Hugging Face's logo

mlx-community
/
DeepSeek-R1-Distill-Qwen-14B-MLX

Text Generation
MLX
Safetensors
English
chat
conversations
Model card Files Community
2
sealad886 commited on
Commit
9774084
·
verified ·
1 Parent(s): 10a0e48

Add files using upload-large-folder tool

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 +11 -0
  2. DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/config.json +3029 -0
  3. DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/model.safetensors +3 -0
  4. DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/model.safetensors.index.json +1262 -0
  5. DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/special_tokens_map.json +23 -0
  6. DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/tokenizer.json +3 -0
  7. DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/tokenizer_config.json +195 -0
  8. DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/config.json +3029 -0
  9. DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/model-00001-of-00002.safetensors +3 -0
  10. DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/model-00002-of-00002.safetensors +3 -0
  11. DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/model.safetensors.index.json +0 -0
  12. DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/special_tokens_map.json +23 -0
  13. DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/tokenizer.json +3 -0
  14. DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/tokenizer_config.json +195 -0
  15. DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/config.json +3029 -0
  16. DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/model-00001-of-00002.safetensors +3 -0
  17. DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/model-00002-of-00002.safetensors +3 -0
  18. DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/model.safetensors.index.json +0 -0
  19. DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/special_tokens_map.json +23 -0
  20. DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/tokenizer.json +3 -0
  21. DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/tokenizer_config.json +195 -0
  22. DeepSeek-R1-Distill-Qwen-14B-3bit/config.json +35 -0
  23. DeepSeek-R1-Distill-Qwen-14B-3bit/model-00001-of-00002.safetensors +3 -0
  24. DeepSeek-R1-Distill-Qwen-14B-3bit/model-00002-of-00002.safetensors +3 -0
  25. DeepSeek-R1-Distill-Qwen-14B-3bit/model.safetensors.index.json +0 -0
  26. DeepSeek-R1-Distill-Qwen-14B-3bit/special_tokens_map.json +23 -0
  27. DeepSeek-R1-Distill-Qwen-14B-3bit/tokenizer.json +3 -0
  28. DeepSeek-R1-Distill-Qwen-14B-3bit/tokenizer_config.json +195 -0
  29. DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/config.json +3029 -0
  30. DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/model-00001-of-00002.safetensors +3 -0
  31. DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/model-00002-of-00002.safetensors +3 -0
  32. DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/model.safetensors.index.json +0 -0
  33. DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/special_tokens_map.json +23 -0
  34. DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/tokenizer.json +3 -0
  35. DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/tokenizer_config.json +195 -0
  36. DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/config.json +3029 -0
  37. DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/model-00001-of-00002.safetensors +3 -0
  38. DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/model-00002-of-00002.safetensors +3 -0
  39. DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/model.safetensors.index.json +0 -0
  40. DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/special_tokens_map.json +23 -0
  41. DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/tokenizer.json +3 -0
  42. DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/tokenizer_config.json +195 -0
  43. DeepSeek-R1-Distill-Qwen-14B-4bit/config.json +35 -0
  44. DeepSeek-R1-Distill-Qwen-14B-4bit/model-00001-of-00002.safetensors +3 -0
  45. DeepSeek-R1-Distill-Qwen-14B-4bit/model-00002-of-00002.safetensors +3 -0
  46. DeepSeek-R1-Distill-Qwen-14B-4bit/model.safetensors.index.json +0 -0
  47. DeepSeek-R1-Distill-Qwen-14B-4bit/special_tokens_map.json +23 -0
  48. DeepSeek-R1-Distill-Qwen-14B-4bit/tokenizer.json +3 -0
  49. DeepSeek-R1-Distill-Qwen-14B-4bit/tokenizer_config.json +195 -0
  50. DeepSeek-R1-Distill-Qwen-14B-6bit/config.json +35 -0
.gitattributes CHANGED
@@ -42,3 +42,14 @@ DeepSeek-R1-Distill-Qwen-14B_3bit/tokenizer.json filter=lfs diff=lfs merge=lfs -
42
  DeepSeek-R1-Distill-Qwen-14B_2,6_mixed/tokenizer.json filter=lfs diff=lfs merge=lfs -text
43
  DeepSeek-R1-Distill-Qwen-14B_3,6_mixed/tokenizer.json filter=lfs diff=lfs merge=lfs -text
44
  DeepSeek-R1-Distill-Qwen-14B_6bit/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
42
  DeepSeek-R1-Distill-Qwen-14B_2,6_mixed/tokenizer.json filter=lfs diff=lfs merge=lfs -text
43
  DeepSeek-R1-Distill-Qwen-14B_3,6_mixed/tokenizer.json filter=lfs diff=lfs merge=lfs -text
44
  DeepSeek-R1-Distill-Qwen-14B_6bit/tokenizer.json filter=lfs diff=lfs merge=lfs -text
45
+ DeepSeek-R1-Distill-Qwen-14B-float16/tokenizer.json filter=lfs diff=lfs merge=lfs -text
46
+ DeepSeek-R1-Distill-Qwen-14B-4bit/tokenizer.json filter=lfs diff=lfs merge=lfs -text
47
+ DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/tokenizer.json filter=lfs diff=lfs merge=lfs -text
48
+ DeepSeek-R1-Distill-Qwen-14B-6bit/tokenizer.json filter=lfs diff=lfs merge=lfs -text
49
+ DeepSeek-R1-Distill-Qwen-14B-bfloat16/tokenizer.json filter=lfs diff=lfs merge=lfs -text
50
+ DeepSeek-R1-Distill-Qwen-14B-3bit/tokenizer.json filter=lfs diff=lfs merge=lfs -text
51
+ DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/tokenizer.json filter=lfs diff=lfs merge=lfs -text
52
+ DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/tokenizer.json filter=lfs diff=lfs merge=lfs -text
53
+ DeepSeek-R1-Distill-Qwen-14B-8bit/tokenizer.json filter=lfs diff=lfs merge=lfs -text
54
+ DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/tokenizer.json filter=lfs diff=lfs merge=lfs -text
55
+ DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/tokenizer.json filter=lfs diff=lfs merge=lfs -text
DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/config.json ADDED
@@ -0,0 +1,3029 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151643,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 5120,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 13824,
12
+ "max_position_embeddings": 131072,
13
+ "max_window_layers": 48,
14
+ "model_type": "qwen2",
15
+ "num_attention_heads": 40,
16
+ "num_hidden_layers": 48,
17
+ "num_key_value_heads": 8,
18
+ "quantization": {
19
+ "group_size": 64,
20
+ "bits": null,
21
+ "model.embed_tokens": {
22
+ "group_size": 64,
23
+ "bits": 2
24
+ },
25
+ "model.layers.0.self_attn.q_proj": {
26
+ "group_size": 64,
27
+ "bits": 2
28
+ },
29
+ "model.layers.0.self_attn.k_proj": {
30
+ "group_size": 64,
31
+ "bits": 2
32
+ },
33
+ "model.layers.0.self_attn.v_proj": {
34
+ "group_size": 64,
35
+ "bits": 4
36
+ },
37
+ "model.layers.0.self_attn.o_proj": {
38
+ "group_size": 64,
39
+ "bits": 2
40
+ },
41
+ "model.layers.0.self_attn.rope": false,
42
+ "model.layers.0.mlp.gate_proj": {
43
+ "group_size": 64,
44
+ "bits": 2
45
+ },
46
+ "model.layers.0.mlp.down_proj": {
47
+ "group_size": 64,
48
+ "bits": 4
49
+ },
50
+ "model.layers.0.mlp.up_proj": {
51
+ "group_size": 64,
52
+ "bits": 2
53
+ },
54
+ "model.layers.0.input_layernorm": false,
55
+ "model.layers.0.post_attention_layernorm": false,
56
+ "model.layers.1.self_attn.q_proj": {
57
+ "group_size": 64,
58
+ "bits": 2
59
+ },
60
+ "model.layers.1.self_attn.k_proj": {
61
+ "group_size": 64,
62
+ "bits": 2
63
+ },
64
+ "model.layers.1.self_attn.v_proj": {
65
+ "group_size": 64,
66
+ "bits": 4
67
+ },
68
+ "model.layers.1.self_attn.o_proj": {
69
+ "group_size": 64,
70
+ "bits": 2
71
+ },
72
+ "model.layers.1.self_attn.rope": false,
73
+ "model.layers.1.mlp.gate_proj": {
74
+ "group_size": 64,
75
+ "bits": 2
76
+ },
77
+ "model.layers.1.mlp.down_proj": {
78
+ "group_size": 64,
79
+ "bits": 4
80
+ },
81
+ "model.layers.1.mlp.up_proj": {
82
+ "group_size": 64,
83
+ "bits": 2
84
+ },
85
+ "model.layers.1.input_layernorm": false,
86
+ "model.layers.1.post_attention_layernorm": false,
87
+ "model.layers.2.self_attn.q_proj": {
88
+ "group_size": 64,
89
+ "bits": 2
90
+ },
91
+ "model.layers.2.self_attn.k_proj": {
92
+ "group_size": 64,
93
+ "bits": 2
94
+ },
95
+ "model.layers.2.self_attn.v_proj": {
96
+ "group_size": 64,
97
+ "bits": 4
98
+ },
99
+ "model.layers.2.self_attn.o_proj": {
100
+ "group_size": 64,
101
+ "bits": 2
102
+ },
103
+ "model.layers.2.self_attn.rope": false,
104
+ "model.layers.2.mlp.gate_proj": {
105
+ "group_size": 64,
106
+ "bits": 2
107
+ },
108
+ "model.layers.2.mlp.down_proj": {
109
+ "group_size": 64,
110
+ "bits": 4
111
+ },
112
+ "model.layers.2.mlp.up_proj": {
113
+ "group_size": 64,
114
+ "bits": 2
115
+ },
116
+ "model.layers.2.input_layernorm": false,
117
+ "model.layers.2.post_attention_layernorm": false,
118
+ "model.layers.3.self_attn.q_proj": {
119
+ "group_size": 64,
120
+ "bits": 2
121
+ },
122
+ "model.layers.3.self_attn.k_proj": {
123
+ "group_size": 64,
124
+ "bits": 2
125
+ },
126
+ "model.layers.3.self_attn.v_proj": {
127
+ "group_size": 64,
128
+ "bits": 4
129
+ },
130
+ "model.layers.3.self_attn.o_proj": {
131
+ "group_size": 64,
132
+ "bits": 2
133
+ },
134
+ "model.layers.3.self_attn.rope": false,
135
+ "model.layers.3.mlp.gate_proj": {
136
+ "group_size": 64,
137
+ "bits": 2
138
+ },
139
+ "model.layers.3.mlp.down_proj": {
140
+ "group_size": 64,
141
+ "bits": 4
142
+ },
143
+ "model.layers.3.mlp.up_proj": {
144
+ "group_size": 64,
145
+ "bits": 2
146
+ },
147
+ "model.layers.3.input_layernorm": false,
148
+ "model.layers.3.post_attention_layernorm": false,
149
+ "model.layers.4.self_attn.q_proj": {
150
+ "group_size": 64,
151
+ "bits": 2
152
+ },
153
+ "model.layers.4.self_attn.k_proj": {
154
+ "group_size": 64,
155
+ "bits": 2
156
+ },
157
+ "model.layers.4.self_attn.v_proj": {
158
+ "group_size": 64,
159
+ "bits": 4
160
+ },
161
+ "model.layers.4.self_attn.o_proj": {
162
+ "group_size": 64,
163
+ "bits": 2
164
+ },
165
+ "model.layers.4.self_attn.rope": false,
166
+ "model.layers.4.mlp.gate_proj": {
167
+ "group_size": 64,
168
+ "bits": 2
169
+ },
170
+ "model.layers.4.mlp.down_proj": {
