Hugging Face's logo

INC4AI
/
GLM-5-int4-mixed-AutoRound

Text Generation
Safetensors
glm_moe_dsa
conversational
4-bit precision
auto-round
Model card Files
xet
Community
1
INC4AI commited on
Commit
3a6e223
·
verified ·
1 Parent(s): a6dfa65

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +2 -0
  2. chat_template.jinja +86 -0
  3. config.json +2726 -0
  4. model-00001-of-00076.safetensors +3 -0
  5. model-00002-of-00076.safetensors +3 -0
  6. model-00003-of-00076.safetensors +3 -0
  7. model-00004-of-00076.safetensors +3 -0
  8. model-00005-of-00076.safetensors +3 -0
  9. model-00006-of-00076.safetensors +3 -0
  10. model-00007-of-00076.safetensors +3 -0
  11. model-00008-of-00076.safetensors +3 -0
  12. model-00009-of-00076.safetensors +3 -0
  13. model-00010-of-00076.safetensors +3 -0
  14. model-00011-of-00076.safetensors +3 -0
  15. model-00012-of-00076.safetensors +3 -0
  16. model-00013-of-00076.safetensors +3 -0
  17. model-00014-of-00076.safetensors +3 -0
  18. model-00015-of-00076.safetensors +3 -0
  19. model-00016-of-00076.safetensors +3 -0
  20. model-00017-of-00076.safetensors +3 -0
  21. model-00018-of-00076.safetensors +3 -0
  22. model-00019-of-00076.safetensors +3 -0
  23. model-00020-of-00076.safetensors +3 -0
  24. model-00021-of-00076.safetensors +3 -0
  25. model-00022-of-00076.safetensors +3 -0
  26. model-00023-of-00076.safetensors +3 -0
  27. model-00024-of-00076.safetensors +3 -0
  28. model-00025-of-00076.safetensors +3 -0
  29. model-00026-of-00076.safetensors +3 -0
  30. model-00027-of-00076.safetensors +3 -0
  31. model-00028-of-00076.safetensors +3 -0
  32. model-00029-of-00076.safetensors +3 -0
  33. model-00030-of-00076.safetensors +3 -0
  34. model-00031-of-00076.safetensors +3 -0
  35. model-00032-of-00076.safetensors +3 -0
  36. model-00033-of-00076.safetensors +3 -0
  37. model-00034-of-00076.safetensors +3 -0
  38. model-00035-of-00076.safetensors +3 -0
  39. model-00036-of-00076.safetensors +3 -0
  40. model-00037-of-00076.safetensors +3 -0
  41. model-00038-of-00076.safetensors +3 -0
  42. model-00039-of-00076.safetensors +3 -0
  43. model-00040-of-00076.safetensors +3 -0
  44. model-00041-of-00076.safetensors +3 -0
  45. model-00042-of-00076.safetensors +3 -0
  46. model-00043-of-00076.safetensors +3 -0
  47. model-00044-of-00076.safetensors +3 -0
  48. model-00045-of-00076.safetensors +3 -0
  49. model-00046-of-00076.safetensors +3 -0
  50. model-00047-of-00076.safetensors +3 -0
.gitattributes CHANGED
@@ -33,3 +33,5 @@ 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
+ model.safetensors.index.json filter=lfs diff=lfs merge=lfs -text
37
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
chat_template.jinja ADDED
@@ -0,0 +1,86 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [gMASK]<sop>
2
+ {%- if tools -%}
3
+ <|system|>
4
+ # Tools
5
+
6
+ You may call one or more functions to assist with the user query.
7
+
8
+ You are provided with function signatures within <tools></tools> XML tags:
9
+ <tools>
10
+ {% for tool in tools %}
11
+ {{ tool | tojson(ensure_ascii=False) }}
12
+ {% endfor %}
13
+ </tools>
14
+
15
+ For each function call, output the function name and arguments within the following XML format:
16
+ <tool_call>{function-name}<arg_key>{arg-key-1}</arg_key><arg_value>{arg-value-1}</arg_value><arg_key>{arg-key-2}</arg_key><arg_value>{arg-value-2}</arg_value>...</tool_call>{%- endif -%}
17
+ {%- macro visible_text(content) -%}
18
+ {%- if content is string -%}
19
+ {{- content }}
20
+ {%- elif content is iterable and content is not mapping -%}
21
+ {%- for item in content -%}
22
+ {%- if item is mapping and item.type == 'text' -%}
23
+ {{- item.text }}
24
+ {%- elif item is string -%}
25
+ {{- item }}
26
+ {%- endif -%}
27
+ {%- endfor -%}
28
+ {%- else -%}
29
+ {{- content }}
30
+ {%- endif -%}
31
+ {%- endmacro -%}
32
+ {%- set ns = namespace(last_user_index=-1) %}
33
+ {%- for m in messages %}
34
+ {%- if m.role == 'user' %}
35
+ {% set ns.last_user_index = loop.index0 -%}
36
+ {%- endif %}
37
+ {%- endfor %}
38
+ {% for m in messages %}
39
+ {%- if m.role == 'user' -%}<|user|>{{ visible_text(m.content) }}
40
+ {%- elif m.role == 'assistant' -%}
41
+ <|assistant|>
42
+ {%- set reasoning_content = '' %}
43
+ {%- set content = visible_text(m.content) %}
44
+ {%- if m.reasoning_content is string %}
45
+ {%- set reasoning_content = m.reasoning_content %}
46
+ {%- else %}
47
+ {%- if '</think>' in content %}
48
+ {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
49
+ {%- set content = content.split('</think>')[-1].lstrip('\n') %}
50
+ {%- endif %}
51
+ {%- endif %}
52
+ {%- if ((clear_thinking is defined and not clear_thinking) or loop.index0 > ns.last_user_index) and reasoning_content -%}
53
+ {{ '<think>' + reasoning_content.strip() + '</think>'}}
54
+ {%- else -%}
55
+ {{ '</think>' }}
56
+ {%- endif -%}
57
+ {%- if content.strip() -%}
58
+ {{ content.strip() }}
59
+ {%- endif -%}
60
+ {% if m.tool_calls %}
61
+ {% for tc in m.tool_calls %}
62
+ {%- if tc.function %}
63
+ {%- set tc = tc.function %}
64
+ {%- endif %}
65
+ {{- '<tool_call>' + tc.name -}}
66
+ {% set _args = tc.arguments %}{% for k, v in _args.items() %}<arg_key>{{ k }}</arg_key><arg_value>{{ v | tojson(ensure_ascii=False) if v is not string else v }}</arg_value>{% endfor %}</tool_call>{% endfor %}
67
+ {% endif %}
68
+ {%- elif m.role == 'tool' -%}
69
+ {%- if m.content is string -%}
70
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
71
+ {{- '<|observation|>' }}
72
+ {%- endif %}
73
+ {{- '<tool_response>' }}
74
+ {{- m.content }}
75
+ {{- '</tool_response>' }}
76
+ {%- else -%}
77
+ <|observation|>{% for tr in m.content %}
78
+ <tool_response>{{ tr.output if tr.output is defined else tr }}</tool_response>{% endfor -%}
79
+ {% endif -%}
80
+ {%- elif m.role == 'system' -%}
81
+ <|system|>{{ visible_text(m.content) }}
82
+ {%- endif -%}
83
+ {%- endfor -%}
84
+ {%- if add_generation_prompt -%}
85
+ <|assistant|>{{- '</think>' if (enable_thinking is defined and not enable_thinking) else '<think>' -}}
86
+ {%- endif -%}
config.json ADDED
@@ -0,0 +1,2726 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "GlmMoeDsaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 0,
8
+ "dtype": "bfloat16",
9
+ "eos_token_id": [
10
+ 154820,
11
+ 154827,
12
+ 154829
13
+ ],
14
+ "ep_size": 1,
15
+ "first_k_dense_replace": 3,
16
+ "head_dim": 64,
17
+ "hidden_act": "silu",
18
+ "hidden_size": 6144,
19
+ "index_head_dim": 128,
20
+ "index_n_heads": 32,
21
+ "index_topk": 2048,
22
+ "indexer_rope_interleave": true,
23
+ "initializer_range": 0.02,
24
+ "intermediate_size": 12288,
25
+ "kv_lora_rank": 512,
26
+ "max_position_embeddings": 202752,
27
+ "mlp_layer_types": [
28
+ "dense",
29
+ "dense",
30
+ "dense",
31
+ "sparse",
32
+ "sparse",
33
+ "sparse",
34
+ "sparse",
35
+ "sparse",
36
+ "sparse",
37
+ "sparse",
38
+ "sparse",
39
+ "sparse",
40
+ "sparse",
41
+ "sparse",
42
+ "sparse",
43
+ "sparse",
44
+ "sparse",
45
+ "sparse",
46
+ "sparse",
47
+ "sparse",
48
+ "sparse",
49
+ "sparse",
50
+ "sparse",
51
+ "sparse",
52
+ "sparse",
53
+ "sparse",
54
+ "sparse",
55
+ "sparse",
56
+ "sparse",
57
+ "sparse",
58
+ "sparse",
59
+ "sparse",
60
+ "sparse",
61
+ "sparse",
62
+ "sparse",
63
+ "sparse",
64
+ "sparse",
65
+ "sparse",
66
+ "sparse",
67
+ "sparse",
68
+ "sparse",
69
+ "sparse",
70
+ "sparse",
71
+ "sparse",
72
+ "sparse",
73
+ "sparse",
74
+ "sparse",
75
+ "sparse",
76
+ "sparse",
77
+ "sparse",
78
+ "sparse",
79
+ "sparse",
80
+ "sparse",
81
+ "sparse",
82
+ "sparse",
83
+ "sparse",
84
+ "sparse",
85
+ "sparse",
86
+ "sparse",
87
+ "sparse",
88
+ "sparse",
89
+ "sparse",
90
+ "sparse",
91
+ "sparse",
92
+ "sparse",
93
+ "sparse",
94
+ "sparse",
95
+ "sparse",
96
+ "sparse",
97
+ "sparse",
98
+ "sparse",
99
+ "sparse",
100
+ "sparse",
101
+ "sparse",
102
+ "sparse",
103
+ "sparse",
104
+ "sparse",
105
+ "sparse"
106
+ ],
107
+ "model_type": "glm_moe_dsa",
108
+ "moe_intermediate_size": 2048,
109
+ "moe_layer_freq": 1,
110
+ "n_group": 1,
111
+ "n_routed_experts": 256,
112
+ "n_shared_experts": 1,
113
+ "norm_topk_prob": true,
114
+ "num_attention_heads": 64,
115
+ "num_experts_per_tok": 8,
116
+ "num_hidden_layers": 78,
117
+ "num_key_value_heads": 64,
118
+ "num_nextn_predict_layers": 1,
119
+ "pad_token_id": 154820,
120
+ "pretraining_tp": 1,
121
+ "q_lora_rank": 2048,
122
+ "qk_head_dim": 256,
123
+ "qk_nope_head_dim": 192,
124
+ "qk_rope_head_dim": 64,
125
+ "quantization_config": {
126
+ "autoround_version": "0.10.0",
127
+ "bits": 4,
128
+ "data_type": "int",
129
+ "extra_config": {
130
+ "model.layers.0.mlp.down_proj": {
131
+ "bits": 8
132
+ },
133
+ "model.layers.0.mlp.gate_proj": {
134
+ "bits": 8
135
+ },
136
+ "model.layers.0.mlp.up_proj": {
137
+ "bits": 8
138
+ },
139
+ "model.layers.0.self_attn.kv_a_proj_with_mqa": {
140
+ "bits": 8
141
+ },
142
+ "model.layers.0.self_attn.kv_b_proj": {
143
+ "bits": 8
144
+ },
145
+ "model.layers.0.self_attn.o_proj": {
146
+ "bits": 8
147
+ },
148
+ "model.layers.0.self_attn.q_a_proj": {
149
+ "bits": 8
150
+ },
151
+ "model.layers.0.self_attn.q_b_proj": {
152
+ "bits": 8
153
+ },
154
+ "model.layers.0.self_attn.weights_proj": {
155
+ "bits": 8
156
+ },
157
+ "model.layers.0.self_attn.wk": {
158
+ "bits": 8
159
+ },
160
+ "model.layers.0.self_attn.wq_b": {
161
+ "bits": 8
162
+ },
163
+ "model.layers.1.mlp.down_proj": {
164
+ "bits": 8
165
+ },
166
+ "model.layers.1.mlp.gate_proj": {
167
+ "bits": 8
168
+ },
169
+ "model.layers.1.mlp.up_proj": {
170
+ "bits": 8
171
+ },
172
+ "model.layers.1.self_attn.kv_a_proj_with_mqa": {
173
+ "bits": 8
174
+ },
175
+ "model.layers.1.self_attn.kv_b_proj": {
176
+ "bits": 8
177
+ },
178
+ "model.layers.1.self_attn.o_proj": {
179
+ "bits": 8
180
+ },
181
+ "model.layers.1.self_attn.q_a_proj": {
182
+ "bits": 8
183
+ },
184
+ "model.layers.1.self_attn.q_b_proj": {
185
+ "bits": 8
186
+ },
187
+ "model.layers.1.self_attn.weights_proj": {
188
+ "bits": 8
189
+ },
190
+ "model.layers.1.self_attn.wk": {
191
+ "bits": 8
192
+ },
193
+ "model.layers.1.self_attn.wq_b": {
194
+ "bits": 8
195
+ },
196
+ "model.layers.10.mlp.shared_experts.down_proj": {
