64
|
| 194 |
+
},
|
| 195 |
+
"language_model.model.layers.17.mlp.up_proj": {
|
| 196 |
+
"bits": 2,
|
| 197 |
+
"group_size": 64
|
| 198 |
+
},
|
| 199 |
+
"language_model.model.layers.17.mlp.down_proj": {
|
| 200 |
+
"bits": 2,
|
| 201 |
+
"group_size": 64
|
| 202 |
+
},
|
| 203 |
+
"language_model.model.layers.17.mlp.gate_proj": {
|
| 204 |
+
"bits": 2,
|
| 205 |
+
"group_size": 64
|
| 206 |
+
},
|
| 207 |
+
"language_model.model.layers.17.linear_attn.out_proj": {
|
| 208 |
+
"bits": 3,
|
| 209 |
+
"group_size": 64
|
| 210 |
+
},
|
| 211 |
+
"language_model.model.layers.17.linear_attn.in_proj_a": {
|
| 212 |
+
"bits": 4,
|
| 213 |
+
"group_size": 64
|
| 214 |
+
},
|
| 215 |
+
"language_model.model.layers.17.linear_attn.in_proj_b": {
|
| 216 |
+
"bits": 5,
|
| 217 |
+
"group_size": 64
|
| 218 |
+
},
|
| 219 |
+
"language_model.model.layers.17.linear_attn.in_proj_z": {
|
| 220 |
+
"bits": 2,
|
| 221 |
+
"group_size": 64
|
| 222 |
+
},
|
| 223 |
+
"language_model.model.layers.17.linear_attn.in_proj_qkv": {
|
| 224 |
+
"bits": 3,
|
| 225 |
+
"group_size": 64
|
| 226 |
+
},
|
| 227 |
+
"language_model.model.layers.16.mlp.up_proj": {
|
| 228 |
+
"bits": 2,
|
| 229 |
+
"group_size": 64
|
| 230 |
+
},
|
| 231 |
+
"language_model.model.layers.16.mlp.down_proj": {
|
| 232 |
+
"bits": 3,
|
| 233 |
+
"group_size": 64
|
| 234 |
+
},
|
| 235 |
+
"language_model.model.layers.16.mlp.gate_proj": {
|
| 236 |
+
"bits": 2,
|
| 237 |
+
"group_size": 64
|
| 238 |
+
},
|
| 239 |
+
"language_model.model.layers.16.linear_attn.out_proj": {
|
| 240 |
+
"bits": 3,
|
| 241 |
+
"group_size": 64
|
| 242 |
+
},
|
| 243 |
+
"language_model.model.layers.16.linear_attn.in_proj_a": {
|
| 244 |
+
"bits": 4,
|
| 245 |
+
"group_size": 64
|
| 246 |
+
},
|
| 247 |
+
"language_model.model.layers.16.linear_attn.in_proj_b": {
|
| 248 |
+
"bits": 5,
|
| 249 |
+
"group_size": 64
|
| 250 |
+
},
|
| 251 |
+
"language_model.model.layers.16.linear_attn.in_proj_z": {
|
| 252 |
+
"bits": 2,
|
| 253 |
+
"group_size": 64
|
| 254 |
+
},
|
| 255 |
+
"language_model.model.layers.16.linear_attn.in_proj_qkv": {
|
| 256 |
+
"bits": 3,
|
| 257 |
+
"group_size": 64
|
| 258 |
+
},
|
| 259 |
+
"language_model.model.layers.15.mlp.up_proj": {
|
| 260 |
+
"bits": 3,
|
| 261 |
+
"group_size": 64
|
| 262 |
+
},
|
| 263 |
+
"language_model.model.layers.15.mlp.down_proj": {
|
| 264 |
+
"bits": 3,
|
| 265 |
+
"group_size": 64
|
| 266 |
+
},
|
| 267 |
+
"language_model.model.layers.15.mlp.gate_proj": {
|
| 268 |
+
"bits": 2,
|
| 269 |
+
"group_size": 64
|
| 270 |
+
},
|
| 271 |
+
"language_model.model.layers.15.self_attn.o_proj": {
|
| 272 |
+
"bits": 4,
|
| 273 |
+
"group_size": 64
|
| 274 |
+
},
|
| 275 |
+
"language_model.model.layers.15.self_attn.v_proj": {
|
| 276 |
+
"bits": 4,
|
| 277 |
+
"group_size": 64
|
| 278 |
+
},
|
| 279 |
+
"language_model.model.layers.15.self_attn.k_proj": {
|
| 280 |
+
"bits": 3,
|
| 281 |
+
"group_size": 64
|
| 282 |
+
},
|
| 283 |
+
"language_model.model.layers.15.self_attn.q_proj": {
|
| 284 |
+
"bits": 2,
|
| 285 |
+
"group_size": 64
|
| 286 |
+
},
|
| 287 |
+
"language_model.model.layers.14.mlp.up_proj": {
|
| 288 |
+
"bits": 3,
|
| 289 |
+
"group_size": 64
|
| 290 |
+
},
|
| 291 |
+
"language_model.model.layers.14.mlp.down_proj": {
|
| 292 |
+
"bits": 3,
|
| 293 |
+
"group_size": 64
|
| 294 |
+
},
|
| 295 |
+
"language_model.model.layers.14.mlp.gate_proj": {
|
| 296 |
+
"bits": 2,
|
| 297 |
+
"group_size": 64
|
| 298 |
+
},
|
| 299 |
+
"language_model.model.layers.14.linear_attn.out_proj": {
|
| 300 |
+
"bits": 3,
|
| 301 |
+
"group_size": 64
|
| 302 |
+
},
|
| 303 |
+
"language_model.model.layers.14.linear_attn.in_proj_a": {
|
| 304 |
+
"bits": 4,
|
| 305 |
+
"group_size": 64
|
| 306 |
+
},
|
| 307 |
+
"language_model.model.layers.14.linear_attn.in_proj_b": {
|
| 308 |
+
"bits": 4,
|
| 309 |
+
"group_size": 64
|
| 310 |
+
},
|
| 311 |
+
"language_model.model.layers.14.linear_attn.in_proj_z": {
|
| 312 |
+
"bits": 2,
|
| 313 |
+
"group_size": 64
|
| 314 |
+
},
|
| 315 |
+
"language_model.model.layers.14.linear_attn.in_proj_qkv": {
|
| 316 |
+
"bits": 3,
|
| 317 |
+
"group_size": 64
|
| 318 |
+
},
|
| 319 |
+
"language_model.model.layers.13.mlp.up_proj": {
|
| 320 |
+
"bits": 2,
|
| 321 |
+
"group_size": 64
|
| 322 |
+
},
|
| 323 |
+
"language_model.model.layers.13.mlp.down_proj": {
|
| 324 |
+
"bits": 2,
|
| 325 |
+
"group_size": 64
|
| 326 |
+
},
|
| 327 |
+
"language_model.model.layers.13.mlp.gate_proj": {
|
| 328 |
+
"bits": 2,
|
| 329 |
+
"group_size": 64
|
| 330 |
+
},
|
| 331 |
+
"language_model.model.layers.13.linear_attn.out_proj": {
|
| 332 |
+
"bits": 3,
|
| 333 |
+
"group_size": 64
|
| 334 |
+
},
|
| 335 |
+
"language_model.model.layers.13.linear_attn.in_proj_a": {
|
| 336 |
+
"bits": 4,
|
| 337 |
+
"group_size": 64
|
| 338 |
+
},
|
| 339 |
+
"language_model.model.layers.13.linear_attn.in_proj_b": {
|
| 340 |
+
"bits": 5,
|
| 341 |
+
"group_size": 64
|
| 342 |
+
},
|
| 343 |
+
"language_model.model.layers.13.linear_attn.in_proj_z": {
|
| 344 |
+
"bits": 2,
|
| 345 |
+
"group_size": 64
|
| 346 |
+
},
|
| 347 |
+
"language_model.model.layers.13.linear_attn.in_proj_qkv": {
|
| 348 |
+
"bits": 3,
|
| 349 |
+
"group_size": 64
|
| 350 |
+
},
|
| 351 |
+
"language_model.model.layers.12.mlp.up_proj": {
|
| 352 |
+
"bits": 2,
|
| 353 |
+
"group_size": 64
|
| 354 |
+
},
|
| 355 |
+
"language_model.model.layers.12.mlp.down_proj": {
|
| 356 |
+
"bits": 2,
|
| 357 |
+
"group_size": 64
|
| 358 |
+
},
|
| 359 |
+
"language_model.model.layers.12.mlp.gate_proj": {
|
| 360 |
+
"bits": 2,
|
| 361 |
+
"group_size": 64
|
| 362 |
+
},
|
| 363 |
+
"language_model.model.layers.12.linear_attn.out_proj": {
|
| 364 |
+
"bits": 3,
|
| 365 |
+
"group_size": 64
|
| 366 |
+
},
|
| 367 |
+
"language_model.model.layers.12.linear_attn.in_proj_a": {
|
| 368 |
+
"bits": 4,
|
| 369 |
+
"group_size": 64
|
| 370 |
+
},
|
| 371 |
+
"language_model.model.layers.12.linear_attn.in_proj_b": {
|
| 372 |
+
"bits": 5,
|
| 373 |
+
"group_size": 64
|
| 374 |
+
},
|
| 375 |
+
"language_model.model.layers.12.linear_attn.in_proj_z": {
|
| 376 |
+
"bits": 3,
|
| 377 |
+
"group_size": 64
|
| 378 |
+
},
|
| 379 |
+
"language_model.model.layers.12.linear_attn.in_proj_qkv": {
|
| 380 |
+
"bits": 3,
|
| 381 |
+
"group_size": 64
|
| 382 |
+
},
|
| 383 |
+
"language_model.model.layers.11.mlp.up_proj": {
|
| 384 |
+
"bits": 3,
|
| 385 |
+
"group_size": 64
|
| 386 |
+
},
|
| 387 |
+
"language_model.model.layers.11.mlp.down_proj": {
|
| 388 |
+
"bits": 2,
|
| 389 |
+
"group_size": 64
|
| 390 |
+
},
|
| 391 |
+
"language_model.model.layers.11.mlp.gate_proj": {
|
| 392 |
+
"bits": 2,
|
| 393 |
+
"group_size": 64
|
| 394 |
+
},
|
| 395 |
+
"language_model.model.layers.11.self_attn.o_proj": {
|
| 396 |
+
"bits": 4,
|
| 397 |
+
"group_size": 64
|
| 398 |
+
},
|
| 399 |
+
"language_model.model.layers.11.self_attn.v_proj": {
|
| 400 |
+
"bits": 4,
|
| 401 |
+
"group_size": 64
|
| 402 |
+
},
|
| 403 |
+
"language_model.model.layers.11.self_attn.k_proj": {
|
| 404 |
+
"bits": 3,
|
| 405 |
+
"group_size": 64
|
| 406 |
+
},
|
| 407 |
+
"language_model.model.layers.11.self_attn.q_proj": {
|
| 408 |
+
"bits": 3,
|
| 409 |
+
"group_size": 64
|
| 410 |
+
},
|
| 411 |
+
"language_model.model.layers.10.mlp.up_proj": {
|
| 412 |
+
"bits": 3,
|
| 413 |
+
"group_size": 64
|
| 414 |
+
},
|
| 415 |
+
"language_model.model.layers.10.mlp.down_proj": {
|
| 416 |
+
"bits": 3,
|
| 417 |
+
"group_size": 64
|
| 418 |
+
},
|
| 419 |
+
"language_model.model.layers.10.mlp.gate_proj": {
|
| 420 |
+
"bits": 2,
|
| 421 |
+
"group_size": 64
|
| 422 |
+
},
|
| 423 |
+
"language_model.model.layers.10.linear_attn.out_proj": {
|
| 424 |
+
"bits": 3,
|
| 425 |
+
"group_size": 64
|
| 426 |
+
},
|
| 427 |
+
"language_model.model.layers.10.linear_attn.in_proj_a": {
|
| 428 |
+
"bits": 4,
|
| 429 |
+
"group_size": 64
|
| 430 |
+
},
|
| 431 |
+
"language_model.model.layers.10.linear_attn.in_proj_b": {
|
| 432 |
+
"bits": 4,
|
| 433 |
+
"group_size": 64
|
| 434 |
+
},
|
| 435 |
+
"language_model.model.layers.10.linear_attn.in_proj_z": {
|
| 436 |
+
"bits": 2,
|
| 437 |
+
"group_size": 64
|
| 438 |
+
},
|
| 439 |
+
"language_model.model.layers.10.linear_attn.in_proj_qkv": {
|
| 440 |
+
"bits": 3,
|
| 441 |
+
"group_size": 64
|
| 442 |
+
},
|
| 443 |
+
"language_model.model.layers.9.mlp.up_proj": {
|
| 444 |
+
"bits": 2,
|
| 445 |
+
"group_size": 64
|
| 446 |
+
},
|
| 447 |
+
"language_model.model.layers.9.mlp.down_proj": {
|
| 448 |
+
"bits": 2,
|
| 449 |
+
"group_size": 64
|
| 450 |
+
},
|
| 451 |
+
"language_model.model.layers.9.mlp.gate_proj": {
|
| 452 |
+
"bits": 2,
|
| 453 |
+
"group_size": 64
|
| 454 |
+
},
|
| 455 |
+
"language_model.model.layers.9.linear_attn.out_proj": {
|
| 456 |
+
"bits": 3,
|
| 457 |
+
"group_size": 64
|
| 458 |
+
},
|
| 459 |
+
"language_model.model.layers.9.linear_attn.in_proj_a": {
|
| 460 |
+
"bits": 4,
|
| 461 |
+
"group_size": 64
|
| 462 |
+
},
|
| 463 |
+
"language_model.model.layers.9.linear_attn.in_proj_b": {
|
| 464 |
+
"bits": 5,
|
| 465 |
+
"group_size": 64
|
| 466 |
+
},
|
| 467 |
+
"language_model.model.layers.9.linear_attn.in_proj_z": {
|
| 468 |
+
"bits": 3,
|
| 469 |
+
"group_size": 64
|
| 470 |
+
},
|
| 471 |
+
"language_model.model.layers.9.linear_attn.in_proj_qkv": {
|
| 472 |
+
"bits": 3,
|
| 473 |
+
"group_size": 64
|
| 474 |
+
},
|
| 475 |
+
"language_model.model.layers.8.mlp.up_proj": {
|
| 476 |
+
"bits": 3,
|
| 477 |
+
"group_size": 64
|
| 478 |
+
},
|
| 479 |
+
"language_model.model.layers.8.mlp.down_proj": {
|
| 480 |
+
"bits": 3,
|
| 481 |
+
"group_size": 64
|
| 482 |
+
},
|
| 483 |
+
"language_model.model.layers.8.mlp.gate_proj": {
|
| 484 |
+
"bits": 2,
|
| 485 |
+
"group_size": 64
|
| 486 |
+
},
|
| 487 |
+
"language_model.model.layers.8.linear_attn.out_proj": {
|
| 488 |
+
"bits": 3,
|
| 489 |
+
"group_size": 64
|
| 490 |
+
},
|
| 491 |
+
"language_model.model.layers.8.linear_attn.in_proj_a": {
|
| 492 |
+
"bits": 4,
|
| 493 |
+
"group_size": 64
|
| 494 |
+
},
|
| 495 |
+
"language_model.model.layers.8.linear_attn.in_proj_b": {
|
| 496 |
+
"bits": 4,
|
| 497 |
+
"group_size": 64
|
| 498 |
+
},
|
| 499 |
+
"language_model.model.layers.8.linear_attn.in_proj_z": {
|
| 500 |
+
"bits": 3,
|
| 501 |
+
"group_size": 64
|
| 502 |
+
},
|
| 503 |
+
"language_model.model.layers.8.linear_attn.in_proj_qkv": {
|
| 504 |
+
"bits": 3,
|
| 505 |
+
"group_size": 64
|
| 506 |
+
},
|
| 507 |
+
"language_model.model.layers.7.mlp.up_proj": {
|
| 508 |
+
"bits": 3,
|
| 509 |
+
"group_size": 64
|
| 510 |
+
},
|
| 511 |
+
"language_model.model.layers.7.mlp.down_proj": {
|
| 512 |
+
"bits": 3,
|
| 513 |
+
"group_size": 64
|
| 514 |
+
},
|
| 515 |
+
"language_model.model.layers.7.mlp.gate_proj": {
|
| 516 |
+
"bits": 2,
|
| 517 |
+
"group_size": 64
|
| 518 |
+
},
|
| 519 |
+
"language_model.model.layers.7.self_attn.o_proj": {
|
| 520 |
+
"bits": 4,
|
| 521 |
+
"group_size": 64
|
| 522 |
+
},
|
| 523 |
+
"language_model.model.layers.7.self_attn.v_proj": {
|
| 524 |
+
"bits": 5,
|
| 525 |
+
"group_size": 64
|
| 526 |
+
},
|
| 527 |
+
"language_model.model.layers.7.self_attn.k_proj": {
|
| 528 |
+
"bits": 3,
|
| 529 |
+
"group_size": 64
|
| 530 |
+
},
|
| 531 |
+
"language_model.model.layers.7.self_attn.q_proj": {
|
| 532 |
+
"bits": 3,
|
| 533 |
+
"group_size": 64
|
| 534 |
+
},
|
| 535 |
+
"language_model.model.layers.6.mlp.up_proj": {
|
| 536 |
+
"bits": 3,
|
| 537 |
+
"group_size": 64
|
| 538 |
+
},
|
| 539 |
+
"language_model.model.layers.6.mlp.down_proj": {
|
| 540 |
+
"bits": 3,
|
| 541 |
+
"group_size": 64
|
| 542 |
+
},
|
| 543 |
+
"language_model.model.layers.6.mlp.gate_proj": {
|
| 544 |
+
"bits": 3,
|
| 545 |
+
"group_size": 64
|
| 546 |
+
},
|
| 547 |
+
"language_model.model.layers.6.linear_attn.out_proj": {
|
| 548 |
+
"bits": 4,
|
| 549 |
+
"group_size": 64
|
| 550 |
+
},
|
| 551 |
+
"language_model.model.layers.6.linear_attn.in_proj_a": {
|
| 552 |
+
"bits": 4,
|
| 553 |
+
"group_size": 64
|
| 554 |
+
},
|
| 555 |
+
"language_model.model.layers.6.linear_attn.in_proj_b": {
|
| 556 |
+
"bits": 5,
|
| 557 |
+
"group_size": 64
|
| 558 |
+
},
|
| 559 |
+
"language_model.model.layers.6.linear_attn.in_proj_z": {
|
| 560 |
+
"bits": 3,
|
| 561 |
+
"group_size": 64
|
| 562 |
+
},
|
| 563 |
+
"language_model.model.layers.6.linear_attn.in_proj_qkv": {
|
| 564 |
+
"bits": 3,
|
| 565 |
+
"group_size": 64
|
| 566 |
+
},
|
| 567 |
+
"language_model.model.layers.5.mlp.up_proj": {
|
| 568 |
+
"bits": 3,
|
| 569 |
+
"group_size": 64