171
+ "group_size": 64,
172
+ "bits": 4
173
+ },
174
+ "model.layers.4.mlp.up_proj": {
175
+ "group_size": 64,
176
+ "bits": 2
177
+ },
178
+ "model.layers.4.input_layernorm": false,
179
+ "model.layers.4.post_attention_layernorm": false,
180
+ "model.layers.5.self_attn.q_proj": {
181
+ "group_size": 64,
182
+ "bits": 2
183
+ },
184
+ "model.layers.5.self_attn.k_proj": {
185
+ "group_size": 64,
186
+ "bits": 2
187
+ },
188
+ "model.layers.5.self_attn.v_proj": {
189
+ "group_size": 64,
190
+ "bits": 4
191
+ },
192
+ "model.layers.5.self_attn.o_proj": {
193
+ "group_size": 64,
194
+ "bits": 2
195
+ },
196
+ "model.layers.5.self_attn.rope": false,
197
+ "model.layers.5.mlp.gate_proj": {
198
+ "group_size": 64,
199
+ "bits": 2
200
+ },
201
+ "model.layers.5.mlp.down_proj": {
202
+ "group_size": 64,
203
+ "bits": 4
204
+ },
205
+ "model.layers.5.mlp.up_proj": {
206
+ "group_size": 64,
207
+ "bits": 2
208
+ },
209
+ "model.layers.5.input_layernorm": false,
210
+ "model.layers.5.post_attention_layernorm": false,
211
+ "model.layers.6.self_attn.q_proj": {
212
+ "group_size": 64,
213
+ "bits": 2
214
+ },
215
+ "model.layers.6.self_attn.k_proj": {
216
+ "group_size": 64,
217
+ "bits": 2
218
+ },
219
+ "model.layers.6.self_attn.v_proj": {
220
+ "group_size": 64,
221
+ "bits": 2
222
+ },
223
+ "model.layers.6.self_attn.o_proj": {
224
+ "group_size": 64,
225
+ "bits": 2
226
+ },
227
+ "model.layers.6.self_attn.rope": false,
228
+ "model.layers.6.mlp.gate_proj": {
229
+ "group_size": 64,
230
+ "bits": 2
231
+ },
232
+ "model.layers.6.mlp.down_proj": {
233
+ "group_size": 64,
234
+ "bits": 2
235
+ },
236
+ "model.layers.6.mlp.up_proj": {
237
+ "group_size": 64,
238
+ "bits": 2
239
+ },
240
+ "model.layers.6.input_layernorm": false,
241
+ "model.layers.6.post_attention_layernorm": false,
242
+ "model.layers.7.self_attn.q_proj": {
243
+ "group_size": 64,
244
+ "bits": 2
245
+ },
246
+ "model.layers.7.self_attn.k_proj": {
247
+ "group_size": 64,
248
+ "bits": 2
249
+ },
250
+ "model.layers.7.self_attn.v_proj": {
251
+ "group_size": 64,
252
+ "bits": 2
253
+ },
254
+ "model.layers.7.self_attn.o_proj": {
255
+ "group_size": 64,
256
+ "bits": 2
257
+ },
258
+ "model.layers.7.self_attn.rope": false,
259
+ "model.layers.7.mlp.gate_proj": {
260
+ "group_size": 64,
261
+ "bits": 2
262
+ },
263
+ "model.layers.7.mlp.down_proj": {
264
+ "group_size": 64,
265
+ "bits": 2
266
+ },
267
+ "model.layers.7.mlp.up_proj": {
268
+ "group_size": 64,
269
+ "bits": 2
270
+ },
271
+ "model.layers.7.input_layernorm": false,
272
+ "model.layers.7.post_attention_layernorm": false,
273
+ "model.layers.8.self_attn.q_proj": {
274
+ "group_size": 64,
275
+ "bits": 2
276
+ },
277
+ "model.layers.8.self_attn.k_proj": {
278
+ "group_size": 64,
279
+ "bits": 2
280
+ },
281
+ "model.layers.8.self_attn.v_proj": {
282
+ "group_size": 64,
283
+ "bits": 4
284
+ },
285
+ "model.layers.8.self_attn.o_proj": {
286
+ "group_size": 64,
287
+ "bits": 2
288
+ },
289
+ "model.layers.8.self_attn.rope": false,
290
+ "model.layers.8.mlp.gate_proj": {
291
+ "group_size": 64,
292
+ "bits": 2
293
+ },
294
+ "model.layers.8.mlp.down_proj": {
295
+ "group_size": 64,
296
+ "bits": 4
297
+ },
298
+ "model.layers.8.mlp.up_proj": {
299
+ "group_size": 64,
300
+ "bits": 2
301
+ },
302
+ "model.layers.8.input_layernorm": false,
303
+ "model.layers.8.post_attention_layernorm": false,
304
+ "model.layers.9.self_attn.q_proj": {
305
+ "group_size": 64,
306
+ "bits": 2
307
+ },
308
+ "model.layers.9.self_attn.k_proj": {
309
+ "group_size": 64,
310
+ "bits": 2
311
+ },
312
+ "model.layers.9.self_attn.v_proj": {
313
+ "group_size": 64,
314
+ "bits": 2
315
+ },
316
+ "model.layers.9.self_attn.o_proj": {
317
+ "group_size": 64,
318
+ "bits": 2
319
+ },
320
+ "model.layers.9.self_attn.rope": false,
321
+ "model.layers.9.mlp.gate_proj": {
322
+ "group_size": 64,
323
+ "bits": 2
324
+ },
325
+ "model.layers.9.mlp.down_proj": {
326
+ "group_size": 64,
327
+ "bits": 2
328
+ },
329
+ "model.layers.9.mlp.up_proj": {
330
+ "group_size": 64,
331
+ "bits": 2
332
+ },
333
+ "model.layers.9.input_layernorm": false,
334
+ "model.layers.9.post_attention_layernorm": false,
335
+ "model.layers.10.self_attn.q_proj": {
336
+ "group_size": 64,
337
+ "bits": 2
338
+ },
339
+ "model.layers.10.self_attn.k_proj": {
340
+ "group_size": 64,
341
+ "bits": 2
342
+ },
343
+ "model.layers.10.self_attn.v_proj": {
344
+ "group_size": 64,
345
+ "bits": 2
346
+ },
347
+ "model.layers.10.self_attn.o_proj": {
348
+ "group_size": 64,
349
+ "bits": 2
350
+ },
351
+ "model.layers.10.self_attn.rope": false,
352
+ "model.layers.10.mlp.gate_proj": {
353
+ "group_size": 64,
354
+ "bits": 2
355
+ },
356
+ "model.layers.10.mlp.down_proj": {
357
+ "group_size": 64,
358
+ "bits": 2
359
+ },
360
+ "model.layers.10.mlp.up_proj": {
361
+ "group_size": 64,
362
+ "bits": 2
363
+ },
364
+ "model.layers.10.input_layernorm": false,
365
+ "model.layers.10.post_attention_layernorm": false,
366
+ "model.layers.11.self_attn.q_proj": {
367
+ "group_size": 64,
368
+ "bits": 2
369
+ },
370
+ "model.layers.11.self_attn.k_proj": {
371
+ "group_size": 64,
372
+ "bits": 2
373
+ },
374
+ "model.layers.11.self_attn.v_proj": {
375
+ "group_size": 64,
376
+ "bits": 4
377
+ },
378
+ "model.layers.11.self_attn.o_proj": {
379
+ "group_size": 64,
380
+ "bits": 2
381
+ },
382
+ "model.layers.11.self_attn.rope": false,
383
+ "model.layers.11.mlp.gate_proj": {
384
+ "group_size": 64,
385
+ "bits": 2
386
+ },
387
+ "model.layers.11.mlp.down_proj": {
388
+ "group_size": 64,
389
+ "bits": 4
390
+ },
391
+ "model.layers.11.mlp.up_proj": {
392
+ "group_size": 64,
393
+ "bits": 2
394
+ },
395
+ "model.layers.11.input_layernorm": false,
396
+ "model.layers.11.post_attention_layernorm": false,
397
+ "model.layers.12.self_attn.q_proj": {
398
+ "group_size": 64,
399
+ "bits": 2
400
+ },
401
+ "model.layers.12.self_attn.k_proj": {
402
+ "group_size": 64,
403
+ "bits": 2
404
+ },
405
+ "model.layers.12.self_attn.v_proj": {
406
+ "group_size": 64,
407
+ "bits": 2
408
+ },
409
+ "model.layers.12.self_attn.o_proj": {
410
+ "group_size": 64,
411
+ "bits": 2
412
+ },
413
+ "model.layers.12.self_attn.rope": false,
414
+ "model.layers.12.mlp.gate_proj": {
415
+ "group_size": 64,
416
+ "bits": 2
417
+ },
418
+ "model.layers.12.mlp.down_proj": {
419
+ "group_size": 64,
420
+ "bits": 2
421
+ },
422
+ "model.layers.12.mlp.up_proj": {
423
+ "group_size": 64,
424
+ "bits": 2
425
+ },
426
+ "model.layers.12.input_layernorm": false,
427
+ "model.layers.12.post_attention_layernorm": false,
428
+ "model.layers.13.self_attn.q_proj": {
429
+ "group_size": 64,
430
+ "bits": 2
431
+ },
432
+ "model.layers.13.self_attn.k_proj": {
433
+ "group_size": 64,
434
+ "bits": 2
435
+ },
436
+ "model.layers.13.self_attn.v_proj": {
437
+ "group_size": 64,
438
+ "bits": 2
439
+ },
440
+ "model.layers.13.self_attn.o_proj": {
441
+ "group_size": 64,
442
+ "bits": 2
443
+ },
444
+ "model.layers.13.self_attn.rope": false,
445
+ "model.layers.13.mlp.gate_proj": {
446
+ "group_size": 64,
447
+ "bits": 2
448
+ },
449
+ "model.layers.13.mlp.down_proj": {
450
+ "group_size": 64,
451
+ "bits": 2
452
+ },
453
+ "model.layers.13.mlp.up_proj": {
454
+ "group_size": 64,
455
+ "bits": 2
456
+ },
457
+ "model.layers.13.input_layernorm": false,
458
+ "model.layers.13.post_attention_layernorm": false,
459
+ "model.layers.14.self_attn.q_proj": {
460
+ "group_size": 64,
461
+ "bits": 2
462
+ },
463
+ "model.layers.14.self_attn.k_proj": {
464
+ "group_size": 64,
465
+ "bits": 2
466
+ },
467
+ "model.layers.14.self_attn.v_proj": {
468
+ "group_size": 64,
469
+ "bits": 4
470
+ },
471
+ "model.layers.14.self_attn.o_proj": {
472
+ "group_size": 64,
473
+ "bits": 2
474
+ },
475
+ "model.layers.14.self_attn.rope": false,
476
+ "model.layers.14.mlp.gate_proj": {
477
+ "group_size": 64,
478
+ "bits": 2
479
+ },
480
+ "model.layers.14.mlp.down_proj": {
481
+ "group_size": 64,
482
+ "bits": 4
483
+ },
484
+ "model.layers.14.mlp.up_proj": {
485
+ "group_size": 64,
486
+ "bits": 2
487
+ },
488
+ "model.layers.14.input_layernorm": false,
489
+ "model.layers.14.post_attention_layernorm": false,
490
+ "model.layers.15.self_attn.q_proj": {
491
+ "group_size": 64,
492
+ "bits": 2
493
+ },
494
+ "model.layers.15.self_attn.k_proj": {
495
+ "group_size": 64,
496
+ "bits": 2
497
+ },
498
+ "model.layers.15.self_attn.v_proj": {
499
+ "group_size": 64,
500
+ "bits": 2
501
+ },
502
+ "model.layers.15.self_attn.o_proj": {
503
+ "group_size": 64,
504
+ "bits": 2
505
+ },
506
+ "model.layers.15.self_attn.rope": false,
507
+ "model.layers.15.mlp.gate_proj": {
508
+ "group_size": 64,
509
+ "bits": 2
510
+ },
511
+ "model.layers.15.mlp.down_proj": {
512
+ "group_size": 64,
513
+ "bits": 2
514
+ },
515
+ "model.layers.15.mlp.up_proj": {
516
+ "group_size": 64,
517
+ "bits": 2
518
+ },
519
+ "model.layers.15.input_layernorm": false,
520
+ "model.layers.15.post_attention_layernorm": false,
521
+ "model.layers.16.self_attn.q_proj": {
522
+ "group_size": 64,
523
+ "bits": 2
524
+ },
525
+ "model.layers.16.self_attn.k_proj": {
526
+ "group_size": 64,
527
+ "bits": 2
528
+ },
529
+ "model.layers.16.self_attn.v_proj": {
530
+ "group_size": 64,
531
+ "bits": 2
532
+ },
533
+ "model.layers.16.self_attn.o_proj": {
534
+ "group_size": 64,
535
+ "bits": 2
536
+ },
537
+ "model.layers.16.self_attn.rope": false,
538
+ "model.layers.16.mlp.gate_proj": {
539
+ "group_size": 64,
540
+ "bits": 2
541
+ },
542
+ "model.layers.16.mlp.down_proj": {
543
+ "group_size": 64,
544
+ "bits": 2
545
+ },
546
+ "model.layers.16.mlp.up_proj": {
547
+ "group_size": 64,
548
+ "bits": 2
549
+ },
550
+ "model.layers.16.input_layernorm": false,
551
+ "model.layers.16.post_attention_layernorm": false,
552
+ "model.layers.17.self_attn.q_proj": {
553
+ "group_size": 64,
554
+ "bits": 2
555
+ },
556
+ "model.layers.17.self_attn.k_proj": {
557
+ "group_size": 64,
558
+ "bits": 2
559
+ },
560
+ "model.layers.17.self_attn.v_proj": {
561
+ "group_size": 64,
562
+ "bits": 4
563
+ },
564
+ "model.layers.17.self_attn.o_proj": {
565
+ "group_size": 64,
566
+ "bits": 2
567
+ },
568
+ "model.layers.17.self_attn.rope": false,
569
+ "model.layers.17.mlp.gate_proj": {
570
+ "group_size": 64,
571
+ "bits": 2
572
+ },
573
+ "model.layers.17.mlp.down_proj": {
574
+ "group_size": 64,
575
+ "bits": 4
576
+ },
577
+ "model.layers.17.mlp.up_proj": {
578
+ "group_size": 64,
579
+ "bits": 2
580
+ },
581
+ "model.layers.17.input_layernorm": false,