197
+ "bits": 8
198
+ },
199
+ "model.layers.10.mlp.shared_experts.gate_proj": {
200
+ "bits": 8
201
+ },
202
+ "model.layers.10.mlp.shared_experts.up_proj": {
203
+ "bits": 8
204
+ },
205
+ "model.layers.10.self_attn.kv_a_proj_with_mqa": {
206
+ "bits": 8
207
+ },
208
+ "model.layers.10.self_attn.kv_b_proj": {
209
+ "bits": 8
210
+ },
211
+ "model.layers.10.self_attn.o_proj": {
212
+ "bits": 8
213
+ },
214
+ "model.layers.10.self_attn.q_a_proj": {
215
+ "bits": 8
216
+ },
217
+ "model.layers.10.self_attn.q_b_proj": {
218
+ "bits": 8
219
+ },
220
+ "model.layers.10.self_attn.weights_proj": {
221
+ "bits": 8
222
+ },
223
+ "model.layers.10.self_attn.wk": {
224
+ "bits": 8
225
+ },
226
+ "model.layers.10.self_attn.wq_b": {
227
+ "bits": 8
228
+ },
229
+ "model.layers.11.mlp.shared_experts.down_proj": {
230
+ "bits": 8
231
+ },
232
+ "model.layers.11.mlp.shared_experts.gate_proj": {
233
+ "bits": 8
234
+ },
235
+ "model.layers.11.mlp.shared_experts.up_proj": {
236
+ "bits": 8
237
+ },
238
+ "model.layers.11.self_attn.kv_a_proj_with_mqa": {
239
+ "bits": 8
240
+ },
241
+ "model.layers.11.self_attn.kv_b_proj": {
242
+ "bits": 8
243
+ },
244
+ "model.layers.11.self_attn.o_proj": {
245
+ "bits": 8
246
+ },
247
+ "model.layers.11.self_attn.q_a_proj": {
248
+ "bits": 8
249
+ },
250
+ "model.layers.11.self_attn.q_b_proj": {
251
+ "bits": 8
252
+ },
253
+ "model.layers.11.self_attn.weights_proj": {
254
+ "bits": 8
255
+ },
256
+ "model.layers.11.self_attn.wk": {
257
+ "bits": 8
258
+ },
259
+ "model.layers.11.self_attn.wq_b": {
260
+ "bits": 8
261
+ },
262
+ "model.layers.12.mlp.shared_experts.down_proj": {
263
+ "bits": 8
264
+ },
265
+ "model.layers.12.mlp.shared_experts.gate_proj": {
266
+ "bits": 8
267
+ },
268
+ "model.layers.12.mlp.shared_experts.up_proj": {
269
+ "bits": 8
270
+ },
271
+ "model.layers.12.self_attn.kv_a_proj_with_mqa": {
272
+ "bits": 8
273
+ },
274
+ "model.layers.12.self_attn.kv_b_proj": {
275
+ "bits": 8
276
+ },
277
+ "model.layers.12.self_attn.o_proj": {
278
+ "bits": 8
279
+ },
280
+ "model.layers.12.self_attn.q_a_proj": {
281
+ "bits": 8
282
+ },
283
+ "model.layers.12.self_attn.q_b_proj": {
284
+ "bits": 8
285
+ },
286
+ "model.layers.12.self_attn.weights_proj": {
287
+ "bits": 8
288
+ },
289
+ "model.layers.12.self_attn.wk": {
290
+ "bits": 8
291
+ },
292
+ "model.layers.12.self_attn.wq_b": {
293
+ "bits": 8
294
+ },
295
+ "model.layers.13.mlp.shared_experts.down_proj": {
296
+ "bits": 8
297
+ },
298
+ "model.layers.13.mlp.shared_experts.gate_proj": {
299
+ "bits": 8
300
+ },
301
+ "model.layers.13.mlp.shared_experts.up_proj": {
302
+ "bits": 8
303
+ },
304
+ "model.layers.13.self_attn.kv_a_proj_with_mqa": {
305
+ "bits": 8
306
+ },
307
+ "model.layers.13.self_attn.kv_b_proj": {
308
+ "bits": 8
309
+ },
310
+ "model.layers.13.self_attn.o_proj": {
311
+ "bits": 8
312
+ },
313
+ "model.layers.13.self_attn.q_a_proj": {
314
+ "bits": 8
315
+ },
316
+ "model.layers.13.self_attn.q_b_proj": {
317
+ "bits": 8
318
+ },
319
+ "model.layers.13.self_attn.weights_proj": {
320
+ "bits": 8
321
+ },
322
+ "model.layers.13.self_attn.wk": {
323
+ "bits": 8
324
+ },
325
+ "model.layers.13.self_attn.wq_b": {
326
+ "bits": 8
327
+ },
328
+ "model.layers.14.mlp.shared_experts.down_proj": {
329
+ "bits": 8
330
+ },
331
+ "model.layers.14.mlp.shared_experts.gate_proj": {
332
+ "bits": 8
333
+ },
334
+ "model.layers.14.mlp.shared_experts.up_proj": {
335
+ "bits": 8
336
+ },
337
+ "model.layers.14.self_attn.kv_a_proj_with_mqa": {
338
+ "bits": 8
339
+ },
340
+ "model.layers.14.self_attn.kv_b_proj": {
341
+ "bits": 8
342
+ },
343
+ "model.layers.14.self_attn.o_proj": {
344
+ "bits": 8
345
+ },
346
+ "model.layers.14.self_attn.q_a_proj": {
347
+ "bits": 8
348
+ },
349
+ "model.layers.14.self_attn.q_b_proj": {
350
+ "bits": 8
351
+ },
352
+ "model.layers.14.self_attn.weights_proj": {
353
+ "bits": 8
354
+ },
355
+ "model.layers.14.self_attn.wk": {
356
+ "bits": 8
357
+ },
358
+ "model.layers.14.self_attn.wq_b": {
359
+ "bits": 8
360
+ },
361
+ "model.layers.15.mlp.shared_experts.down_proj": {
362
+ "bits": 8
363
+ },
364
+ "model.layers.15.mlp.shared_experts.gate_proj": {
365
+ "bits": 8
366
+ },
367
+ "model.layers.15.mlp.shared_experts.up_proj": {
368
+ "bits": 8
369
+ },
370
+ "model.layers.15.self_attn.kv_a_proj_with_mqa": {
371
+ "bits": 8
372
+ },
373
+ "model.layers.15.self_attn.kv_b_proj": {
374
+ "bits": 8
375
+ },
376
+ "model.layers.15.self_attn.o_proj": {
377
+ "bits": 8
378
+ },
379
+ "model.layers.15.self_attn.q_a_proj": {
380
+ "bits": 8
381
+ },
382
+ "model.layers.15.self_attn.q_b_proj": {
383
+ "bits": 8
384
+ },
385
+ "model.layers.15.self_attn.weights_proj": {
386
+ "bits": 8
387
+ },
388
+ "model.layers.15.self_attn.wk": {
389
+ "bits": 8
390
+ },
391
+ "model.layers.15.self_attn.wq_b": {
392
+ "bits": 8
393
+ },
394
+ "model.layers.16.mlp.shared_experts.down_proj": {
395
+ "bits": 8
396
+ },
397
+ "model.layers.16.mlp.shared_experts.gate_proj": {
398
+ "bits": 8
399
+ },
400
+ "model.layers.16.mlp.shared_experts.up_proj": {
401
+ "bits": 8
402
+ },
403
+ "model.layers.16.self_attn.kv_a_proj_with_mqa": {
404
+ "bits": 8
405
+ },
406
+ "model.layers.16.self_attn.kv_b_proj": {
407
+ "bits": 8
408
+ },
409
+ "model.layers.16.self_attn.o_proj": {
410
+ "bits": 8
411
+ },
412
+ "model.layers.16.self_attn.q_a_proj": {
413
+ "bits": 8
414
+ },
415
+ "model.layers.16.self_attn.q_b_proj": {
416
+ "bits": 8
417
+ },
418
+ "model.layers.16.self_attn.weights_proj": {
419
+ "bits": 8
420
+ },
421
+ "model.layers.16.self_attn.wk": {
422
+ "bits": 8
423
+ },
424
+ "model.layers.16.self_attn.wq_b": {
425
+ "bits": 8
426
+ },
427
+ "model.layers.17.mlp.shared_experts.down_proj": {
428
+ "bits": 8
429
+ },
430
+ "model.layers.17.mlp.shared_experts.gate_proj": {
431
+ "bits": 8
432
+ },
433
+ "model.layers.17.mlp.shared_experts.up_proj": {
434
+ "bits": 8
435
+ },
436
+ "model.layers.17.self_attn.kv_a_proj_with_mqa": {
437
+ "bits": 8
438
+ },
439
+ "model.layers.17.self_attn.kv_b_proj": {
440
+ "bits": 8
441
+ },
442
+ "model.layers.17.self_attn.o_proj": {
443
+ "bits": 8
444
+ },
445
+ "model.layers.17.self_attn.q_a_proj": {
446
+ "bits": 8
447
+ },
448
+ "model.layers.17.self_attn.q_b_proj": {
449
+ "bits": 8
450
+ },
451
+ "model.layers.17.self_attn.weights_proj": {
452
+ "bits": 8
453
+ },
454
+ "model.layers.17.self_attn.wk": {
455
+ "bits": 8
456
+ },
457
+ "model.layers.17.self_attn.wq_b": {
458
+ "bits": 8
459
+ },
460
+ "model.layers.18.mlp.shared_experts.down_proj": {
461
+ "bits": 8
462
+ },
463
+ "model.layers.18.mlp.shared_experts.gate_proj": {
464
+ "bits": 8
465
+ },
466
+ "model.layers.18.mlp.shared_experts.up_proj": {
467
+ "bits": 8
468
+ },
469
+ "model.layers.18.self_attn.kv_a_proj_with_mqa": {
470
+ "bits": 8
471
+ },
472
+ "model.layers.18.self_attn.kv_b_proj": {
473
+ "bits": 8
474
+ },
475
+ "model.layers.18.self_attn.o_proj": {
476
+ "bits": 8
477
+ },
478
+ "model.layers.18.self_attn.q_a_proj": {
479
+ "bits": 8
480
+ },
481
+ "model.layers.18.self_attn.q_b_proj": {
482
+ "bits": 8
483
+ },
484
+ "model.layers.18.self_attn.weights_proj": {
485
+ "bits": 8
486
+ },
487
+ "model.layers.18.self_attn.wk": {
488
+ "bits": 8
489
+ },
490
+ "model.layers.18.self_attn.wq_b": {
491
+ "bits": 8
492
+ },
493
+ "model.layers.19.mlp.shared_experts.down_proj": {
494
+ "bits": 8
495
+ },
496
+ "model.layers.19.mlp.shared_experts.gate_proj": {
497
+ "bits": 8
498
+ },
499
+ "model.layers.19.mlp.shared_experts.up_proj": {
500
+ "bits": 8
501
+ },
502
+ "model.layers.19.self_attn.kv_a_proj_with_mqa": {
503
+ "bits": 8
504
+ },
505
+ "model.layers.19.self_attn.kv_b_proj": {
506
+ "bits": 8
507
+ },
508
+ "model.layers.19.self_attn.o_proj": {
509
+ "bits": 8
510
+ },
511
+ "model.layers.19.self_attn.q_a_proj": {
512
+ "bits": 8
513
+ },
514
+ "model.layers.19.self_attn.q_b_proj": {
515
+ "bits": 8
516
+ },
517
+ "model.layers.19.self_attn.weights_proj": {
518
+ "bits": 8
519
+ },
520
+ "model.layers.19.self_attn.wk": {
521
+ "bits": 8
522
+ },
523
+ "model.layers.19.self_attn.wq_b": {
524
+ "bits": 8
525
+ },
526
+ "model.layers.2.mlp.down_proj": {
527
+ "bits": 8
528
+ },
529
+ "model.layers.2.mlp.gate_proj": {
530
+ "bits": 8
531
+ },
532
+ "model.layers.2.mlp.up_proj": {
533
+ "bits": 8
534
+ },
535
+ "model.layers.2.self_attn.kv_a_proj_with_mqa": {
536
+ "bits": 8
537
+ },
538
+ "model.layers.2.self_attn.kv_b_proj": {
539
+ "bits": 8
540
+ },
541
+ "model.layers.2.self_attn.o_proj": {
542
+ "bits": 8
543
+ },
544
+ "model.layers.2.self_attn.q_a_proj": {
545
+ "bits": 8
546
+ },
547
+ "model.layers.2.self_attn.q_b_proj": {
548
+ "bits": 8
549
+ },
550
+ "model.layers.2.self_attn.weights_proj": {
551
+ "bits": 8
552
+ },
553
+ "model.layers.2.self_attn.wk": {
554
+ "bits": 8
555
+ },
556
+ "model.layers.2.self_attn.wq_b": {
557
+ "bits": 8
558
+ },
559
+ "model.layers.20.mlp.shared_experts.down_proj": {
560
+ "bits": 8
561
+ },
562
+ "model.layers.20.mlp.shared_experts.gate_proj": {
563
+ "bits": 8
564
+ },
565
+ "model.layers.20.mlp.shared_experts.up_proj": {
566
+ "bits": 8
567
+ },
568
+ "model.layers.20.self_attn.kv_a_proj_with_mqa": {
569
+ "bits": 8
570
+ },
571
+ "model.layers.20.self_attn.kv_b_proj": {
572
+ "bits": 8
573
+ },
574
+ "model.layers.20.self_attn.o_proj": {
575
+ "bits": 8
576
+ },
577
+ "model.layers.20.self_attn.q_a_proj": {
578
+ "bits": 8
579
+ },
580
+ "model.layers.20.self_attn.q_b_proj": {
581
+ "bits": 8
582
+ },
583
+ "model.layers.20.self_attn.weights_proj": {
584
+ "bits": 8
585
+ },
586
+ "model.layers.20.self_attn.wk": {
587
+ "bits": 8
588
+ },
589
+ "model.layers.20.self_attn.wq_b": {
590
+ "bits": 8
591
+ },
592
+ "model.layers.21.mlp.shared_experts.down_proj": {
593
+ "bits": 8
594
+ },
595
+ "model.layers.21.mlp.shared_experts.gate_proj": {
596
+ "bits": 8
597
+ },
598
+ "model.layers.21.mlp.shared_experts.up_proj": {
599
+ "bits": 8
600
+ },
601
+ "model.layers.21.self_attn.kv_a_proj_with_mqa": {
602
+ "bits": 8
603
+ },
604
+ "model.layers.21.self_attn.kv_b_proj": {
605
+ "bits": 8
606
+ },
607
+ "model.layers.21.self_attn.o_proj": {
608
+ "bits": 8
609
+ },
610