|
| 570 |
+
},
|
| 571 |
+
"language_model.model.layers.5.mlp.down_proj": {
|
| 572 |
+
"bits": 3,
|
| 573 |
+
"group_size": 64
|
| 574 |
+
},
|
| 575 |
+
"language_model.model.layers.5.mlp.gate_proj": {
|
| 576 |
+
"bits": 3,
|
| 577 |
+
"group_size": 64
|
| 578 |
+
},
|
| 579 |
+
"language_model.model.layers.5.linear_attn.out_proj": {
|
| 580 |
+
"bits": 3,
|
| 581 |
+
"group_size": 64
|
| 582 |
+
},
|
| 583 |
+
"language_model.model.layers.5.linear_attn.in_proj_a": {
|
| 584 |
+
"bits": 4,
|
| 585 |
+
"group_size": 64
|
| 586 |
+
},
|
| 587 |
+
"language_model.model.layers.5.linear_attn.in_proj_b": {
|
| 588 |
+
"bits": 5,
|
| 589 |
+
"group_size": 64
|
| 590 |
+
},
|
| 591 |
+
"language_model.model.layers.5.linear_attn.in_proj_z": {
|
| 592 |
+
"bits": 3,
|
| 593 |
+
"group_size": 64
|
| 594 |
+
},
|
| 595 |
+
"language_model.model.layers.5.linear_attn.in_proj_qkv": {
|
| 596 |
+
"bits": 3,
|
| 597 |
+
"group_size": 64
|
| 598 |
+
},
|
| 599 |
+
"language_model.model.layers.4.mlp.up_proj": {
|
| 600 |
+
"bits": 3,
|
| 601 |
+
"group_size": 64
|
| 602 |
+
},
|
| 603 |
+
"language_model.model.layers.4.mlp.down_proj": {
|
| 604 |
+
"bits": 3,
|
| 605 |
+
"group_size": 64
|
| 606 |
+
},
|
| 607 |
+
"language_model.model.layers.4.mlp.gate_proj": {
|
| 608 |
+
"bits": 3,
|
| 609 |
+
"group_size": 64
|
| 610 |
+
},
|
| 611 |
+
"language_model.model.layers.4.linear_attn.out_proj": {
|
| 612 |
+
"bits": 3,
|
| 613 |
+
"group_size": 64
|
| 614 |
+
},
|
| 615 |
+
"language_model.model.layers.4.linear_attn.in_proj_a": {
|
| 616 |
+
"bits": 4,
|
| 617 |
+
"group_size": 64
|
| 618 |
+
},
|
| 619 |
+
"language_model.model.layers.4.linear_attn.in_proj_b": {
|
| 620 |
+
"bits": 5,
|
| 621 |
+
"group_size": 64
|
| 622 |
+
},
|
| 623 |
+
"language_model.model.layers.4.linear_attn.in_proj_z": {
|
| 624 |
+
"bits": 3,
|
| 625 |
+
"group_size": 64
|
| 626 |
+
},
|
| 627 |
+
"language_model.model.layers.4.linear_attn.in_proj_qkv": {
|
| 628 |
+
"bits": 3,
|
| 629 |
+
"group_size": 64
|
| 630 |
+
},
|
| 631 |
+
"language_model.model.layers.3.mlp.up_proj": {
|
| 632 |
+
"bits": 3,
|
| 633 |
+
"group_size": 64
|
| 634 |
+
},
|
| 635 |
+
"language_model.model.layers.3.mlp.down_proj": {
|
| 636 |
+
"bits": 3,
|
| 637 |
+
"group_size": 64
|
| 638 |
+
},
|
| 639 |
+
"language_model.model.layers.3.mlp.gate_proj": {
|
| 640 |
+
"bits": 3,
|
| 641 |
+
"group_size": 64
|
| 642 |
+
},
|
| 643 |
+
"language_model.model.layers.3.self_attn.o_proj": {
|
| 644 |
+
"bits": 4,
|
| 645 |
+
"group_size": 64
|
| 646 |
+
},
|
| 647 |
+
"language_model.model.layers.3.self_attn.v_proj": {
|
| 648 |
+
"bits": 5,
|
| 649 |
+
"group_size": 64
|
| 650 |
+
},
|
| 651 |
+
"language_model.model.layers.3.self_attn.k_proj": {
|
| 652 |
+
"bits": 3,
|
| 653 |
+
"group_size": 64
|
| 654 |
+
},
|
| 655 |
+
"language_model.model.layers.3.self_attn.q_proj": {
|
| 656 |
+
"bits": 3,
|
| 657 |
+
"group_size": 64
|
| 658 |
+
},
|
| 659 |
+
"language_model.model.layers.2.mlp.up_proj": {
|
| 660 |
+
"bits": 3,
|
| 661 |
+
"group_size": 64
|
| 662 |
+
},
|
| 663 |
+
"language_model.model.layers.2.mlp.down_proj": {
|
| 664 |
+
"bits": 3,
|
| 665 |
+
"group_size": 64
|
| 666 |
+
},
|
| 667 |
+
"language_model.model.layers.2.mlp.gate_proj": {
|
| 668 |
+
"bits": 3,
|
| 669 |
+
"group_size": 64
|
| 670 |
+
},
|
| 671 |
+
"language_model.model.layers.2.linear_attn.out_proj": {
|
| 672 |
+
"bits": 4,
|
| 673 |
+
"group_size": 64
|
| 674 |
+
},
|
| 675 |
+
"language_model.model.layers.2.linear_attn.in_proj_a": {
|
| 676 |
+
"bits": 4,
|
| 677 |
+
"group_size": 64
|
| 678 |
+
},
|
| 679 |
+
"language_model.model.layers.2.linear_attn.in_proj_b": {
|
| 680 |
+
"bits": 5,
|
| 681 |
+
"group_size": 64
|
| 682 |
+
},
|
| 683 |
+
"language_model.model.layers.2.linear_attn.in_proj_z": {
|
| 684 |
+
"bits": 3,
|
| 685 |
+
"group_size": 64
|
| 686 |
+
},
|
| 687 |
+
"language_model.model.layers.2.linear_attn.in_proj_qkv": {
|
| 688 |
+
"bits": 3,
|
| 689 |
+
"group_size": 64
|
| 690 |
+
},
|
| 691 |
+
"language_model.model.layers.1.mlp.up_proj": {
|
| 692 |
+
"bits": 3,
|
| 693 |
+
"group_size": 64
|
| 694 |
+
},
|
| 695 |
+
"language_model.model.layers.1.mlp.down_proj": {
|
| 696 |
+
"bits": 3,
|
| 697 |
+
"group_size": 64
|
| 698 |
+
},
|
| 699 |
+
"language_model.model.layers.1.mlp.gate_proj": {
|
| 700 |
+
"bits": 3,
|
| 701 |
+
"group_size": 64
|
| 702 |
+
},
|
| 703 |
+
"language_model.model.layers.1.linear_attn.out_proj": {
|
| 704 |
+
"bits": 4,
|
| 705 |
+
"group_size": 64
|
| 706 |
+
},
|
| 707 |
+
"language_model.model.layers.1.linear_attn.in_proj_a": {
|
| 708 |
+
"bits": 6,
|
| 709 |
+
"group_size": 64
|
| 710 |
+
},
|
| 711 |
+
"language_model.model.layers.1.linear_attn.in_proj_b": {
|
| 712 |
+
"bits": 4,
|
| 713 |
+
"group_size": 64
|
| 714 |
+
},
|
| 715 |
+
"language_model.model.layers.1.linear_attn.in_proj_z": {
|
| 716 |
+
"bits": 3,
|
| 717 |
+
"group_size": 64
|
| 718 |
+
},
|
| 719 |
+
"language_model.model.layers.1.linear_attn.in_proj_qkv": {
|
| 720 |
+
"bits": 3,
|
| 721 |
+
"group_size": 64
|
| 722 |
+
},
|
| 723 |
+
"language_model.model.layers.0.mlp.up_proj": {
|
| 724 |
+
"bits": 6,
|
| 725 |
+
"group_size": 64
|
| 726 |
+
},
|
| 727 |
+
"language_model.model.layers.0.mlp.down_proj": {
|
| 728 |
+
"bits": 6,
|
| 729 |
+
"group_size": 64
|
| 730 |
+
},
|
| 731 |
+
"language_model.model.layers.0.mlp.gate_proj": {
|
| 732 |
+
"bits": 6,
|
| 733 |
+
"group_size": 64
|
| 734 |
+
},
|
| 735 |
+
"language_model.model.layers.0.linear_attn.out_proj": {
|
| 736 |
+
"bits": 6,
|
| 737 |
+
"group_size": 64
|
| 738 |
+
},
|
| 739 |
+
"language_model.model.layers.0.linear_attn.in_proj_a": {
|
| 740 |
+
"bits": 6,
|
| 741 |
+
"group_size": 64
|
| 742 |
+
},
|
| 743 |
+
"language_model.model.layers.0.linear_attn.in_proj_b": {
|
| 744 |
+
"bits": 6,
|
| 745 |
+
"group_size": 64
|
| 746 |
+
},
|
| 747 |
+
"language_model.model.layers.0.linear_attn.in_proj_z": {
|
| 748 |
+
"bits": 6,
|
| 749 |
+
"group_size": 64
|
| 750 |
+
},
|
| 751 |
+
"language_model.model.layers.0.linear_attn.in_proj_qkv": {
|
| 752 |
+
"bits": 6,
|
| 753 |
+
"group_size": 64
|
| 754 |
+
}
|
| 755 |
+
}
|
| 756 |
+
}
|
optiq/mtp.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:93cfb5bf5ecb39c894ef1ba6ee2637e1791ca0ede037e689f9ad68eeb6d98a5c
|
| 3 |
+
size 46292949
|
optiq/optiq_vision.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f58447c755767fa222dcc07d129e80373a23b3444b05c3ca5f1433b6bf1f7d4
|
| 3 |
+
size 662866147
|
optiq/sensitivity.json
ADDED
|
@@ -0,0 +1,2246 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_model": "empero-ai/Qwen3.8-2B-Distill",
|
| 3 |
+
"candidate_bits": [
|
| 4 |
+
2,
|
| 5 |
+
3,
|
| 6 |
+
4,
|
| 7 |
+
5,
|
| 8 |
+
6
|
| 9 |
+
],
|
| 10 |
+
"group_size": 64,
|
| 11 |
+
"metric": "kl_divergence_vs_reference",
|
| 12 |
+
"layers": [
|
| 13 |
+
{
|
| 14 |
+
"layer_name": "language_model.model.layers.23.mlp.up_proj",
|
| 15 |
+
"sensitivities": {
|
| 16 |
+
"2": 0.022889396796623867,
|
| 17 |
+
"3": 0.005864370134077035,
|
| 18 |
+
"4": 0.0016542644370929338,
|
| 19 |
+
"5": 0.000738907325285254,
|
| 20 |
+
"6": 0.0004634729915172405
|
| 21 |
+
},
|
| 22 |
+
"param_count": 12582912,
|
| 23 |
+
"metadata": {}
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"layer_name": "language_model.model.layers.23.mlp.down_proj",
|
| 27 |
+
"sensitivities": {
|
| 28 |
+
"2": 0.07524260692298412,
|
| 29 |
+
"3": 0.009423329504594827,
|
| 30 |
+
"4": 0.0024585577484685928,
|
| 31 |
+
"5": 0.001010144214281657,
|
| 32 |
+
"6": 0.0005051019118885355
|
| 33 |
+
},
|
| 34 |
+
"param_count": 12582912,
|
| 35 |
+
"metadata": {}
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"layer_name": "language_model.model.layers.23.mlp.gate_proj",
|
| 39 |
+
"sensitivities": {
|
| 40 |
+
"2": 0.01319051988927337,
|
| 41 |
+
"3": 0.002864282197454789,
|
| 42 |
+
"4": 0.0009605420612691281,
|
| 43 |
+
"5": 0.0004922099130150551,
|
| 44 |
+
"6": 0.000311081681199236
|
| 45 |
+
},
|
| 46 |
+
"param_count": 12582912,
|
| 47 |
+
"metadata": {}
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"layer_name": "language_model.model.layers.23.self_attn.o_proj",
|
| 51 |
+
"sensitivities": {
|
| 52 |
+
"2": 0.021424493830030162,
|
| 53 |
+
"3": 0.005750213972836112,
|
| 54 |
+
"4": 0.0016112725315906573,
|
| 55 |
+
"5": 0.000732784224965144,
|
| 56 |
+
"6": 0.0004464034039604788
|
| 57 |
+
},
|
| 58 |
+
"param_count": 4194304,
|
| 59 |
+
"metadata": {}
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"layer_name": "language_model.model.layers.23.self_attn.v_proj",
|
| 63 |
+
"sensitivities": {
|
| 64 |
+
"2": 0.004863989490938063,
|
| 65 |
+
"3": 0.0016981545923044905,
|
| 66 |
+
"4": 0.0008714322581605908,
|
| 67 |
+
"5": 0.00048651552545682836,
|
| 68 |
+
"6": 0.0003411814747475243
|
| 69 |
+
},
|
| 70 |
+
"param_count": 1048576,
|
| 71 |
+
"metadata": {}
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"layer_name": "language_model.model.layers.23.self_attn.k_proj",
|
| 75 |
+
"sensitivities": {
|
| 76 |
+
"2": 0.0015306950323671724,
|
| 77 |
+
"3": 0.0006018209405738162,
|
| 78 |
+
"4": 0.0003679009278130252,
|
| 79 |
+
"5": 0.00026836768544550676,
|
| 80 |
+
"6": 0.00023285415863938397
|
| 81 |
+
},
|
| 82 |
+
"param_count": 1048576,
|
| 83 |
+
"metadata": {}
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"layer_name": "language_model.model.layers.23.self_attn.q_proj",
|
| 87 |
+
"sensitivities": {
|
| 88 |
+
"2": 0.004132667784385073,
|
| 89 |
+
"3": 0.0012825351380646073,
|
| 90 |
+
"4": 0.0006327221829754611,
|
| 91 |
+
"5": 0.0004078486238843955,
|
| 92 |
+
"6": 0.0003100313927764849
|
| 93 |
+
},
|
| 94 |
+
"param_count": 8388608,
|
| 95 |
+
"metadata": {}
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"layer_name": "language_model.model.layers.22.mlp.up_proj",
|
| 99 |
+
"sensitivities": {
|
| 100 |
+
"2": 0.0072227599448524415,
|
| 101 |
+
"3": 0.002076016823896983,
|
| 102 |
+
"4": 0.0008285000791753797,
|
| 103 |
+
"5": 0.000498536613425434,
|
| 104 |
+
"6": 0.0003487009056091968
|
| 105 |
+
},
|
| 106 |
+
"param_count": 12582912,
|
| 107 |
+
"metadata": {}
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"layer_name": "language_model.model.layers.22.mlp.down_proj",
|
| 111 |
+
"sensitivities": {
|
| 112 |
+
"2": 0.007340312843249801,
|
| 113 |
+
"3": 0.002193007746730776,
|
| 114 |
+
"4": 0.0008596682164352387,
|
| 115 |
+
"5": 0.00048073946739653667,
|
| 116 |
+
"6": 0.00034690553517672623
|
| 117 |
+
},
|
| 118 |
+
"param_count": 12582912,
|
| 119 |
+
"metadata": {}
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"layer_name": "language_model.model.layers.22.mlp.gate_proj",
|
| 123 |
+
"sensitivities": {
|
| 124 |
+
"2": 0.0054251158580882475,
|
| 125 |
+
"3": 0.0014932176369863253,
|
| 126 |
+
"4": 0.0006634455057792366,
|
| 127 |
+
"5": 0.0004091161657318783,
|
| 128 |
+
"6": 0.0003211103297265557
|
| 129 |
+
},
|
| 130 |
+
"param_count": 12582912,
|
| 131 |
+
"metadata": {}
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"layer_name": "language_model.model.layers.22.linear_attn.out_proj",
|
| 135 |
+
"sensitivities": {
|
| 136 |
+
"2": 0.005855952826095745,
|
| 137 |
+
"3": 0.0016401179006303817,
|
| 138 |
+
"4": 0.0007672165096058355,
|
| 139 |
+
"5": 0.00044733987548776594,
|
| 140 |
+
"6": 0.0003548673924645603
|
| 141 |
+
},
|
| 142 |
+
"param_count": 4194304,
|
| 143 |
+
"metadata": {}
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"layer_name": "language_model.model.layers.22.linear_attn.in_proj_a",
|
| 147 |
+
"sensitivities": {
|
| 148 |
+
"2": 0.00040877657647797605,
|
| 149 |
+
"3": 0.00029402160665388993,
|
| 150 |
+
"4": 0.0002607208892489628,
|
| 151 |
+
"5": 0.00025171608012897195,
|
| 152 |
+
"6": 0.00024811304532098194
|
| 153 |
+
},
|
| 154 |
+
"param_count": 32768,
|
| 155 |
+
"metadata": {}
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"layer_name": "language_model.model.layers.22.linear_attn.in_proj_b",
|
| 159 |
+
"sensitivities": {
|
| 160 |
+
"2": 0.0005506053760958215,
|
| 161 |
+
"3": 0.00032314512463926803,
|
| 162 |
+
"4": 0.0002879587749096875,
|
| 163 |
+
"5": 0.000259493415645314,
|
| 164 |
+
"6": 0.0002492442405734134
|
| 165 |
+
},
|
| 166 |
+
"param_count": 32768,
|
| 167 |
+
"metadata": {}
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"layer_name": "language_model.model.layers.22.linear_attn.in_proj_z",
|
| 171 |
+
"sensitivities": {
|
| 172 |
+
"2": 0.0019737020714577134,
|
| 173 |
+
"3": 0.0009088053423814321,