582
+ "model.layers.17.post_attention_layernorm": false,
583
+ "model.layers.18.self_attn.q_proj": {
584
+ "group_size": 64,
585
+ "bits": 2
586
+ },
587
+ "model.layers.18.self_attn.k_proj": {
588
+ "group_size": 64,
589
+ "bits": 2
590
+ },
591
+ "model.layers.18.self_attn.v_proj": {
592
+ "group_size": 64,
593
+ "bits": 2
594
+ },
595
+ "model.layers.18.self_attn.o_proj": {
596
+ "group_size": 64,
597
+ "bits": 2
598
+ },
599
+ "model.layers.18.self_attn.rope": false,
600
+ "model.layers.18.mlp.gate_proj": {
601
+ "group_size": 64,
602
+ "bits": 2
603
+ },
604
+ "model.layers.18.mlp.down_proj": {
605
+ "group_size": 64,
606
+ "bits": 2
607
+ },
608
+ "model.layers.18.mlp.up_proj": {
609
+ "group_size": 64,
610
+ "bits": 2
611
+ },
612
+ "model.layers.18.input_layernorm": false,
613
+ "model.layers.18.post_attention_layernorm": false,
614
+ "model.layers.19.self_attn.q_proj": {
615
+ "group_size": 64,
616
+ "bits": 2
617
+ },
618
+ "model.layers.19.self_attn.k_proj": {
619
+ "group_size": 64,
620
+ "bits": 2
621
+ },
622
+ "model.layers.19.self_attn.v_proj": {
623
+ "group_size": 64,
624
+ "bits": 2
625
+ },
626
+ "model.layers.19.self_attn.o_proj": {
627
+ "group_size": 64,
628
+ "bits": 2
629
+ },
630
+ "model.layers.19.self_attn.rope": false,
631
+ "model.layers.19.mlp.gate_proj": {
632
+ "group_size": 64,
633
+ "bits": 2
634
+ },
635
+ "model.layers.19.mlp.down_proj": {
636
+ "group_size": 64,
637
+ "bits": 2
638
+ },
639
+ "model.layers.19.mlp.up_proj": {
640
+ "group_size": 64,
641
+ "bits": 2
642
+ },
643
+ "model.layers.19.input_layernorm": false,
644
+ "model.layers.19.post_attention_layernorm": false,
645
+ "model.layers.20.self_attn.q_proj": {
646
+ "group_size": 64,
647
+ "bits": 2
648
+ },
649
+ "model.layers.20.self_attn.k_proj": {
650
+ "group_size": 64,
651
+ "bits": 2
652
+ },
653
+ "model.layers.20.self_attn.v_proj": {
654
+ "group_size": 64,
655
+ "bits": 4
656
+ },
657
+ "model.layers.20.self_attn.o_proj": {
658
+ "group_size": 64,
659
+ "bits": 2
660
+ },
661
+ "model.layers.20.self_attn.rope": false,
662
+ "model.layers.20.mlp.gate_proj": {
663
+ "group_size": 64,
664
+ "bits": 2
665
+ },
666
+ "model.layers.20.mlp.down_proj": {
667
+ "group_size": 64,
668
+ "bits": 4
669
+ },
670
+ "model.layers.20.mlp.up_proj": {
671
+ "group_size": 64,
672
+ "bits": 2
673
+ },
674
+ "model.layers.20.input_layernorm": false,
675
+ "model.layers.20.post_attention_layernorm": false,
676
+ "model.layers.21.self_attn.q_proj": {
677
+ "group_size": 64,
678
+ "bits": 2
679
+ },
680
+ "model.layers.21.self_attn.k_proj": {
681
+ "group_size": 64,
682
+ "bits": 2
683
+ },
684
+ "model.layers.21.self_attn.v_proj": {
685
+ "group_size": 64,
686
+ "bits": 2
687
+ },
688
+ "model.layers.21.self_attn.o_proj": {
689
+ "group_size": 64,
690
+ "bits": 2
691
+ },
692
+ "model.layers.21.self_attn.rope": false,
693
+ "model.layers.21.mlp.gate_proj": {
694
+ "group_size": 64,
695
+ "bits": 2
696
+ },
697
+ "model.layers.21.mlp.down_proj": {
698
+ "group_size": 64,
699
+ "bits": 2
700
+ },
701
+ "model.layers.21.mlp.up_proj": {
702
+ "group_size": 64,
703
+ "bits": 2
704
+ },
705
+ "model.layers.21.input_layernorm": false,
706
+ "model.layers.21.post_attention_layernorm": false,
707
+ "model.layers.22.self_attn.q_proj": {
708
+ "group_size": 64,
709
+ "bits": 2
710
+ },
711
+ "model.layers.22.self_attn.k_proj": {
712
+ "group_size": 64,
713
+ "bits": 2
714
+ },
715
+ "model.layers.22.self_attn.v_proj": {
716
+ "group_size": 64,
717
+ "bits": 2
718
+ },
719
+ "model.layers.22.self_attn.o_proj": {
720
+ "group_size": 64,
721
+ "bits": 2
722
+ },
723
+ "model.layers.22.self_attn.rope": false,
724
+ "model.layers.22.mlp.gate_proj": {
725
+ "group_size": 64,
726
+ "bits": 2
727
+ },
728
+ "model.layers.22.mlp.down_proj": {
729
+ "group_size": 64,
730
+ "bits": 2
731
+ },
732
+ "model.layers.22.mlp.up_proj": {
733
+ "group_size": 64,
734
+ "bits": 2
735
+ },
736
+ "model.layers.22.input_layernorm": false,
737
+ "model.layers.22.post_attention_layernorm": false,
738
+ "model.layers.23.self_attn.q_proj": {
739
+ "group_size": 64,
740
+ "bits": 2
741
+ },
742
+ "model.layers.23.self_attn.k_proj": {
743
+ "group_size": 64,
744
+ "bits": 2
745
+ },
746
+ "model.layers.23.self_attn.v_proj": {
747
+ "group_size": 64,
748
+ "bits": 4
749
+ },
750
+ "model.layers.23.self_attn.o_proj": {
751
+ "group_size": 64,
752
+ "bits": 2
753
+ },
754
+ "model.layers.23.self_attn.rope": false,
755
+ "model.layers.23.mlp.gate_proj": {
756
+ "group_size": 64,
757
+ "bits": 2
758
+ },
759
+ "model.layers.23.mlp.down_proj": {
760
+ "group_size": 64,
761
+ "bits": 4
762
+ },
763
+ "model.layers.23.mlp.up_proj": {
764
+ "group_size": 64,
765
+ "bits": 2
766
+ },
767
+ "model.layers.23.input_layernorm": false,
768
+ "model.layers.23.post_attention_layernorm": false,
769
+ "model.layers.24.self_attn.q_proj": {
770
+ "group_size": 64,
771
+ "bits": 2
772
+ },
773
+ "model.layers.24.self_attn.k_proj": {
774
+ "group_size": 64,
775
+ "bits": 2
776
+ },
777
+ "model.layers.24.self_attn.v_proj": {
778
+ "group_size": 64,
779
+ "bits": 2
780
+ },
781
+ "model.layers.24.self_attn.o_proj": {
782
+ "group_size": 64,
783
+ "bits": 2
784
+ },
785
+ "model.layers.24.self_attn.rope": false,
786
+ "model.layers.24.mlp.gate_proj": {
787
+ "group_size": 64,
788
+ "bits": 2
789
+ },
790
+ "model.layers.24.mlp.down_proj": {
791
+ "group_size": 64,
792
+ "bits": 2
793
+ },
794
+ "model.layers.24.mlp.up_proj": {
795
+ "group_size": 64,
796
+ "bits": 2
797
+ },
798
+ "model.layers.24.input_layernorm": false,
799
+ "model.layers.24.post_attention_layernorm": false,
800
+ "model.layers.25.self_attn.q_proj": {
801
+ "group_size": 64,
802
+ "bits": 2
803
+ },
804
+ "model.layers.25.self_attn.k_proj": {
805
+ "group_size": 64,
806
+ "bits": 2
807
+ },
808
+ "model.layers.25.self_attn.v_proj": {
809
+ "group_size": 64,
810
+ "bits": 2
811
+ },
812
+ "model.layers.25.self_attn.o_proj": {
813
+ "group_size": 64,
814
+ "bits": 2
815
+ },
816
+ "model.layers.25.self_attn.rope": false,
817
+ "model.layers.25.mlp.gate_proj": {
818
+ "group_size": 64,
819
+ "bits": 2
820
+ },
821
+ "model.layers.25.mlp.down_proj": {
822
+ "group_size": 64,
823
+ "bits": 2
824
+ },
825
+ "model.layers.25.mlp.up_proj": {
826
+ "group_size": 64,
827
+ "bits": 2
828
+ },
829
+ "model.layers.25.input_layernorm": false,
830
+ "model.layers.25.post_attention_layernorm": false,
831
+ "model.layers.26.self_attn.q_proj": {
832
+ "group_size": 64,
833
+ "bits": 2
834
+ },
835
+ "model.layers.26.self_attn.k_proj": {
836
+ "group_size": 64,
837
+ "bits": 2
838
+ },
839
+ "model.layers.26.self_attn.v_proj": {
840
+ "group_size": 64,
841
+ "bits": 4
842
+ },
843
+ "model.layers.26.self_attn.o_proj": {
844
+ "group_size": 64,
845
+ "bits": 2
846
+ },
847
+ "model.layers.26.self_attn.rope": false,
848
+ "model.layers.26.mlp.gate_proj": {
849
+ "group_size": 64,
850
+ "bits": 2
851
+ },
852
+ "model.layers.26.mlp.down_proj": {
853
+ "group_size": 64,
854
+ "bits": 4
855
+ },
856
+ "model.layers.26.mlp.up_proj": {
857
+ "group_size": 64,
858
+ "bits": 2
859
+ },
860
+ "model.layers.26.input_layernorm": false,
861
+ "model.layers.26.post_attention_layernorm": false,
862
+ "model.layers.27.self_attn.q_proj": {
863
+ "group_size": 64,
864
+ "bits": 2
865
+ },
866
+ "model.layers.27.self_attn.k_proj": {
867
+ "group_size": 64,
868
+ "bits": 2
869
+ },
870
+ "model.layers.27.self_attn.v_proj": {
871
+ "group_size": 64,
872
+ "bits": 2
873
+ },
874
+ "model.layers.27.self_attn.o_proj": {
875
+ "group_size": 64,
876
+ "bits": 2
877
+ },
878
+ "model.layers.27.self_attn.rope": false,
879
+ "model.layers.27.mlp.gate_proj": {
880
+ "group_size": 64,
881
+ "bits": 2
882
+ },
883
+ "model.layers.27.mlp.down_proj": {
884
+ "group_size": 64,
885
+ "bits": 2
886
+ },
887
+ "model.layers.27.mlp.up_proj": {
888
+ "group_size": 64,
889
+ "bits": 2
890
+ },
891
+ "model.layers.27.input_layernorm": false,
892
+ "model.layers.27.post_attention_layernorm": false,
893
+ "model.layers.28.self_attn.q_proj": {
894
+ "group_size": 64,
895
+ "bits": 2
896
+ },
897
+ "model.layers.28.self_attn.k_proj": {
898
+ "group_size": 64,
899
+ "bits": 2
900
+ },
901
+ "model.layers.28.self_attn.v_proj": {
902
+ "group_size": 64,
903
+ "bits": 2
904
+ },
905
+ "model.layers.28.self_attn.o_proj": {
906
+ "group_size": 64,
907
+ "bits": 2
908
+ },
909
+ "model.layers.28.self_attn.rope": false,
910
+ "model.layers.28.mlp.gate_proj": {
911
+ "group_size": 64,
912
+ "bits": 2
913
+ },
914
+ "model.layers.28.mlp.down_proj": {
915
+ "group_size": 64,
916
+ "bits": 2
917
+ },
918
+ "model.layers.28.mlp.up_proj": {
919
+ "group_size": 64,
920
+ "bits": 2
921
+ },
922
+ "model.layers.28.input_layernorm": false,
923
+ "model.layers.28.post_attention_layernorm": false,
924
+ "model.layers.29.self_attn.q_proj": {
925
+ "group_size": 64,
926
+ "bits": 2
927
+ },
928
+ "model.layers.29.self_attn.k_proj": {
929
+ "group_size": 64,
930
+ "bits": 2
931
+ },
932
+ "model.layers.29.self_attn.v_proj": {
933
+ "group_size": 64,
934
+ "bits": 4
935
+ },
936
+ "model.layers.29.self_attn.o_proj": {
937
+ "group_size": 64,
938
+ "bits": 2
939
+ },
940
+ "model.layers.29.self_attn.rope": false,
941
+ "model.layers.29.mlp.gate_proj": {
942
+ "group_size": 64,
943
+ "bits": 2
944
+ },
945
+ "model.layers.29.mlp.down_proj": {
946
+ "group_size": 64,
947
+ "bits": 4
948
+ },
949
+ "model.layers.29.mlp.up_proj": {
950
+ "group_size": 64,
951
+ "bits": 2
952
+ },
953
+ "model.layers.29.input_layernorm": false,
954
+ "model.layers.29.post_attention_layernorm": false,
955
+ "model.layers.30.self_attn.q_proj": {
956
+ "group_size": 64,
957
+ "bits": 2
958
+ },
959
+ "model.layers.30.self_attn.k_proj": {
960
+ "group_size": 64,
961
+ "bits": 2
962
+ },
963
+ "model.layers.30.self_attn.v_proj": {
964
+ "group_size": 64,
965
+ "bits": 2
966
+ },
967
+ "model.layers.30.self_attn.o_proj": {
968
+ "group_size": 64,
969
+ "bits": 2
970
+ },
971
+ "model.layers.30.self_attn.rope": false,
972
+ "model.layers.30.mlp.gate_proj": {
973
+ "group_size": 64,
974
+ "bits": 2
975
+ },
976
+ "model.layers.30.mlp.down_proj": {
977
+ "group_size": 64,
978
+ "bits": 2
979
+ },
980
+ "model.layers.30.mlp.up_proj": {
981
+ "group_size": 64,
982
+ "bits": 2
983
+ },
984
+ "model.layers.30.input_layernorm": false,
985
+ "model.layers.30.post_attention_layernorm": false,
986
+ "model.layers.31.self_attn.q_proj": {
987
+ "group_size": 64,
988
+ "bits": 2
989
+ },