+ "model.layers.21.self_attn.q_a_proj": {
611
+ "bits": 8
612
+ },
613
+ "model.layers.21.self_attn.q_b_proj": {
614
+ "bits": 8
615
+ },
616
+ "model.layers.21.self_attn.weights_proj": {
617
+ "bits": 8
618
+ },
619
+ "model.layers.21.self_attn.wk": {
620
+ "bits": 8
621
+ },
622
+ "model.layers.21.self_attn.wq_b": {
623
+ "bits": 8
624
+ },
625
+ "model.layers.22.mlp.shared_experts.down_proj": {
626
+ "bits": 8
627
+ },
628
+ "model.layers.22.mlp.shared_experts.gate_proj": {
629
+ "bits": 8
630
+ },
631
+ "model.layers.22.mlp.shared_experts.up_proj": {
632
+ "bits": 8
633
+ },
634
+ "model.layers.22.self_attn.kv_a_proj_with_mqa": {
635
+ "bits": 8
636
+ },
637
+ "model.layers.22.self_attn.kv_b_proj": {
638
+ "bits": 8
639
+ },
640
+ "model.layers.22.self_attn.o_proj": {
641
+ "bits": 8
642
+ },
643
+ "model.layers.22.self_attn.q_a_proj": {
644
+ "bits": 8
645
+ },
646
+ "model.layers.22.self_attn.q_b_proj": {
647
+ "bits": 8
648
+ },
649
+ "model.layers.22.self_attn.weights_proj": {
650
+ "bits": 8
651
+ },
652
+ "model.layers.22.self_attn.wk": {
653
+ "bits": 8
654
+ },
655
+ "model.layers.22.self_attn.wq_b": {
656
+ "bits": 8
657
+ },
658
+ "model.layers.23.mlp.shared_experts.down_proj": {
659
+ "bits": 8
660
+ },
661
+ "model.layers.23.mlp.shared_experts.gate_proj": {
662
+ "bits": 8
663
+ },
664
+ "model.layers.23.mlp.shared_experts.up_proj": {
665
+ "bits": 8
666
+ },
667
+ "model.layers.23.self_attn.kv_a_proj_with_mqa": {
668
+ "bits": 8
669
+ },
670
+ "model.layers.23.self_attn.kv_b_proj": {
671
+ "bits": 8
672
+ },
673
+ "model.layers.23.self_attn.o_proj": {
674
+ "bits": 8
675
+ },
676
+ "model.layers.23.self_attn.q_a_proj": {
677
+ "bits": 8
678
+ },
679
+ "model.layers.23.self_attn.q_b_proj": {
680
+ "bits": 8
681
+ },
682
+ "model.layers.23.self_attn.weights_proj": {
683
+ "bits": 8
684
+ },
685
+ "model.layers.23.self_attn.wk": {
686
+ "bits": 8
687
+ },
688
+ "model.layers.23.self_attn.wq_b": {
689
+ "bits": 8
690
+ },
691
+ "model.layers.24.mlp.shared_experts.down_proj": {
692
+ "bits": 8
693
+ },
694
+ "model.layers.24.mlp.shared_experts.gate_proj": {
695
+ "bits": 8
696
+ },
697
+ "model.layers.24.mlp.shared_experts.up_proj": {
698
+ "bits": 8
699
+ },
700
+ "model.layers.24.self_attn.kv_a_proj_with_mqa": {
701
+ "bits": 8
702
+ },
703
+ "model.layers.24.self_attn.kv_b_proj": {
704
+ "bits": 8
705
+ },
706
+ "model.layers.24.self_attn.o_proj": {
707
+ "bits": 8
708
+ },
709
+ "model.layers.24.self_attn.q_a_proj": {
710
+ "bits": 8
711
+ },
712
+ "model.layers.24.self_attn.q_b_proj": {
713
+ "bits": 8
714
+ },
715
+ "model.layers.24.self_attn.weights_proj": {
716
+ "bits": 8
717
+ },
718
+ "model.layers.24.self_attn.wk": {
719
+ "bits": 8
720
+ },
721
+ "model.layers.24.self_attn.wq_b": {
722
+ "bits": 8
723
+ },
724
+ "model.layers.25.mlp.shared_experts.down_proj": {
725
+ "bits": 8
726
+ },
727
+ "model.layers.25.mlp.shared_experts.gate_proj": {
728
+ "bits": 8
729
+ },
730
+ "model.layers.25.mlp.shared_experts.up_proj": {
731
+ "bits": 8
732
+ },
733
+ "model.layers.25.self_attn.kv_a_proj_with_mqa": {
734
+ "bits": 8
735
+ },
736
+ "model.layers.25.self_attn.kv_b_proj": {
737
+ "bits": 8
738
+ },
739
+ "model.layers.25.self_attn.o_proj": {
740
+ "bits": 8
741
+ },
742
+ "model.layers.25.self_attn.q_a_proj": {
743
+ "bits": 8
744
+ },
745
+ "model.layers.25.self_attn.q_b_proj": {
746
+ "bits": 8
747
+ },
748
+ "model.layers.25.self_attn.weights_proj": {
749
+ "bits": 8
750
+ },
751
+ "model.layers.25.self_attn.wk": {
752
+ "bits": 8
753
+ },
754
+ "model.layers.25.self_attn.wq_b": {
755
+ "bits": 8
756
+ },
757
+ "model.layers.26.mlp.shared_experts.down_proj": {
758
+ "bits": 8
759
+ },
760
+ "model.layers.26.mlp.shared_experts.gate_proj": {
761
+ "bits": 8
762
+ },
763
+ "model.layers.26.mlp.shared_experts.up_proj": {
764
+ "bits": 8
765
+ },
766
+ "model.layers.26.self_attn.kv_a_proj_with_mqa": {
767
+ "bits": 8
768
+ },
769
+ "model.layers.26.self_attn.kv_b_proj": {
770
+ "bits": 8
771
+ },
772
+ "model.layers.26.self_attn.o_proj": {
773
+ "bits": 8
774
+ },
775
+ "model.layers.26.self_attn.q_a_proj": {
776
+ "bits": 8
777
+ },
778
+ "model.layers.26.self_attn.q_b_proj": {
779
+ "bits": 8
780
+ },
781
+ "model.layers.26.self_attn.weights_proj": {
782
+ "bits": 8
783
+ },
784
+ "model.layers.26.self_attn.wk": {
785
+ "bits": 8
786
+ },
787
+ "model.layers.26.self_attn.wq_b": {
788
+ "bits": 8
789
+ },
790
+ "model.layers.27.mlp.shared_experts.down_proj": {
791
+ "bits": 8
792
+ },
793
+ "model.layers.27.mlp.shared_experts.gate_proj": {
794
+ "bits": 8
795
+ },
796
+ "model.layers.27.mlp.shared_experts.up_proj": {
797
+ "bits": 8
798
+ },
799
+ "model.layers.27.self_attn.kv_a_proj_with_mqa": {
800
+ "bits": 8
801
+ },
802
+ "model.layers.27.self_attn.kv_b_proj": {
803
+ "bits": 8
804
+ },
805
+ "model.layers.27.self_attn.o_proj": {
806
+ "bits": 8
807
+ },
808
+ "model.layers.27.self_attn.q_a_proj": {
809
+ "bits": 8
810
+ },
811
+ "model.layers.27.self_attn.q_b_proj": {
812
+ "bits": 8
813
+ },
814
+ "model.layers.27.self_attn.weights_proj": {
815
+ "bits": 8
816
+ },
817
+ "model.layers.27.self_attn.wk": {
818
+ "bits": 8
819
+ },
820
+ "model.layers.27.self_attn.wq_b": {
821
+ "bits": 8
822
+ },
823
+ "model.layers.28.mlp.shared_experts.down_proj": {
824
+ "bits": 8
825
+ },
826
+ "model.layers.28.mlp.shared_experts.gate_proj": {
827
+ "bits": 8
828
+ },
829
+ "model.layers.28.mlp.shared_experts.up_proj": {
830
+ "bits": 8
831
+ },
832
+ "model.layers.28.self_attn.kv_a_proj_with_mqa": {
833
+ "bits": 8
834
+ },
835
+ "model.layers.28.self_attn.kv_b_proj": {
836
+ "bits": 8
837
+ },
838
+ "model.layers.28.self_attn.o_proj": {
839
+ "bits": 8
840
+ },
841
+ "model.layers.28.self_attn.q_a_proj": {
842
+ "bits": 8
843
+ },
844
+ "model.layers.28.self_attn.q_b_proj": {
845
+ "bits": 8
846
+ },
847
+ "model.layers.28.self_attn.weights_proj": {
848
+ "bits": 8
849
+ },
850
+ "model.layers.28.self_attn.wk": {
851
+ "bits": 8
852
+ },
853
+ "model.layers.28.self_attn.wq_b": {
854
+ "bits": 8
855
+ },
856
+ "model.layers.29.mlp.shared_experts.down_proj": {
857
+ "bits": 8
858
+ },
859
+ "model.layers.29.mlp.shared_experts.gate_proj": {
860
+ "bits": 8
861
+ },
862
+ "model.layers.29.mlp.shared_experts.up_proj": {
863
+ "bits": 8
864
+ },
865
+ "model.layers.29.self_attn.kv_a_proj_with_mqa": {
866
+ "bits": 8
867
+ },
868
+ "model.layers.29.self_attn.kv_b_proj": {
869
+ "bits": 8
870
+ },
871
+ "model.layers.29.self_attn.o_proj": {
872
+ "bits": 8
873
+ },
874
+ "model.layers.29.self_attn.q_a_proj": {
875
+ "bits": 8
876
+ },
877
+ "model.layers.29.self_attn.q_b_proj": {
878
+ "bits": 8
879
+ },
880
+ "model.layers.29.self_attn.weights_proj": {
881
+ "bits": 8
882
+ },
883
+ "model.layers.29.self_attn.wk": {
884
+ "bits": 8
885
+ },
886
+ "model.layers.29.self_attn.wq_b": {
887
+ "bits": 8
888
+ },
889
+ "model.layers.3.mlp.shared_experts.down_proj": {
890
+ "bits": 8
891
+ },
892
+ "model.layers.3.mlp.shared_experts.gate_proj": {
893
+ "bits": 8
894
+ },
895
+ "model.layers.3.mlp.shared_experts.up_proj": {
896
+ "bits": 8
897
+ },
898
+ "model.layers.3.self_attn.kv_a_proj_with_mqa": {
899
+ "bits": 8
900
+ },
901
+ "model.layers.3.self_attn.kv_b_proj": {
902
+ "bits": 8
903
+ },
904
+ "model.layers.3.self_attn.o_proj": {
905
+ "bits": 8
906
+ },
907
+ "model.layers.3.self_attn.q_a_proj": {
908
+ "bits": 8
909
+ },
910
+ "model.layers.3.self_attn.q_b_proj": {
911
+ "bits": 8
912
+ },
913
+ "model.layers.3.self_attn.weights_proj": {
914
+ "bits": 8
915
+ },
916
+ "model.layers.3.self_attn.wk": {
917
+ "bits": 8
918
+ },
919
+ "model.layers.3.self_attn.wq_b": {
920
+ "bits": 8
921
+ },
922
+ "model.layers.30.mlp.shared_experts.down_proj": {
923
+ "bits": 8
924
+ },
925
+ "model.layers.30.mlp.shared_experts.gate_proj": {
926
+ "bits": 8
927
+ },
928
+ "model.layers.30.mlp.shared_experts.up_proj": {
929
+ "bits": 8
930
+ },
931
+ "model.layers.30.self_attn.kv_a_proj_with_mqa": {
932
+ "bits": 8
933
+ },
934
+ "model.layers.30.self_attn.kv_b_proj": {
935
+ "bits": 8
936
+ },
937
+ "model.layers.30.self_attn.o_proj": {
938
+ "bits": 8
939
+ },
940
+ "model.layers.30.self_attn.q_a_proj": {
941
+ "bits": 8
942
+ },
943
+ "model.layers.30.self_attn.q_b_proj": {
944
+ "bits": 8
945
+ },
946
+ "model.layers.30.self_attn.weights_proj": {
947
+ "bits": 8
948
+ },
949
+ "model.layers.30.self_attn.wk": {
950
+ "bits": 8
951
+ },
952
+ "model.layers.30.self_attn.wq_b": {
953
+ "bits": 8
954
+ },
955
+ "model.layers.31.mlp.shared_experts.down_proj": {
956
+ "bits": 8
957
+ },
958
+ "model.layers.31.mlp.shared_experts.gate_proj": {
959
+ "bits": 8
960
+ },
961
+ "model.layers.31.mlp.shared_experts.up_proj": {
962
+ "bits": 8
963
+ },
964
+ "model.layers.31.self_attn.kv_a_proj_with_mqa": {
965
+ "bits": 8
966
+ },
967
+ "model.layers.31.self_attn.kv_b_proj": {
968
+ "bits": 8
969
+ },
970
+ "model.layers.31.self_attn.o_proj": {
971
+ "bits": 8
972
+ },
973
+ "model.layers.31.self_attn.q_a_proj": {
974
+ "bits": 8
975
+ },
976
+ "model.layers.31.self_attn.q_b_proj": {
977
+ "bits": 8
978
+ },
979
+ "model.layers.31.self_attn.weights_proj": {
980
+ "bits": 8
981
+ },
982
+ "model.layers.31.self_attn.wk": {
983
+ "bits": 8
984
+ },
985
+ "model.layers.31.self_attn.wq_b": {
986
+ "bits": 8
987
+ },
988
+ "model.layers.32.mlp.shared_experts.down_proj": {
989
+ "bits": 8
990
+ },
991
+ "model.layers.32.mlp.shared_experts.gate_proj": {
992
+ "bits": 8
993
+ },
994
+ "model.layers.32.mlp.shared_experts.up_proj": {
995
+ "bits": 8
996
+ },
997
+ "model.layers.32.self_attn.kv_a_proj_with_mqa": {
998
+ "bits": 8
999
+ },
1000
+ "model.layers.32.self_attn.kv_b_proj": {
1001
+ "bits": 8
1002
+ },
1003
+ "model.layers.32.self_attn.o_proj": {
1004
+ "bits": 8
1005
+ },
1006
+ "model.layers.32.self_attn.q_a_proj": {
1007
+ "bits": 8
1008
+ },
1009
+ "model.layers.32.self_attn.q_b_proj": {
1010
+ "bits": 8
1011
+ },
1012
+ "model.layers.32.self_attn.weights_proj": {
1013
+ "bits": 8
1014
+ },
1015
+ "model.layers.32.self_attn.wk": {
1016
+ "bits": 8
1017
+ },
1018
+ "model.layers.32.self_attn.wq_b": {
1019
+ "bits": 8
1020
+ },
1021
+ "model.layers.33.mlp.shared_experts.down_proj": {