|
| 174 |
+
"4": 0.0004970280545724867,
|
| 175 |
+
"5": 0.0003711752257610594,
|
| 176 |
+
"6": 0.00029614466893690405
|
| 177 |
+
},
|
| 178 |
+
"param_count": 4194304,
|
| 179 |
+
"metadata": {}
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"layer_name": "language_model.model.layers.22.linear_attn.in_proj_qkv",
|
| 183 |
+
"sensitivities": {
|
| 184 |
+
"2": 0.0032001393022558964,
|
| 185 |
+
"3": 0.001184641027066391,
|
| 186 |
+
"4": 0.0005809050489915535,
|
| 187 |
+
"5": 0.00040679439492426656,
|
| 188 |
+
"6": 0.00032513435629274073
|
| 189 |
+
},
|
| 190 |
+
"param_count": 12582912,
|
| 191 |
+
"metadata": {}
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"layer_name": "language_model.model.layers.21.mlp.up_proj",
|
| 195 |
+
"sensitivities": {
|
| 196 |
+
"2": 0.005692710304477562,
|
| 197 |
+
"3": 0.0016777145719970576,
|
| 198 |
+
"4": 0.0007404645706022469,
|
| 199 |
+
"5": 0.00045306347965379246,
|
| 200 |
+
"6": 0.0003520853721662813
|
| 201 |
+
},
|
| 202 |
+
"param_count": 12582912,
|
| 203 |
+
"metadata": {}
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"layer_name": "language_model.model.layers.21.mlp.down_proj",
|
| 207 |
+
"sensitivities": {
|
| 208 |
+
"2": 0.005179674718723011,
|
| 209 |
+
"3": 0.001543702664397036,
|
| 210 |
+
"4": 0.0007065836313510468,
|
| 211 |
+
"5": 0.0004472691277139044,
|
| 212 |
+
"6": 0.00035275551878536743
|
| 213 |
+
},
|
| 214 |
+
"param_count": 12582912,
|
| 215 |
+
"metadata": {}
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"layer_name": "language_model.model.layers.21.mlp.gate_proj",
|
| 219 |
+
"sensitivities": {
|
| 220 |
+
"2": 0.004566296862321906,
|
| 221 |
+
"3": 0.0013602330861128091,
|
| 222 |
+
"4": 0.000633769187212844,
|
| 223 |
+
"5": 0.0004157839854694127,
|
| 224 |
+
"6": 0.0003285163171919218
|
| 225 |
+
},
|
| 226 |
+
"param_count": 12582912,
|
| 227 |
+
"metadata": {}
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"layer_name": "language_model.model.layers.21.linear_attn.out_proj",
|
| 231 |
+
"sensitivities": {
|
| 232 |
+
"2": 0.0031479030755387307,
|
| 233 |
+
"3": 0.0011545070095356398,
|
| 234 |
+
"4": 0.0006043641005817335,
|
| 235 |
+
"5": 0.00041223811300975893,
|
| 236 |
+
"6": 0.0003307170457749938
|
| 237 |
+
},
|
| 238 |
+
"param_count": 4194304,
|
| 239 |
+
"metadata": {}
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"layer_name": "language_model.model.layers.21.linear_attn.in_proj_a",
|
| 243 |
+
"sensitivities": {
|
| 244 |
+
"2": 0.00055213918252169,
|
| 245 |
+
"3": 0.0003129309661744628,
|
| 246 |
+
"4": 0.00030220249633809243,
|
| 247 |
+
"5": 0.0002868863903131569,
|
| 248 |
+
"6": 0.00028064851691548637
|
| 249 |
+
},
|
| 250 |
+
"param_count": 32768,
|
| 251 |
+
"metadata": {}
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"layer_name": "language_model.model.layers.21.linear_attn.in_proj_b",
|
| 255 |
+
"sensitivities": {
|
| 256 |
+
"2": 0.0004433992732325957,
|
| 257 |
+
"3": 0.00034464218697394244,
|
| 258 |
+
"4": 0.00031512617473102483,
|
| 259 |
+
"5": 0.00028607156067058287,
|
| 260 |
+
"6": 0.0002804598070724751
|
| 261 |
+
},
|
| 262 |
+
"param_count": 32768,
|
| 263 |
+
"metadata": {}
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"layer_name": "language_model.model.layers.21.linear_attn.in_proj_z",
|
| 267 |
+
"sensitivities": {
|
| 268 |
+
"2": 0.0019399424830529217,
|
| 269 |
+
"3": 0.0007732317450669749,
|
| 270 |
+
"4": 0.00048270892087505973,
|
| 271 |
+
"5": 0.0003598842843833457,
|
| 272 |
+
"6": 0.00031587992028410855
|
| 273 |
+
},
|
| 274 |
+
"param_count": 4194304,
|
| 275 |
+
"metadata": {}
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"layer_name": "language_model.model.layers.21.linear_attn.in_proj_qkv",
|
| 279 |
+
"sensitivities": {
|
| 280 |
+
"2": 0.0032669137605504752,
|
| 281 |
+
"3": 0.0011020807820993166,
|
| 282 |
+
"4": 0.0005553400199763322,
|
| 283 |
+
"5": 0.00040502061710867565,
|
| 284 |
+
"6": 0.0003327313115732977
|
| 285 |
+
},
|
| 286 |
+
"param_count": 12582912,
|
| 287 |
+
"metadata": {}
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"layer_name": "language_model.model.layers.20.mlp.up_proj",
|
| 291 |
+
"sensitivities": {
|
| 292 |
+
"2": 0.005118766595842317,
|
| 293 |
+
"3": 0.001696106031886302,
|
| 294 |
+
"4": 0.0007137344097524571,
|
| 295 |
+
"5": 0.00045575186171239085,
|
| 296 |
+
"6": 0.00036402301278333954
|
| 297 |
+
},
|
| 298 |
+
"param_count": 12582912,
|
| 299 |
+
"metadata": {}
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"layer_name": "language_model.model.layers.20.mlp.down_proj",
|
| 303 |
+
"sensitivities": {
|
| 304 |
+
"2": 0.0051167806959711015,
|
| 305 |
+
"3": 0.0016198170184604048,
|
| 306 |
+
"4": 0.0007170970687487473,
|
| 307 |
+
"5": 0.000445386258434155,
|
| 308 |
+
"6": 0.0003572724429735293
|
| 309 |
+
},
|
| 310 |
+
"param_count": 12582912,
|
| 311 |
+
"metadata": {}
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"layer_name": "language_model.model.layers.20.mlp.gate_proj",
|
| 315 |
+
"sensitivities": {
|
| 316 |
+
"2": 0.004149125641561113,
|
| 317 |
+
"3": 0.0013489226621459238,
|
| 318 |
+
"4": 0.0006299914142194515,
|
| 319 |
+
"5": 0.00041615614372858545,
|
| 320 |
+
"6": 0.00035084451640917297
|
| 321 |
+
},
|
| 322 |
+
"param_count": 12582912,
|
| 323 |
+
"metadata": {}
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"layer_name": "language_model.model.layers.20.linear_attn.out_proj",
|
| 327 |
+
"sensitivities": {
|
| 328 |
+
"2": 0.004336462673866966,
|
| 329 |
+
"3": 0.0014099645753352281,
|
| 330 |
+
"4": 0.000662311435007723,
|
| 331 |
+
"5": 0.0004438879495864967,
|
| 332 |
+
"6": 0.00036200557769916486
|
| 333 |
+
},
|
| 334 |
+
"param_count": 4194304,
|
| 335 |
+
"metadata": {}
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"layer_name": "language_model.model.layers.20.linear_attn.in_proj_a",
|
| 339 |
+
"sensitivities": {
|
| 340 |
+
"2": 0.00036831060606346,
|
| 341 |
+
"3": 0.0003377934878396142,
|
| 342 |
+
"4": 0.00031962499603347777,
|
| 343 |
+
"5": 0.00030977015133733704,
|
| 344 |
+
"6": 0.000303664040681421
|
| 345 |
+
},
|
| 346 |
+
"param_count": 32768,
|
| 347 |
+
"metadata": {}
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"layer_name": "language_model.model.layers.20.linear_attn.in_proj_b",
|
| 351 |
+
"sensitivities": {
|
| 352 |
+
"2": 0.00044047730625607073,
|
| 353 |
+
"3": 0.0003544873816281324,
|
| 354 |
+
"4": 0.0003192826983043536,
|
| 355 |
+
"5": 0.00031075441256689373,
|
| 356 |
+
"6": 0.00030823834701247205
|
| 357 |
+
},
|
| 358 |
+
"param_count": 32768,
|
| 359 |
+
"metadata": {}
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"layer_name": "language_model.model.layers.20.linear_attn.in_proj_z",
|
| 363 |
+
"sensitivities": {
|
| 364 |
+
"2": 0.002011087407784847,
|
| 365 |
+
"3": 0.0008220997212144235,
|
| 366 |
+
"4": 0.0004801454203213022,
|
| 367 |
+
"5": 0.00038073316803396057,
|
| 368 |
+
"6": 0.00034256719724605017
|
| 369 |
+
},
|
| 370 |
+
"param_count": 4194304,
|
| 371 |
+
"metadata": {}
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"layer_name": "language_model.model.layers.20.linear_attn.in_proj_qkv",
|
| 375 |
+
"sensitivities": {
|
| 376 |
+
"2": 0.003023376402173502,
|
| 377 |
+
"3": 0.0010550292196664184,
|
| 378 |
+
"4": 0.0005431359829041563,
|
| 379 |
+
"5": 0.0004069235892529832,
|
| 380 |
+
"6": 0.0003491182988000219
|
| 381 |
+
},
|
| 382 |
+
"param_count": 12582912,
|
| 383 |
+
"metadata": {}
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"layer_name": "language_model.model.layers.19.mlp.up_proj",
|
| 387 |
+
"sensitivities": {
|
| 388 |
+
"2": 0.005886373352647449,
|
| 389 |
+
"3": 0.0018109687201407116,
|
| 390 |
+
"4": 0.0007755547540000407,
|
| 391 |
+
"5": 0.00048699448840731446,
|
| 392 |
+
"6": 0.00038368474300417194
|
| 393 |
+
},
|
| 394 |
+
"param_count": 12582912,
|
| 395 |
+
"metadata": {}
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"layer_name": "language_model.model.layers.19.mlp.down_proj",
|
| 399 |
+
"sensitivities": {
|
| 400 |
+
"2": 0.007172049920579108,
|
| 401 |
+
"3": 0.0020441845917957835,
|
| 402 |
+
"4": 0.0008472401520217924,
|
| 403 |
+
"5": 0.0005016297497301517,
|
| 404 |
+
"6": 0.00039514202520270675
|
| 405 |
+
},
|
| 406 |
+
"param_count": 12582912,
|
| 407 |
+
"metadata": {}
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"layer_name": "language_model.model.layers.19.mlp.gate_proj",
|
| 411 |
+
"sensitivities": {
|
| 412 |
+
"2": 0.004987381422931018,
|
| 413 |
+
"3": 0.0014731915701607552,
|
| 414 |
+
"4": 0.0006816160360661646,
|
| 415 |
+
"5": 0.00044166905354359187,
|
| 416 |
+
"6": 0.00037795773399314686
|
| 417 |
+
},
|
| 418 |
+
"param_count": 12582912,
|
| 419 |
+
"metadata": {}
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"layer_name": "language_model.model.layers.19.self_attn.o_proj",
|
| 423 |
+
"sensitivities": {
|
| 424 |
+
"2": 0.010323332423771111,
|
| 425 |
+
"3": 0.0032136654626810923,
|
| 426 |
+
"4": 0.0010781680612126365,
|
| 427 |
+
"5": 0.0006089774500045072,
|
| 428 |
+
"6": 0.00044127777315831435
|
| 429 |
+
},
|
| 430 |
+
"param_count": 4194304,
|
| 431 |
+
"metadata": {}
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"layer_name": "language_model.model.layers.19.self_attn.v_proj",
|
| 435 |
+
"sensitivities": {
|
| 436 |
+
"2": 0.003855544695397839,
|
| 437 |
+
"3": 0.001653690558062711,
|
| 438 |
+
"4": 0.0008302755777549464,
|
| 439 |
+
"5": 0.0005433602066962825,
|
| 440 |
+
"6": 0.0004065270968567347
|
| 441 |
+
},
|
| 442 |
+
"param_count": 1048576,
|
| 443 |
+
"metadata": {}
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"layer_name": "language_model.model.layers.19.self_attn.k_proj",
|
| 447 |
+
"sensitivities": {
|
| 448 |
+
"2": 0.0025537921998572224,
|
| 449 |
+
"3": 0.000902174956233163,
|
| 450 |
+
"4": 0.0005116393197871124,
|
| 451 |
+
"5": 0.0004076674013049342,
|
| 452 |
+
"6": 0.00038209764049194445
|
| 453 |
+
},
|
| 454 |
+
"param_count": 1048576,
|
| 455 |
+
"metadata": {}
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"layer_name": "language_model.model.layers.19.self_attn.q_proj",
|
| 459 |
+
"sensitivities": {
|
| 460 |
+
"2": 0.0034364661939131715,
|
| 461 |
+
"3": 0.001170218226131207,
|
| 462 |
+
"4": 0.0005953245830217687,
|
| 463 |
+
"5": 0.00044516982961795293,
|
| 464 |
+
"6": 0.0003910335053660674
|
| 465 |
+
},
|
| 466 |
+
"param_count": 8388608,
|
| 467 |
+
"metadata": {}
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"layer_name": "language_model.model.layers.18.mlp.up_proj",
|
| 471 |
+
"sensitivities": {
|
| 472 |
+
"2": 0.007433839510971059,
|
| 473 |
+
"3": 0.002203709173045354,
|
| 474 |
+
"4": 0.0008663502788598029,
|
| 475 |
+
"5": 0.0005335220412234776,
|
| 476 |
+
"6": 0.00041495203610490233
|
| 477 |
+
},
|
| 478 |
+
"param_count": 12582912,
|
| 479 |
+
"metadata": {}
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"layer_name": "language_model.model.layers.18.mlp.down_proj",
|
| 483 |
+
"sensitivities": {
|
| 484 |
+
"2": 0.007957058556106253,
|
| 485 |
+
"3": 0.002336282789959417,
|
| 486 |
+
"4": 0.0009055571341984129,
|
| 487 |
+
"5": 0.0005533959726259733,
|
| 488 |
+
"6": 0.0004284993253046802
|
| 489 |
+
},
|
| 490 |
+
"param_count": 12582912,
|
| 491 |
+
"metadata": {}
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"layer_name": "language_model.model.layers.18.mlp.gate_proj",
|
| 495 |
+
"sensitivities": {
|
| 496 |
+
"2": 0.006411260415916331,
|
| 497 |
+
"3": 0.001770203957372966,
|
| 498 |
+
"4": 0.0007664468248549383,
|
| 499 |
+
"5": 0.000498591760636676,
|
| 500 |
+
"6": 0.0004128660633189914
|
| 501 |
+
},
|
| 502 |
+
"param_count": 12582912,
|
| 503 |
+
"metadata": {}
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"layer_name": "language_model.model.layers.18.linear_attn.out_proj",
|
| 507 |
+
"sensitivities": {
|
| 508 |
+
"2": 0.0044861592129260925,
|
| 509 |
+
"3": 0.0014801480695799303,
|
| 510 |
+
"4": 0.0007210178694852706,
|
| 511 |
+
"5": 0.000481834295593823,
|
| 512 |
+
"6": 0.00041032154695130885
|
| 513 |
+
},
|
| 514 |
+
"param_count": 4194304,
|
| 515 |
+
"metadata": {}
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"layer_name": "language_model.model.layers.18.linear_attn.in_proj_a",
|
| 519 |
+
"sensitivities": {
|
| 520 |
+
"2": 0.0004181485255685402,
|
| 521 |
+
"3": 0.0003870961572829401,
|
| 522 |
+
"4": 0.0003712768575496739,
|
| 523 |
+
"5": 0.00036131578751034493,