990
+ "model.layers.31.self_attn.k_proj": {
991
+ "group_size": 64,
992
+ "bits": 2
993
+ },
994
+ "model.layers.31.self_attn.v_proj": {
995
+ "group_size": 64,
996
+ "bits": 2
997
+ },
998
+ "model.layers.31.self_attn.o_proj": {
999
+ "group_size": 64,
1000
+ "bits": 2
1001
+ },
1002
+ "model.layers.31.self_attn.rope": false,
1003
+ "model.layers.31.mlp.gate_proj": {
1004
+ "group_size": 64,
1005
+ "bits": 2
1006
+ },
1007
+ "model.layers.31.mlp.down_proj": {
1008
+ "group_size": 64,
1009
+ "bits": 2
1010
+ },
1011
+ "model.layers.31.mlp.up_proj": {
1012
+ "group_size": 64,
1013
+ "bits": 2
1014
+ },
1015
+ "model.layers.31.input_layernorm": false,
1016
+ "model.layers.31.post_attention_layernorm": false,
1017
+ "model.layers.32.self_attn.q_proj": {
1018
+ "group_size": 64,
1019
+ "bits": 2
1020
+ },
1021
+ "model.layers.32.self_attn.k_proj": {
1022
+ "group_size": 64,
1023
+ "bits": 2
1024
+ },
1025
+ "model.layers.32.self_attn.v_proj": {
1026
+ "group_size": 64,
1027
+ "bits": 4
1028
+ },
1029
+ "model.layers.32.self_attn.o_proj": {
1030
+ "group_size": 64,
1031
+ "bits": 2
1032
+ },
1033
+ "model.layers.32.self_attn.rope": false,
1034
+ "model.layers.32.mlp.gate_proj": {
1035
+ "group_size": 64,
1036
+ "bits": 2
1037
+ },
1038
+ "model.layers.32.mlp.down_proj": {
1039
+ "group_size": 64,
1040
+ "bits": 4
1041
+ },
1042
+ "model.layers.32.mlp.up_proj": {
1043
+ "group_size": 64,
1044
+ "bits": 2
1045
+ },
1046
+ "model.layers.32.input_layernorm": false,
1047
+ "model.layers.32.post_attention_layernorm": false,
1048
+ "model.layers.33.self_attn.q_proj": {
1049
+ "group_size": 64,
1050
+ "bits": 2
1051
+ },
1052
+ "model.layers.33.self_attn.k_proj": {
1053
+ "group_size": 64,
1054
+ "bits": 2
1055
+ },
1056
+ "model.layers.33.self_attn.v_proj": {
1057
+ "group_size": 64,
1058
+ "bits": 2
1059
+ },
1060
+ "model.layers.33.self_attn.o_proj": {
1061
+ "group_size": 64,
1062
+ "bits": 2
1063
+ },
1064
+ "model.layers.33.self_attn.rope": false,
1065
+ "model.layers.33.mlp.gate_proj": {
1066
+ "group_size": 64,
1067
+ "bits": 2
1068
+ },
1069
+ "model.layers.33.mlp.down_proj": {
1070
+ "group_size": 64,
1071
+ "bits": 2
1072
+ },
1073
+ "model.layers.33.mlp.up_proj": {
1074
+ "group_size": 64,
1075
+ "bits": 2
1076
+ },
1077
+ "model.layers.33.input_layernorm": false,
1078
+ "model.layers.33.post_attention_layernorm": false,
1079
+ "model.layers.34.self_attn.q_proj": {
1080
+ "group_size": 64,
1081
+ "bits": 2
1082
+ },
1083
+ "model.layers.34.self_attn.k_proj": {
1084
+ "group_size": 64,
1085
+ "bits": 2
1086
+ },
1087
+ "model.layers.34.self_attn.v_proj": {
1088
+ "group_size": 64,
1089
+ "bits": 2
1090
+ },
1091
+ "model.layers.34.self_attn.o_proj": {
1092
+ "group_size": 64,
1093
+ "bits": 2
1094
+ },
1095
+ "model.layers.34.self_attn.rope": false,
1096
+ "model.layers.34.mlp.gate_proj": {
1097
+ "group_size": 64,
1098
+ "bits": 2
1099
+ },
1100
+ "model.layers.34.mlp.down_proj": {
1101
+ "group_size": 64,
1102
+ "bits": 2
1103
+ },
1104
+ "model.layers.34.mlp.up_proj": {
1105
+ "group_size": 64,
1106
+ "bits": 2
1107
+ },
1108
+ "model.layers.34.input_layernorm": false,
1109
+ "model.layers.34.post_attention_layernorm": false,
1110
+ "model.layers.35.self_attn.q_proj": {
1111
+ "group_size": 64,
1112
+ "bits": 2
1113
+ },
1114
+ "model.layers.35.self_attn.k_proj": {
1115
+ "group_size": 64,
1116
+ "bits": 2
1117
+ },
1118
+ "model.layers.35.self_attn.v_proj": {
1119
+ "group_size": 64,
1120
+ "bits": 4
1121
+ },
1122
+ "model.layers.35.self_attn.o_proj": {
1123
+ "group_size": 64,
1124
+ "bits": 2
1125
+ },
1126
+ "model.layers.35.self_attn.rope": false,
1127
+ "model.layers.35.mlp.gate_proj": {
1128
+ "group_size": 64,
1129
+ "bits": 2
1130
+ },
1131
+ "model.layers.35.mlp.down_proj": {
1132
+ "group_size": 64,
1133
+ "bits": 4
1134
+ },
1135
+ "model.layers.35.mlp.up_proj": {
1136
+ "group_size": 64,
1137
+ "bits": 2
1138
+ },
1139
+ "model.layers.35.input_layernorm": false,
1140
+ "model.layers.35.post_attention_layernorm": false,
1141
+ "model.layers.36.self_attn.q_proj": {
1142
+ "group_size": 64,
1143
+ "bits": 2
1144
+ },
1145
+ "model.layers.36.self_attn.k_proj": {
1146
+ "group_size": 64,
1147
+ "bits": 2
1148
+ },
1149
+ "model.layers.36.self_attn.v_proj": {
1150
+ "group_size": 64,
1151
+ "bits": 2
1152
+ },
1153
+ "model.layers.36.self_attn.o_proj": {
1154
+ "group_size": 64,
1155
+ "bits": 2
1156
+ },
1157
+ "model.layers.36.self_attn.rope": false,
1158
+ "model.layers.36.mlp.gate_proj": {
1159
+ "group_size": 64,
1160
+ "bits": 2
1161
+ },
1162
+ "model.layers.36.mlp.down_proj": {
1163
+ "group_size": 64,
1164
+ "bits": 2
1165
+ },
1166
+ "model.layers.36.mlp.up_proj": {
1167
+ "group_size": 64,
1168
+ "bits": 2
1169
+ },
1170
+ "model.layers.36.input_layernorm": false,
1171
+ "model.layers.36.post_attention_layernorm": false,
1172
+ "model.layers.37.self_attn.q_proj": {
1173
+ "group_size": 64,
1174
+ "bits": 2
1175
+ },
1176
+ "model.layers.37.self_attn.k_proj": {
1177
+ "group_size": 64,
1178
+ "bits": 2
1179
+ },
1180
+ "model.layers.37.self_attn.v_proj": {
1181
+ "group_size": 64,
1182
+ "bits": 2
1183
+ },
1184
+ "model.layers.37.self_attn.o_proj": {
1185
+ "group_size": 64,
1186
+ "bits": 2
1187
+ },
1188
+ "model.layers.37.self_attn.rope": false,
1189
+ "model.layers.37.mlp.gate_proj": {
1190
+ "group_size": 64,
1191
+ "bits": 2
1192
+ },
1193
+ "model.layers.37.mlp.down_proj": {
1194
+ "group_size": 64,
1195
+ "bits": 2
1196
+ },
1197
+ "model.layers.37.mlp.up_proj": {
1198
+ "group_size": 64,
1199
+ "bits": 2
1200
+ },
1201
+ "model.layers.37.input_layernorm": false,
1202
+ "model.layers.37.post_attention_layernorm": false,
1203
+ "model.layers.38.self_attn.q_proj": {
1204
+ "group_size": 64,
1205
+ "bits": 2
1206
+ },
1207
+ "model.layers.38.self_attn.k_proj": {
1208
+ "group_size": 64,
1209
+ "bits": 2
1210
+ },
1211
+ "model.layers.38.self_attn.v_proj": {
1212
+ "group_size": 64,
1213
+ "bits": 4
1214
+ },
1215
+ "model.layers.38.self_attn.o_proj": {
1216
+ "group_size": 64,
1217
+ "bits": 2
1218
+ },
1219
+ "model.layers.38.self_attn.rope": false,
1220
+ "model.layers.38.mlp.gate_proj": {
1221
+ "group_size": 64,
1222
+ "bits": 2
1223
+ },
1224
+ "model.layers.38.mlp.down_proj": {
1225
+ "group_size": 64,
1226
+ "bits": 4
1227
+ },
1228
+ "model.layers.38.mlp.up_proj": {
1229
+ "group_size": 64,
1230
+ "bits": 2
1231
+ },
1232
+ "model.layers.38.input_layernorm": false,
1233
+ "model.layers.38.post_attention_layernorm": false,
1234
+ "model.layers.39.self_attn.q_proj": {
1235
+ "group_size": 64,
1236
+ "bits": 2
1237
+ },
1238
+ "model.layers.39.self_attn.k_proj": {
1239
+ "group_size": 64,
1240
+ "bits": 2
1241
+ },
1242
+ "model.layers.39.self_attn.v_proj": {
1243
+ "group_size": 64,
1244
+ "bits": 2
1245
+ },
1246
+ "model.layers.39.self_attn.o_proj": {
1247
+ "group_size": 64,
1248
+ "bits": 2
1249
+ },
1250
+ "model.layers.39.self_attn.rope": false,
1251
+ "model.layers.39.mlp.gate_proj": {
1252
+ "group_size": 64,
1253
+ "bits": 2
1254
+ },
1255
+ "model.layers.39.mlp.down_proj": {
1256
+ "group_size": 64,
1257
+ "bits": 2
1258
+ },
1259
+ "model.layers.39.mlp.up_proj": {
1260
+ "group_size": 64,
1261
+ "bits": 2
1262
+ },
1263
+ "model.layers.39.input_layernorm": false,
1264
+ "model.layers.39.post_attention_layernorm": false,
1265
+ "model.layers.40.self_attn.q_proj": {
1266
+ "group_size": 64,
1267
+ "bits": 2
1268
+ },
1269
+ "model.layers.40.self_attn.k_proj": {
1270
+ "group_size": 64,
1271
+ "bits": 2
1272
+ },
1273
+ "model.layers.40.self_attn.v_proj": {
1274
+ "group_size": 64,
1275
+ "bits": 2
1276
+ },
1277
+ "model.layers.40.self_attn.o_proj": {
1278
+ "group_size": 64,
1279
+ "bits": 2
1280
+ },
1281
+ "model.layers.40.self_attn.rope": false,
1282
+ "model.layers.40.mlp.gate_proj": {
1283
+ "group_size": 64,
1284
+ "bits": 2
1285
+ },
1286
+ "model.layers.40.mlp.down_proj": {
1287
+ "group_size": 64,
1288
+ "bits": 2
1289
+ },
1290
+ "model.layers.40.mlp.up_proj": {
1291
+ "group_size": 64,
1292
+ "bits": 2
1293
+ },
1294
+ "model.layers.40.input_layernorm": false,
1295
+ "model.layers.40.post_attention_layernorm": false,
1296
+ "model.layers.41.self_attn.q_proj": {
1297
+ "group_size": 64,
1298
+ "bits": 2
1299
+ },
1300
+ "model.layers.41.self_attn.k_proj": {
1301
+ "group_size": 64,
1302
+ "bits": 2
1303
+ },
1304
+ "model.layers.41.self_attn.v_proj": {
1305
+ "group_size": 64,
1306
+ "bits": 4
1307
+ },
1308
+ "model.layers.41.self_attn.o_proj": {
1309
+ "group_size": 64,
1310
+ "bits": 2
1311
+ },
1312
+ "model.layers.41.self_attn.rope": false,
1313
+ "model.layers.41.mlp.gate_proj": {
1314
+ "group_size": 64,
1315
+ "bits": 2
1316
+ },
1317
+ "model.layers.41.mlp.down_proj": {
1318
+ "group_size": 64,
1319
+ "bits": 4
1320
+ },
1321
+ "model.layers.41.mlp.up_proj": {
1322
+ "group_size": 64,
1323
+ "bits": 2
1324
+ },
1325
+ "model.layers.41.input_layernorm": false,
1326
+ "model.layers.41.post_attention_layernorm": false,
1327
+ "model.layers.42.self_attn.q_proj": {
1328
+ "group_size": 64,
1329
+ "bits": 2
1330
+ },
1331
+ "model.layers.42.self_attn.k_proj": {
1332
+ "group_size": 64,
1333
+ "bits": 2
1334
+ },
1335
+ "model.layers.42.self_attn.v_proj": {
1336
+ "group_size": 64,
1337
+ "bits": 4
1338
+ },
1339
+ "model.layers.42.self_attn.o_proj": {
1340
+ "group_size": 64,
1341
+ "bits": 2
1342
+ },
1343
+ "model.layers.42.self_attn.rope": false,
1344
+ "model.layers.42.mlp.gate_proj": {
1345
+ "group_size": 64,
1346
+ "bits": 2
1347
+ },
1348
+ "model.layers.42.mlp.down_proj": {
1349
+ "group_size": 64,
1350
+ "bits": 4
1351
+ },
1352
+ "model.layers.42.mlp.up_proj": {
1353
+ "group_size": 64,
1354
+ "bits": 2
1355
+ },
1356
+ "model.layers.42.input_layernorm": false,
1357
+ "model.layers.42.post_attention_layernorm": false,
1358
+ "model.layers.43.self_attn.q_proj": {
1359
+ "group_size": 64,
1360
+ "bits": 2
1361
+ },
1362
+ "model.layers.43.self_attn.k_proj": {
1363
+ "group_size": 64,
1364
+ "bits": 2
1365
+ },
1366
+ "model.layers.43.self_attn.v_proj": {
1367
+ "group_size": 64,
1368
+ "bits": 4
1369
+ },
1370
+ "model.layers.43.self_attn.o_proj": {
1371
+ "group_size": 64,
1372
+ "bits": 2
1373
+ },
1374
+ "model.layers.43.self_attn.rope": false,
1375
+ "model.layers.43.mlp.gate_proj": {
1376
+ "group_size": 64,
1377
+ "bits": 2
1378
+ },
1379
+ "model.layers.43.mlp.down_proj": {
1380
+ "group_size": 64,
1381
+ "bits": 4
1382
+ },
1383
+ "model.layers.43.mlp.up_proj": {
1384