1022
+ "bits": 8
1023
+ },
1024
+ "model.layers.33.mlp.shared_experts.gate_proj": {
1025
+ "bits": 8
1026
+ },
1027
+ "model.layers.33.mlp.shared_experts.up_proj": {
1028
+ "bits": 8
1029
+ },
1030
+ "model.layers.33.self_attn.kv_a_proj_with_mqa": {
1031
+ "bits": 8
1032
+ },
1033
+ "model.layers.33.self_attn.kv_b_proj": {
1034
+ "bits": 8
1035
+ },
1036
+ "model.layers.33.self_attn.o_proj": {
1037
+ "bits": 8
1038
+ },
1039
+ "model.layers.33.self_attn.q_a_proj": {
1040
+ "bits": 8
1041
+ },
1042
+ "model.layers.33.self_attn.q_b_proj": {
1043
+ "bits": 8
1044
+ },
1045
+ "model.layers.33.self_attn.weights_proj": {
1046
+ "bits": 8
1047
+ },
1048
+ "model.layers.33.self_attn.wk": {
1049
+ "bits": 8
1050
+ },
1051
+ "model.layers.33.self_attn.wq_b": {
1052
+ "bits": 8
1053
+ },
1054
+ "model.layers.34.mlp.shared_experts.down_proj": {
1055
+ "bits": 8
1056
+ },
1057
+ "model.layers.34.mlp.shared_experts.gate_proj": {
1058
+ "bits": 8
1059
+ },
1060
+ "model.layers.34.mlp.shared_experts.up_proj": {
1061
+ "bits": 8
1062
+ },
1063
+ "model.layers.34.self_attn.kv_a_proj_with_mqa": {
1064
+ "bits": 8
1065
+ },
1066
+ "model.layers.34.self_attn.kv_b_proj": {
1067
+ "bits": 8
1068
+ },
1069
+ "model.layers.34.self_attn.o_proj": {
1070
+ "bits": 8
1071
+ },
1072
+ "model.layers.34.self_attn.q_a_proj": {
1073
+ "bits": 8
1074
+ },
1075
+ "model.layers.34.self_attn.q_b_proj": {
1076
+ "bits": 8
1077
+ },
1078
+ "model.layers.34.self_attn.weights_proj": {
1079
+ "bits": 8
1080
+ },
1081
+ "model.layers.34.self_attn.wk": {
1082
+ "bits": 8
1083
+ },
1084
+ "model.layers.34.self_attn.wq_b": {
1085
+ "bits": 8
1086
+ },
1087
+ "model.layers.35.mlp.shared_experts.down_proj": {
1088
+ "bits": 8
1089
+ },
1090
+ "model.layers.35.mlp.shared_experts.gate_proj": {
1091
+ "bits": 8
1092
+ },
1093
+ "model.layers.35.mlp.shared_experts.up_proj": {
1094
+ "bits": 8
1095
+ },
1096
+ "model.layers.35.self_attn.kv_a_proj_with_mqa": {
1097
+ "bits": 8
1098
+ },
1099
+ "model.layers.35.self_attn.kv_b_proj": {
1100
+ "bits": 8
1101
+ },
1102
+ "model.layers.35.self_attn.o_proj": {
1103
+ "bits": 8
1104
+ },
1105
+ "model.layers.35.self_attn.q_a_proj": {
1106
+ "bits": 8
1107
+ },
1108
+ "model.layers.35.self_attn.q_b_proj": {
1109
+ "bits": 8
1110
+ },
1111
+ "model.layers.35.self_attn.weights_proj": {
1112
+ "bits": 8
1113
+ },
1114
+ "model.layers.35.self_attn.wk": {
1115
+ "bits": 8
1116
+ },
1117
+ "model.layers.35.self_attn.wq_b": {
1118
+ "bits": 8
1119
+ },
1120
+ "model.layers.36.mlp.shared_experts.down_proj": {
1121
+ "bits": 8
1122
+ },
1123
+ "model.layers.36.mlp.shared_experts.gate_proj": {
1124
+ "bits": 8
1125
+ },
1126
+ "model.layers.36.mlp.shared_experts.up_proj": {
1127
+ "bits": 8
1128
+ },
1129
+ "model.layers.36.self_attn.kv_a_proj_with_mqa": {
1130
+ "bits": 8
1131
+ },
1132
+ "model.layers.36.self_attn.kv_b_proj": {
1133
+ "bits": 8
1134
+ },
1135
+ "model.layers.36.self_attn.o_proj": {
1136
+ "bits": 8
1137
+ },
1138
+ "model.layers.36.self_attn.q_a_proj": {
1139
+ "bits": 8
1140
+ },
1141
+ "model.layers.36.self_attn.q_b_proj": {
1142
+ "bits": 8
1143
+ },
1144
+ "model.layers.36.self_attn.weights_proj": {
1145
+ "bits": 8
1146
+ },
1147
+ "model.layers.36.self_attn.wk": {
1148
+ "bits": 8
1149
+ },
1150
+ "model.layers.36.self_attn.wq_b": {
1151
+ "bits": 8
1152
+ },
1153
+ "model.layers.37.mlp.shared_experts.down_proj": {
1154
+ "bits": 8
1155
+ },
1156
+ "model.layers.37.mlp.shared_experts.gate_proj": {
1157
+ "bits": 8
1158
+ },
1159
+ "model.layers.37.mlp.shared_experts.up_proj": {
1160
+ "bits": 8
1161
+ },
1162
+ "model.layers.37.self_attn.kv_a_proj_with_mqa": {
1163
+ "bits": 8
1164
+ },
1165
+ "model.layers.37.self_attn.kv_b_proj": {
1166
+ "bits": 8
1167
+ },
1168
+ "model.layers.37.self_attn.o_proj": {
1169
+ "bits": 8
1170
+ },
1171
+ "model.layers.37.self_attn.q_a_proj": {
1172
+ "bits": 8
1173
+ },
1174
+ "model.layers.37.self_attn.q_b_proj": {
1175
+ "bits": 8
1176
+ },
1177
+ "model.layers.37.self_attn.weights_proj": {
1178
+ "bits": 8
1179
+ },
1180
+ "model.layers.37.self_attn.wk": {
1181
+ "bits": 8
1182
+ },
1183
+ "model.layers.37.self_attn.wq_b": {
1184
+ "bits": 8
1185
+ },
1186
+ "model.layers.38.mlp.shared_experts.down_proj": {
1187
+ "bits": 8
1188
+ },
1189
+ "model.layers.38.mlp.shared_experts.gate_proj": {
1190
+ "bits": 8
1191
+ },
1192
+ "model.layers.38.mlp.shared_experts.up_proj": {
1193
+ "bits": 8
1194
+ },
1195
+ "model.layers.38.self_attn.kv_a_proj_with_mqa": {
1196
+ "bits": 8
1197
+ },
1198
+ "model.layers.38.self_attn.kv_b_proj": {
1199
+ "bits": 8
1200
+ },
1201
+ "model.layers.38.self_attn.o_proj": {
1202
+ "bits": 8
1203
+ },
1204
+ "model.layers.38.self_attn.q_a_proj": {
1205
+ "bits": 8
1206
+ },
1207
+ "model.layers.38.self_attn.q_b_proj": {
1208
+ "bits": 8
1209
+ },
1210
+ "model.layers.38.self_attn.weights_proj": {
1211
+ "bits": 8
1212
+ },
1213
+ "model.layers.38.self_attn.wk": {
1214
+ "bits": 8
1215
+ },
1216
+ "model.layers.38.self_attn.wq_b": {
1217
+ "bits": 8
1218
+ },
1219
+ "model.layers.39.mlp.shared_experts.down_proj": {
1220
+ "bits": 8
1221
+ },
1222
+ "model.layers.39.mlp.shared_experts.gate_proj": {
1223
+ "bits": 8
1224
+ },
1225
+ "model.layers.39.mlp.shared_experts.up_proj": {
1226
+ "bits": 8
1227
+ },
1228
+ "model.layers.39.self_attn.kv_a_proj_with_mqa": {
1229
+ "bits": 8
1230
+ },
1231
+ "model.layers.39.self_attn.kv_b_proj": {
1232
+ "bits": 8
1233
+ },
1234
+ "model.layers.39.self_attn.o_proj": {
1235
+ "bits": 8
1236
+ },
1237
+ "model.layers.39.self_attn.q_a_proj": {
1238
+ "bits": 8
1239
+ },
1240
+ "model.layers.39.self_attn.q_b_proj": {
1241
+ "bits": 8
1242
+ },
1243
+ "model.layers.39.self_attn.weights_proj": {
1244
+ "bits": 8
1245
+ },
1246
+ "model.layers.39.self_attn.wk": {
1247
+ "bits": 8
1248
+ },
1249
+ "model.layers.39.self_attn.wq_b": {
1250
+ "bits": 8
1251
+ },
1252
+ "model.layers.4.mlp.shared_experts.down_proj": {
1253
+ "bits": 8
1254
+ },
1255
+ "model.layers.4.mlp.shared_experts.gate_proj": {
1256
+ "bits": 8
1257
+ },
1258
+ "model.layers.4.mlp.shared_experts.up_proj": {
1259
+ "bits": 8
1260
+ },
1261
+ "model.layers.4.self_attn.kv_a_proj_with_mqa": {
1262
+ "bits": 8
1263
+ },
1264
+ "model.layers.4.self_attn.kv_b_proj": {
1265
+ "bits": 8
1266
+ },
1267
+ "model.layers.4.self_attn.o_proj": {
1268
+ "bits": 8
1269
+ },
1270
+ "model.layers.4.self_attn.q_a_proj": {
1271
+ "bits": 8
1272
+ },
1273
+ "model.layers.4.self_attn.q_b_proj": {
1274
+ "bits": 8
1275
+ },
1276
+ "model.layers.4.self_attn.weights_proj": {
1277
+ "bits": 8
1278
+ },
1279
+ "model.layers.4.self_attn.wk": {
1280
+ "bits": 8
1281
+ },
1282
+ "model.layers.4.self_attn.wq_b": {
1283
+ "bits": 8
1284
+ },
1285
+ "model.layers.40.mlp.shared_experts.down_proj": {
1286
+ "bits": 8
1287
+ },
1288
+ "model.layers.40.mlp.shared_experts.gate_proj": {
1289
+ "bits": 8
1290
+ },
1291
+ "model.layers.40.mlp.shared_experts.up_proj": {
1292
+ "bits": 8
1293
+ },
1294
+ "model.layers.40.self_attn.kv_a_proj_with_mqa": {
1295
+ "bits": 8
1296
+ },
1297
+ "model.layers.40.self_attn.kv_b_proj": {
1298
+ "bits": 8
1299
+ },
1300
+ "model.layers.40.self_attn.o_proj": {
1301
+ "bits": 8
1302
+ },
1303
+ "model.layers.40.self_attn.q_a_proj": {
1304
+ "bits": 8
1305
+ },
1306
+ "model.layers.40.self_attn.q_b_proj": {
1307
+ "bits": 8
1308
+ },
1309
+ "model.layers.40.self_attn.weights_proj": {
1310
+ "bits": 8
1311
+ },
1312
+ "model.layers.40.self_attn.wk": {
1313
+ "bits": 8
1314
+ },
1315
+ "model.layers.40.self_attn.wq_b": {
1316
+ "bits": 8
1317
+ },
1318
+ "model.layers.41.mlp.shared_experts.down_proj": {
1319
+ "bits": 8
1320
+ },
1321
+ "model.layers.41.mlp.shared_experts.gate_proj": {
1322
+ "bits": 8
1323
+ },
1324
+ "model.layers.41.mlp.shared_experts.up_proj": {
1325
+ "bits": 8
1326
+ },
1327
+ "model.layers.41.self_attn.kv_a_proj_with_mqa": {
1328
+ "bits": 8
1329
+ },
1330
+ "model.layers.41.self_attn.kv_b_proj": {
1331
+ "bits": 8
1332
+ },
1333
+ "model.layers.41.self_attn.o_proj": {
1334
+ "bits": 8
1335
+ },
1336
+ "model.layers.41.self_attn.q_a_proj": {
1337
+ "bits": 8
1338
+ },
1339
+ "model.layers.41.self_attn.q_b_proj": {
1340
+ "bits": 8
1341
+ },
1342
+ "model.layers.41.self_attn.weights_proj": {
1343
+ "bits": 8
1344
+ },
1345
+ "model.layers.41.self_attn.wk": {
1346
+ "bits": 8
1347
+ },
1348
+ "model.layers.41.self_attn.wq_b": {
1349
+ "bits": 8
1350
+ },
1351
+ "model.layers.42.mlp.shared_experts.down_proj": {
1352
+ "bits": 8
1353
+ },
1354
+ "model.layers.42.mlp.shared_experts.gate_proj": {
1355
+ "bits": 8
1356
+ },
1357
+ "model.layers.42.mlp.shared_experts.up_proj": {
1358
+ "bits": 8
1359
+ },
1360
+ "model.layers.42.self_attn.kv_a_proj_with_mqa": {
1361
+ "bits": 8
1362
+ },
1363
+ "model.layers.42.self_attn.kv_b_proj": {
1364
+ "bits": 8
1365
+ },
1366
+ "model.layers.42.self_attn.o_proj": {
1367
+ "bits": 8
1368
+ },
1369
+ "model.layers.42.self_attn.q_a_proj": {
1370
+ "bits": 8
1371
+ },
1372
+ "model.layers.42.self_attn.q_b_proj": {
1373
+ "bits": 8
1374
+ },
1375
+ "model.layers.42.self_attn.weights_proj": {
1376
+ "bits": 8
1377
+ },
1378
+ "model.layers.42.self_attn.wk": {
1379
+ "bits": 8
1380
+ },
1381
+ "model.layers.42.self_attn.wq_b": {
1382
+ "bits": 8
1383
+ },
1384
+ "model.layers.43.mlp.shared_experts.down_proj": {
1385
+ "bits": 8
1386
+ },
1387
+ "model.layers.43.mlp.shared_experts.gate_proj": {
1388
+ "bits": 8
1389
+ },
1390
+ "model.layers.43.mlp.shared_experts.up_proj": {
1391
+ "bits": 8
1392
+ },
1393
+ "model.layers.43.self_attn.kv_a_proj_with_mqa": {
1394
+ "bits": 8
1395
+ },
1396
+ "model.layers.43.self_attn.kv_b_proj": {
1397
+ "bits": 8
1398
+ },
1399
+ "model.layers.43.self_attn.o_proj": {
1400
+ "bits": 8
1401
+ },
1402
+ "model.layers.43.self_attn.q_a_proj": {
1403
+ "bits": 8
1404
+ },
1405
+ "model.layers.43.self_attn.q_b_proj": {
1406
+ "bits": 8
1407
+ },
1408
+ "model.layers.43.self_attn.weights_proj": {
1409
+ "bits": 8
1410
+ },
1411
+ "model.layers.43.self_attn.wk": {
1412
+ "bits": 8
1413
+ },
1414
+ "model.layers.43.self_attn.wq_b": {
1415
+ "bits": 8
1416
+ },
1417
+ "model.layers.44.mlp.shared_experts.down_proj": {