|
| 524 |
+
"6": 0.00036473974129573133
|
| 525 |
+
},
|
| 526 |
+
"param_count": 32768,
|
| 527 |
+
"metadata": {}
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"layer_name": "language_model.model.layers.18.linear_attn.in_proj_b",
|
| 531 |
+
"sensitivities": {
|
| 532 |
+
"2": 0.0004831960620019042,
|
| 533 |
+
"3": 0.00040929536589828786,
|
| 534 |
+
"4": 0.0003825072575030693,
|
| 535 |
+
"5": 0.0003608593712366807,
|
| 536 |
+
"6": 0.0003628972790465923
|
| 537 |
+
},
|
| 538 |
+
"param_count": 32768,
|
| 539 |
+
"metadata": {}
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"layer_name": "language_model.model.layers.18.linear_attn.in_proj_z",
|
| 543 |
+
"sensitivities": {
|
| 544 |
+
"2": 0.002793392338692987,
|
| 545 |
+
"3": 0.0010011650556407403,
|
| 546 |
+
"4": 0.0005588795247604139,
|
| 547 |
+
"5": 0.00043731602454499807,
|
| 548 |
+
"6": 0.0003887747728488951
|
| 549 |
+
},
|
| 550 |
+
"param_count": 4194304,
|
| 551 |
+
"metadata": {}
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"layer_name": "language_model.model.layers.18.linear_attn.in_proj_qkv",
|
| 555 |
+
"sensitivities": {
|
| 556 |
+
"2": 0.0035599808834376745,
|
| 557 |
+
"3": 0.0011301374167184501,
|
| 558 |
+
"4": 0.0006094327545724809,
|
| 559 |
+
"5": 0.0004491992391801129,
|
| 560 |
+
"6": 0.0003983426477134344
|
| 561 |
+
},
|
| 562 |
+
"param_count": 12582912,
|
| 563 |
+
"metadata": {}
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"layer_name": "language_model.model.layers.17.mlp.up_proj",
|
| 567 |
+
"sensitivities": {
|
| 568 |
+
"2": 0.007933031030309698,
|
| 569 |
+
"3": 0.002168619384368261,
|
| 570 |
+
"4": 0.0008952599843420709,
|
| 571 |
+
"5": 0.0005544827757451761,
|
| 572 |
+
"6": 0.0004394544263656523
|
| 573 |
+
},
|
| 574 |
+
"param_count": 12582912,
|
| 575 |
+
"metadata": {}
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"layer_name": "language_model.model.layers.17.mlp.down_proj",
|
| 579 |
+
"sensitivities": {
|
| 580 |
+
"2": 0.0088346138606236,
|
| 581 |
+
"3": 0.0024337218492291868,
|
| 582 |
+
"4": 0.0009611720561224502,
|
| 583 |
+
"5": 0.0005701775662600994,
|
| 584 |
+
"6": 0.0004422725302598944
|
| 585 |
+
},
|
| 586 |
+
"param_count": 12582912,
|
| 587 |
+
"metadata": {}
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"layer_name": "language_model.model.layers.17.mlp.gate_proj",
|
| 591 |
+
"sensitivities": {
|
| 592 |
+
"2": 0.006522652591229416,
|
| 593 |
+
"3": 0.0018156510341214016,
|
| 594 |
+
"4": 0.0007952602815445667,
|
| 595 |
+
"5": 0.0005142078346883258,
|
| 596 |
+
"6": 0.00042525669778115116
|
| 597 |
+
},
|
| 598 |
+
"param_count": 12582912,
|
| 599 |
+
"metadata": {}
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"layer_name": "language_model.model.layers.17.linear_attn.out_proj",
|
| 603 |
+
"sensitivities": {
|
| 604 |
+
"2": 0.006022836207800235,
|
| 605 |
+
"3": 0.0019106197554113653,
|
| 606 |
+
"4": 0.0008096261117316317,
|
| 607 |
+
"5": 0.0005259700707635299,
|
| 608 |
+
"6": 0.00043780878877441864
|
| 609 |
+
},
|
| 610 |
+
"param_count": 4194304,
|
| 611 |
+
"metadata": {}
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"layer_name": "language_model.model.layers.17.linear_attn.in_proj_a",
|
| 615 |
+
"sensitivities": {
|
| 616 |
+
"2": 0.0006068453306700879,
|
| 617 |
+
"3": 0.000419972501428371,
|
| 618 |
+
"4": 0.0003957351321635845,
|
| 619 |
+
"5": 0.00038980332101345994,
|
| 620 |
+
"6": 0.0003830937196956559
|
| 621 |
+
},
|
| 622 |
+
"param_count": 32768,
|
| 623 |
+
"metadata": {}
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"layer_name": "language_model.model.layers.17.linear_attn.in_proj_b",
|
| 627 |
+
"sensitivities": {
|
| 628 |
+
"2": 0.0005948931302555138,
|
| 629 |
+
"3": 0.00046442968778137583,
|
| 630 |
+
"4": 0.0004095059739483986,
|
| 631 |
+
"5": 0.00039187428440830746,
|
| 632 |
+
"6": 0.00038206107819860335
|
| 633 |
+
},
|
| 634 |
+
"param_count": 32768,
|
| 635 |
+
"metadata": {}
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"layer_name": "language_model.model.layers.17.linear_attn.in_proj_z",
|
| 639 |
+
"sensitivities": {
|
| 640 |
+
"2": 0.003036055277334526,
|
| 641 |
+
"3": 0.0010796979543859682,
|
| 642 |
+
"4": 0.0006014708117921449,
|
| 643 |
+
"5": 0.00046434341432662524,
|
| 644 |
+
"6": 0.0004162881208079246
|
| 645 |
+
},
|
| 646 |
+
"param_count": 4194304,
|
| 647 |
+
"metadata": {}
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"layer_name": "language_model.model.layers.17.linear_attn.in_proj_qkv",
|
| 651 |
+
"sensitivities": {
|
| 652 |
+
"2": 0.005522147422501196,
|
| 653 |
+
"3": 0.0016118041627729933,
|
| 654 |
+
"4": 0.0007424685973091982,
|
| 655 |
+
"5": 0.0004988557554194509,
|
| 656 |
+
"6": 0.00042759561317022116
|
| 657 |
+
},
|
| 658 |
+
"param_count": 12582912,
|
| 659 |
+
"metadata": {}
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"layer_name": "language_model.model.layers.16.mlp.up_proj",
|
| 663 |
+
"sensitivities": {
|
| 664 |
+
"2": 0.008788964594714344,
|
| 665 |
+
"3": 0.0024738212220351366,
|
| 666 |
+
"4": 0.0009819428463136621,
|
| 667 |
+
"5": 0.0005860094491557296,
|
| 668 |
+
"6": 0.0004658450015995186
|
| 669 |
+
},
|
| 670 |
+
"param_count": 12582912,
|
| 671 |
+
"metadata": {}
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"layer_name": "language_model.model.layers.16.mlp.down_proj",
|
| 675 |
+
"sensitivities": {
|
| 676 |
+
"2": 0.010837621016738316,
|
| 677 |
+
"3": 0.003065576616791077,
|
| 678 |
+
"4": 0.001088670638637268,
|
| 679 |
+
"5": 0.0006141405671466297,
|
| 680 |
+
"6": 0.0004785644944907593
|
| 681 |
+
},
|
| 682 |
+
"param_count": 12582912,
|
| 683 |
+
"metadata": {}
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"layer_name": "language_model.model.layers.16.mlp.gate_proj",
|
| 687 |
+
"sensitivities": {
|
| 688 |
+
"2": 0.00696840124631611,
|
| 689 |
+
"3": 0.0019108105601238397,
|
| 690 |
+
"4": 0.0008225605609671524,
|
| 691 |
+
"5": 0.0005392715744771218,
|
| 692 |
+
"6": 0.0004435746177477995
|
| 693 |
+
},
|
| 694 |
+
"param_count": 12582912,
|
| 695 |
+
"metadata": {}
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"layer_name": "language_model.model.layers.16.linear_attn.out_proj",
|
| 699 |
+
"sensitivities": {
|
| 700 |
+
"2": 0.008754666234987477,
|
| 701 |
+
"3": 0.0023875193025257127,
|
| 702 |
+
"4": 0.0009868539794600413,
|
| 703 |
+
"5": 0.0005959852248148915,
|
| 704 |
+
"6": 0.0004834424689761363
|
| 705 |
+
},
|
| 706 |
+
"param_count": 4194304,
|
| 707 |
+
"metadata": {}
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"layer_name": "language_model.model.layers.16.linear_attn.in_proj_a",
|
| 711 |
+
"sensitivities": {
|
| 712 |
+
"2": 0.0005215696061592704,
|
| 713 |
+
"3": 0.00044448673058165394,
|
| 714 |
+
"4": 0.0004128416088254501,
|
| 715 |
+
"5": 0.00040226354819120996,
|
| 716 |
+
"6": 0.00039968368704042706
|
| 717 |
+
},
|
| 718 |
+
"param_count": 32768,
|
| 719 |
+
"metadata": {}
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"layer_name": "language_model.model.layers.16.linear_attn.in_proj_b",
|
| 723 |
+
"sensitivities": {
|
| 724 |
+
"2": 0.0006828383872440705,
|
| 725 |
+
"3": 0.0004835047972543786,
|
| 726 |
+
"4": 0.0004320617824002208,
|
| 727 |
+
"5": 0.00040628293390909676,
|
| 728 |
+
"6": 0.0004077235892812799
|
| 729 |
+
},
|
| 730 |
+
"param_count": 32768,
|
| 731 |
+
"metadata": {}
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"layer_name": "language_model.model.layers.16.linear_attn.in_proj_z",
|
| 735 |
+
"sensitivities": {
|
| 736 |
+
"2": 0.0030329171058838256,
|
| 737 |
+
"3": 0.0010533247514104005,
|
| 738 |
+
"4": 0.0006193393164721783,
|
| 739 |
+
"5": 0.00047354081137503573,
|
| 740 |
+
"6": 0.00042552364175207913
|
| 741 |
+
},
|
| 742 |
+
"param_count": 4194304,
|
| 743 |
+
"metadata": {}
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"layer_name": "language_model.model.layers.16.linear_attn.in_proj_qkv",
|
| 747 |
+
"sensitivities": {
|
| 748 |
+
"2": 0.007359471414626266,
|
| 749 |
+
"3": 0.0019950863182505905,
|
| 750 |
+
"4": 0.0008785241128255924,
|
| 751 |
+
"5": 0.0005524645572828982,
|
| 752 |
+
"6": 0.0004577265705544657
|
| 753 |
+
},
|
| 754 |
+
"param_count": 12582912,
|
| 755 |
+
"metadata": {}
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"layer_name": "language_model.model.layers.15.mlp.up_proj",
|
| 759 |
+
"sensitivities": {
|
| 760 |
+
"2": 0.010233937476490004,
|
| 761 |
+
"3": 0.0031141662014609515,
|
| 762 |
+
"4": 0.0011725737519251804,
|
| 763 |
+
"5": 0.0006887060638594752,
|
| 764 |
+
"6": 0.0005054456190312825
|
| 765 |
+
},
|
| 766 |
+
"param_count": 12582912,
|
| 767 |
+
"metadata": {}
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"layer_name": "language_model.model.layers.15.mlp.down_proj",
|
| 771 |
+
"sensitivities": {
|
| 772 |
+
"2": 0.013206733662324647,
|
| 773 |
+
"3": 0.003654784780034485,
|
| 774 |
+
"4": 0.0012385396039462648,
|
| 775 |
+
"5": 0.000670974357490195,
|
| 776 |
+
"6": 0.0005051164310619546
|
| 777 |
+
},
|
| 778 |
+
"param_count": 12582912,
|
| 779 |
+
"metadata": {}
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"layer_name": "language_model.model.layers.15.mlp.gate_proj",
|
| 783 |
+
"sensitivities": {
|
| 784 |
+
"2": 0.007315659934344391,
|
| 785 |
+
"3": 0.002064433962611171,
|
| 786 |
+
"4": 0.0008631094727509966,
|
| 787 |
+
"5": 0.0005584043162040567,
|
| 788 |
+
"6": 0.00046771071235222433
|
| 789 |
+
},
|
| 790 |
+
"param_count": 12582912,
|
| 791 |
+
"metadata": {}
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"layer_name": "language_model.model.layers.15.self_attn.o_proj",
|
| 795 |
+
"sensitivities": {
|
| 796 |
+
"2": 0.020197030874745298,
|
| 797 |
+
"3": 0.005279539754459013,
|
| 798 |
+
"4": 0.0017447570159371633,
|
| 799 |
+
"5": 0.0007989972552119676,
|
| 800 |
+
"6": 0.0005365728805675948
|
| 801 |
+
},
|
| 802 |
+
"param_count": 4194304,
|
| 803 |
+
"metadata": {}
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"layer_name": "language_model.model.layers.15.self_attn.v_proj",
|
| 807 |
+
"sensitivities": {
|
| 808 |
+
"2": 0.00628529015618066,
|
| 809 |
+
"3": 0.002153019406250678,
|
| 810 |
+
"4": 0.000977495800422427,
|
| 811 |
+
"5": 0.0005969984413241036,
|
| 812 |
+
"6": 0.000479581856779987
|
| 813 |
+
},
|
| 814 |
+
"param_count": 1048576,
|
| 815 |
+
"metadata": {}
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"layer_name": "language_model.model.layers.15.self_attn.k_proj",
|
| 819 |
+
"sensitivities": {
|
| 820 |
+
"2": 0.004149579370277934,
|
| 821 |
+
"3": 0.0011555929280196626,
|
| 822 |
+
"4": 0.0006336492309249783,
|
| 823 |
+
"5": 0.0005009591908068008,
|
| 824 |
+
"6": 0.00046580559319409076
|
| 825 |
+
},
|
| 826 |
+
"param_count": 1048576,
|
| 827 |
+
"metadata": {}
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"layer_name": "language_model.model.layers.15.self_attn.q_proj",
|
| 831 |
+
"sensitivities": {
|
| 832 |
+
"2": 0.006053162137201677,
|
| 833 |
+
"3": 0.0018404361932577256,
|
| 834 |
+
"4": 0.000852241023191406,
|
| 835 |
+
"5": 0.0005620284803929584,
|
| 836 |
+
"6": 0.0004819470947647157
|
| 837 |
+
},
|
| 838 |
+
"param_count": 8388608,
|
| 839 |
+
"metadata": {}
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"layer_name": "language_model.model.layers.14.mlp.up_proj",
|
| 843 |
+
"sensitivities": {
|
| 844 |
+
"2": 0.01064420904731378,
|
| 845 |
+
"3": 0.003169933489213387,
|
| 846 |
+
"4": 0.0011876187757782948,
|
| 847 |
+
"5": 0.0006820645697492486,
|
| 848 |
+
"6": 0.0005270268963310324
|
| 849 |
+
},
|
| 850 |
+
"param_count": 12582912,
|
| 851 |
+
"metadata": {}
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"layer_name": "language_model.model.layers.14.mlp.down_proj",
|
| 855 |
+
"sensitivities": {
|
| 856 |
+
"2": 0.018537963818137843,
|
| 857 |
+
"3": 0.006525251636048779,
|
| 858 |
+
"4": 0.002059532484660546,
|
| 859 |
+
"5": 0.0008675048823837036,
|
| 860 |
+
"6": 0.0005608681640296709
|
| 861 |
+
},
|
| 862 |
+
"param_count": 12582912,
|
| 863 |
+
"metadata": {}
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"layer_name": "language_model.model.layers.14.mlp.gate_proj",
|
| 867 |
+
"sensitivities": {
|
| 868 |
+
"2": 0.00910614157813446,
|
| 869 |
+
"3": 0.002463228052268581,
|
| 870 |
+
"4": 0.0009815030013366293,
|
| 871 |
+
"5": 0.000615959899732843,
|
| 872 |
+
"6": 0.0004921673765541831
|
| 873 |
+
},
|
| 874 |
+
"param_count": 12582912,
|
| 875 |
+
"metadata": {}