+ "group_size": 64,
1385
+ "bits": 2
1386
+ },
1387
+ "model.layers.43.input_layernorm": false,
1388
+ "model.layers.43.post_attention_layernorm": false,
1389
+ "model.layers.44.self_attn.q_proj": {
1390
+ "group_size": 64,
1391
+ "bits": 2
1392
+ },
1393
+ "model.layers.44.self_attn.k_proj": {
1394
+ "group_size": 64,
1395
+ "bits": 2
1396
+ },
1397
+ "model.layers.44.self_attn.v_proj": {
1398
+ "group_size": 64,
1399
+ "bits": 4
1400
+ },
1401
+ "model.layers.44.self_attn.o_proj": {
1402
+ "group_size": 64,
1403
+ "bits": 2
1404
+ },
1405
+ "model.layers.44.self_attn.rope": false,
1406
+ "model.layers.44.mlp.gate_proj": {
1407
+ "group_size": 64,
1408
+ "bits": 2
1409
+ },
1410
+ "model.layers.44.mlp.down_proj": {
1411
+ "group_size": 64,
1412
+ "bits": 4
1413
+ },
1414
+ "model.layers.44.mlp.up_proj": {
1415
+ "group_size": 64,
1416
+ "bits": 2
1417
+ },
1418
+ "model.layers.44.input_layernorm": false,
1419
+ "model.layers.44.post_attention_layernorm": false,
1420
+ "model.layers.45.self_attn.q_proj": {
1421
+ "group_size": 64,
1422
+ "bits": 2
1423
+ },
1424
+ "model.layers.45.self_attn.k_proj": {
1425
+ "group_size": 64,
1426
+ "bits": 2
1427
+ },
1428
+ "model.layers.45.self_attn.v_proj": {
1429
+ "group_size": 64,
1430
+ "bits": 4
1431
+ },
1432
+ "model.layers.45.self_attn.o_proj": {
1433
+ "group_size": 64,
1434
+ "bits": 2
1435
+ },
1436
+ "model.layers.45.self_attn.rope": false,
1437
+ "model.layers.45.mlp.gate_proj": {
1438
+ "group_size": 64,
1439
+ "bits": 2
1440
+ },
1441
+ "model.layers.45.mlp.down_proj": {
1442
+ "group_size": 64,
1443
+ "bits": 4
1444
+ },
1445
+ "model.layers.45.mlp.up_proj": {
1446
+ "group_size": 64,
1447
+ "bits": 2
1448
+ },
1449
+ "model.layers.45.input_layernorm": false,
1450
+ "model.layers.45.post_attention_layernorm": false,
1451
+ "model.layers.46.self_attn.q_proj": {
1452
+ "group_size": 64,
1453
+ "bits": 2
1454
+ },
1455
+ "model.layers.46.self_attn.k_proj": {
1456
+ "group_size": 64,
1457
+ "bits": 2
1458
+ },
1459
+ "model.layers.46.self_attn.v_proj": {
1460
+ "group_size": 64,
1461
+ "bits": 4
1462
+ },
1463
+ "model.layers.46.self_attn.o_proj": {
1464
+ "group_size": 64,
1465
+ "bits": 2
1466
+ },
1467
+ "model.layers.46.self_attn.rope": false,
1468
+ "model.layers.46.mlp.gate_proj": {
1469
+ "group_size": 64,
1470
+ "bits": 2
1471
+ },
1472
+ "model.layers.46.mlp.down_proj": {
1473
+ "group_size": 64,
1474
+ "bits": 4
1475
+ },
1476
+ "model.layers.46.mlp.up_proj": {
1477
+ "group_size": 64,
1478
+ "bits": 2
1479
+ },
1480
+ "model.layers.46.input_layernorm": false,
1481
+ "model.layers.46.post_attention_layernorm": false,
1482
+ "model.layers.47.self_attn.q_proj": {
1483
+ "group_size": 64,
1484
+ "bits": 2
1485
+ },
1486
+ "model.layers.47.self_attn.k_proj": {
1487
+ "group_size": 64,
1488
+ "bits": 2
1489
+ },
1490
+ "model.layers.47.self_attn.v_proj": {
1491
+ "group_size": 64,
1492
+ "bits": 4
1493
+ },
1494
+ "model.layers.47.self_attn.o_proj": {
1495
+ "group_size": 64,
1496
+ "bits": 2
1497
+ },
1498
+ "model.layers.47.self_attn.rope": false,
1499
+ "model.layers.47.mlp.gate_proj": {
1500
+ "group_size": 64,
1501
+ "bits": 2
1502
+ },
1503
+ "model.layers.47.mlp.down_proj": {
1504
+ "group_size": 64,
1505
+ "bits": 4
1506
+ },
1507
+ "model.layers.47.mlp.up_proj": {
1508
+ "group_size": 64,
1509
+ "bits": 2
1510
+ },
1511
+ "model.layers.47.input_layernorm": false,
1512
+ "model.layers.47.post_attention_layernorm": false,
1513
+ "model.norm": false,
1514
+ "lm_head": {
1515
+ "group_size": 64,
1516
+ "bits": 4
1517
+ }
1518
+ },
1519
+ "quantization_config": {
1520
+ "group_size": 64,
1521
+ "bits": null,
1522
+ "model.embed_tokens": {
1523
+ "group_size": 64,
1524
+ "bits": 2
1525
+ },
1526
+ "model.layers.0.self_attn.q_proj": {
1527
+ "group_size": 64,
1528
+ "bits": 2
1529
+ },
1530
+ "model.layers.0.self_attn.k_proj": {
1531
+ "group_size": 64,
1532
+ "bits": 2
1533
+ },
1534
+ "model.layers.0.self_attn.v_proj": {
1535
+ "group_size": 64,
1536
+ "bits": 4
1537
+ },
1538
+ "model.layers.0.self_attn.o_proj": {
1539
+ "group_size": 64,
1540
+ "bits": 2
1541
+ },
1542
+ "model.layers.0.self_attn.rope": false,
1543
+ "model.layers.0.mlp.gate_proj": {
1544
+ "group_size": 64,
1545
+ "bits": 2
1546
+ },
1547
+ "model.layers.0.mlp.down_proj": {
1548
+ "group_size": 64,
1549
+ "bits": 4
1550
+ },
1551
+ "model.layers.0.mlp.up_proj": {
1552
+ "group_size": 64,
1553
+ "bits": 2
1554
+ },
1555
+ "model.layers.0.input_layernorm": false,
1556
+ "model.layers.0.post_attention_layernorm": false,
1557
+ "model.layers.1.self_attn.q_proj": {
1558
+ "group_size": 64,
1559
+ "bits": 2
1560
+ },
1561
+ "model.layers.1.self_attn.k_proj": {
1562
+ "group_size": 64,
1563
+ "bits": 2
1564
+ },
1565
+ "model.layers.1.self_attn.v_proj": {
1566
+ "group_size": 64,
1567
+ "bits": 4
1568
+ },
1569
+ "model.layers.1.self_attn.o_proj": {
1570
+ "group_size": 64,
1571
+ "bits": 2
1572
+ },
1573
+ "model.layers.1.self_attn.rope": false,
1574
+ "model.layers.1.mlp.gate_proj": {
1575
+ "group_size": 64,
1576
+ "bits": 2
1577
+ },
1578
+ "model.layers.1.mlp.down_proj": {
1579
+ "group_size": 64,
1580
+ "bits": 4
1581
+ },
1582
+ "model.layers.1.mlp.up_proj": {
1583
+ "group_size": 64,
1584
+ "bits": 2
1585
+ },
1586
+ "model.layers.1.input_layernorm": false,
1587
+ "model.layers.1.post_attention_layernorm": false,
1588
+ "model.layers.2.self_attn.q_proj": {
1589
+ "group_size": 64,
1590
+ "bits": 2
1591
+ },
1592
+ "model.layers.2.self_attn.k_proj": {
1593
+ "group_size": 64,
1594
+ "bits": 2
1595
+ },
1596
+ "model.layers.2.self_attn.v_proj": {
1597
+ "group_size": 64,
1598
+ "bits": 4
1599
+ },
1600
+ "model.layers.2.self_attn.o_proj": {
1601
+ "group_size": 64,
1602
+ "bits": 2
1603
+ },
1604
+ "model.layers.2.self_attn.rope": false,
1605
+ "model.layers.2.mlp.gate_proj": {
1606
+ "group_size": 64,
1607
+ "bits": 2
1608
+ },
1609
+ "model.layers.2.mlp.down_proj": {
1610
+ "group_size": 64,
1611
+ "bits": 4
1612
+ },
1613
+ "model.layers.2.mlp.up_proj": {
1614
+ "group_size": 64,
1615
+ "bits": 2
1616
+ },
1617
+ "model.layers.2.input_layernorm": false,
1618
+ "model.layers.2.post_attention_layernorm": false,
1619
+ "model.layers.3.self_attn.q_proj": {
1620
+ "group_size": 64,
1621
+ "bits": 2
1622
+ },
1623
+ "model.layers.3.self_attn.k_proj": {
1624
+ "group_size": 64,
1625
+ "bits": 2
1626
+ },
1627
+ "model.layers.3.self_attn.v_proj": {
1628
+ "group_size": 64,
1629
+ "bits": 4
1630
+ },
1631
+ "model.layers.3.self_attn.o_proj": {
1632
+ "group_size": 64,
1633
+ "bits": 2
1634
+ },
1635
+ "model.layers.3.self_attn.rope": false,
1636
+ "model.layers.3.mlp.gate_proj": {
1637
+ "group_size": 64,
1638
+ "bits": 2
1639
+ },
1640
+ "model.layers.3.mlp.down_proj": {
1641
+ "group_size": 64,
1642
+ "bits": 4
1643
+ },
1644
+ "model.layers.3.mlp.up_proj": {
1645
+ "group_size": 64,
1646
+ "bits": 2
1647
+ },
1648
+ "model.layers.3.input_layernorm": false,
1649
+ "model.layers.3.post_attention_layernorm": false,
1650
+ "model.layers.4.self_attn.q_proj": {
1651
+ "group_size": 64,
1652
+ "bits": 2
1653
+ },
1654
+ "model.layers.4.self_attn.k_proj": {
1655
+ "group_size": 64,
1656
+ "bits": 2
1657
+ },
1658
+ "model.layers.4.self_attn.v_proj": {
1659
+ "group_size": 64,
1660
+ "bits": 4
1661
+ },
1662
+ "model.layers.4.self_attn.o_proj": {
1663
+ "group_size": 64,
1664
+ "bits": 2
1665
+ },
1666
+ "model.layers.4.self_attn.rope": false,
1667
+ "model.layers.4.mlp.gate_proj": {
1668
+ "group_size": 64,
1669
+ "bits": 2
1670
+ },
1671
+ "model.layers.4.mlp.down_proj": {
1672
+ "group_size": 64,
1673
+ "bits": 4
1674
+ },
1675
+ "model.layers.4.mlp.up_proj": {
1676
+ "group_size": 64,
1677
+ "bits": 2
1678
+ },
1679
+ "model.layers.4.input_layernorm": false,
1680
+ "model.layers.4.post_attention_layernorm": false,
1681
+ "model.layers.5.self_attn.q_proj": {
1682
+ "group_size": 64,
1683
+ "bits": 2
1684
+ },
1685
+ "model.layers.5.self_attn.k_proj": {
1686
+ "group_size": 64,
1687
+ "bits": 2
1688
+ },
1689
+ "model.layers.5.self_attn.v_proj": {
1690
+ "group_size": 64,
1691
+ "bits": 4
1692
+ },
1693
+ "model.layers.5.self_attn.o_proj": {
1694
+ "group_size": 64,
1695
+ "bits": 2
1696
+ },
1697
+ "model.layers.5.self_attn.rope": false,
1698
+ "model.layers.5.mlp.gate_proj": {
1699
+ "group_size": 64,
1700
+ "bits": 2
1701
+ },
1702
+ "model.layers.5.mlp.down_proj": {
1703
+ "group_size": 64,
1704
+ "bits": 4
1705
+ },
1706
+ "model.layers.5.mlp.up_proj": {
1707
+ "group_size": 64,
1708
+ "bits": 2
1709
+ },
1710
+ "model.layers.5.input_layernorm": false,
1711
+ "model.layers.5.post_attention_layernorm": false,
1712
+ "model.layers.6.self_attn.q_proj": {
1713
+ "group_size": 64,
1714
+ "bits": 2
1715
+ },
1716
+ "model.layers.6.self_attn.k_proj": {
1717
+ "group_size": 64,
1718
+ "bits": 2
1719
+ },
1720
+ "model.layers.6.self_attn.v_proj": {
1721
+ "group_size": 64,
1722
+ "bits": 2
1723
+ },
1724
+ "model.layers.6.self_attn.o_proj": {
1725
+ "group_size": 64,
1726
+ "bits": 2
1727
+ },
1728
+ "model.layers.6.self_attn.rope": false,
1729
+ "model.layers.6.mlp.gate_proj": {
1730
+ "group_size": 64,
1731
+ "bits": 2
1732
+ },
1733
+ "model.layers.6.mlp.down_proj": {
1734
+ "group_size": 64,
1735
+ "bits": 2
1736
+ },
1737
+ "model.layers.6.mlp.up_proj": {
1738
+ "group_size": 64,
1739
+ "bits": 2
1740
+ },
1741
+ "model.layers.6.input_layernorm": false,
1742
+ "model.layers.6.post_attention_layernorm": false,
1743
+ "model.layers.7.self_attn.q_proj": {
1744
+ "group_size": 64,
1745
+ "bits": 2
1746
+ },
1747
+ "model.layers.7.self_attn.k_proj": {
1748
+ "group_size": 64,
1749
+ "bits": 2
1750
+ },
1751
+ "model.layers.7.self_attn.v_proj": {
1752
+ "group_size": 64,
1753
+ "bits": 2
1754
+ },
1755
+ "model.layers.7.self_attn.o_proj": {
1756
+ "group_size": 64,
1757
+ "bits": 2
1758
+ },
1759
+ "model.layers.7.self_attn.rope": false,
1760
+ "model.layers.7.mlp.gate_proj": {
1761
+ "group_size": 64,
1762
+ "bits": 2
1763
+ },
1764
+ "model.layers.7.mlp.down_proj": {
1765
+ "group_size": 64,
1766
+ "bits": 2
1767
+ },
1768
+ "model.layers.7.mlp.up_proj": {
1769
+ "group_size": 64,
1770
+ "bits": 2
1771
+ },
1772
+ "model.layers.7.input_layernorm": false,
1773
+ "model.layers.7.post_attention_layernorm": false,
1774
+ "model.layers.8.self_attn.q_proj": {
1775
+ "group_size": 64,
1776
+ "bits": 2
1777
+ },
1778
+ "model.layers.8.self_attn.k_proj": {
1779
+ "group_size": 64,
1780
+ "bits": 2
1781
+ },
1782
+ "model.layers.8.self_attn.v_proj": {