1418
+ "bits": 8
1419
+ },
1420
+ "model.layers.44.mlp.shared_experts.gate_proj": {
1421
+ "bits": 8
1422
+ },
1423
+ "model.layers.44.mlp.shared_experts.up_proj": {
1424
+ "bits": 8
1425
+ },
1426
+ "model.layers.44.self_attn.kv_a_proj_with_mqa": {
1427
+ "bits": 8
1428
+ },
1429
+ "model.layers.44.self_attn.kv_b_proj": {
1430
+ "bits": 8
1431
+ },
1432
+ "model.layers.44.self_attn.o_proj": {
1433
+ "bits": 8
1434
+ },
1435
+ "model.layers.44.self_attn.q_a_proj": {
1436
+ "bits": 8
1437
+ },
1438
+ "model.layers.44.self_attn.q_b_proj": {
1439
+ "bits": 8
1440
+ },
1441
+ "model.layers.44.self_attn.weights_proj": {
1442
+ "bits": 8
1443
+ },
1444
+ "model.layers.44.self_attn.wk": {
1445
+ "bits": 8
1446
+ },
1447
+ "model.layers.44.self_attn.wq_b": {
1448
+ "bits": 8
1449
+ },
1450
+ "model.layers.45.mlp.shared_experts.down_proj": {
1451
+ "bits": 8
1452
+ },
1453
+ "model.layers.45.mlp.shared_experts.gate_proj": {
1454
+ "bits": 8
1455
+ },
1456
+ "model.layers.45.mlp.shared_experts.up_proj": {
1457
+ "bits": 8
1458
+ },
1459
+ "model.layers.45.self_attn.kv_a_proj_with_mqa": {
1460
+ "bits": 8
1461
+ },
1462
+ "model.layers.45.self_attn.kv_b_proj": {
1463
+ "bits": 8
1464
+ },
1465
+ "model.layers.45.self_attn.o_proj": {
1466
+ "bits": 8
1467
+ },
1468
+ "model.layers.45.self_attn.q_a_proj": {
1469
+ "bits": 8
1470
+ },
1471
+ "model.layers.45.self_attn.q_b_proj": {
1472
+ "bits": 8
1473
+ },
1474
+ "model.layers.45.self_attn.weights_proj": {
1475
+ "bits": 8
1476
+ },
1477
+ "model.layers.45.self_attn.wk": {
1478
+ "bits": 8
1479
+ },
1480
+ "model.layers.45.self_attn.wq_b": {
1481
+ "bits": 8
1482
+ },
1483
+ "model.layers.46.mlp.shared_experts.down_proj": {
1484
+ "bits": 8
1485
+ },
1486
+ "model.layers.46.mlp.shared_experts.gate_proj": {
1487
+ "bits": 8
1488
+ },
1489
+ "model.layers.46.mlp.shared_experts.up_proj": {
1490
+ "bits": 8
1491
+ },
1492
+ "model.layers.46.self_attn.kv_a_proj_with_mqa": {
1493
+ "bits": 8
1494
+ },
1495
+ "model.layers.46.self_attn.kv_b_proj": {
1496
+ "bits": 8
1497
+ },
1498
+ "model.layers.46.self_attn.o_proj": {
1499
+ "bits": 8
1500
+ },
1501
+ "model.layers.46.self_attn.q_a_proj": {
1502
+ "bits": 8
1503
+ },
1504
+ "model.layers.46.self_attn.q_b_proj": {
1505
+ "bits": 8
1506
+ },
1507
+ "model.layers.46.self_attn.weights_proj": {
1508
+ "bits": 8
1509
+ },
1510
+ "model.layers.46.self_attn.wk": {
1511
+ "bits": 8
1512
+ },
1513
+ "model.layers.46.self_attn.wq_b": {
1514
+ "bits": 8
1515
+ },
1516
+ "model.layers.47.mlp.shared_experts.down_proj": {
1517
+ "bits": 8
1518
+ },
1519
+ "model.layers.47.mlp.shared_experts.gate_proj": {
1520
+ "bits": 8
1521
+ },
1522
+ "model.layers.47.mlp.shared_experts.up_proj": {
1523
+ "bits": 8
1524
+ },
1525
+ "model.layers.47.self_attn.kv_a_proj_with_mqa": {
1526
+ "bits": 8
1527
+ },
1528
+ "model.layers.47.self_attn.kv_b_proj": {
1529
+ "bits": 8
1530
+ },
1531
+ "model.layers.47.self_attn.o_proj": {
1532
+ "bits": 8
1533
+ },
1534
+ "model.layers.47.self_attn.q_a_proj": {
1535
+ "bits": 8
1536
+ },
1537
+ "model.layers.47.self_attn.q_b_proj": {
1538
+ "bits": 8
1539
+ },
1540
+ "model.layers.47.self_attn.weights_proj": {
1541
+ "bits": 8
1542
+ },
1543
+ "model.layers.47.self_attn.wk": {
1544
+ "bits": 8
1545
+ },
1546
+ "model.layers.47.self_attn.wq_b": {
1547
+ "bits": 8
1548
+ },
1549
+ "model.layers.48.mlp.shared_experts.down_proj": {
1550
+ "bits": 8
1551
+ },
1552
+ "model.layers.48.mlp.shared_experts.gate_proj": {
1553
+ "bits": 8
1554
+ },
1555
+ "model.layers.48.mlp.shared_experts.up_proj": {
1556
+ "bits": 8
1557
+ },
1558
+ "model.layers.48.self_attn.kv_a_proj_with_mqa": {
1559
+ "bits": 8
1560
+ },
1561
+ "model.layers.48.self_attn.kv_b_proj": {
1562
+ "bits": 8
1563
+ },
1564
+ "model.layers.48.self_attn.o_proj": {
1565
+ "bits": 8
1566
+ },
1567
+ "model.layers.48.self_attn.q_a_proj": {
1568
+ "bits": 8
1569
+ },
1570
+ "model.layers.48.self_attn.q_b_proj": {
1571
+ "bits": 8
1572
+ },
1573
+ "model.layers.48.self_attn.weights_proj": {
1574
+ "bits": 8
1575
+ },
1576
+ "model.layers.48.self_attn.wk": {
1577
+ "bits": 8
1578
+ },
1579
+ "model.layers.48.self_attn.wq_b": {
1580
+ "bits": 8
1581
+ },
1582
+ "model.layers.49.mlp.shared_experts.down_proj": {
1583
+ "bits": 8
1584
+ },
1585
+ "model.layers.49.mlp.shared_experts.gate_proj": {
1586
+ "bits": 8
1587
+ },
1588
+ "model.layers.49.mlp.shared_experts.up_proj": {
1589
+ "bits": 8
1590
+ },
1591
+ "model.layers.49.self_attn.kv_a_proj_with_mqa": {
1592
+ "bits": 8
1593
+ },
1594
+ "model.layers.49.self_attn.kv_b_proj": {
1595
+ "bits": 8
1596
+ },
1597
+ "model.layers.49.self_attn.o_proj": {
1598
+ "bits": 8
1599
+ },
1600
+ "model.layers.49.self_attn.q_a_proj": {
1601
+ "bits": 8
1602
+ },
1603
+ "model.layers.49.self_attn.q_b_proj": {
1604
+ "bits": 8
1605
+ },
1606
+ "model.layers.49.self_attn.weights_proj": {
1607
+ "bits": 8
1608
+ },
1609
+ "model.layers.49.self_attn.wk": {
1610
+ "bits": 8
1611
+ },
1612
+ "model.layers.49.self_attn.wq_b": {
1613
+ "bits": 8
1614
+ },
1615
+ "model.layers.5.mlp.shared_experts.down_proj": {
1616
+ "bits": 8
1617
+ },
1618
+ "model.layers.5.mlp.shared_experts.gate_proj": {
1619
+ "bits": 8
1620
+ },
1621
+ "model.layers.5.mlp.shared_experts.up_proj": {
1622
+ "bits": 8
1623
+ },
1624
+ "model.layers.5.self_attn.kv_a_proj_with_mqa": {
1625
+ "bits": 8
1626
+ },
1627
+ "model.layers.5.self_attn.kv_b_proj": {
1628
+ "bits": 8
1629
+ },
1630
+ "model.layers.5.self_attn.o_proj": {
1631
+ "bits": 8
1632
+ },
1633
+ "model.layers.5.self_attn.q_a_proj": {
1634
+ "bits": 8
1635
+ },
1636
+ "model.layers.5.self_attn.q_b_proj": {
1637
+ "bits": 8
1638
+ },
1639
+ "model.layers.5.self_attn.weights_proj": {
1640
+ "bits": 8
1641
+ },
1642
+ "model.layers.5.self_attn.wk": {
1643
+ "bits": 8
1644
+ },
1645
+ "model.layers.5.self_attn.wq_b": {
1646
+ "bits": 8
1647
+ },
1648
+ "model.layers.50.mlp.shared_experts.down_proj": {
1649
+ "bits": 8
1650
+ },
1651
+ "model.layers.50.mlp.shared_experts.gate_proj": {
1652
+ "bits": 8
1653
+ },
1654
+ "model.layers.50.mlp.shared_experts.up_proj": {
1655
+ "bits": 8
1656
+ },
1657
+ "model.layers.50.self_attn.kv_a_proj_with_mqa": {
1658
+ "bits": 8
1659
+ },
1660
+ "model.layers.50.self_attn.kv_b_proj": {
1661
+ "bits": 8
1662
+ },
1663
+ "model.layers.50.self_attn.o_proj": {
1664
+ "bits": 8
1665
+ },
1666
+ "model.layers.50.self_attn.q_a_proj": {
1667
+ "bits": 8
1668
+ },
1669
+ "model.layers.50.self_attn.q_b_proj": {
1670
+ "bits": 8
1671
+ },
1672
+ "model.layers.50.self_attn.weights_proj": {
1673
+ "bits": 8
1674
+ },
1675
+ "model.layers.50.self_attn.wk": {
1676
+ "bits": 8
1677
+ },
1678
+ "model.layers.50.self_attn.wq_b": {
1679
+ "bits": 8
1680
+ },
1681
+ "model.layers.51.mlp.shared_experts.down_proj": {
1682
+ "bits": 8
1683
+ },
1684
+ "model.layers.51.mlp.shared_experts.gate_proj": {
1685
+ "bits": 8
1686
+ },
1687
+ "model.layers.51.mlp.shared_experts.up_proj": {
1688
+ "bits": 8
1689
+ },
1690
+ "model.layers.51.self_attn.kv_a_proj_with_mqa": {
1691
+ "bits": 8
1692
+ },
1693
+ "model.layers.51.self_attn.kv_b_proj": {
1694
+ "bits": 8
1695
+ },
1696
+ "model.layers.51.self_attn.o_proj": {
1697
+ "bits": 8
1698
+ },
1699
+ "model.layers.51.self_attn.q_a_proj": {
1700
+ "bits": 8
1701
+ },
1702
+ "model.layers.51.self_attn.q_b_proj": {
1703
+ "bits": 8
1704
+ },
1705
+ "model.layers.51.self_attn.weights_proj": {
1706
+ "bits": 8
1707
+ },
1708
+ "model.layers.51.self_attn.wk": {
1709
+ "bits": 8
1710
+ },
1711
+ "model.layers.51.self_attn.wq_b": {
1712
+ "bits": 8
1713
+ },
1714
+ "model.layers.52.mlp.shared_experts.down_proj": {
1715
+ "bits": 8
1716
+ },
1717
+ "model.layers.52.mlp.shared_experts.gate_proj": {
1718
+ "bits": 8
1719
+ },
1720
+ "model.layers.52.mlp.shared_experts.up_proj": {
1721
+ "bits": 8
1722
+ },
1723
+ "model.layers.52.self_attn.kv_a_proj_with_mqa": {
1724
+ "bits": 8
1725
+ },
1726
+ "model.layers.52.self_attn.kv_b_proj": {
1727
+ "bits": 8
1728
+ },
1729
+ "model.layers.52.self_attn.o_proj": {
1730
+ "bits": 8
1731
+ },
1732
+ "model.layers.52.self_attn.q_a_proj": {
1733
+ "bits": 8
1734
+ },
1735
+ "model.layers.52.self_attn.q_b_proj": {
1736
+ "bits": 8
1737
+ },
1738
+ "model.layers.52.self_attn.weights_proj": {
1739
+ "bits": 8
1740
+ },
1741
+ "model.layers.52.self_attn.wk": {
1742
+ "bits": 8
1743
+ },
1744
+ "model.layers.52.self_attn.wq_b": {
1745
+ "bits": 8
1746
+ },
1747
+ "model.layers.53.mlp.shared_experts.down_proj": {
1748
+ "bits": 8
1749
+ },
1750
+ "model.layers.53.mlp.shared_experts.gate_proj": {
1751
+ "bits": 8
1752
+ },
1753
+ "model.layers.53.mlp.shared_experts.up_proj": {
1754
+ "bits": 8
1755
+ },
1756
+ "model.layers.53.self_attn.kv_a_proj_with_mqa": {
1757
+ "bits": 8
1758
+ },
1759
+ "model.layers.53.self_attn.kv_b_proj": {
1760
+ "bits": 8
1761
+ },
1762
+ "model.layers.53.self_attn.o_proj": {
1763
+ "bits": 8
1764
+ },
1765
+ "model.layers.53.self_attn.q_a_proj": {
1766
+ "bits": 8
1767
+ },
1768
+ "model.layers.53.self_attn.q_b_proj": {
1769
+ "bits": 8
1770
+ },
1771
+ "model.layers.53.self_attn.weights_proj": {
1772
+ "bits": 8
1773
+ },
1774
+ "model.layers.53.self_attn.wk": {
1775
+ "bits": 8
1776
+ },
1777
+ "model.layers.53.self_attn.wq_b": {
1778
+ "bits": 8
1779
+ },
1780
+ "model.layers.54.mlp.shared_experts.down_proj": {
1781
+ "bits": 8
1782
+ },
1783
+ "model.layers.54.mlp.shared_experts.gate_proj": {
1784
+ "bits": 8
1785
+ },
1786
+ "model.layers.54.mlp.shared_experts.up_proj": {
1787
+ "bits": 8
1788
+ },
1789
+ "model.layers.54.self_attn.kv_a_proj_with_mqa": {
1790
+ "bits": 8
1791
+ },
1792
+ "model.layers.54.self_attn.kv_b_proj": {
1793
+ "bits": 8
1794
+ },
1795
+ "model.layers.54.self_attn.o_proj": {
1796
+ "bits": 8
1797
+ },
1798
+ "model.layers.54.self_attn.q_a_proj": {
1799
+ "bits": 8
1800
+ },
1801
+ "model.layers.54.self_attn.q_b_proj": {
1802
+ "bits": 8
1803
+ },
1804
+ "model.layers.54.self_attn.weights_proj": {
1805
+ "bits": 8
1806
+ },
1807
+ "model.layers.54.self_attn.wk": {
1808
+ "bits": 8
1809
+ },
1810
+ "model.layers.54.self_attn.wq_b": {
1811
+ "bits": 8
1812
+ },
1813
+ "model.layers.55.mlp.shared_experts.down_proj": {