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"layer_name": "language_model.model.layers.14.linear_attn.out_proj",
|
| 879 |
+
"sensitivities": {
|
| 880 |
+
"2": 0.009095845598494634,
|
| 881 |
+
"3": 0.0025505881761394753,
|
| 882 |
+
"4": 0.0010254524580280606,
|
| 883 |
+
"5": 0.0006508909452046888,
|
| 884 |
+
"6": 0.00052572624917957
|
| 885 |
+
},
|
| 886 |
+
"param_count": 4194304,
|
| 887 |
+
"metadata": {}
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"layer_name": "language_model.model.layers.14.linear_attn.in_proj_a",
|
| 891 |
+
"sensitivities": {
|
| 892 |
+
"2": 0.0006181676162668737,
|
| 893 |
+
"3": 0.000514552876969295,
|
| 894 |
+
"4": 0.0004711608665578145,
|
| 895 |
+
"5": 0.0004585139719589885,
|
| 896 |
+
"6": 0.00044568776532590465
|
| 897 |
+
},
|
| 898 |
+
"param_count": 32768,
|
| 899 |
+
"metadata": {}
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"layer_name": "language_model.model.layers.14.linear_attn.in_proj_b",
|
| 903 |
+
"sensitivities": {
|
| 904 |
+
"2": 0.0007466161381065225,
|
| 905 |
+
"3": 0.0005496387439052342,
|
| 906 |
+
"4": 0.00047531792491402786,
|
| 907 |
+
"5": 0.00046330486657097936,
|
| 908 |
+
"6": 0.00045183106218852725
|
| 909 |
+
},
|
| 910 |
+
"param_count": 32768,
|
| 911 |
+
"metadata": {}
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"layer_name": "language_model.model.layers.14.linear_attn.in_proj_z",
|
| 915 |
+
"sensitivities": {
|
| 916 |
+
"2": 0.0028095459080456444,
|
| 917 |
+
"3": 0.0010408348437825528,
|
| 918 |
+
"4": 0.0006203524584028249,
|
| 919 |
+
"5": 0.0005101718749453236,
|
| 920 |
+
"6": 0.00047207688476191834
|
| 921 |
+
},
|
| 922 |
+
"param_count": 4194304,
|
| 923 |
+
"metadata": {}
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"layer_name": "language_model.model.layers.14.linear_attn.in_proj_qkv",
|
| 927 |
+
"sensitivities": {
|
| 928 |
+
"2": 0.009474482445511967,
|
| 929 |
+
"3": 0.0028602693346329033,
|
| 930 |
+
"4": 0.0010903566993268516,
|
| 931 |
+
"5": 0.0006791557013154185,
|
| 932 |
+
"6": 0.0005352472326194402
|
| 933 |
+
},
|
| 934 |
+
"param_count": 12582912,
|
| 935 |
+
"metadata": {}
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"layer_name": "language_model.model.layers.13.mlp.up_proj",
|
| 939 |
+
"sensitivities": {
|
| 940 |
+
"2": 0.009484339165889347,
|
| 941 |
+
"3": 0.002816848699391509,
|
| 942 |
+
"4": 0.0010881778392407189,
|
| 943 |
+
"5": 0.000661783457568769,
|
| 944 |
+
"6": 0.0005288143584039062
|
| 945 |
+
},
|
| 946 |
+
"param_count": 12582912,
|
| 947 |
+
"metadata": {}
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"layer_name": "language_model.model.layers.13.mlp.down_proj",
|
| 951 |
+
"sensitivities": {
|
| 952 |
+
"2": 0.009029947279486805,
|
| 953 |
+
"3": 0.002571757742164967,
|
| 954 |
+
"4": 0.0010094155010544152,
|
| 955 |
+
"5": 0.0006373814618806742,
|
| 956 |
+
"6": 0.0005178330029593781
|
| 957 |
+
},
|
| 958 |
+
"param_count": 12582912,
|
| 959 |
+
"metadata": {}
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"layer_name": "language_model.model.layers.13.mlp.gate_proj",
|
| 963 |
+
"sensitivities": {
|
| 964 |
+
"2": 0.007235499962310617,
|
| 965 |
+
"3": 0.0020598815123472982,
|
| 966 |
+
"4": 0.0009292609696179474,
|
| 967 |
+
"5": 0.0006082836201433869,
|
| 968 |
+
"6": 0.0005122598928816539
|
| 969 |
+
},
|
| 970 |
+
"param_count": 12582912,
|
| 971 |
+
"metadata": {}
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"layer_name": "language_model.model.layers.13.linear_attn.out_proj",
|
| 975 |
+
"sensitivities": {
|
| 976 |
+
"2": 0.0072143913760858895,
|
| 977 |
+
"3": 0.0022142395185558903,
|
| 978 |
+
"4": 0.0009465841042886799,
|
| 979 |
+
"5": 0.0006277133300803447,
|
| 980 |
+
"6": 0.0005239191195869353
|
| 981 |
+
},
|
| 982 |
+
"param_count": 4194304,
|
| 983 |
+
"metadata": {}
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"layer_name": "language_model.model.layers.13.linear_attn.in_proj_a",
|
| 987 |
+
"sensitivities": {
|
| 988 |
+
"2": 0.0006744608369141739,
|
| 989 |
+
"3": 0.0005369241516746115,
|
| 990 |
+
"4": 0.0004775992174472776,
|
| 991 |
+
"5": 0.0004706283853010973,
|
| 992 |
+
"6": 0.00046884310177119914
|
| 993 |
+
},
|
| 994 |
+
"param_count": 32768,
|
| 995 |
+
"metadata": {}
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"layer_name": "language_model.model.layers.13.linear_attn.in_proj_b",
|
| 999 |
+
"sensitivities": {
|
| 1000 |
+
"2": 0.0006943414925141648,
|
| 1001 |
+
"3": 0.0005466579690012926,
|
| 1002 |
+
"4": 0.0004936007853757474,
|
| 1003 |
+
"5": 0.00047111589719861513,
|
| 1004 |
+
"6": 0.00047404949873453006
|
| 1005 |
+
},
|
| 1006 |
+
"param_count": 32768,
|
| 1007 |
+
"metadata": {}
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"layer_name": "language_model.model.layers.13.linear_attn.in_proj_z",
|
| 1011 |
+
"sensitivities": {
|
| 1012 |
+
"2": 0.0027744716959811435,
|
| 1013 |
+
"3": 0.0010011472792636293,
|
| 1014 |
+
"4": 0.0006195107125677168,
|
| 1015 |
+
"5": 0.0005226716275501531,
|
| 1016 |
+
"6": 0.00048261822121276055
|
| 1017 |
+
},
|
| 1018 |
+
"param_count": 4194304,
|
| 1019 |
+
"metadata": {}
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"layer_name": "language_model.model.layers.13.linear_attn.in_proj_qkv",
|
| 1023 |
+
"sensitivities": {
|
| 1024 |
+
"2": 0.010185367403513132,
|
| 1025 |
+
"3": 0.0030891039471801682,
|
| 1026 |
+
"4": 0.0011918399892844416,
|
| 1027 |
+
"5": 0.0006985126832053842,
|
| 1028 |
+
"6": 0.0005484195471581188
|
| 1029 |
+
},
|
| 1030 |
+
"param_count": 12582912,
|
| 1031 |
+
"metadata": {}
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"layer_name": "language_model.model.layers.12.mlp.up_proj",
|
| 1035 |
+
"sensitivities": {
|
| 1036 |
+
"2": 0.00956919279997237,
|
| 1037 |
+
"3": 0.0029483760881703347,
|
| 1038 |
+
"4": 0.0011338346339471173,
|
| 1039 |
+
"5": 0.0006820013732067309,
|
| 1040 |
+
"6": 0.000535990642674733
|
| 1041 |
+
},
|
| 1042 |
+
"param_count": 12582912,
|
| 1043 |
+
"metadata": {}
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"layer_name": "language_model.model.layers.12.mlp.down_proj",
|
| 1047 |
+
"sensitivities": {
|
| 1048 |
+
"2": 0.008944620669353753,
|
| 1049 |
+
"3": 0.002576602618015992,
|
| 1050 |
+
"4": 0.0010209293565518844,
|
| 1051 |
+
"5": 0.0006494528267163938,
|
| 1052 |
+
"6": 0.0005404256359421803
|
| 1053 |
+
},
|
| 1054 |
+
"param_count": 12582912,
|
| 1055 |
+
"metadata": {}
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"layer_name": "language_model.model.layers.12.mlp.gate_proj",
|
| 1059 |
+
"sensitivities": {
|
| 1060 |
+
"2": 0.006942726744455285,
|
| 1061 |
+
"3": 0.0020259803204680793,
|
| 1062 |
+
"4": 0.000895948465768015,
|
| 1063 |
+
"5": 0.0006146884625195526,
|
| 1064 |
+
"6": 0.0005144178461099121
|
| 1065 |
+
},
|
| 1066 |
+
"param_count": 12582912,
|
| 1067 |
+
"metadata": {}
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"layer_name": "language_model.model.layers.12.linear_attn.out_proj",
|
| 1071 |
+
"sensitivities": {
|
| 1072 |
+
"2": 0.007762689322892887,
|
| 1073 |
+
"3": 0.002528752333698018,
|
| 1074 |
+
"4": 0.0010718785839950822,
|
| 1075 |
+
"5": 0.0006674994074273854,
|
| 1076 |
+
"6": 0.0005475697489600861
|
| 1077 |
+
},
|
| 1078 |
+
"param_count": 4194304,
|
| 1079 |
+
"metadata": {}
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"layer_name": "language_model.model.layers.12.linear_attn.in_proj_a",
|
| 1083 |
+
"sensitivities": {
|
| 1084 |
+
"2": 0.0006075096322698906,
|
| 1085 |
+
"3": 0.0005196756477137873,
|
| 1086 |
+
"4": 0.0004799997368536424,
|
| 1087 |
+
"5": 0.000488825125406341,
|
| 1088 |
+
"6": 0.00047868184386364493
|
| 1089 |
+
},
|
| 1090 |
+
"param_count": 32768,
|
| 1091 |
+
"metadata": {}
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"layer_name": "language_model.model.layers.12.linear_attn.in_proj_b",
|
| 1095 |
+
"sensitivities": {
|
| 1096 |
+
"2": 0.0008044106089073466,
|
| 1097 |
+
"3": 0.0005662464791384991,
|
| 1098 |
+
"4": 0.0005076531803448839,
|
| 1099 |
+
"5": 0.0004887327001294276,
|
| 1100 |
+
"6": 0.00048145316395675763
|
| 1101 |
+
},
|
| 1102 |
+
"param_count": 32768,
|
| 1103 |
+
"metadata": {}
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"layer_name": "language_model.model.layers.12.linear_attn.in_proj_z",
|
| 1107 |
+
"sensitivities": {
|
| 1108 |
+
"2": 0.0035705460444053947,
|
| 1109 |
+
"3": 0.001175227193622656,
|
| 1110 |
+
"4": 0.0006960412750534791,
|
| 1111 |
+
"5": 0.000547294958475201,
|
| 1112 |
+
"6": 0.0005025641406367262
|
| 1113 |
+
},
|
| 1114 |
+
"param_count": 4194304,
|
| 1115 |
+
"metadata": {}
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"layer_name": "language_model.model.layers.12.linear_attn.in_proj_qkv",
|
| 1119 |
+
"sensitivities": {
|
| 1120 |
+
"2": 0.008573977616227543,
|
| 1121 |
+
"3": 0.0026332102958501005,
|
| 1122 |
+
"4": 0.001086155939750218,
|
| 1123 |
+
"5": 0.0006843514735616433,
|
| 1124 |
+
"6": 0.0005395782415386444
|
| 1125 |
+
},
|
| 1126 |
+
"param_count": 12582912,
|
| 1127 |
+
"metadata": {}
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"layer_name": "language_model.model.layers.11.mlp.up_proj",
|
| 1131 |
+
"sensitivities": {
|
| 1132 |
+
"2": 0.009940196827907736,
|
| 1133 |
+
"3": 0.00306264282213912,
|
| 1134 |
+
"4": 0.0012071711801885006,
|
| 1135 |
+
"5": 0.0007089850278134691,
|
| 1136 |
+
"6": 0.0005638869927982645
|
| 1137 |
+
},
|
| 1138 |
+
"param_count": 12582912,
|
| 1139 |
+
"metadata": {}
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"layer_name": "language_model.model.layers.11.mlp.down_proj",
|
| 1143 |
+
"sensitivities": {
|
| 1144 |
+
"2": 0.008750642936017053,
|
| 1145 |
+
"3": 0.0025069877738133073,
|
| 1146 |
+
"4": 0.001027650596370222,
|
| 1147 |
+
"5": 0.0006526668494188925,
|
| 1148 |
+
"6": 0.0005392560002898487
|
| 1149 |
+
},
|
| 1150 |
+
"param_count": 12582912,
|
| 1151 |
+
"metadata": {}
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"layer_name": "language_model.model.layers.11.mlp.gate_proj",
|
| 1155 |
+
"sensitivities": {
|
| 1156 |
+
"2": 0.006927876136614941,
|
| 1157 |
+
"3": 0.0020506323999143206,
|
| 1158 |
+
"4": 0.0009167596144834533,
|
| 1159 |
+
"5": 0.0006210874247093064,
|
| 1160 |
+
"6": 0.0005296924734163136
|
| 1161 |
+
},
|
| 1162 |
+
"param_count": 12582912,
|
| 1163 |
+
"metadata": {}
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"layer_name": "language_model.model.layers.11.self_attn.o_proj",
|
| 1167 |
+
"sensitivities": {
|
| 1168 |
+
"2": 0.023077569940748315,
|
| 1169 |
+
"3": 0.006392306997440755,
|
| 1170 |
+
"4": 0.0019076643414640178,
|
| 1171 |
+
"5": 0.0009177549751863504,
|
| 1172 |
+
"6": 0.0006178014003429174
|
| 1173 |
+
},
|
| 1174 |
+
"param_count": 4194304,
|
| 1175 |
+
"metadata": {}
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"layer_name": "language_model.model.layers.11.self_attn.v_proj",
|
| 1179 |
+
"sensitivities": {
|
| 1180 |
+
"2": 0.007021534334247311,
|
| 1181 |
+
"3": 0.0030780980062748617,
|
| 1182 |
+
"4": 0.0011376967983475577,
|
| 1183 |
+
"5": 0.0008481908917019609,
|
| 1184 |
+
"6": 0.000558946126451095
|
| 1185 |
+
},
|
| 1186 |
+
"param_count": 1048576,
|
| 1187 |
+
"metadata": {}
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"layer_name": "language_model.model.layers.11.self_attn.k_proj",
|
| 1191 |
+
"sensitivities": {
|
| 1192 |
+
"2": 0.0031676394040308273,
|
| 1193 |
+
"3": 0.0011211750212775466,
|
| 1194 |
+
"4": 0.0006759981806681026,
|
| 1195 |
+
"5": 0.0005636316630746782,
|
| 1196 |
+
"6": 0.0005131019158094811
|
| 1197 |
+
},
|
| 1198 |
+
"param_count": 1048576,
|
| 1199 |
+
"metadata": {}
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"layer_name": "language_model.model.layers.11.self_attn.q_proj",
|
| 1203 |
+
"sensitivities": {
|
| 1204 |
+
"2": 0.00889339933443504,
|
| 1205 |
+
"3": 0.0023529300415248144,
|
| 1206 |
+
"4": 0.0009685491313575767,
|
| 1207 |
+
"5": 0.0006583462673006579,
|
| 1208 |
+
"6": 0.0005561165683805788
|
| 1209 |
+
},
|
| 1210 |
+
"param_count": 8388608,
|
| 1211 |
+
"metadata": {}
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"layer_name": "language_model.model.layers.10.mlp.up_proj",
|
| 1215 |
+
"sensitivities": {
|
| 1216 |
+
"2": 0.01171053342598801,
|
| 1217 |
+
"3": 0.003378561098846452,
|
| 1218 |
+
"4": 0.0012651239606687643,
|
| 1219 |
+
"5": 0.0007462199173460249,
|
| 1220 |
+
"6": 0.0005754768183881728
|
| 1221 |
+
},