1783
+ "group_size": 64,
1784
+ "bits": 4
1785
+ },
1786
+ "model.layers.8.self_attn.o_proj": {
1787
+ "group_size": 64,
1788
+ "bits": 2
1789
+ },
1790
+ "model.layers.8.self_attn.rope": false,
1791
+ "model.layers.8.mlp.gate_proj": {
1792
+ "group_size": 64,
1793
+ "bits": 2
1794
+ },
1795
+ "model.layers.8.mlp.down_proj": {
1796
+ "group_size": 64,
1797
+ "bits": 4
1798
+ },
1799
+ "model.layers.8.mlp.up_proj": {
1800
+ "group_size": 64,
1801
+ "bits": 2
1802
+ },
1803
+ "model.layers.8.input_layernorm": false,
1804
+ "model.layers.8.post_attention_layernorm": false,
1805
+ "model.layers.9.self_attn.q_proj": {
1806
+ "group_size": 64,
1807
+ "bits": 2
1808
+ },
1809
+ "model.layers.9.self_attn.k_proj": {
1810
+ "group_size": 64,
1811
+ "bits": 2
1812
+ },
1813
+ "model.layers.9.self_attn.v_proj": {
1814
+ "group_size": 64,
1815
+ "bits": 2
1816
+ },
1817
+ "model.layers.9.self_attn.o_proj": {
1818
+ "group_size": 64,
1819
+ "bits": 2
1820
+ },
1821
+ "model.layers.9.self_attn.rope": false,
1822
+ "model.layers.9.mlp.gate_proj": {
1823
+ "group_size": 64,
1824
+ "bits": 2
1825
+ },
1826
+ "model.layers.9.mlp.down_proj": {
1827
+ "group_size": 64,
1828
+ "bits": 2
1829
+ },
1830
+ "model.layers.9.mlp.up_proj": {
1831
+ "group_size": 64,
1832
+ "bits": 2
1833
+ },
1834
+ "model.layers.9.input_layernorm": false,
1835
+ "model.layers.9.post_attention_layernorm": false,
1836
+ "model.layers.10.self_attn.q_proj": {
1837
+ "group_size": 64,
1838
+ "bits": 2
1839
+ },
1840
+ "model.layers.10.self_attn.k_proj": {
1841
+ "group_size": 64,
1842
+ "bits": 2
1843
+ },
1844
+ "model.layers.10.self_attn.v_proj": {
1845
+ "group_size": 64,
1846
+ "bits": 2
1847
+ },
1848
+ "model.layers.10.self_attn.o_proj": {
1849
+ "group_size": 64,
1850
+ "bits": 2
1851
+ },
1852
+ "model.layers.10.self_attn.rope": false,
1853
+ "model.layers.10.mlp.gate_proj": {
1854
+ "group_size": 64,
1855
+ "bits": 2
1856
+ },
1857
+ "model.layers.10.mlp.down_proj": {
1858
+ "group_size": 64,
1859
+ "bits": 2
1860
+ },
1861
+ "model.layers.10.mlp.up_proj": {
1862
+ "group_size": 64,
1863
+ "bits": 2
1864
+ },
1865
+ "model.layers.10.input_layernorm": false,
1866
+ "model.layers.10.post_attention_layernorm": false,
1867
+ "model.layers.11.self_attn.q_proj": {
1868
+ "group_size": 64,
1869
+ "bits": 2
1870
+ },
1871
+ "model.layers.11.self_attn.k_proj": {
1872
+ "group_size": 64,
1873
+ "bits": 2
1874
+ },
1875
+ "model.layers.11.self_attn.v_proj": {
1876
+ "group_size": 64,
1877
+ "bits": 4
1878
+ },
1879
+ "model.layers.11.self_attn.o_proj": {
1880
+ "group_size": 64,
1881
+ "bits": 2
1882
+ },
1883
+ "model.layers.11.self_attn.rope": false,
1884
+ "model.layers.11.mlp.gate_proj": {
1885
+ "group_size": 64,
1886
+ "bits": 2
1887
+ },
1888
+ "model.layers.11.mlp.down_proj": {
1889
+ "group_size": 64,
1890
+ "bits": 4
1891
+ },
1892
+ "model.layers.11.mlp.up_proj": {
1893
+ "group_size": 64,
1894
+ "bits": 2
1895
+ },
1896
+ "model.layers.11.input_layernorm": false,
1897
+ "model.layers.11.post_attention_layernorm": false,
1898
+ "model.layers.12.self_attn.q_proj": {
1899
+ "group_size": 64,
1900
+ "bits": 2
1901
+ },
1902
+ "model.layers.12.self_attn.k_proj": {
1903
+ "group_size": 64,
1904
+ "bits": 2
1905
+ },
1906
+ "model.layers.12.self_attn.v_proj": {
1907
+ "group_size": 64,
1908
+ "bits": 2
1909
+ },
1910
+ "model.layers.12.self_attn.o_proj": {
1911
+ "group_size": 64,
1912
+ "bits": 2
1913
+ },
1914
+ "model.layers.12.self_attn.rope": false,
1915
+ "model.layers.12.mlp.gate_proj": {
1916
+ "group_size": 64,
1917
+ "bits": 2
1918
+ },
1919
+ "model.layers.12.mlp.down_proj": {
1920
+ "group_size": 64,
1921
+ "bits": 2
1922
+ },
1923
+ "model.layers.12.mlp.up_proj": {
1924
+ "group_size": 64,
1925
+ "bits": 2
1926
+ },
1927
+ "model.layers.12.input_layernorm": false,
1928
+ "model.layers.12.post_attention_layernorm": false,
1929
+ "model.layers.13.self_attn.q_proj": {
1930
+ "group_size": 64,
1931
+ "bits": 2
1932
+ },
1933
+ "model.layers.13.self_attn.k_proj": {
1934
+ "group_size": 64,
1935
+ "bits": 2
1936
+ },
1937
+ "model.layers.13.self_attn.v_proj": {
1938
+ "group_size": 64,
1939
+ "bits": 2
1940
+ },
1941
+ "model.layers.13.self_attn.o_proj": {
1942
+ "group_size": 64,
1943
+ "bits": 2
1944
+ },
1945
+ "model.layers.13.self_attn.rope": false,
1946
+ "model.layers.13.mlp.gate_proj": {
1947
+ "group_size": 64,
1948
+ "bits": 2
1949
+ },
1950
+ "model.layers.13.mlp.down_proj": {
1951
+ "group_size": 64,
1952
+ "bits": 2
1953
+ },
1954
+ "model.layers.13.mlp.up_proj": {
1955
+ "group_size": 64,
1956
+ "bits": 2
1957
+ },
1958
+ "model.layers.13.input_layernorm": false,
1959
+ "model.layers.13.post_attention_layernorm": false,
1960
+ "model.layers.14.self_attn.q_proj": {
1961
+ "group_size": 64,
1962
+ "bits": 2
1963
+ },
1964
+ "model.layers.14.self_attn.k_proj": {
1965
+ "group_size": 64,
1966
+ "bits": 2
1967
+ },
1968
+ "model.layers.14.self_attn.v_proj": {
1969
+ "group_size": 64,
1970
+ "bits": 4
1971
+ },
1972
+ "model.layers.14.self_attn.o_proj": {
1973
+ "group_size": 64,
1974
+ "bits": 2
1975
+ },
1976
+ "model.layers.14.self_attn.rope": false,
1977
+ "model.layers.14.mlp.gate_proj": {
1978
+ "group_size": 64,
1979
+ "bits": 2
1980
+ },
1981
+ "model.layers.14.mlp.down_proj": {
1982
+ "group_size": 64,
1983
+ "bits": 4
1984
+ },
1985
+ "model.layers.14.mlp.up_proj": {
1986
+ "group_size": 64,
1987
+ "bits": 2
1988
+ },
1989
+ "model.layers.14.input_layernorm": false,
1990
+ "model.layers.14.post_attention_layernorm": false,
1991
+ "model.layers.15.self_attn.q_proj": {
1992
+ "group_size": 64,
1993
+ "bits": 2
1994
+ },
1995
+ "model.layers.15.self_attn.k_proj": {
1996
+ "group_size": 64,
1997
+ "bits": 2
1998
+ },
1999
+ "model.layers.15.self_attn.v_proj": {
2000
+ "group_size": 64,
2001
+ "bits": 2
2002
+ },
2003
+ "model.layers.15.self_attn.o_proj": {
2004
+ "group_size": 64,
2005
+ "bits": 2
2006
+ },
2007
+ "model.layers.15.self_attn.rope": false,
2008
+ "model.layers.15.mlp.gate_proj": {
2009
+ "group_size": 64,
2010
+ "bits": 2
2011
+ },
2012
+ "model.layers.15.mlp.down_proj": {
2013
+ "group_size": 64,
2014
+ "bits": 2
2015
+ },
2016
+ "model.layers.15.mlp.up_proj": {
2017
+ "group_size": 64,
2018
+ "bits": 2
2019
+ },
2020
+ "model.layers.15.input_layernorm": false,
2021
+ "model.layers.15.post_attention_layernorm": false,
2022
+ "model.layers.16.self_attn.q_proj": {
2023
+ "group_size": 64,
2024
+ "bits": 2
2025
+ },
2026
+ "model.layers.16.self_attn.k_proj": {
2027
+ "group_size": 64,
2028
+ "bits": 2
2029
+ },
2030
+ "model.layers.16.self_attn.v_proj": {
2031
+ "group_size": 64,
2032
+ "bits": 2
2033
+ },
2034
+ "model.layers.16.self_attn.o_proj": {
2035
+ "group_size": 64,
2036
+ "bits": 2
2037
+ },
2038
+ "model.layers.16.self_attn.rope": false,
2039
+ "model.layers.16.mlp.gate_proj": {
2040
+ "group_size": 64,
2041
+ "bits": 2
2042
+ },
2043
+ "model.layers.16.mlp.down_proj": {
2044
+ "group_size": 64,
2045
+ "bits": 2
2046
+ },
2047
+ "model.layers.16.mlp.up_proj": {
2048
+ "group_size": 64,
2049
+ "bits": 2
2050
+ },
2051
+ "model.layers.16.input_layernorm": false,
2052
+ "model.layers.16.post_attention_layernorm": false,
2053
+ "model.layers.17.self_attn.q_proj": {
2054
+ "group_size": 64,
2055
+ "bits": 2
2056
+ },
2057
+ "model.layers.17.self_attn.k_proj": {
2058
+ "group_size": 64,
2059
+ "bits": 2
2060
+ },
2061
+ "model.layers.17.self_attn.v_proj": {
2062
+ "group_size": 64,
2063
+ "bits": 4
2064
+ },
2065
+ "model.layers.17.self_attn.o_proj": {
2066
+ "group_size": 64,
2067
+ "bits": 2
2068
+ },
2069
+ "model.layers.17.self_attn.rope": false,
2070
+ "model.layers.17.mlp.gate_proj": {
2071
+ "group_size": 64,
2072
+ "bits": 2
2073
+ },
2074
+ "model.layers.17.mlp.down_proj": {
2075
+ "group_size": 64,
2076
+ "bits": 4
2077
+ },
2078
+ "model.layers.17.mlp.up_proj": {
2079
+ "group_size": 64,
2080
+ "bits": 2
2081
+ },
2082
+ "model.layers.17.input_layernorm": false,
2083
+ "model.layers.17.post_attention_layernorm": false,
2084
+ "model.layers.18.self_attn.q_proj": {
2085
+ "group_size": 64,
2086
+ "bits": 2
2087
+ },
2088
+ "model.layers.18.self_attn.k_proj": {
2089
+ "group_size": 64,
2090
+ "bits": 2
2091
+ },
2092
+ "model.layers.18.self_attn.v_proj": {
2093
+ "group_size": 64,
2094
+ "bits": 2
2095
+ },
2096
+ "model.layers.18.self_attn.o_proj": {
2097
+ "group_size": 64,
2098
+ "bits": 2
2099
+ },
2100
+ "model.layers.18.self_attn.rope": false,
2101
+ "model.layers.18.mlp.gate_proj": {
2102
+ "group_size": 64,
2103
+ "bits": 2
2104
+ },
2105
+ "model.layers.18.mlp.down_proj": {
2106
+ "group_size": 64,
2107
+ "bits": 2
2108
+ },
2109
+ "model.layers.18.mlp.up_proj": {
2110
+ "group_size": 64,
2111
+ "bits": 2
2112
+ },
2113
+ "model.layers.18.input_layernorm": false,
2114
+ "model.layers.18.post_attention_layernorm": false,
2115
+ "model.layers.19.self_attn.q_proj": {
2116
+ "group_size": 64,
2117
+ "bits": 2
2118
+ },
2119
+ "model.layers.19.self_attn.k_proj": {
2120
+ "group_size": 64,
2121
+ "bits": 2
2122
+ },
2123
+ "model.layers.19.self_attn.v_proj": {
2124
+ "group_size": 64,
2125
+ "bits": 2
2126
+ },
2127
+ "model.layers.19.self_attn.o_proj": {
2128
+ "group_size": 64,
2129
+ "bits": 2
2130
+ },
2131
+ "model.layers.19.self_attn.rope": false,
2132
+ "model.layers.19.mlp.gate_proj": {
2133
+ "group_size": 64,
2134
+ "bits": 2
2135
+ },
2136
+ "model.layers.19.mlp.down_proj": {
2137
+ "group_size": 64,
2138
+ "bits": 2
2139
+ },
2140
+ "model.layers.19.mlp.up_proj": {
2141
+ "group_size": 64,
2142
+ "bits": 2
2143
+ },
2144
+ "model.layers.19.input_layernorm": false,
2145
+ "model.layers.19.post_attention_layernorm": false,
2146
+ "model.layers.20.self_attn.q_proj": {
2147
+ "group_size": 64,
2148
+ "bits": 2
2149
+ },
2150
+ "model.layers.20.self_attn.k_proj": {
2151
+ "group_size": 64,
2152
+ "bits": 2
2153
+ },
2154
+ "model.layers.20.self_attn.v_proj": {
2155
+ "group_size": 64,
2156
+ "bits": 4
2157
+ },
2158
+ "model.layers.20.self_attn.o_proj": {
2159
+ "group_size": 64,
2160
+ "bits": 2
2161
+ },
2162
+ "model.layers.20.self_attn.rope": false,
2163
+ "model.layers.20.mlp.gate_proj": {
2164
+ "group_size": 64,
2165
+ "bits": 2
2166
+ },
2167
+ "model.layers.20.mlp.down_proj": {
2168
+ "group_size": 64,
2169
+ "bits": 4
2170
+ },
2171
+ "model.layers.20.mlp.up_proj": {
2172
+ "group_size": 64,
2173
+ "bits": 2
2174
+ },
2175
+ "model.layers.20.input_layernorm": false,
2176
+ "model.layers.20.post_attention_layernorm": false,