1814
+ "bits": 8
1815
+ },
1816
+ "model.layers.55.mlp.shared_experts.gate_proj": {
1817
+ "bits": 8
1818
+ },
1819
+ "model.layers.55.mlp.shared_experts.up_proj": {
1820
+ "bits": 8
1821
+ },
1822
+ "model.layers.55.self_attn.kv_a_proj_with_mqa": {
1823
+ "bits": 8
1824
+ },
1825
+ "model.layers.55.self_attn.kv_b_proj": {
1826
+ "bits": 8
1827
+ },
1828
+ "model.layers.55.self_attn.o_proj": {
1829
+ "bits": 8
1830
+ },
1831
+ "model.layers.55.self_attn.q_a_proj": {
1832
+ "bits": 8
1833
+ },
1834
+ "model.layers.55.self_attn.q_b_proj": {
1835
+ "bits": 8
1836
+ },
1837
+ "model.layers.55.self_attn.weights_proj": {
1838
+ "bits": 8
1839
+ },
1840
+ "model.layers.55.self_attn.wk": {
1841
+ "bits": 8
1842
+ },
1843
+ "model.layers.55.self_attn.wq_b": {
1844
+ "bits": 8
1845
+ },
1846
+ "model.layers.56.mlp.shared_experts.down_proj": {
1847
+ "bits": 8
1848
+ },
1849
+ "model.layers.56.mlp.shared_experts.gate_proj": {
1850
+ "bits": 8
1851
+ },
1852
+ "model.layers.56.mlp.shared_experts.up_proj": {
1853
+ "bits": 8
1854
+ },
1855
+ "model.layers.56.self_attn.kv_a_proj_with_mqa": {
1856
+ "bits": 8
1857
+ },
1858
+ "model.layers.56.self_attn.kv_b_proj": {
1859
+ "bits": 8
1860
+ },
1861
+ "model.layers.56.self_attn.o_proj": {
1862
+ "bits": 8
1863
+ },
1864
+ "model.layers.56.self_attn.q_a_proj": {
1865
+ "bits": 8
1866
+ },
1867
+ "model.layers.56.self_attn.q_b_proj": {
1868
+ "bits": 8
1869
+ },
1870
+ "model.layers.56.self_attn.weights_proj": {
1871
+ "bits": 8
1872
+ },
1873
+ "model.layers.56.self_attn.wk": {
1874
+ "bits": 8
1875
+ },
1876
+ "model.layers.56.self_attn.wq_b": {
1877
+ "bits": 8
1878
+ },
1879
+ "model.layers.57.mlp.shared_experts.down_proj": {
1880
+ "bits": 8
1881
+ },
1882
+ "model.layers.57.mlp.shared_experts.gate_proj": {
1883
+ "bits": 8
1884
+ },
1885
+ "model.layers.57.mlp.shared_experts.up_proj": {
1886
+ "bits": 8
1887
+ },
1888
+ "model.layers.57.self_attn.kv_a_proj_with_mqa": {
1889
+ "bits": 8
1890
+ },
1891
+ "model.layers.57.self_attn.kv_b_proj": {
1892
+ "bits": 8
1893
+ },
1894
+ "model.layers.57.self_attn.o_proj": {
1895
+ "bits": 8
1896
+ },
1897
+ "model.layers.57.self_attn.q_a_proj": {
1898
+ "bits": 8
1899
+ },
1900
+ "model.layers.57.self_attn.q_b_proj": {
1901
+ "bits": 8
1902
+ },
1903
+ "model.layers.57.self_attn.weights_proj": {
1904
+ "bits": 8
1905
+ },
1906
+ "model.layers.57.self_attn.wk": {
1907
+ "bits": 8
1908
+ },
1909
+ "model.layers.57.self_attn.wq_b": {
1910
+ "bits": 8
1911
+ },
1912
+ "model.layers.58.mlp.shared_experts.down_proj": {
1913
+ "bits": 8
1914
+ },
1915
+ "model.layers.58.mlp.shared_experts.gate_proj": {
1916
+ "bits": 8
1917
+ },
1918
+ "model.layers.58.mlp.shared_experts.up_proj": {
1919
+ "bits": 8
1920
+ },
1921
+ "model.layers.58.self_attn.kv_a_proj_with_mqa": {
1922
+ "bits": 8
1923
+ },
1924
+ "model.layers.58.self_attn.kv_b_proj": {
1925
+ "bits": 8
1926
+ },
1927
+ "model.layers.58.self_attn.o_proj": {
1928
+ "bits": 8
1929
+ },
1930
+ "model.layers.58.self_attn.q_a_proj": {
1931
+ "bits": 8
1932
+ },
1933
+ "model.layers.58.self_attn.q_b_proj": {
1934
+ "bits": 8
1935
+ },
1936
+ "model.layers.58.self_attn.weights_proj": {
1937
+ "bits": 8
1938
+ },
1939
+ "model.layers.58.self_attn.wk": {
1940
+ "bits": 8
1941
+ },
1942
+ "model.layers.58.self_attn.wq_b": {
1943
+ "bits": 8
1944
+ },
1945
+ "model.layers.59.mlp.shared_experts.down_proj": {
1946
+ "bits": 8
1947
+ },
1948
+ "model.layers.59.mlp.shared_experts.gate_proj": {
1949
+ "bits": 8
1950
+ },
1951
+ "model.layers.59.mlp.shared_experts.up_proj": {
1952
+ "bits": 8
1953
+ },
1954
+ "model.layers.59.self_attn.kv_a_proj_with_mqa": {
1955
+ "bits": 8
1956
+ },
1957
+ "model.layers.59.self_attn.kv_b_proj": {
1958
+ "bits": 8
1959
+ },
1960
+ "model.layers.59.self_attn.o_proj": {
1961
+ "bits": 8
1962
+ },
1963
+ "model.layers.59.self_attn.q_a_proj": {
1964
+ "bits": 8
1965
+ },
1966
+ "model.layers.59.self_attn.q_b_proj": {
1967
+ "bits": 8
1968
+ },
1969
+ "model.layers.59.self_attn.weights_proj": {
1970
+ "bits": 8
1971
+ },
1972
+ "model.layers.59.self_attn.wk": {
1973
+ "bits": 8
1974
+ },
1975
+ "model.layers.59.self_attn.wq_b": {
1976
+ "bits": 8
1977
+ },
1978
+ "model.layers.6.mlp.shared_experts.down_proj": {
1979
+ "bits": 8
1980
+ },
1981
+ "model.layers.6.mlp.shared_experts.gate_proj": {
1982
+ "bits": 8
1983
+ },
1984
+ "model.layers.6.mlp.shared_experts.up_proj": {
1985
+ "bits": 8
1986
+ },
1987
+ "model.layers.6.self_attn.kv_a_proj_with_mqa": {
1988
+ "bits": 8
1989
+ },
1990
+ "model.layers.6.self_attn.kv_b_proj": {
1991
+ "bits": 8
1992
+ },
1993
+ "model.layers.6.self_attn.o_proj": {
1994
+ "bits": 8
1995
+ },
1996
+ "model.layers.6.self_attn.q_a_proj": {
1997
+ "bits": 8
1998
+ },
1999
+ "model.layers.6.self_attn.q_b_proj": {
2000
+ "bits": 8
2001
+ },
2002
+ "model.layers.6.self_attn.weights_proj": {
2003
+ "bits": 8
2004
+ },
2005
+ "model.layers.6.self_attn.wk": {
2006
+ "bits": 8
2007
+ },
2008
+ "model.layers.6.self_attn.wq_b": {
2009
+ "bits": 8
2010
+ },
2011
+ "model.layers.60.mlp.shared_experts.down_proj": {
2012
+ "bits": 8
2013
+ },
2014
+ "model.layers.60.mlp.shared_experts.gate_proj": {
2015
+ "bits": 8
2016
+ },
2017
+ "model.layers.60.mlp.shared_experts.up_proj": {
2018
+ "bits": 8
2019
+ },
2020
+ "model.layers.60.self_attn.kv_a_proj_with_mqa": {
2021
+ "bits": 8
2022
+ },
2023
+ "model.layers.60.self_attn.kv_b_proj": {
2024
+ "bits": 8
2025
+ },
2026
+ "model.layers.60.self_attn.o_proj": {
2027
+ "bits": 8
2028
+ },
2029
+ "model.layers.60.self_attn.q_a_proj": {
2030
+ "bits": 8
2031
+ },
2032
+ "model.layers.60.self_attn.q_b_proj": {
2033
+ "bits": 8
2034
+ },
2035
+ "model.layers.60.self_attn.weights_proj": {
2036
+ "bits": 8
2037
+ },
2038
+ "model.layers.60.self_attn.wk": {
2039
+ "bits": 8
2040
+ },
2041
+ "model.layers.60.self_attn.wq_b": {
2042
+ "bits": 8
2043
+ },
2044
+ "model.layers.61.mlp.shared_experts.down_proj": {
2045
+ "bits": 8
2046
+ },
2047
+ "model.layers.61.mlp.shared_experts.gate_proj": {
2048
+ "bits": 8
2049
+ },
2050
+ "model.layers.61.mlp.shared_experts.up_proj": {
2051
+ "bits": 8
2052
+ },
2053
+ "model.layers.61.self_attn.kv_a_proj_with_mqa": {
2054
+ "bits": 8
2055
+ },
2056
+ "model.layers.61.self_attn.kv_b_proj": {
2057
+ "bits": 8
2058
+ },
2059
+ "model.layers.61.self_attn.o_proj": {
2060
+ "bits": 8
2061
+ },
2062
+ "model.layers.61.self_attn.q_a_proj": {
2063
+ "bits": 8
2064
+ },
2065
+ "model.layers.61.self_attn.q_b_proj": {
2066
+ "bits": 8
2067
+ },
2068
+ "model.layers.61.self_attn.weights_proj": {
2069
+ "bits": 8
2070
+ },
2071
+ "model.layers.61.self_attn.wk": {
2072
+ "bits": 8
2073
+ },
2074
+ "model.layers.61.self_attn.wq_b": {
2075
+ "bits": 8
2076
+ },
2077
+ "model.layers.62.mlp.shared_experts.down_proj": {
2078
+ "bits": 8
2079
+ },
2080
+ "model.layers.62.mlp.shared_experts.gate_proj": {
2081
+ "bits": 8
2082
+ },
2083
+ "model.layers.62.mlp.shared_experts.up_proj": {
2084
+ "bits": 8
2085
+ },
2086
+ "model.layers.62.self_attn.kv_a_proj_with_mqa": {
2087
+ "bits": 8
2088
+ },
2089
+ "model.layers.62.self_attn.kv_b_proj": {
2090
+ "bits": 8
2091
+ },
2092
+ "model.layers.62.self_attn.o_proj": {
2093
+ "bits": 8
2094
+ },
2095
+ "model.layers.62.self_attn.q_a_proj": {
2096
+ "bits": 8
2097
+ },
2098
+ "model.layers.62.self_attn.q_b_proj": {
2099
+ "bits": 8
2100
+ },
2101
+ "model.layers.62.self_attn.weights_proj": {
2102
+ "bits": 8
2103
+ },
2104
+ "model.layers.62.self_attn.wk": {
2105
+ "bits": 8
2106
+ },
2107
+ "model.layers.62.self_attn.wq_b": {
2108
+ "bits": 8
2109
+ },
2110
+ "model.layers.63.mlp.shared_experts.down_proj": {
2111
+ "bits": 8
2112
+ },
2113
+ "model.layers.63.mlp.shared_experts.gate_proj": {
2114
+ "bits": 8
2115
+ },
2116
+ "model.layers.63.mlp.shared_experts.up_proj": {
2117
+ "bits": 8
2118
+ },
2119
+ "model.layers.63.self_attn.kv_a_proj_with_mqa": {
2120
+ "bits": 8
2121
+ },
2122
+ "model.layers.63.self_attn.kv_b_proj": {
2123
+ "bits": 8
2124
+ },
2125
+ "model.layers.63.self_attn.o_proj": {
2126
+ "bits": 8
2127
+ },
2128
+ "model.layers.63.self_attn.q_a_proj": {
2129
+ "bits": 8
2130
+ },
2131
+ "model.layers.63.self_attn.q_b_proj": {
2132
+ "bits": 8
2133
+ },
2134
+ "model.layers.63.self_attn.weights_proj": {
2135
+ "bits": 8
2136
+ },
2137
+ "model.layers.63.self_attn.wk": {
2138
+ "bits": 8
2139
+ },
2140
+ "model.layers.63.self_attn.wq_b": {
2141
+ "bits": 8
2142
+ },
2143
+ "model.layers.64.mlp.shared_experts.down_proj": {
2144
+ "bits": 8
2145
+ },
2146
+ "model.layers.64.mlp.shared_experts.gate_proj": {
2147
+ "bits": 8
2148
+ },
2149
+ "model.layers.64.mlp.shared_experts.up_proj": {
2150
+ "bits": 8
2151
+ },
2152
+ "model.layers.64.self_attn.kv_a_proj_with_mqa": {
2153
+ "bits": 8
2154
+ },
2155
+ "model.layers.64.self_attn.kv_b_proj": {
2156
+ "bits": 8
2157
+ },
2158
+ "model.layers.64.self_attn.o_proj": {
2159
+ "bits": 8
2160
+ },
2161
+ "model.layers.64.self_attn.q_a_proj": {
2162
+ "bits": 8
2163
+ },
2164
+ "model.layers.64.self_attn.q_b_proj": {
2165
+ "bits": 8
2166
+ },
2167
+ "model.layers.64.self_attn.weights_proj": {
2168
+ "bits": 8
2169
+ },
2170
+ "model.layers.64.self_attn.wk": {
2171
+ "bits": 8
2172
+ },
2173
+ "model.layers.64.self_attn.wq_b": {
2174
+ "bits": 8
2175
+ },
2176
+ "model.layers.65.mlp.shared_experts.down_proj": {
2177
+ "bits": 8
2178
+ },
2179
+ "model.layers.65.mlp.shared_experts.gate_proj": {
2180
+ "bits": 8
2181
+ },
2182
+ "model.layers.65.mlp.shared_experts.up_proj": {
2183
+ "bits": 8
2184
+ },
2185
+ "model.layers.65.self_attn.kv_a_proj_with_mqa": {
2186
+ "bits": 8
2187
+ },
2188
+ "model.layers.65.self_attn.kv_b_proj": {
2189
+ "bits": 8
2190
+ },
2191
+ "model.layers.65.self_attn.o_proj": {
2192
+ "bits": 8
2193
+ },
2194
+ "model.layers.65.self_attn.q_a_proj": {
2195
+ "bits": 8
2196
+ },
2197
+ "model.layers.65.self_attn.q_b_proj": {
2198
+ "bits": 8
2199
+ },
2200
+ "model.layers.65.self_attn.weights_proj": {
2201
+ "bits": 8
2202
+ },
2203
+ "model.layers.65.self_attn.wk": {
2204
+ "bits": 8
2205
+ },
2206
+ "model.layers.65.self_attn.wq_b": {
2207
+ "bits": 8
2208
+ },
2209
+ "model.layers.66.mlp.shared_experts.down_proj": {