|
| 1222 |
+
"param_count": 12582912,
|
| 1223 |
+
"metadata": {}
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"layer_name": "language_model.model.layers.10.mlp.down_proj",
|
| 1227 |
+
"sensitivities": {
|
| 1228 |
+
"2": 0.009295990079408512,
|
| 1229 |
+
"3": 0.0025527389079798013,
|
| 1230 |
+
"4": 0.0010988656601208884,
|
| 1231 |
+
"5": 0.0006704258673077371,
|
| 1232 |
+
"6": 0.0005630586553403797
|
| 1233 |
+
},
|
| 1234 |
+
"param_count": 12582912,
|
| 1235 |
+
"metadata": {}
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"layer_name": "language_model.model.layers.10.mlp.gate_proj",
|
| 1239 |
+
"sensitivities": {
|
| 1240 |
+
"2": 0.007563161920794907,
|
| 1241 |
+
"3": 0.0022115032394746472,
|
| 1242 |
+
"4": 0.000959240165684605,
|
| 1243 |
+
"5": 0.0006442307506707342,
|
| 1244 |
+
"6": 0.0005550391754998903
|
| 1245 |
+
},
|
| 1246 |
+
"param_count": 12582912,
|
| 1247 |
+
"metadata": {}
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"layer_name": "language_model.model.layers.10.linear_attn.out_proj",
|
| 1251 |
+
"sensitivities": {
|
| 1252 |
+
"2": 0.009969159984999957,
|
| 1253 |
+
"3": 0.0028809758926702975,
|
| 1254 |
+
"4": 0.0011171083691200086,
|
| 1255 |
+
"5": 0.0007031722107058158,
|
| 1256 |
+
"6": 0.0005819206041148087
|
| 1257 |
+
},
|
| 1258 |
+
"param_count": 4194304,
|
| 1259 |
+
"metadata": {}
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"layer_name": "language_model.model.layers.10.linear_attn.in_proj_a",
|
| 1263 |
+
"sensitivities": {
|
| 1264 |
+
"2": 0.0008483172156653988,
|
| 1265 |
+
"3": 0.0005963146204521763,
|
| 1266 |
+
"4": 0.00054041267564268,
|
| 1267 |
+
"5": 0.0005290641978111429,
|
| 1268 |
+
"6": 0.0005184566747023686
|
| 1269 |
+
},
|
| 1270 |
+
"param_count": 32768,
|
| 1271 |
+
"metadata": {}
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"layer_name": "language_model.model.layers.10.linear_attn.in_proj_b",
|
| 1275 |
+
"sensitivities": {
|
| 1276 |
+
"2": 0.0008245474346040282,
|
| 1277 |
+
"3": 0.0006114042898843763,
|
| 1278 |
+
"4": 0.0005362270385376178,
|
| 1279 |
+
"5": 0.0005275378559114566,
|
| 1280 |
+
"6": 0.0005221250345736431
|
| 1281 |
+
},
|
| 1282 |
+
"param_count": 32768,
|
| 1283 |
+
"metadata": {}
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"layer_name": "language_model.model.layers.10.linear_attn.in_proj_z",
|
| 1287 |
+
"sensitivities": {
|
| 1288 |
+
"2": 0.003159514303357961,
|
| 1289 |
+
"3": 0.0011535318080859724,
|
| 1290 |
+
"4": 0.0007077336570849487,
|
| 1291 |
+
"5": 0.0005676809426707526,
|
| 1292 |
+
"6": 0.0005411098494126539
|
| 1293 |
+
},
|
| 1294 |
+
"param_count": 4194304,
|
| 1295 |
+
"metadata": {}
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"layer_name": "language_model.model.layers.10.linear_attn.in_proj_qkv",
|
| 1299 |
+
"sensitivities": {
|
| 1300 |
+
"2": 0.014127769779103497,
|
| 1301 |
+
"3": 0.004052434126303221,
|
| 1302 |
+
"4": 0.0014557715876435395,
|
| 1303 |
+
"5": 0.00084266806091667,
|
| 1304 |
+
"6": 0.0006317190958119075
|
| 1305 |
+
},
|
| 1306 |
+
"param_count": 12582912,
|
| 1307 |
+
"metadata": {}
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"layer_name": "language_model.model.layers.9.mlp.up_proj",
|
| 1311 |
+
"sensitivities": {
|
| 1312 |
+
"2": 0.009802988856487596,
|
| 1313 |
+
"3": 0.003209887772148553,
|
| 1314 |
+
"4": 0.0012759688179357909,
|
| 1315 |
+
"5": 0.0007477236819492342,
|
| 1316 |
+
"6": 0.0005987250715406844
|
| 1317 |
+
},
|
| 1318 |
+
"param_count": 12582912,
|
| 1319 |
+
"metadata": {}
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"layer_name": "language_model.model.layers.9.mlp.down_proj",
|
| 1323 |
+
"sensitivities": {
|
| 1324 |
+
"2": 0.008586498476991741,
|
| 1325 |
+
"3": 0.0025802549498621374,
|
| 1326 |
+
"4": 0.001066393441457573,
|
| 1327 |
+
"5": 0.0006842801643263859,
|
| 1328 |
+
"6": 0.0005846280212911855
|
| 1329 |
+
},
|
| 1330 |
+
"param_count": 12582912,
|
| 1331 |
+
"metadata": {}
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"layer_name": "language_model.model.layers.9.mlp.gate_proj",
|
| 1335 |
+
"sensitivities": {
|
| 1336 |
+
"2": 0.008234954628278501,
|
| 1337 |
+
"3": 0.0023020816879579797,
|
| 1338 |
+
"4": 0.0009838916897327483,
|
| 1339 |
+
"5": 0.000659322244246141,
|
| 1340 |
+
"6": 0.0005766963762046847
|
| 1341 |
+
},
|
| 1342 |
+
"param_count": 12582912,
|
| 1343 |
+
"metadata": {}
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"layer_name": "language_model.model.layers.9.linear_attn.out_proj",
|
| 1347 |
+
"sensitivities": {
|
| 1348 |
+
"2": 0.00832961519578627,
|
| 1349 |
+
"3": 0.002562749041923477,
|
| 1350 |
+
"4": 0.0010880117163954612,
|
| 1351 |
+
"5": 0.0006956170915752106,
|
| 1352 |
+
"6": 0.0005846008631730607
|
| 1353 |
+
},
|
| 1354 |
+
"param_count": 4194304,
|
| 1355 |
+
"metadata": {}
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"layer_name": "language_model.model.layers.9.linear_attn.in_proj_a",
|
| 1359 |
+
"sensitivities": {
|
| 1360 |
+
"2": 0.0006802463758504018,
|
| 1361 |
+
"3": 0.0005880223579879384,
|
| 1362 |
+
"4": 0.0005409358130539962,
|
| 1363 |
+
"5": 0.0005362698023721654,
|
| 1364 |
+
"6": 0.0005336923907937793
|
| 1365 |
+
},
|
| 1366 |
+
"param_count": 32768,
|
| 1367 |
+
"metadata": {}
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"layer_name": "language_model.model.layers.9.linear_attn.in_proj_b",
|
| 1371 |
+
"sensitivities": {
|
| 1372 |
+
"2": 0.0014235061535146087,
|
| 1373 |
+
"3": 0.0006368821220045598,
|
| 1374 |
+
"4": 0.0005888378109375481,
|
| 1375 |
+
"5": 0.0005439762911313059,
|
| 1376 |
+
"6": 0.0005406089003372472
|
| 1377 |
+
},
|
| 1378 |
+
"param_count": 32768,
|
| 1379 |
+
"metadata": {}
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"layer_name": "language_model.model.layers.9.linear_attn.in_proj_z",
|
| 1383 |
+
"sensitivities": {
|
| 1384 |
+
"2": 0.004212484794455425,
|
| 1385 |
+
"3": 0.0013634171718877042,
|
| 1386 |
+
"4": 0.0007595829677787455,
|
| 1387 |
+
"5": 0.000612661143046959,
|
| 1388 |
+
"6": 0.0005601726831325019
|
| 1389 |
+
},
|
| 1390 |
+
"param_count": 4194304,
|
| 1391 |
+
"metadata": {}
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"layer_name": "language_model.model.layers.9.linear_attn.in_proj_qkv",
|
| 1395 |
+
"sensitivities": {
|
| 1396 |
+
"2": 0.01137153763556853,
|
| 1397 |
+
"3": 0.003403712665506949,
|
| 1398 |
+
"4": 0.0013760158447742772,
|
| 1399 |
+
"5": 0.0007479057124631557,
|
| 1400 |
+
"6": 0.0006162834494413497
|
| 1401 |
+
},
|
| 1402 |
+
"param_count": 12582912,
|
| 1403 |
+
"metadata": {}
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"layer_name": "language_model.model.layers.8.mlp.up_proj",
|
| 1407 |
+
"sensitivities": {
|
| 1408 |
+
"2": 0.011812777263306392,
|
| 1409 |
+
"3": 0.0035483468673191965,
|
| 1410 |
+
"4": 0.0013862119255160603,
|
| 1411 |
+
"5": 0.0007884381296510886,
|
| 1412 |
+
"6": 0.0006269097642264873
|
| 1413 |
+
},
|
| 1414 |
+
"param_count": 12582912,
|
| 1415 |
+
"metadata": {}
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"layer_name": "language_model.model.layers.8.mlp.down_proj",
|
| 1419 |
+
"sensitivities": {
|
| 1420 |
+
"2": 0.010095076247428855,
|
| 1421 |
+
"3": 0.002905143960864128,
|
| 1422 |
+
"4": 0.0011411030269907012,
|
| 1423 |
+
"5": 0.0007183388127790143,
|
| 1424 |
+
"6": 0.0005977752128577171
|
| 1425 |
+
},
|
| 1426 |
+
"param_count": 12582912,
|
| 1427 |
+
"metadata": {}
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"layer_name": "language_model.model.layers.8.mlp.gate_proj",
|
| 1431 |
+
"sensitivities": {
|
| 1432 |
+
"2": 0.008105803276218163,
|
| 1433 |
+
"3": 0.002335677752853371,
|
| 1434 |
+
"4": 0.001002017873058018,
|
| 1435 |
+
"5": 0.0006891131373170841,
|
| 1436 |
+
"6": 0.000583234908238713
|
| 1437 |
+
},
|
| 1438 |
+
"param_count": 12582912,
|
| 1439 |
+
"metadata": {}
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"layer_name": "language_model.model.layers.8.linear_attn.out_proj",
|
| 1443 |
+
"sensitivities": {
|
| 1444 |
+
"2": 0.007769564292781676,
|
| 1445 |
+
"3": 0.0023873365425970405,
|
| 1446 |
+
"4": 0.0010491295918958106,
|
| 1447 |
+
"5": 0.0007060536160376311,
|
| 1448 |
+
"6": 0.0005976917091174982
|
| 1449 |
+
},
|
| 1450 |
+
"param_count": 4194304,
|
| 1451 |
+
"metadata": {}
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"layer_name": "language_model.model.layers.8.linear_attn.in_proj_a",
|
| 1455 |
+
"sensitivities": {
|
| 1456 |
+
"2": 0.0007147051170856381,
|
| 1457 |
+
"3": 0.0006110187217321558,
|
| 1458 |
+
"4": 0.0005620356951112626,
|
| 1459 |
+
"5": 0.0005511301957691709,
|
| 1460 |
+
"6": 0.000540618979963862
|
| 1461 |
+
},
|
| 1462 |
+
"param_count": 32768,
|
| 1463 |
+
"metadata": {}
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"layer_name": "language_model.model.layers.8.linear_attn.in_proj_b",
|
| 1467 |
+
"sensitivities": {
|
| 1468 |
+
"2": 0.0007433431407359118,
|
| 1469 |
+
"3": 0.000587749562934429,
|
| 1470 |
+
"4": 0.0005583754518738715,
|
| 1471 |
+
"5": 0.0005536570952244801,
|
| 1472 |
+
"6": 0.000541589858282047
|
| 1473 |
+
},
|
| 1474 |
+
"param_count": 32768,
|
| 1475 |
+
"metadata": {}
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"layer_name": "language_model.model.layers.8.linear_attn.in_proj_z",
|
| 1479 |
+
"sensitivities": {
|
| 1480 |
+
"2": 0.003614719108251544,
|
| 1481 |
+
"3": 0.0012706745680285774,
|
| 1482 |
+
"4": 0.0007406099848594749,
|
| 1483 |
+
"5": 0.0005969127087155357,
|
| 1484 |
+
"6": 0.0005754416949154498
|
| 1485 |
+
},
|
| 1486 |
+
"param_count": 4194304,
|
| 1487 |
+
"metadata": {}
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"layer_name": "language_model.model.layers.8.linear_attn.in_proj_qkv",
|
| 1491 |
+
"sensitivities": {
|
| 1492 |
+
"2": 0.03481895992687593,
|
| 1493 |
+
"3": 0.0028332900425690846,
|
| 1494 |
+
"4": 0.0012336299102268338,
|
| 1495 |
+
"5": 0.0007504084751417395,
|
| 1496 |
+
"6": 0.000613556703683571
|
| 1497 |
+
},
|
| 1498 |
+
"param_count": 12582912,
|
| 1499 |
+
"metadata": {}
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"layer_name": "language_model.model.layers.7.mlp.up_proj",
|
| 1503 |
+
"sensitivities": {
|
| 1504 |
+
"2": 0.012778486358001828,
|
| 1505 |
+
"3": 0.004118884285465659,
|
| 1506 |
+
"4": 0.001445508195805208,
|
| 1507 |
+
"5": 0.000816462587560333,
|
| 1508 |
+
"6": 0.0006318808688471714
|
| 1509 |
+
},
|
| 1510 |
+
"param_count": 12582912,
|
| 1511 |
+
"metadata": {}
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"layer_name": "language_model.model.layers.7.mlp.down_proj",
|
| 1515 |
+
"sensitivities": {
|
| 1516 |
+
"2": 0.010875919368118048,
|
| 1517 |
+
"3": 0.0031571896882572523,
|
| 1518 |
+
"4": 0.0011847067326016258,
|
| 1519 |
+
"5": 0.0007436109760116475,
|
| 1520 |
+
"6": 0.0006064725991260881
|
| 1521 |
+
},
|
| 1522 |
+
"param_count": 12582912,
|
| 1523 |
+
"metadata": {}
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"layer_name": "language_model.model.layers.7.mlp.gate_proj",
|
| 1527 |
+
"sensitivities": {
|
| 1528 |
+
"2": 0.008173152329012131,
|
| 1529 |
+
"3": 0.0022785763467254583,
|
| 1530 |
+
"4": 0.0010116253646022717,
|
| 1531 |
+
"5": 0.0006819089261019448,
|
| 1532 |
+
"6": 0.0005953435344660344
|
| 1533 |
+
},
|
| 1534 |
+
"param_count": 12582912,
|
| 1535 |
+
"metadata": {}
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"layer_name": "language_model.model.layers.7.self_attn.o_proj",
|
| 1539 |
+
"sensitivities": {
|
| 1540 |
+
"2": 0.024533623616055895,
|
| 1541 |
+
"3": 0.006196063516351084,
|
| 1542 |
+
"4": 0.001970649887274097,
|
| 1543 |
+
"5": 0.0009310314580943668,
|
| 1544 |
+
"6": 0.0006725482204880487
|
| 1545 |
+
},
|
| 1546 |
+
"param_count": 4194304,
|
| 1547 |
+
"metadata": {}
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"layer_name": "language_model.model.layers.7.self_attn.v_proj",
|
| 1551 |
+
"sensitivities": {
|
| 1552 |
+
"2": 0.01883278515500327,
|
| 1553 |
+
"3": 0.006277569259206454,
|
| 1554 |
+
"4": 0.0019011234059386577,
|
| 1555 |
+
"5": 0.0008133406681736233,
|
| 1556 |
+
"6": 0.0006298560583672952
|
| 1557 |
+
},
|
| 1558 |
+
"param_count": 1048576,
|
| 1559 |
+
"metadata": {}
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"layer_name": "language_model.model.layers.7.self_attn.k_proj",
|
| 1563 |
+
"sensitivities": {
|
| 1564 |
+
"2": 0.003162360151084916,
|
| 1565 |
+
"3": 0.0011603820142530215,
|
| 1566 |
+