2177
+ "model.layers.21.self_attn.q_proj": {
2178
+ "group_size": 64,
2179
+ "bits": 2
2180
+ },
2181
+ "model.layers.21.self_attn.k_proj": {
2182
+ "group_size": 64,
2183
+ "bits": 2
2184
+ },
2185
+ "model.layers.21.self_attn.v_proj": {
2186
+ "group_size": 64,
2187
+ "bits": 2
2188
+ },
2189
+ "model.layers.21.self_attn.o_proj": {
2190
+ "group_size": 64,
2191
+ "bits": 2
2192
+ },
2193
+ "model.layers.21.self_attn.rope": false,
2194
+ "model.layers.21.mlp.gate_proj": {
2195
+ "group_size": 64,
2196
+ "bits": 2
2197
+ },
2198
+ "model.layers.21.mlp.down_proj": {
2199
+ "group_size": 64,
2200
+ "bits": 2
2201
+ },
2202
+ "model.layers.21.mlp.up_proj": {
2203
+ "group_size": 64,
2204
+ "bits": 2
2205
+ },
2206
+ "model.layers.21.input_layernorm": false,
2207
+ "model.layers.21.post_attention_layernorm": false,
2208
+ "model.layers.22.self_attn.q_proj": {
2209
+ "group_size": 64,
2210
+ "bits": 2
2211
+ },
2212
+ "model.layers.22.self_attn.k_proj": {
2213
+ "group_size": 64,
2214
+ "bits": 2
2215
+ },
2216
+ "model.layers.22.self_attn.v_proj": {
2217
+ "group_size": 64,
2218
+ "bits": 2
2219
+ },
2220
+ "model.layers.22.self_attn.o_proj": {
2221
+ "group_size": 64,
2222
+ "bits": 2
2223
+ },
2224
+ "model.layers.22.self_attn.rope": false,
2225
+ "model.layers.22.mlp.gate_proj": {
2226
+ "group_size": 64,
2227
+ "bits": 2
2228
+ },
2229
+ "model.layers.22.mlp.down_proj": {
2230
+ "group_size": 64,
2231
+ "bits": 2
2232
+ },
2233
+ "model.layers.22.mlp.up_proj": {
2234
+ "group_size": 64,
2235
+ "bits": 2
2236
+ },
2237
+ "model.layers.22.input_layernorm": false,
2238
+ "model.layers.22.post_attention_layernorm": false,
2239
+ "model.layers.23.self_attn.q_proj": {
2240
+ "group_size": 64,
2241
+ "bits": 2
2242
+ },
2243
+ "model.layers.23.self_attn.k_proj": {
2244
+ "group_size": 64,
2245
+ "bits": 2
2246
+ },
2247
+ "model.layers.23.self_attn.v_proj": {
2248
+ "group_size": 64,
2249
+ "bits": 4
2250
+ },
2251
+ "model.layers.23.self_attn.o_proj": {
2252
+ "group_size": 64,
2253
+ "bits": 2
2254
+ },
2255
+ "model.layers.23.self_attn.rope": false,
2256
+ "model.layers.23.mlp.gate_proj": {
2257
+ "group_size": 64,
2258
+ "bits": 2
2259
+ },
2260
+ "model.layers.23.mlp.down_proj": {
2261
+ "group_size": 64,
2262
+ "bits": 4
2263
+ },
2264
+ "model.layers.23.mlp.up_proj": {
2265
+ "group_size": 64,
2266
+ "bits": 2
2267
+ },
2268
+ "model.layers.23.input_layernorm": false,
2269
+ "model.layers.23.post_attention_layernorm": false,
2270
+ "model.layers.24.self_attn.q_proj": {
2271
+ "group_size": 64,
2272
+ "bits": 2
2273
+ },
2274
+ "model.layers.24.self_attn.k_proj": {
2275
+ "group_size": 64,
2276
+ "bits": 2
2277
+ },
2278
+ "model.layers.24.self_attn.v_proj": {
2279
+ "group_size": 64,
2280
+ "bits": 2
2281
+ },
2282
+ "model.layers.24.self_attn.o_proj": {
2283
+ "group_size": 64,
2284
+ "bits": 2
2285
+ },
2286
+ "model.layers.24.self_attn.rope": false,
2287
+ "model.layers.24.mlp.gate_proj": {
2288
+ "group_size": 64,
2289
+ "bits": 2
2290
+ },
2291
+ "model.layers.24.mlp.down_proj": {
2292
+ "group_size": 64,
2293
+ "bits": 2
2294
+ },
2295
+ "model.layers.24.mlp.up_proj": {
2296
+ "group_size": 64,
2297
+ "bits": 2
2298
+ },
2299
+ "model.layers.24.input_layernorm": false,
2300
+ "model.layers.24.post_attention_layernorm": false,
2301
+ "model.layers.25.self_attn.q_proj": {
2302
+ "group_size": 64,
2303
+ "bits": 2
2304
+ },
2305
+ "model.layers.25.self_attn.k_proj": {
2306
+ "group_size": 64,
2307
+ "bits": 2
2308
+ },
2309
+ "model.layers.25.self_attn.v_proj": {
2310
+ "group_size": 64,
2311
+ "bits": 2
2312
+ },
2313
+ "model.layers.25.self_attn.o_proj": {
2314
+ "group_size": 64,
2315
+ "bits": 2
2316
+ },
2317
+ "model.layers.25.self_attn.rope": false,
2318
+ "model.layers.25.mlp.gate_proj": {
2319
+ "group_size": 64,
2320
+ "bits": 2
2321
+ },
2322
+ "model.layers.25.mlp.down_proj": {
2323
+ "group_size": 64,
2324
+ "bits": 2
2325
+ },
2326
+ "model.layers.25.mlp.up_proj": {
2327
+ "group_size": 64,
2328
+ "bits": 2
2329
+ },
2330
+ "model.layers.25.input_layernorm": false,
2331
+ "model.layers.25.post_attention_layernorm": false,
2332
+ "model.layers.26.self_attn.q_proj": {
2333
+ "group_size": 64,
2334
+ "bits": 2
2335
+ },
2336
+ "model.layers.26.self_attn.k_proj": {
2337
+ "group_size": 64,
2338
+ "bits": 2
2339
+ },
2340
+ "model.layers.26.self_attn.v_proj": {
2341
+ "group_size": 64,
2342
+ "bits": 4
2343
+ },
2344
+ "model.layers.26.self_attn.o_proj": {
2345
+ "group_size": 64,
2346
+ "bits": 2
2347
+ },
2348
+ "model.layers.26.self_attn.rope": false,
2349
+ "model.layers.26.mlp.gate_proj": {
2350
+ "group_size": 64,
2351
+ "bits": 2
2352
+ },
2353
+ "model.layers.26.mlp.down_proj": {
2354
+ "group_size": 64,
2355
+ "bits": 4
2356
+ },
2357
+ "model.layers.26.mlp.up_proj": {
2358
+ "group_size": 64,
2359
+ "bits": 2
2360
+ },
2361
+ "model.layers.26.input_layernorm": false,
2362
+ "model.layers.26.post_attention_layernorm": false,
2363
+ "model.layers.27.self_attn.q_proj": {
2364
+ "group_size": 64,
2365
+ "bits": 2
2366
+ },
2367
+ "model.layers.27.self_attn.k_proj": {
2368
+ "group_size": 64,
2369
+ "bits": 2
2370
+ },
2371
+ "model.layers.27.self_attn.v_proj": {
2372
+ "group_size": 64,
2373
+ "bits": 2
2374
+ },
2375
+ "model.layers.27.self_attn.o_proj": {
2376
+ "group_size": 64,
2377
+ "bits": 2
2378
+ },
2379
+ "model.layers.27.self_attn.rope": false,
2380
+ "model.layers.27.mlp.gate_proj": {
2381
+ "group_size": 64,
2382
+ "bits": 2
2383
+ },
2384
+ "model.layers.27.mlp.down_proj": {
2385
+ "group_size": 64,
2386
+ "bits": 2
2387
+ },
2388
+ "model.layers.27.mlp.up_proj": {
2389
+ "group_size": 64,
2390
+ "bits": 2
2391
+ },
2392
+ "model.layers.27.input_layernorm": false,
2393
+ "model.layers.27.post_attention_layernorm": false,
2394
+ "model.layers.28.self_attn.q_proj": {
2395
+ "group_size": 64,
2396
+ "bits": 2
2397
+ },
2398
+ "model.layers.28.self_attn.k_proj": {
2399
+ "group_size": 64,
2400
+ "bits": 2
2401
+ },
2402
+ "model.layers.28.self_attn.v_proj": {
2403
+ "group_size": 64,
2404
+ "bits": 2
2405
+ },
2406
+ "model.layers.28.self_attn.o_proj": {
2407
+ "group_size": 64,
2408
+ "bits": 2
2409
+ },
2410
+ "model.layers.28.self_attn.rope": false,
2411
+ "model.layers.28.mlp.gate_proj": {
2412
+ "group_size": 64,
2413
+ "bits": 2
2414
+ },
2415
+ "model.layers.28.mlp.down_proj": {
2416
+ "group_size": 64,
2417
+ "bits": 2
2418
+ },
2419
+ "model.layers.28.mlp.up_proj": {
2420
+ "group_size": 64,
2421
+ "bits": 2
2422
+ },
2423
+ "model.layers.28.input_layernorm": false,
2424
+ "model.layers.28.post_attention_layernorm": false,
2425
+ "model.layers.29.self_attn.q_proj": {
2426
+ "group_size": 64,
2427
+ "bits": 2
2428
+ },
2429
+ "model.layers.29.self_attn.k_proj": {
2430
+ "group_size": 64,
2431
+ "bits": 2
2432
+ },
2433
+ "model.layers.29.self_attn.v_proj": {
2434
+ "group_size": 64,
2435
+ "bits": 4
2436
+ },
2437
+ "model.layers.29.self_attn.o_proj": {
2438
+ "group_size": 64,
2439
+ "bits": 2
2440
+ },
2441
+ "model.layers.29.self_attn.rope": false,
2442
+ "model.layers.29.mlp.gate_proj": {
2443
+ "group_size": 64,
2444
+ "bits": 2
2445
+ },
2446
+ "model.layers.29.mlp.down_proj": {
2447
+ "group_size": 64,
2448
+ "bits": 4
2449
+ },
2450
+ "model.layers.29.mlp.up_proj": {
2451
+ "group_size": 64,
2452
+ "bits": 2
2453
+ },
2454
+ "model.layers.29.input_layernorm": false,
2455
+ "model.layers.29.post_attention_layernorm": false,
2456
+ "model.layers.30.self_attn.q_proj": {
2457
+ "group_size": 64,
2458
+ "bits": 2
2459
+ },
2460
+ "model.layers.30.self_attn.k_proj": {
2461
+ "group_size": 64,
2462
+ "bits": 2
2463
+ },
2464
+ "model.layers.30.self_attn.v_proj": {
2465
+ "group_size": 64,
2466
+ "bits": 2
2467
+ },
2468
+ "model.layers.30.self_attn.o_proj": {
2469
+ "group_size": 64,
2470
+ "bits": 2
2471
+ },
2472
+ "model.layers.30.self_attn.rope": false,
2473
+ "model.layers.30.mlp.gate_proj": {
2474
+ "group_size": 64,
2475
+ "bits": 2
2476
+ },
2477
+ "model.layers.30.mlp.down_proj": {
2478
+ "group_size": 64,
2479
+ "bits": 2
2480
+ },
2481
+ "model.layers.30.mlp.up_proj": {
2482
+ "group_size": 64,
2483
+ "bits": 2
2484
+ },
2485
+ "model.layers.30.input_layernorm": false,
2486
+ "model.layers.30.post_attention_layernorm": false,
2487
+ "model.layers.31.self_attn.q_proj": {
2488
+ "group_size": 64,
2489
+ "bits": 2
2490
+ },
2491
+ "model.layers.31.self_attn.k_proj": {
2492
+ "group_size": 64,
2493
+ "bits": 2
2494
+ },
2495
+ "model.layers.31.self_attn.v_proj": {
2496
+ "group_size": 64,
2497
+ "bits": 2
2498
+ },
2499
+ "model.layers.31.self_attn.o_proj": {
2500
+ "group_size": 64,
2501
+ "bits": 2
2502
+ },
2503
+ "model.layers.31.self_attn.rope": false,
2504
+ "model.layers.31.mlp.gate_proj": {
2505
+ "group_size": 64,
2506
+ "bits": 2
2507
+ },
2508
+ "model.layers.31.mlp.down_proj": {
2509
+ "group_size": 64,
2510
+ "bits": 2
2511
+ },
2512
+ "model.layers.31.mlp.up_proj": {
2513
+ "group_size": 64,
2514
+ "bits": 2
2515
+ },
2516
+ "model.layers.31.input_layernorm": false,
2517
+ "model.layers.31.post_attention_layernorm": false,
2518
+ "model.layers.32.self_attn.q_proj": {
2519
+ "group_size": 64,
2520
+ "bits": 2
2521
+ },
2522
+ "model.layers.32.self_attn.k_proj": {
2523
+ "group_size": 64,
2524
+ "bits": 2
2525
+ },
2526
+ "model.layers.32.self_attn.v_proj": {
2527
+ "group_size": 64,
2528
+ "bits": 4
2529
+ },
2530
+ "model.layers.32.self_attn.o_proj": {
2531
+ "group_size": 64,
2532
+ "bits": 2
2533
+ },
2534
+ "model.layers.32.self_attn.rope": false,
2535
+ "model.layers.32.mlp.gate_proj": {
2536
+ "group_size": 64,
2537
+ "bits": 2
2538
+ },
2539
+ "model.layers.32.mlp.down_proj": {
2540
+ "group_size": 64,
2541
+ "bits": 4
2542
+ },
2543
+ "model.layers.32.mlp.up_proj": {
2544
+ "group_size": 64,
2545
+ "bits": 2
2546
+ },
2547
+ "model.layers.32.input_layernorm": false,
2548
+ "model.layers.32.post_attention_layernorm": false,
2549
+ "model.layers.33.self_attn.q_proj": {
2550
+ "group_size": 64,
2551
+ "bits": 2
2552
+ },
2553
+ "model.layers.33.self_attn.k_proj": {
2554
+ "group_size": 64,
2555
+ "bits": 2
2556
+ },
2557
+ "model.layers.33.self_attn.v_proj": {
2558
+ "group_size": 64,
2559
+ "bits": 2
2560
+ },
2561
+ "model.layers.33.self_attn.o_proj": {
2562
+ "group_size": 64,
2563
+ "bits": 2
2564
+ },
2565
+ "model.layers.33.self_attn.rope": false,
2566
+ "model.layers.33.mlp.gate_proj": {
2567
+ "group_size": 64,
2568
+ "bits": 2
2569
+ },
2570
+ "model.layers.33.mlp.down_proj": {