2210
+ "bits": 8
2211
+ },
2212
+ "model.layers.66.mlp.shared_experts.gate_proj": {
2213
+ "bits": 8
2214
+ },
2215
+ "model.layers.66.mlp.shared_experts.up_proj": {
2216
+ "bits": 8
2217
+ },
2218
+ "model.layers.66.self_attn.kv_a_proj_with_mqa": {
2219
+ "bits": 8
2220
+ },
2221
+ "model.layers.66.self_attn.kv_b_proj": {
2222
+ "bits": 8
2223
+ },
2224
+ "model.layers.66.self_attn.o_proj": {
2225
+ "bits": 8
2226
+ },
2227
+ "model.layers.66.self_attn.q_a_proj": {
2228
+ "bits": 8
2229
+ },
2230
+ "model.layers.66.self_attn.q_b_proj": {
2231
+ "bits": 8
2232
+ },
2233
+ "model.layers.66.self_attn.weights_proj": {
2234
+ "bits": 8
2235
+ },
2236
+ "model.layers.66.self_attn.wk": {
2237
+ "bits": 8
2238
+ },
2239
+ "model.layers.66.self_attn.wq_b": {
2240
+ "bits": 8
2241
+ },
2242
+ "model.layers.67.mlp.shared_experts.down_proj": {
2243
+ "bits": 8
2244
+ },
2245
+ "model.layers.67.mlp.shared_experts.gate_proj": {
2246
+ "bits": 8
2247
+ },
2248
+ "model.layers.67.mlp.shared_experts.up_proj": {
2249
+ "bits": 8
2250
+ },
2251
+ "model.layers.67.self_attn.kv_a_proj_with_mqa": {
2252
+ "bits": 8
2253
+ },
2254
+ "model.layers.67.self_attn.kv_b_proj": {
2255
+ "bits": 8
2256
+ },
2257
+ "model.layers.67.self_attn.o_proj": {
2258
+ "bits": 8
2259
+ },
2260
+ "model.layers.67.self_attn.q_a_proj": {
2261
+ "bits": 8
2262
+ },
2263
+ "model.layers.67.self_attn.q_b_proj": {
2264
+ "bits": 8
2265
+ },
2266
+ "model.layers.67.self_attn.weights_proj": {
2267
+ "bits": 8
2268
+ },
2269
+ "model.layers.67.self_attn.wk": {
2270
+ "bits": 8
2271
+ },
2272
+ "model.layers.67.self_attn.wq_b": {
2273
+ "bits": 8
2274
+ },
2275
+ "model.layers.68.mlp.shared_experts.down_proj": {
2276
+ "bits": 8
2277
+ },
2278
+ "model.layers.68.mlp.shared_experts.gate_proj": {
2279
+ "bits": 8
2280
+ },
2281
+ "model.layers.68.mlp.shared_experts.up_proj": {
2282
+ "bits": 8
2283
+ },
2284
+ "model.layers.68.self_attn.kv_a_proj_with_mqa": {
2285
+ "bits": 8
2286
+ },
2287
+ "model.layers.68.self_attn.kv_b_proj": {
2288
+ "bits": 8
2289
+ },
2290
+ "model.layers.68.self_attn.o_proj": {
2291
+ "bits": 8
2292
+ },
2293
+ "model.layers.68.self_attn.q_a_proj": {
2294
+ "bits": 8
2295
+ },
2296
+ "model.layers.68.self_attn.q_b_proj": {
2297
+ "bits": 8
2298
+ },
2299
+ "model.layers.68.self_attn.weights_proj": {
2300
+ "bits": 8
2301
+ },
2302
+ "model.layers.68.self_attn.wk": {
2303
+ "bits": 8
2304
+ },
2305
+ "model.layers.68.self_attn.wq_b": {
2306
+ "bits": 8
2307
+ },
2308
+ "model.layers.69.mlp.shared_experts.down_proj": {
2309
+ "bits": 8
2310
+ },
2311
+ "model.layers.69.mlp.shared_experts.gate_proj": {
2312
+ "bits": 8
2313
+ },
2314
+ "model.layers.69.mlp.shared_experts.up_proj": {
2315
+ "bits": 8
2316
+ },
2317
+ "model.layers.69.self_attn.kv_a_proj_with_mqa": {
2318
+ "bits": 8
2319
+ },
2320
+ "model.layers.69.self_attn.kv_b_proj": {
2321
+ "bits": 8
2322
+ },
2323
+ "model.layers.69.self_attn.o_proj": {
2324
+ "bits": 8
2325
+ },
2326
+ "model.layers.69.self_attn.q_a_proj": {
2327
+ "bits": 8
2328
+ },
2329
+ "model.layers.69.self_attn.q_b_proj": {
2330
+ "bits": 8
2331
+ },
2332
+ "model.layers.69.self_attn.weights_proj": {
2333
+ "bits": 8
2334
+ },
2335
+ "model.layers.69.self_attn.wk": {
2336
+ "bits": 8
2337
+ },
2338
+ "model.layers.69.self_attn.wq_b": {
2339
+ "bits": 8
2340
+ },
2341
+ "model.layers.7.mlp.shared_experts.down_proj": {
2342
+ "bits": 8
2343
+ },
2344
+ "model.layers.7.mlp.shared_experts.gate_proj": {
2345
+ "bits": 8
2346
+ },
2347
+ "model.layers.7.mlp.shared_experts.up_proj": {
2348
+ "bits": 8
2349
+ },
2350
+ "model.layers.7.self_attn.kv_a_proj_with_mqa": {
2351
+ "bits": 8
2352
+ },
2353
+ "model.layers.7.self_attn.kv_b_proj": {
2354
+ "bits": 8
2355
+ },
2356
+ "model.layers.7.self_attn.o_proj": {
2357
+ "bits": 8
2358
+ },
2359
+ "model.layers.7.self_attn.q_a_proj": {
2360
+ "bits": 8
2361
+ },
2362
+ "model.layers.7.self_attn.q_b_proj": {
2363
+ "bits": 8
2364
+ },
2365
+ "model.layers.7.self_attn.weights_proj": {
2366
+ "bits": 8
2367
+ },
2368
+ "model.layers.7.self_attn.wk": {
2369
+ "bits": 8
2370
+ },
2371
+ "model.layers.7.self_attn.wq_b": {
2372
+ "bits": 8
2373
+ },
2374
+ "model.layers.70.mlp.shared_experts.down_proj": {
2375
+ "bits": 8
2376
+ },
2377
+ "model.layers.70.mlp.shared_experts.gate_proj": {
2378
+ "bits": 8
2379
+ },
2380
+ "model.layers.70.mlp.shared_experts.up_proj": {
2381
+ "bits": 8
2382
+ },
2383
+ "model.layers.70.self_attn.kv_a_proj_with_mqa": {
2384
+ "bits": 8
2385
+ },
2386
+ "model.layers.70.self_attn.kv_b_proj": {
2387
+ "bits": 8
2388
+ },
2389
+ "model.layers.70.self_attn.o_proj": {
2390
+ "bits": 8
2391
+ },
2392
+ "model.layers.70.self_attn.q_a_proj": {
2393
+ "bits": 8
2394
+ },
2395
+ "model.layers.70.self_attn.q_b_proj": {
2396
+ "bits": 8
2397
+ },
2398
+ "model.layers.70.self_attn.weights_proj": {
2399
+ "bits": 8
2400
+ },
2401
+ "model.layers.70.self_attn.wk": {
2402
+ "bits": 8
2403
+ },
2404
+ "model.layers.70.self_attn.wq_b": {
2405
+ "bits": 8
2406
+ },
2407
+ "model.layers.71.mlp.shared_experts.down_proj": {
2408
+ "bits": 8
2409
+ },
2410
+ "model.layers.71.mlp.shared_experts.gate_proj": {
2411
+ "bits": 8
2412
+ },
2413
+ "model.layers.71.mlp.shared_experts.up_proj": {
2414
+ "bits": 8
2415
+ },
2416
+ "model.layers.71.self_attn.kv_a_proj_with_mqa": {
2417
+ "bits": 8
2418
+ },
2419
+ "model.layers.71.self_attn.kv_b_proj": {
2420
+ "bits": 8
2421
+ },
2422
+ "model.layers.71.self_attn.o_proj": {
2423
+ "bits": 8
2424
+ },
2425
+ "model.layers.71.self_attn.q_a_proj": {
2426
+ "bits": 8
2427
+ },
2428
+ "model.layers.71.self_attn.q_b_proj": {
2429
+ "bits": 8
2430
+ },
2431
+ "model.layers.71.self_attn.weights_proj": {
2432
+ "bits": 8
2433
+ },
2434
+ "model.layers.71.self_attn.wk": {
2435
+ "bits": 8
2436
+ },
2437
+ "model.layers.71.self_attn.wq_b": {
2438
+ "bits": 8
2439
+ },
2440
+ "model.layers.72.mlp.shared_experts.down_proj": {
2441
+ "bits": 8
2442
+ },
2443
+ "model.layers.72.mlp.shared_experts.gate_proj": {
2444
+ "bits": 8
2445
+ },
2446
+ "model.layers.72.mlp.shared_experts.up_proj": {
2447
+ "bits": 8
2448
+ },
2449
+ "model.layers.72.self_attn.kv_a_proj_with_mqa": {
2450
+ "bits": 8
2451
+ },
2452
+ "model.layers.72.self_attn.kv_b_proj": {
2453
+ "bits": 8
2454
+ },
2455
+ "model.layers.72.self_attn.o_proj": {
2456
+ "bits": 8
2457
+ },
2458
+ "model.layers.72.self_attn.q_a_proj": {
2459
+ "bits": 8
2460
+ },
2461
+ "model.layers.72.self_attn.q_b_proj": {
2462
+ "bits": 8
2463
+ },
2464
+ "model.layers.72.self_attn.weights_proj": {
2465
+ "bits": 8
2466
+ },
2467
+ "model.layers.72.self_attn.wk": {
2468
+ "bits": 8
2469
+ },
2470
+ "model.layers.72.self_attn.wq_b": {
2471
+ "bits": 8
2472
+ },
2473
+ "model.layers.73.mlp.shared_experts.down_proj": {
2474
+ "bits": 8
2475
+ },
2476
+ "model.layers.73.mlp.shared_experts.gate_proj": {
2477
+ "bits": 8
2478
+ },
2479
+ "model.layers.73.mlp.shared_experts.up_proj": {
2480
+ "bits": 8
2481
+ },
2482
+ "model.layers.73.self_attn.kv_a_proj_with_mqa": {
2483
+ "bits": 8
2484
+ },
2485
+ "model.layers.73.self_attn.kv_b_proj": {
2486
+ "bits": 8
2487
+ },
2488
+ "model.layers.73.self_attn.o_proj": {
2489
+ "bits": 8
2490
+ },
2491
+ "model.layers.73.self_attn.q_a_proj": {
2492
+ "bits": 8
2493
+ },
2494
+ "model.layers.73.self_attn.q_b_proj": {
2495
+ "bits": 8
2496
+ },
2497
+ "model.layers.73.self_attn.weights_proj": {
2498
+ "bits": 8
2499
+ },
2500
+ "model.layers.73.self_attn.wk": {
2501
+ "bits": 8
2502
+ },
2503
+ "model.layers.73.self_attn.wq_b": {
2504
+ "bits": 8
2505
+ },
2506
+ "model.layers.74.mlp.shared_experts.down_proj": {
2507
+ "bits": 8
2508
+ },
2509
+ "model.layers.74.mlp.shared_experts.gate_proj": {
2510
+ "bits": 8
2511
+ },
2512
+ "model.layers.74.mlp.shared_experts.up_proj": {
2513
+ "bits": 8
2514
+ },
2515
+ "model.layers.74.self_attn.kv_a_proj_with_mqa": {
2516
+ "bits": 8
2517
+ },
2518
+ "model.layers.74.self_attn.kv_b_proj": {
2519
+ "bits": 8
2520
+ },
2521
+ "model.layers.74.self_attn.o_proj": {
2522
+ "bits": 8
2523
+ },
2524
+ "model.layers.74.self_attn.q_a_proj": {
2525
+ "bits": 8
2526
+ },
2527
+ "model.layers.74.self_attn.q_b_proj": {
2528
+ "bits": 8
2529
+ },
2530
+ "model.layers.74.self_attn.weights_proj": {
2531
+ "bits": 8
2532
+ },
2533
+ "model.layers.74.self_attn.wk": {
2534
+ "bits": 8
2535
+ },
2536
+ "model.layers.74.self_attn.wq_b": {
2537
+ "bits": 8
2538
+ },
2539
+ "model.layers.75.mlp.shared_experts.down_proj": {
2540
+ "bits": 8
2541
+ },
2542
+ "model.layers.75.mlp.shared_experts.gate_proj": {
2543
+ "bits": 8
2544
+ },
2545
+ "model.layers.75.mlp.shared_experts.up_proj": {
2546
+ "bits": 8
2547
+ },
2548
+ "model.layers.75.self_attn.kv_a_proj_with_mqa": {
2549
+ "bits": 8
2550
+ },
2551
+ "model.layers.75.self_attn.kv_b_proj": {
2552
+ "bits": 8
2553
+ },
2554
+ "model.layers.75.self_attn.o_proj": {
2555
+ "bits": 8
2556
+ },
2557
+ "model.layers.75.self_attn.q_a_proj": {
2558
+ "bits": 8
2559
+ },
2560
+ "model.layers.75.self_attn.q_b_proj": {
2561
+ "bits": 8
2562
+ },
2563
+ "model.layers.75.self_attn.weights_proj": {
2564
+ "bits": 8
2565
+ },
2566
+ "model.layers.75.self_attn.wk": {
2567
+ "bits": 8
2568
+ },
2569
+ "model.layers.75.self_attn.wq_b": {
2570
+ "bits": 8
2571
+ },
2572
+ "model.layers.76.mlp.shared_experts.down_proj": {
2573
+ "bits": 8
2574
+ },
2575
+ "model.layers.76.mlp.shared_experts.gate_proj": {
2576
+ "bits": 8
2577
+ },
2578
+ "model.layers.76.mlp.shared_experts.up_proj": {
2579
+ "bits": 8
2580
+ },
2581
+ "model.layers.76.self_attn.kv_a_proj_with_mqa": {
2582
+ "bits": 8
2583
+ },
2584
+ "model.layers.76.self_attn.kv_b_proj": {
2585
+ "bits": 8
2586
+ },
2587
+ "model.layers.76.self_attn.o_proj": {
2588
+ "bits": 8
2589
+ },
2590
+ "model.layers.76.self_attn.q_a_proj": {
2591
+ "bits": 8
2592
+ },
2593
+ "model.layers.76.self_attn.q_b_proj": {
2594
+ "bits": 8
2595
+ },
2596
+ "model.layers.76.self_attn.weights_proj": {
2597
+ "bits": 8
2598
+ },
2599
+ "model.layers.76.self_attn.wk": {
2600
+ "bits": 8
2601
+ },
2602
+ "model.layers.76.self_attn.wq_b": {
2603
+ "bits": 8
2604
+ },
2605
+ "model.layers.77.mlp.shared_experts.down_proj": {
2606