"4": 0.0007000078264051505,
|
| 1567 |
+
"5": 0.0006039587127209719,
|
| 1568 |
+
"6": 0.000581321763699331
|
| 1569 |
+
},
|
| 1570 |
+
"param_count": 1048576,
|
| 1571 |
+
"metadata": {}
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"layer_name": "language_model.model.layers.7.self_attn.q_proj",
|
| 1575 |
+
"sensitivities": {
|
| 1576 |
+
"2": 0.010100869743231064,
|
| 1577 |
+
"3": 0.0024180926023594416,
|
| 1578 |
+
"4": 0.001061241258867085,
|
| 1579 |
+
"5": 0.0007230274338022961,
|
| 1580 |
+
"6": 0.0006375790623375602
|
| 1581 |
+
},
|
| 1582 |
+
"param_count": 8388608,
|
| 1583 |
+
"metadata": {}
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"layer_name": "language_model.model.layers.6.mlp.up_proj",
|
| 1587 |
+
"sensitivities": {
|
| 1588 |
+
"2": 0.014851690929693481,
|
| 1589 |
+
"3": 0.004340703725271548,
|
| 1590 |
+
"4": 0.0015263961771173247,
|
| 1591 |
+
"5": 0.0008469799395243172,
|
| 1592 |
+
"6": 0.0006638854635336126
|
| 1593 |
+
},
|
| 1594 |
+
"param_count": 12582912,
|
| 1595 |
+
"metadata": {}
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"layer_name": "language_model.model.layers.6.mlp.down_proj",
|
| 1599 |
+
"sensitivities": {
|
| 1600 |
+
"2": 0.012329060458190119,
|
| 1601 |
+
"3": 0.0035062268070760183,
|
| 1602 |
+
"4": 0.0013031964226684067,
|
| 1603 |
+
"5": 0.0007835034957679454,
|
| 1604 |
+
"6": 0.0006416480652357374
|
| 1605 |
+
},
|
| 1606 |
+
"param_count": 12582912,
|
| 1607 |
+
"metadata": {}
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"layer_name": "language_model.model.layers.6.mlp.gate_proj",
|
| 1611 |
+
"sensitivities": {
|
| 1612 |
+
"2": 0.011155722507586082,
|
| 1613 |
+
"3": 0.0028249319851359664,
|
| 1614 |
+
"4": 0.0011791956543068711,
|
| 1615 |
+
"5": 0.000755706494601327,
|
| 1616 |
+
"6": 0.000640374508293462
|
| 1617 |
+
},
|
| 1618 |
+
"param_count": 12582912,
|
| 1619 |
+
"metadata": {}
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"layer_name": "language_model.model.layers.6.linear_attn.out_proj",
|
| 1623 |
+
"sensitivities": {
|
| 1624 |
+
"2": 0.01423639280255884,
|
| 1625 |
+
"3": 0.004122252110391855,
|
| 1626 |
+
"4": 0.001393909499408134,
|
| 1627 |
+
"5": 0.0008301619260843532,
|
| 1628 |
+
"6": 0.0006609533878266424
|
| 1629 |
+
},
|
| 1630 |
+
"param_count": 4194304,
|
| 1631 |
+
"metadata": {}
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"layer_name": "language_model.model.layers.6.linear_attn.in_proj_a",
|
| 1635 |
+
"sensitivities": {
|
| 1636 |
+
"2": 0.0008324757509399205,
|
| 1637 |
+
"3": 0.0006505880652791044,
|
| 1638 |
+
"4": 0.0005931716247384126,
|
| 1639 |
+
"5": 0.0005826339808360596,
|
| 1640 |
+
"6": 0.000584431477667143
|
| 1641 |
+
},
|
| 1642 |
+
"param_count": 32768,
|
| 1643 |
+
"metadata": {}
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"layer_name": "language_model.model.layers.6.linear_attn.in_proj_b",
|
| 1647 |
+
"sensitivities": {
|
| 1648 |
+
"2": 0.0008604249802980727,
|
| 1649 |
+
"3": 0.0006834920453305434,
|
| 1650 |
+
"4": 0.0006201505542170102,
|
| 1651 |
+
"5": 0.0005876055183762219,
|
| 1652 |
+
"6": 0.000595989764406113
|
| 1653 |
+
},
|
| 1654 |
+
"param_count": 32768,
|
| 1655 |
+
"metadata": {}
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"layer_name": "language_model.model.layers.6.linear_attn.in_proj_z",
|
| 1659 |
+
"sensitivities": {
|
| 1660 |
+
"2": 0.005671261734581397,
|
| 1661 |
+
"3": 0.001567219995195046,
|
| 1662 |
+
"4": 0.0008609001603569292,
|
| 1663 |
+
"5": 0.0006723137400210059,
|
| 1664 |
+
"6": 0.0006132452084178416
|
| 1665 |
+
},
|
| 1666 |
+
"param_count": 4194304,
|
| 1667 |
+
"metadata": {}
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"layer_name": "language_model.model.layers.6.linear_attn.in_proj_qkv",
|
| 1671 |
+
"sensitivities": {
|
| 1672 |
+
"2": 0.021044176964399714,
|
| 1673 |
+
"3": 0.004874579164606985,
|
| 1674 |
+
"4": 0.0013679137409781106,
|
| 1675 |
+
"5": 0.0008129978535483436,
|
| 1676 |
+
"6": 0.0006661872248514555
|
| 1677 |
+
},
|
| 1678 |
+
"param_count": 12582912,
|
| 1679 |
+
"metadata": {}
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"layer_name": "language_model.model.layers.5.mlp.up_proj",
|
| 1683 |
+
"sensitivities": {
|
| 1684 |
+
"2": 0.01312018238240853,
|
| 1685 |
+
"3": 0.004335268065915443,
|
| 1686 |
+
"4": 0.0015145789766393136,
|
| 1687 |
+
"5": 0.0008729634864721447,
|
| 1688 |
+
"6": 0.0006741304696333827
|
| 1689 |
+
},
|
| 1690 |
+
"param_count": 12582912,
|
| 1691 |
+
"metadata": {}
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"layer_name": "language_model.model.layers.5.mlp.down_proj",
|
| 1695 |
+
"sensitivities": {
|
| 1696 |
+
"2": 0.012572033558778154,
|
| 1697 |
+
"3": 0.003419627396700283,
|
| 1698 |
+
"4": 0.0012641735993383918,
|
| 1699 |
+
"5": 0.0008190781033287445,
|
| 1700 |
+
"6": 0.0006624283887504134
|
| 1701 |
+
},
|
| 1702 |
+
"param_count": 12582912,
|
| 1703 |
+
"metadata": {}
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"layer_name": "language_model.model.layers.5.mlp.gate_proj",
|
| 1707 |
+
"sensitivities": {
|
| 1708 |
+
"2": 0.01041741516867963,
|
| 1709 |
+
"3": 0.0027252345219797767,
|
| 1710 |
+
"4": 0.0011102490437527497,
|
| 1711 |
+
"5": 0.0007528255616004268,
|
| 1712 |
+
"6": 0.0006544071408522237
|
| 1713 |
+
},
|
| 1714 |
+
"param_count": 12582912,
|
| 1715 |
+
"metadata": {}
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"layer_name": "language_model.model.layers.5.linear_attn.out_proj",
|
| 1719 |
+
"sensitivities": {
|
| 1720 |
+
"2": 0.011017731497607505,
|
| 1721 |
+
"3": 0.0034647406888931678,
|
| 1722 |
+
"4": 0.001350575104879681,
|
| 1723 |
+
"5": 0.0008285286894533783,
|
| 1724 |
+
"6": 0.0006665428245469229
|
| 1725 |
+
},
|
| 1726 |
+
"param_count": 4194304,
|
| 1727 |
+
"metadata": {}
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"layer_name": "language_model.model.layers.5.linear_attn.in_proj_a",
|
| 1731 |
+
"sensitivities": {
|
| 1732 |
+
"2": 0.0008551792016078252,
|
| 1733 |
+
"3": 0.0006563680969217481,
|
| 1734 |
+
"4": 0.000619638953745986,
|
| 1735 |
+
"5": 0.0006085646861417141,
|
| 1736 |
+
"6": 0.0006044759502401575
|
| 1737 |
+
},
|
| 1738 |
+
"param_count": 32768,
|
| 1739 |
+
"metadata": {}
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"layer_name": "language_model.model.layers.5.linear_attn.in_proj_b",
|
| 1743 |
+
"sensitivities": {
|
| 1744 |
+
"2": 0.0008767704063454099,
|
| 1745 |
+
"3": 0.0006979444545625787,
|
| 1746 |
+
"4": 0.0006383983454725239,
|
| 1747 |
+
"5": 0.0006122347009901811,
|
| 1748 |
+
"6": 0.0006040568065751964
|
| 1749 |
+
},
|
| 1750 |
+
"param_count": 32768,
|
| 1751 |
+
"metadata": {}
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"layer_name": "language_model.model.layers.5.linear_attn.in_proj_z",
|
| 1755 |
+
"sensitivities": {
|
| 1756 |
+
"2": 0.007329763170370522,
|
| 1757 |
+
"3": 0.0019751639459476187,
|
| 1758 |
+
"4": 0.0009540419365900258,
|
| 1759 |
+
"5": 0.0007089421766674301,
|
| 1760 |
+
"6": 0.0006324446073146343
|
| 1761 |
+
},
|
| 1762 |
+
"param_count": 4194304,
|
| 1763 |
+
"metadata": {}
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"layer_name": "language_model.model.layers.5.linear_attn.in_proj_qkv",
|
| 1767 |
+
"sensitivities": {
|
| 1768 |
+
"2": 0.018860064999898896,
|
| 1769 |
+
"3": 0.004769341636953565,
|
| 1770 |
+
"4": 0.0013334317506329778,
|
| 1771 |
+
"5": 0.0008236669364123372,
|
| 1772 |
+
"6": 0.0006886378541821614
|
| 1773 |
+
},
|
| 1774 |
+
"param_count": 12582912,
|
| 1775 |
+
"metadata": {}
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"layer_name": "language_model.model.layers.4.mlp.up_proj",
|
| 1779 |
+
"sensitivities": {
|
| 1780 |
+
"2": 0.013942821737146005,
|
| 1781 |
+
"3": 0.004070183177342794,
|
| 1782 |
+
"4": 0.0015625901699725848,
|
| 1783 |
+
"5": 0.000862095345534423,
|
| 1784 |
+
"6": 0.0006955904609640129
|
| 1785 |
+
},
|
| 1786 |
+
"param_count": 12582912,
|
| 1787 |
+
"metadata": {}
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"layer_name": "language_model.model.layers.4.mlp.down_proj",
|
| 1791 |
+
"sensitivities": {
|
| 1792 |
+
"2": 0.013054531921322146,
|
| 1793 |
+
"3": 0.003575960741727613,
|
| 1794 |
+
"4": 0.0013440125609728664,
|
| 1795 |
+
"5": 0.0008165046153105019,
|
| 1796 |
+
"6": 0.0006804468721384183
|
| 1797 |
+
},
|
| 1798 |
+
"param_count": 12582912,
|
| 1799 |
+
"metadata": {}
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"layer_name": "language_model.model.layers.4.mlp.gate_proj",
|
| 1803 |
+
"sensitivities": {
|
| 1804 |
+
"2": 0.010083985291809464,
|
| 1805 |
+
"3": 0.002680814837124975,
|
| 1806 |
+
"4": 0.0012012788004843362,
|
| 1807 |
+
"5": 0.0007784871747086678,
|
| 1808 |
+
"6": 0.0006645032384161217
|
| 1809 |
+
},
|
| 1810 |
+
"param_count": 12582912,
|
| 1811 |
+
"metadata": {}
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"layer_name": "language_model.model.layers.4.linear_attn.out_proj",
|
| 1815 |
+
"sensitivities": {
|
| 1816 |
+
"2": 0.0102939090381066,
|
| 1817 |
+
"3": 0.003040899983413207,
|
| 1818 |
+
"4": 0.001273585400970963,
|
| 1819 |
+
"5": 0.0008076543987651045,
|
| 1820 |
+
"6": 0.000675606049602114
|
| 1821 |
+
},
|
| 1822 |
+
"param_count": 4194304,
|
| 1823 |
+
"metadata": {}
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"layer_name": "language_model.model.layers.4.linear_attn.in_proj_a",
|
| 1827 |
+
"sensitivities": {
|
| 1828 |
+
"2": 0.0008688874865280619,
|
| 1829 |
+
"3": 0.0006868385219907699,
|
| 1830 |
+
"4": 0.0006339784789209565,
|
| 1831 |
+
"5": 0.000622270085538427,
|
| 1832 |
+
"6": 0.0006230382265736504
|
| 1833 |
+
},
|
| 1834 |
+
"param_count": 32768,
|
| 1835 |
+
"metadata": {}
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"layer_name": "language_model.model.layers.4.linear_attn.in_proj_b",
|
| 1839 |
+
"sensitivities": {
|
| 1840 |
+
"2": 0.0009992057166527957,
|
| 1841 |
+
"3": 0.0007282946996080378,
|
| 1842 |
+
"4": 0.0006512685019212464,
|
| 1843 |
+
"5": 0.000633465586361126,
|
| 1844 |
+
"6": 0.0006264543717406923
|
| 1845 |
+
},
|
| 1846 |
+
"param_count": 32768,
|
| 1847 |
+
"metadata": {}
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"layer_name": "language_model.model.layers.4.linear_attn.in_proj_z",
|
| 1851 |
+
"sensitivities": {
|
| 1852 |
+
"2": 0.007340592409794529,
|
| 1853 |
+
"3": 0.001825619498655821,
|
| 1854 |
+
"4": 0.0009170382982119918,
|
| 1855 |
+
"5": 0.0007215852674562484,
|
| 1856 |
+
"6": 0.000655765561532462
|
| 1857 |
+
},
|
| 1858 |
+
"param_count": 4194304,
|
| 1859 |
+
"metadata": {}
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"layer_name": "language_model.model.layers.4.linear_attn.in_proj_qkv",
|
| 1863 |
+
"sensitivities": {
|
| 1864 |
+
"2": 0.009975035345026603,
|
| 1865 |
+
"3": 0.003287324382351168,
|
| 1866 |
+
"4": 0.001244251444101489,
|
| 1867 |
+
"5": 0.0007899436956601372,
|
| 1868 |
+
"6": 0.0006714652202693591
|
| 1869 |
+
},
|
| 1870 |
+
"param_count": 12582912,
|
| 1871 |
+
"metadata": {}
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"layer_name": "language_model.model.layers.3.mlp.up_proj",
|
| 1875 |
+
"sensitivities": {
|
| 1876 |
+
"2": 0.013523875056610754,
|
| 1877 |
+
"3": 0.004135167269851081,
|
| 1878 |
+
"4": 0.0014811170597871144,
|
| 1879 |
+
"5": 0.0008584310326114064,
|
| 1880 |
+
"6": 0.0007105843069439288
|
| 1881 |
+
},
|
| 1882 |
+
"param_count": 12582912,
|
| 1883 |
+
"metadata": {}
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"layer_name": "language_model.model.layers.3.mlp.down_proj",
|
| 1887 |
+
"sensitivities": {
|
| 1888 |
+
"2": 0.01369342571706511,
|
| 1889 |
+
"3": 0.0035236425440719663,
|
| 1890 |
+
"4": 0.0013495567012190197,
|
| 1891 |
+
"5": 0.0008396878498994435,
|
| 1892 |
+
"6": 0.0006908999475854216
|
| 1893 |
+
},
|
| 1894 |
+
"param_count": 12582912,
|
| 1895 |
+
"metadata": {}
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"layer_name": "language_model.model.layers.3.mlp.gate_proj",
|
| 1899 |
+
"sensitivities": {
|
| 1900 |
+
"2": 0.010460648850615447,
|
| 1901 |
+
"3": 0.0028929833618652387,
|
| 1902 |
+
"4": 0.001210855414683465,
|
| 1903 |
+
"5": 0.0007948397245248392,
|
| 1904 |
+
"6": 0.0006870741660047012
|
| 1905 |
+
},
|
| 1906 |
+
"param_count": 12582912,
|
| 1907 |
+
"metadata": {}
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"layer_name": "language_model.model.layers.3.self_attn.o_proj",
|
| 1911 |
+
"sensitivities": {
|
| 1912 |
+