+ "bits": 8
2607
+ },
2608
+ "model.layers.77.mlp.shared_experts.gate_proj": {
2609
+ "bits": 8
2610
+ },
2611
+ "model.layers.77.mlp.shared_experts.up_proj": {
2612
+ "bits": 8
2613
+ },
2614
+ "model.layers.77.self_attn.kv_a_proj_with_mqa": {
2615
+ "bits": 8
2616
+ },
2617
+ "model.layers.77.self_attn.kv_b_proj": {
2618
+ "bits": 8
2619
+ },
2620
+ "model.layers.77.self_attn.o_proj": {
2621
+ "bits": 8
2622
+ },
2623
+ "model.layers.77.self_attn.q_a_proj": {
2624
+ "bits": 8
2625
+ },
2626
+ "model.layers.77.self_attn.q_b_proj": {
2627
+ "bits": 8
2628
+ },
2629
+ "model.layers.77.self_attn.weights_proj": {
2630
+ "bits": 8
2631
+ },
2632
+ "model.layers.77.self_attn.wk": {
2633
+ "bits": 8
2634
+ },
2635
+ "model.layers.77.self_attn.wq_b": {
2636
+ "bits": 8
2637
+ },
2638
+ "model.layers.8.mlp.shared_experts.down_proj": {
2639
+ "bits": 8
2640
+ },
2641
+ "model.layers.8.mlp.shared_experts.gate_proj": {
2642
+ "bits": 8
2643
+ },
2644
+ "model.layers.8.mlp.shared_experts.up_proj": {
2645
+ "bits": 8
2646
+ },
2647
+ "model.layers.8.self_attn.kv_a_proj_with_mqa": {
2648
+ "bits": 8
2649
+ },
2650
+ "model.layers.8.self_attn.kv_b_proj": {
2651
+ "bits": 8
2652
+ },
2653
+ "model.layers.8.self_attn.o_proj": {
2654
+ "bits": 8
2655
+ },
2656
+ "model.layers.8.self_attn.q_a_proj": {
2657
+ "bits": 8
2658
+ },
2659
+ "model.layers.8.self_attn.q_b_proj": {
2660
+ "bits": 8
2661
+ },
2662
+ "model.layers.8.self_attn.weights_proj": {
2663
+ "bits": 8
2664
+ },
2665
+ "model.layers.8.self_attn.wk": {
2666
+ "bits": 8
2667
+ },
2668
+ "model.layers.8.self_attn.wq_b": {
2669
+ "bits": 8
2670
+ },
2671
+ "model.layers.9.mlp.shared_experts.down_proj": {
2672
+ "bits": 8
2673
+ },
2674
+ "model.layers.9.mlp.shared_experts.gate_proj": {
2675
+ "bits": 8
2676
+ },
2677
+ "model.layers.9.mlp.shared_experts.up_proj": {
2678
+ "bits": 8
2679
+ },
2680
+ "model.layers.9.self_attn.kv_a_proj_with_mqa": {
2681
+ "bits": 8
2682
+ },
2683
+ "model.layers.9.self_attn.kv_b_proj": {
2684
+ "bits": 8
2685
+ },
2686
+ "model.layers.9.self_attn.o_proj": {
2687
+ "bits": 8
2688
+ },
2689
+ "model.layers.9.self_attn.q_a_proj": {
2690
+ "bits": 8
2691
+ },
2692
+ "model.layers.9.self_attn.q_b_proj": {
2693
+ "bits": 8
2694
+ },
2695
+ "model.layers.9.self_attn.weights_proj": {
2696
+ "bits": 8
2697
+ },
2698
+ "model.layers.9.self_attn.wk": {
2699
+ "bits": 8
2700
+ },
2701
+ "model.layers.9.self_attn.wq_b": {
2702
+ "bits": 8
2703
+ }
2704
+ },
2705
+ "group_size": 128,
2706
+ "iters": 0,
2707
+ "packing_format": "auto_round:auto_gptq",
2708
+ "quant_method": "auto-round",
2709
+ "sym": true
2710
+ },
2711
+ "rms_norm_eps": 1e-05,
2712
+ "rope_interleave": true,
2713
+ "rope_parameters": {
2714
+ "rope_theta": 1000000,
2715
+ "rope_type": "default"
2716
+ },
2717
+ "routed_scaling_factor": 2.5,
2718
+ "scoring_func": "sigmoid",
2719
+ "tie_word_embeddings": false,
2720
+ "topk_group": 1,
2721
+ "topk_method": "noaux_tc",
2722
+ "transformers_version": "5.2.0.dev0",
2723
+ "use_cache": true,
2724
+ "v_head_dim": 256,
2725
+ "vocab_size": 154880
2726
+ }
model-00001-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b83e9d0ad0a56dac2770a8af57bf576dce3cb904773a5422be0c5f9bfed25ab
3
+ size 5366923072
model-00002-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e0fdede0711cdc57bd4c34d56034b6680502164ae686febb5a6e7a3de5653534
3
+ size 5362713760
model-00003-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1431458a88bb44a875b48f3d882e21835a9dba6da4440154fba9d59b2566c79d
3
+ size 5362713784
model-00004-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3847a69c2a3b79c1b84e6738d38e8d596cf80981aa838631ec4ff0b99128fe91
3
+ size 5362713816
model-00005-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:99305ad6522849718bb58af19270edc62b9c1271823ca599723a7ee77b11d46f
3
+ size 5362713816
model-00006-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a49bcd95f32ed0db3d10ec7cc5968d2a2dd213da1e4a6d8fc27fbbbe4bc47e71
3
+ size 5362713816
model-00007-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f8a1faf2f6d31e8949884e62ae749d2e8c4b0578540c6375a5b2e10187fb569
3
+ size 5362713816
model-00008-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e157b39397847af0c220f68e6ecde1d99fd681db8a55887e98647f21d06a31f
3
+ size 5362715144
model-00009-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff1e1b5087e555851b2309883a059b40ba660b8e8fb6a418f0939df152c2befb
3
+ size 5362715432
model-00010-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1fdbf5319f5c8c0ea30a2efdcea4795857e977ac297ee32d6b5d3757adbbc98
3
+ size 5362715432
model-00011-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ac7784feb4019ff13f6f60000365aca8b42216ffaf3fcb2bd15fa2221662e648
3
+ size 5362715432
model-00012-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c1194b561e892c9bc098440d0dff6606f2be0254a27dfde18a761f5a551cb57
3
+ size 5362715432
model-00013-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf9e0b38c45a7b9544e4375926eef734bca8b0cb7c489565ca2333c8493977bc
3
+ size 5368859320
model-00014-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff7a47582aeb84db0329cd918c78047028c35b6460b3d8edbb2e1aa29cccdc80
3
+ size 5356470896
model-00015-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:765ed1987eb01eb0e308c982c52985918b8e29f430cf3199cc4675ca1c9dc37f
3
+ size 5367674528
model-00016-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c14f17bcf6c62cda9c2acd7d2dcc453bfb06e8151888c21eff1a15d39a9faf6d
3
+ size 5366379680
model-00017-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ade7714806556ed1250f38bbdbc1101cdb3398b45c82778e0d18f7f92cb9b36a
3
+ size 5368856712
model-00018-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b75c8a14e676558434f89230178c0b40244ae9698eff7e847f922da8d6f13a80
3
+ size 5356568600
model-00019-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1d3cf8e26ec03c8191ad641dfe6ae16812e434d7ed51c5751687cc2ad3080d0e
3
+ size 5368856728
model-00020-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2a06dac3bf746d1b0ba890a68e72c613a3cb076068841cee8f4e8b9dac44a46e
3
+ size 5356568600
model-00021-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94df641d686eadab356110218e69571c8f22813b0f71118a049ebdefce303ac2
3
+ size 5368856728
model-00022-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5b3fea9d60d692a2bd38426bf8fb90b769b836a1a846c46c97b369637e019c9
3
+ size 5356568600
model-00023-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b19b0b0ce7a233006cb6b896c4ce3a2f721113e143794eee2454fcd05477961
3
+ size 5368856728
model-00024-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8cf773480a1d4638f8098e8bc3f47c5eade4aef2c5f24c928f52ab67ad270290
3
+ size 5356568600
model-00025-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0307b2efa9b2086555fee2c9355f1d37c09f04699c4f774a51a0345c371f82ae
3
+ size 5368856728
model-00026-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e6231d0db0663c90be2b0b6693b5cb499900e34e8f829b08c05a4b666b41579
3
+ size 5356568600
model-00027-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ac0df36d2e307cce9d461c594722eba2f46a19f24fd3c8dd59523cac389c85f
3
+ size 5368856728
model-00028-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:216acd17044e6d3da142ee2d560cf11c7e46b582b0c91baf402d0d6e277e909e
3
+ size 5356568600
model-00029-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2259be935a9e8e7732a90bcb53c2064ab05e42be611d246a5be30efbb7392725
3
+ size 5368856736
model-00030-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c2e5e8deac03ddb66e3fbc3feb2796da710d81703d9a2dcae1ebbbc6acb61e1f
3
+ size 5356568624
model-00031-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:589b2668d458c8c9d2c07ac3235da199e560613a2218522db6169e2b2672af86
3
+ size 5368856752
model-00032-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:91568559fe6de526fff2613e340a48945c8c72225c286ba63ca399bcfd283765
3
+ size 5356568624
model-00033-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:233f7746641c895661d3d81d25eabaea200f69ed63b0dbd254418430096052ac
3
+ size 5368856752
model-00034-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3459522256bdf4f13acfc381a7d3f5a7f1787f06943c870c20e1ea723709708
3
+ size 5356568624
model-00035-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:525a9afbc52bbf13fe23bb0d6ef5ed06d0dc44c324775b7b88aa3d29cd25c2a6
3
+ size 5368856752
model-00036-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8593a90dcc457f392be616c184f89698c331967017a0c5c929cb9dcf14817211
3
+ size 5356568624
model-00037-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49ce7aecc9f4b28c5f3e1cfcf585dae5f7c47a9ca3cd0e0e34640437a00b045f
3
+ size 5368856752
model-00038-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3951004bad0f36dc24ab7e54b38cf44b3869a5613bf61f6f9a3badabb8f9236
3
+ size 5356568624
model-00039-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e2c9db6c9ff59673047bc8e7a20e4c18d2f4d926b10169b936154af91da6fe9
3
+ size 5368856752
model-00040-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c37da3518ab4cbcfd5757bd815b3e908441c69e65f6cc177f5dcb9946bc393a
3
+ size 5356568624
model-00041-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f7397a584b016ffc46120481fb864fa3c882b7f0bef8711cbc3c0e9c7e7f921
3
+ size 5368856752
model-00042-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7848d34b3af5defceeb87a1c77ab21ddaccebd4db071148fe5cee4cf3b086c7b
3
+ size 5356568624
model-00043-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7ecf7c5f6d3c4f353508036455df4f2abf286505902dc67c33fe4affee5dd07a
3
+ size 5368856752
model-00044-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e18e5b3098a8380cbc43fc14d764d035eee5dc987a00ce23a0f6bf4cbf36936f
3
+ size 5356568624
model-00045-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8b9cb254a1e1eb1b192d1b10b3410bde1e05187cd02cbf502caa358b0b7286f3
3
+ size 5368856752
model-00046-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:445f7e4ee11539b998415705c161b16a65438ab4fa55edead111f3774a5c7895
3
+ size 5356568624
model-00047-of-00076.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d1d44e3f80307fccbe11a058e04aeb02490e3bfe185effeabd39fb35bcd5df9f
3
+ size 5368856752