"2": 0.018217159600074712,
|
| 1913 |
+
"3": 0.004563965582444022,
|
| 1914 |
+
"4": 0.001636317749216687,
|
| 1915 |
+
"5": 0.0009095968222633625,
|
| 1916 |
+
"6": 0.0007155357707233634
|
| 1917 |
+
},
|
| 1918 |
+
"param_count": 4194304,
|
| 1919 |
+
"metadata": {}
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"layer_name": "language_model.model.layers.3.self_attn.v_proj",
|
| 1923 |
+
"sensitivities": {
|
| 1924 |
+
"2": 0.013098290966202816,
|
| 1925 |
+
"3": 0.006201225400824721,
|
| 1926 |
+
"4": 0.001436049218076126,
|
| 1927 |
+
"5": 0.0008338717210184162,
|
| 1928 |
+
"6": 0.0007284416284771092
|
| 1929 |
+
},
|
| 1930 |
+
"param_count": 1048576,
|
| 1931 |
+
"metadata": {}
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"layer_name": "language_model.model.layers.3.self_attn.k_proj",
|
| 1935 |
+
"sensitivities": {
|
| 1936 |
+
"2": 0.0029975675837098,
|
| 1937 |
+
"3": 0.0011011197627036988,
|
| 1938 |
+
"4": 0.000796304409353373,
|
| 1939 |
+
"5": 0.0006765657029366897,
|
| 1940 |
+
"6": 0.0006609956035390496
|
| 1941 |
+
},
|
| 1942 |
+
"param_count": 1048576,
|
| 1943 |
+
"metadata": {}
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"layer_name": "language_model.model.layers.3.self_attn.q_proj",
|
| 1947 |
+
"sensitivities": {
|
| 1948 |
+
"2": 0.01557940295121322,
|
| 1949 |
+
"3": 0.00246271644209628,
|
| 1950 |
+
"4": 0.0011554938719200436,
|
| 1951 |
+
"5": 0.0007947645766156105,
|
| 1952 |
+
"6": 0.000690136695993715
|
| 1953 |
+
},
|
| 1954 |
+
"param_count": 8388608,
|
| 1955 |
+
"metadata": {}
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"layer_name": "language_model.model.layers.2.mlp.up_proj",
|
| 1959 |
+
"sensitivities": {
|
| 1960 |
+
"2": 0.012601931763735289,
|
| 1961 |
+
"3": 0.004050463384677035,
|
| 1962 |
+
"4": 0.0015326273993802413,
|
| 1963 |
+
"5": 0.0008913539028678011,
|
| 1964 |
+
"6": 0.000725003542053552
|
| 1965 |
+
},
|
| 1966 |
+
"param_count": 12582912,
|
| 1967 |
+
"metadata": {}
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"layer_name": "language_model.model.layers.2.mlp.down_proj",
|
| 1971 |
+
"sensitivities": {
|
| 1972 |
+
"2": 0.015134855813812464,
|
| 1973 |
+
"3": 0.004025612470286433,
|
| 1974 |
+
"4": 0.0015561307882308029,
|
| 1975 |
+
"5": 0.0008897500635309067,
|
| 1976 |
+
"6": 0.0007262391409312841
|
| 1977 |
+
},
|
| 1978 |
+
"param_count": 12582912,
|
| 1979 |
+
"metadata": {}
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"layer_name": "language_model.model.layers.2.mlp.gate_proj",
|
| 1983 |
+
"sensitivities": {
|
| 1984 |
+
"2": 0.011826593952719122,
|
| 1985 |
+
"3": 0.0033633083755072826,
|
| 1986 |
+
"4": 0.0012777378845688265,
|
| 1987 |
+
"5": 0.0008447689724562224,
|
| 1988 |
+
"6": 0.0007181479243930274
|
| 1989 |
+
},
|
| 1990 |
+
"param_count": 12582912,
|
| 1991 |
+
"metadata": {}
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"layer_name": "language_model.model.layers.2.linear_attn.out_proj",
|
| 1995 |
+
"sensitivities": {
|
| 1996 |
+
"2": 0.029565126906769972,
|
| 1997 |
+
"3": 0.006916260569899653,
|
| 1998 |
+
"4": 0.002184954377298709,
|
| 1999 |
+
"5": 0.0010938635447625227,
|
| 2000 |
+
"6": 0.0007869263802907275
|
| 2001 |
+
},
|
| 2002 |
+
"param_count": 4194304,
|
| 2003 |
+
"metadata": {}
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"layer_name": "language_model.model.layers.2.linear_attn.in_proj_a",
|
| 2007 |
+
"sensitivities": {
|
| 2008 |
+
"2": 0.0011151844482810702,
|
| 2009 |
+
"3": 0.0007697061482758727,
|
| 2010 |
+
"4": 0.0006842241618869593,
|
| 2011 |
+
"5": 0.0006782368982385378,
|
| 2012 |
+
"6": 0.0006479376788774971
|
| 2013 |
+
},
|
| 2014 |
+
"param_count": 32768,
|
| 2015 |
+
"metadata": {}
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"layer_name": "language_model.model.layers.2.linear_attn.in_proj_b",
|
| 2019 |
+
"sensitivities": {
|
| 2020 |
+
"2": 0.0011326321206676464,
|
| 2021 |
+
"3": 0.0008187466155504808,
|
| 2022 |
+
"4": 0.0007026880424139866,
|
| 2023 |
+
"5": 0.0006807700956414919,
|
| 2024 |
+
"6": 0.0006712008956431722
|
| 2025 |
+
},
|
| 2026 |
+
"param_count": 32768,
|
| 2027 |
+
"metadata": {}
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"layer_name": "language_model.model.layers.2.linear_attn.in_proj_z",
|
| 2031 |
+
"sensitivities": {
|
| 2032 |
+
"2": 0.00699581209240326,
|
| 2033 |
+
"3": 0.0018451928796518284,
|
| 2034 |
+
"4": 0.0009942074611899443,
|
| 2035 |
+
"5": 0.000743712056040143,
|
| 2036 |
+
"6": 0.0006938851523348907
|
| 2037 |
+
},
|
| 2038 |
+
"param_count": 4194304,
|
| 2039 |
+
"metadata": {}
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"layer_name": "language_model.model.layers.2.linear_attn.in_proj_qkv",
|
| 2043 |
+
"sensitivities": {
|
| 2044 |
+
"2": 0.015811362594831735,
|
| 2045 |
+
"3": 0.004006082000463114,
|
| 2046 |
+
"4": 0.001589768893609289,
|
| 2047 |
+
"5": 0.0009063915398049479,
|
| 2048 |
+
"6": 0.0007461178623392092
|
| 2049 |
+
},
|
| 2050 |
+
"param_count": 12582912,
|
| 2051 |
+
"metadata": {}
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"layer_name": "language_model.model.layers.1.mlp.up_proj",
|
| 2055 |
+
"sensitivities": {
|
| 2056 |
+
"2": 0.010221001779427752,
|
| 2057 |
+
"3": 0.0033970397756396173,
|
| 2058 |
+
"4": 0.0014251155116653536,
|
| 2059 |
+
"5": 0.0008711083901289385,
|
| 2060 |
+
"6": 0.0007333647160218485
|
| 2061 |
+
},
|
| 2062 |
+
"param_count": 12582912,
|
| 2063 |
+
"metadata": {}
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"layer_name": "language_model.model.layers.1.mlp.down_proj",
|
| 2067 |
+
"sensitivities": {
|
| 2068 |
+
"2": 0.011487890680048926,
|
| 2069 |
+
"3": 0.0033733720726255947,
|
| 2070 |
+
"4": 0.001295011108595645,
|
| 2071 |
+
"5": 0.0008564010840927949,
|
| 2072 |
+
"6": 0.000736263552729118
|
| 2073 |
+
},
|
| 2074 |
+
"param_count": 12582912,
|
| 2075 |
+
"metadata": {}
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"layer_name": "language_model.model.layers.1.mlp.gate_proj",
|
| 2079 |
+
"sensitivities": {
|
| 2080 |
+
"2": 0.011139082256704569,
|
| 2081 |
+
"3": 0.00325369113124907,
|
| 2082 |
+
"4": 0.0012731527046222861,
|
| 2083 |
+
"5": 0.0008339915354251085,
|
| 2084 |
+
"6": 0.0007208840500728305
|
| 2085 |
+
},
|
| 2086 |
+
"param_count": 12582912,
|
| 2087 |
+
"metadata": {}
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"layer_name": "language_model.model.layers.1.linear_attn.out_proj",
|
| 2091 |
+
"sensitivities": {
|
| 2092 |
+
"2": 0.012037927575875074,
|
| 2093 |
+
"3": 0.003623796243725034,
|
| 2094 |
+
"4": 0.0013797276284700881,
|
| 2095 |
+
"5": 0.0008657840602002883,
|
| 2096 |
+
"6": 0.0007402556966553675
|
| 2097 |
+
},
|
| 2098 |
+
"param_count": 4194304,
|
| 2099 |
+
"metadata": {}
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"layer_name": "language_model.model.layers.1.linear_attn.in_proj_a",
|
| 2103 |
+
"sensitivities": {
|
| 2104 |
+
"2": 0.0014032348638769083,
|
| 2105 |
+
"3": 0.0007628179797999716,
|
| 2106 |
+
"4": 0.0007194990478941085,
|
| 2107 |
+
"5": 0.000689315559308549,
|
| 2108 |
+
"6": 0.0006716263475633847
|
| 2109 |
+
},
|
| 2110 |
+
"param_count": 32768,
|
| 2111 |
+
"metadata": {}
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"layer_name": "language_model.model.layers.1.linear_attn.in_proj_b",
|
| 2115 |
+
"sensitivities": {
|
| 2116 |
+
"2": 0.0011315337481695071,
|
| 2117 |
+
"3": 0.0007869832079450134,
|
| 2118 |
+
"4": 0.0007085167205029089,
|
| 2119 |
+
"5": 0.0006925270821132775,
|
| 2120 |
+
"6": 0.0006903271817767139
|
| 2121 |
+
},
|
| 2122 |
+
"param_count": 32768,
|
| 2123 |
+
"metadata": {}
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"layer_name": "language_model.model.layers.1.linear_attn.in_proj_z",
|
| 2127 |
+
"sensitivities": {
|
| 2128 |
+
"2": 0.006613240620936267,
|
| 2129 |
+
"3": 0.002027414180095851,
|
| 2130 |
+
"4": 0.0010317080571743038,
|
| 2131 |
+
"5": 0.0007825390287810782,
|
| 2132 |
+
"6": 0.0007113759214310752
|
| 2133 |
+
},
|
| 2134 |
+
"param_count": 4194304,
|
| 2135 |
+
"metadata": {}
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"layer_name": "language_model.model.layers.1.linear_attn.in_proj_qkv",
|
| 2139 |
+
"sensitivities": {
|
| 2140 |
+
"2": 0.014970489151892252,
|
| 2141 |
+
"3": 0.004721548803596913,
|
| 2142 |
+
"4": 0.0016168094249830272,
|
| 2143 |
+
"5": 0.00093699258104607,
|
| 2144 |
+
"6": 0.0007505288364579125
|
| 2145 |
+
},
|
| 2146 |
+
"param_count": 12582912,
|
| 2147 |
+
"metadata": {}
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"layer_name": "language_model.model.layers.0.mlp.up_proj",
|
| 2151 |
+
"sensitivities": {
|
| 2152 |
+
"2": 0.009800014158827253,
|
| 2153 |
+
"3": 0.0029476434016639055,
|
| 2154 |
+
"4": 0.001265500812830093,
|
| 2155 |
+
"5": 0.0008314645674545318,
|
| 2156 |
+
"6": 0.0007488028944256561
|
| 2157 |
+
},
|
| 2158 |
+
"param_count": 12582912,
|
| 2159 |
+
"metadata": {}
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"layer_name": "language_model.model.layers.0.mlp.down_proj",
|
| 2163 |
+
"sensitivities": {
|
| 2164 |
+
"2": 0.01360617058041195,
|
| 2165 |
+
"3": 0.0037330198028939776,
|
| 2166 |
+
"4": 0.0015210995140175025,
|
| 2167 |
+
"5": 0.0009026831327597998,
|
| 2168 |
+
"6": 0.0007709821420576191
|
| 2169 |
+
},
|
| 2170 |
+
"param_count": 12582912,
|
| 2171 |
+
"metadata": {}
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"layer_name": "language_model.model.layers.0.mlp.gate_proj",
|
| 2175 |
+
"sensitivities": {
|
| 2176 |
+
"2": 0.008524882311273055,
|
| 2177 |
+
"3": 0.0023337135498877615,
|
| 2178 |
+
"4": 0.0010995213002994813,
|
| 2179 |
+
"5": 0.0008411641756538302,
|
| 2180 |
+
"6": 0.0007258908663061447
|
| 2181 |
+
},
|
| 2182 |
+
"param_count": 12582912,
|
| 2183 |
+
"metadata": {}
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"layer_name": "language_model.model.layers.0.linear_attn.out_proj",
|
| 2187 |
+
"sensitivities": {
|
| 2188 |
+
"2": 0.2411517578487595,
|
| 2189 |
+
"3": 0.057953598986690245,
|
| 2190 |
+
"4": 0.013033893328004828,
|
| 2191 |
+
"5": 0.003675751887688724,
|
| 2192 |
+
"6": 0.0016291814075278428
|
| 2193 |
+
},
|
| 2194 |
+
"param_count": 4194304,
|
| 2195 |
+
"metadata": {}
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"layer_name": "language_model.model.layers.0.linear_attn.in_proj_a",
|
| 2199 |
+
"sensitivities": {
|
| 2200 |
+
"2": 0.0009306360546664413,
|
| 2201 |
+
"3": 0.0007173148126942882,
|
| 2202 |
+
"4": 0.0007155387381014103,
|
| 2203 |
+
"5": 0.0006989434587012511,
|
| 2204 |
+
"6": 0.0006846034254219072
|
| 2205 |
+
},
|
| 2206 |
+
"param_count": 32768,
|
| 2207 |
+
"metadata": {}
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"layer_name": "language_model.model.layers.0.linear_attn.in_proj_b",
|
| 2211 |
+
"sensitivities": {
|
| 2212 |
+
"2": 0.0013787257266812958,
|
| 2213 |
+
"3": 0.000845362421387108,
|
| 2214 |
+
"4": 0.0007493734998812821,
|
| 2215 |
+
"5": 0.0007040374427257726,
|
| 2216 |
+
"6": 0.0007004143111165225
|
| 2217 |
+
},
|
| 2218 |
+
"param_count": 32768,
|
| 2219 |
+
"metadata": {}
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"layer_name": "language_model.model.layers.0.linear_attn.in_proj_z",
|
| 2223 |
+
"sensitivities": {
|
| 2224 |
+
"2": 0.022454326661924522,
|
| 2225 |
+
"3": 0.004593782073546511,
|
| 2226 |
+
"4": 0.0016777982552108976,
|
| 2227 |
+
"5": 0.0009666465151288625,
|
| 2228 |
+
"6": 0.0007862018719606567
|
| 2229 |
+
},
|
| 2230 |
+
"param_count": 4194304,
|
| 2231 |
+
"metadata": {}
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"layer_name": "language_model.model.layers.0.linear_attn.in_proj_qkv",
|
| 2235 |
+
"sensitivities": {
|
| 2236 |
+
"2": 0.020587159456529964,
|
| 2237 |
+
"3": 0.004562516948984315,
|
| 2238 |
+
"4": 0.0018581711665319744,
|
| 2239 |
+
"5": 0.0009564283991494449,
|
| 2240 |
+
"6": 0.0007860938312660437
|
| 2241 |
+
},
|
| 2242 |
+
"param_count": 12582912,
|
| 2243 |
+
"metadata": {}
|
| 2244 |
+
}
|
| 2245 |
+
]
|
| 2246 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523
|
| 3 |
+
size 19989325
|
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 |
+
"split_special_tokens": false,
|
| 27 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 28 |
+
"tool_parser_type": "qwen3_coder",
|
| 29 |
+
"unk_token": null,
|
| 30 |
+
"video_token": "<|video_pad|>",
|
| 31 |
+
"vision_bos_token": "<|vision_start|>",
|
| 32 |
+
"vision_eos_token": "<|vision_end|>"
|
| 33 |
+
}
|