2571
+ "group_size": 64,
2572
+ "bits": 2
2573
+ },
2574
+ "model.layers.33.mlp.up_proj": {
2575
+ "group_size": 64,
2576
+ "bits": 2
2577
+ },
2578
+ "model.layers.33.input_layernorm": false,
2579
+ "model.layers.33.post_attention_layernorm": false,
2580
+ "model.layers.34.self_attn.q_proj": {
2581
+ "group_size": 64,
2582
+ "bits": 2
2583
+ },
2584
+ "model.layers.34.self_attn.k_proj": {
2585
+ "group_size": 64,
2586
+ "bits": 2
2587
+ },
2588
+ "model.layers.34.self_attn.v_proj": {
2589
+ "group_size": 64,
2590
+ "bits": 2
2591
+ },
2592
+ "model.layers.34.self_attn.o_proj": {
2593
+ "group_size": 64,
2594
+ "bits": 2
2595
+ },
2596
+ "model.layers.34.self_attn.rope": false,
2597
+ "model.layers.34.mlp.gate_proj": {
2598
+ "group_size": 64,
2599
+ "bits": 2
2600
+ },
2601
+ "model.layers.34.mlp.down_proj": {
2602
+ "group_size": 64,
2603
+ "bits": 2
2604
+ },
2605
+ "model.layers.34.mlp.up_proj": {
2606
+ "group_size": 64,
2607
+ "bits": 2
2608
+ },
2609
+ "model.layers.34.input_layernorm": false,
2610
+ "model.layers.34.post_attention_layernorm": false,
2611
+ "model.layers.35.self_attn.q_proj": {
2612
+ "group_size": 64,
2613
+ "bits": 2
2614
+ },
2615
+ "model.layers.35.self_attn.k_proj": {
2616
+ "group_size": 64,
2617
+ "bits": 2
2618
+ },
2619
+ "model.layers.35.self_attn.v_proj": {
2620
+ "group_size": 64,
2621
+ "bits": 4
2622
+ },
2623
+ "model.layers.35.self_attn.o_proj": {
2624
+ "group_size": 64,
2625
+ "bits": 2
2626
+ },
2627
+ "model.layers.35.self_attn.rope": false,
2628
+ "model.layers.35.mlp.gate_proj": {
2629
+ "group_size": 64,
2630
+ "bits": 2
2631
+ },
2632
+ "model.layers.35.mlp.down_proj": {
2633
+ "group_size": 64,
2634
+ "bits": 4
2635
+ },
2636
+ "model.layers.35.mlp.up_proj": {
2637
+ "group_size": 64,
2638
+ "bits": 2
2639
+ },
2640
+ "model.layers.35.input_layernorm": false,
2641
+ "model.layers.35.post_attention_layernorm": false,
2642
+ "model.layers.36.self_attn.q_proj": {
2643
+ "group_size": 64,
2644
+ "bits": 2
2645
+ },
2646
+ "model.layers.36.self_attn.k_proj": {
2647
+ "group_size": 64,
2648
+ "bits": 2
2649
+ },
2650
+ "model.layers.36.self_attn.v_proj": {
2651
+ "group_size": 64,
2652
+ "bits": 2
2653
+ },
2654
+ "model.layers.36.self_attn.o_proj": {
2655
+ "group_size": 64,
2656
+ "bits": 2
2657
+ },
2658
+ "model.layers.36.self_attn.rope": false,
2659
+ "model.layers.36.mlp.gate_proj": {
2660
+ "group_size": 64,
2661
+ "bits": 2
2662
+ },
2663
+ "model.layers.36.mlp.down_proj": {
2664
+ "group_size": 64,
2665
+ "bits": 2
2666
+ },
2667
+ "model.layers.36.mlp.up_proj": {
2668
+ "group_size": 64,
2669
+ "bits": 2
2670
+ },
2671
+ "model.layers.36.input_layernorm": false,
2672
+ "model.layers.36.post_attention_layernorm": false,
2673
+ "model.layers.37.self_attn.q_proj": {
2674
+ "group_size": 64,
2675
+ "bits": 2
2676
+ },
2677
+ "model.layers.37.self_attn.k_proj": {
2678
+ "group_size": 64,
2679
+ "bits": 2
2680
+ },
2681
+ "model.layers.37.self_attn.v_proj": {
2682
+ "group_size": 64,
2683
+ "bits": 2
2684
+ },
2685
+ "model.layers.37.self_attn.o_proj": {
2686
+ "group_size": 64,
2687
+ "bits": 2
2688
+ },
2689
+ "model.layers.37.self_attn.rope": false,
2690
+ "model.layers.37.mlp.gate_proj": {
2691
+ "group_size": 64,
2692
+ "bits": 2
2693
+ },
2694
+ "model.layers.37.mlp.down_proj": {
2695
+ "group_size": 64,
2696
+ "bits": 2
2697
+ },
2698
+ "model.layers.37.mlp.up_proj": {
2699
+ "group_size": 64,
2700
+ "bits": 2
2701
+ },
2702
+ "model.layers.37.input_layernorm": false,
2703
+ "model.layers.37.post_attention_layernorm": false,
2704
+ "model.layers.38.self_attn.q_proj": {
2705
+ "group_size": 64,
2706
+ "bits": 2
2707
+ },
2708
+ "model.layers.38.self_attn.k_proj": {
2709
+ "group_size": 64,
2710
+ "bits": 2
2711
+ },
2712
+ "model.layers.38.self_attn.v_proj": {
2713
+ "group_size": 64,
2714
+ "bits": 4
2715
+ },
2716
+ "model.layers.38.self_attn.o_proj": {
2717
+ "group_size": 64,
2718
+ "bits": 2
2719
+ },
2720
+ "model.layers.38.self_attn.rope": false,
2721
+ "model.layers.38.mlp.gate_proj": {
2722
+ "group_size": 64,
2723
+ "bits": 2
2724
+ },
2725
+ "model.layers.38.mlp.down_proj": {
2726
+ "group_size": 64,
2727
+ "bits": 4
2728
+ },
2729
+ "model.layers.38.mlp.up_proj": {
2730
+ "group_size": 64,
2731
+ "bits": 2
2732
+ },
2733
+ "model.layers.38.input_layernorm": false,
2734
+ "model.layers.38.post_attention_layernorm": false,
2735
+ "model.layers.39.self_attn.q_proj": {
2736
+ "group_size": 64,
2737
+ "bits": 2
2738
+ },
2739
+ "model.layers.39.self_attn.k_proj": {
2740
+ "group_size": 64,
2741
+ "bits": 2
2742
+ },
2743
+ "model.layers.39.self_attn.v_proj": {
2744
+ "group_size": 64,
2745
+ "bits": 2
2746
+ },
2747
+ "model.layers.39.self_attn.o_proj": {
2748
+ "group_size": 64,
2749
+ "bits": 2
2750
+ },
2751
+ "model.layers.39.self_attn.rope": false,
2752
+ "model.layers.39.mlp.gate_proj": {
2753
+ "group_size": 64,
2754
+ "bits": 2
2755
+ },
2756
+ "model.layers.39.mlp.down_proj": {
2757
+ "group_size": 64,
2758
+ "bits": 2
2759
+ },
2760
+ "model.layers.39.mlp.up_proj": {
2761
+ "group_size": 64,
2762
+ "bits": 2
2763
+ },
2764
+ "model.layers.39.input_layernorm": false,
2765
+ "model.layers.39.post_attention_layernorm": false,
2766
+ "model.layers.40.self_attn.q_proj": {
2767
+ "group_size": 64,
2768
+ "bits": 2
2769
+ },
2770
+ "model.layers.40.self_attn.k_proj": {
2771
+ "group_size": 64,
2772
+ "bits": 2
2773
+ },
2774
+ "model.layers.40.self_attn.v_proj": {
2775
+ "group_size": 64,
2776
+ "bits": 2
2777
+ },
2778
+ "model.layers.40.self_attn.o_proj": {
2779
+ "group_size": 64,
2780
+ "bits": 2
2781
+ },
2782
+ "model.layers.40.self_attn.rope": false,
2783
+ "model.layers.40.mlp.gate_proj": {
2784
+ "group_size": 64,
2785
+ "bits": 2
2786
+ },
2787
+ "model.layers.40.mlp.down_proj": {
2788
+ "group_size": 64,
2789
+ "bits": 2
2790
+ },
2791
+ "model.layers.40.mlp.up_proj": {
2792
+ "group_size": 64,
2793
+ "bits": 2
2794
+ },
2795
+ "model.layers.40.input_layernorm": false,
2796
+ "model.layers.40.post_attention_layernorm": false,
2797
+ "model.layers.41.self_attn.q_proj": {
2798
+ "group_size": 64,
2799
+ "bits": 2
2800
+ },
2801
+ "model.layers.41.self_attn.k_proj": {
2802
+ "group_size": 64,
2803
+ "bits": 2
2804
+ },
2805
+ "model.layers.41.self_attn.v_proj": {
2806
+ "group_size": 64,
2807
+ "bits": 4
2808
+ },
2809
+ "model.layers.41.self_attn.o_proj": {
2810
+ "group_size": 64,
2811
+ "bits": 2
2812
+ },
2813
+ "model.layers.41.self_attn.rope": false,
2814
+ "model.layers.41.mlp.gate_proj": {
2815
+ "group_size": 64,
2816
+ "bits": 2
2817
+ },
2818
+ "model.layers.41.mlp.down_proj": {
2819
+ "group_size": 64,
2820
+ "bits": 4
2821
+ },
2822
+ "model.layers.41.mlp.up_proj": {
2823
+ "group_size": 64,
2824
+ "bits": 2
2825
+ },
2826
+ "model.layers.41.input_layernorm": false,
2827
+ "model.layers.41.post_attention_layernorm": false,
2828
+ "model.layers.42.self_attn.q_proj": {
2829
+ "group_size": 64,
2830
+ "bits": 2
2831
+ },
2832
+ "model.layers.42.self_attn.k_proj": {
2833
+ "group_size": 64,
2834
+ "bits": 2
2835
+ },
2836
+ "model.layers.42.self_attn.v_proj": {
2837
+ "group_size": 64,
2838
+ "bits": 4
2839
+ },
2840
+ "model.layers.42.self_attn.o_proj": {
2841
+ "group_size": 64,
2842
+ "bits": 2
2843
+ },
2844
+ "model.layers.42.self_attn.rope": false,
2845
+ "model.layers.42.mlp.gate_proj": {
2846
+ "group_size": 64,
2847
+ "bits": 2
2848
+ },
2849
+ "model.layers.42.mlp.down_proj": {
2850
+ "group_size": 64,
2851
+ "bits": 4
2852
+ },
2853
+ "model.layers.42.mlp.up_proj": {
2854
+ "group_size": 64,
2855
+ "bits": 2
2856
+ },
2857
+ "model.layers.42.input_layernorm": false,
2858
+ "model.layers.42.post_attention_layernorm": false,
2859
+ "model.layers.43.self_attn.q_proj": {
2860
+ "group_size": 64,
2861
+ "bits": 2
2862
+ },
2863
+ "model.layers.43.self_attn.k_proj": {
2864
+ "group_size": 64,
2865
+ "bits": 2
2866
+ },
2867
+ "model.layers.43.self_attn.v_proj": {
2868
+ "group_size": 64,
2869
+ "bits": 4
2870
+ },
2871
+ "model.layers.43.self_attn.o_proj": {
2872
+ "group_size": 64,
2873
+ "bits": 2
2874
+ },
2875
+ "model.layers.43.self_attn.rope": false,
2876
+ "model.layers.43.mlp.gate_proj": {
2877
+ "group_size": 64,
2878
+ "bits": 2
2879
+ },
2880
+ "model.layers.43.mlp.down_proj": {
2881
+ "group_size": 64,
2882
+ "bits": 4
2883
+ },
2884
+ "model.layers.43.mlp.up_proj": {
2885
+ "group_size": 64,
2886
+ "bits": 2
2887
+ },
2888
+ "model.layers.43.input_layernorm": false,
2889
+ "model.layers.43.post_attention_layernorm": false,
2890
+ "model.layers.44.self_attn.q_proj": {
2891
+ "group_size": 64,
2892
+ "bits": 2
2893
+ },
2894
+ "model.layers.44.self_attn.k_proj": {
2895
+ "group_size": 64,
2896
+ "bits": 2
2897
+ },
2898
+ "model.layers.44.self_attn.v_proj": {
2899
+ "group_size": 64,
2900
+ "bits": 4
2901
+ },
2902
+ "model.layers.44.self_attn.o_proj": {
2903
+ "group_size": 64,
2904
+ "bits": 2
2905
+ },
2906
+ "model.layers.44.self_attn.rope": false,
2907
+ "model.layers.44.mlp.gate_proj": {
2908
+ "group_size": 64,
2909
+ "bits": 2
2910
+ },
2911
+ "model.layers.44.mlp.down_proj": {
2912
+ "group_size": 64,
2913
+ "bits": 4
2914
+ },
2915
+ "model.layers.44.mlp.up_proj": {
2916
+ "group_size": 64,
2917
+ "bits": 2
2918
+ },
2919
+ "model.layers.44.input_layernorm": false,
2920
+ "model.layers.44.post_attention_layernorm": false,
2921
+ "model.layers.45.self_attn.q_proj": {
2922
+ "group_size": 64,
2923
+ "bits": 2
2924
+ },
2925
+ "model.layers.45.self_attn.k_proj": {
2926
+ "group_size": 64,
2927
+ "bits": 2
2928
+ },
2929
+ "model.layers.45.self_attn.v_proj": {
2930
+ "group_size": 64,
2931
+ "bits": 4
2932
+ },
2933
+ "model.layers.45.self_attn.o_proj": {
2934
+ "group_size": 64,
2935
+ "bits": 2
2936
+ },
2937
+ "model.layers.45.self_attn.rope": false,
2938
+ "model.layers.45.mlp.gate_proj": {
2939
+ "group_size": 64,
2940
+ "bits": 2
2941
+ },
2942
+ "model.layers.45.mlp.down_proj": {
2943
+ "group_size": 64,
2944
+ "bits": 4
2945
+ },
2946
+ "model.layers.45.mlp.up_proj": {
2947
+ "group_size": 64,
2948
+ "bits": 2
2949
+ },
2950
+ "model.layers.45.input_layernorm": false,
2951
+ "model.layers.45.post_attention_layernorm": false,
2952
+ "model.layers.46.self_attn.q_proj": {
2953
+ "group_size": 64,
2954
+ "bits": 2
2955
+ },
2956
+ "model.layers.46.self_attn.k_proj": {
2957
+ "group_size": 64,
2958
+ "bits": 2
2959
+ },
2960
+ "model.layers.46.self_attn.v_proj": {
2961
+ "group_size": 64,
2962
+ "bits": 4
2963
+ },
2964
+ "model.layers.46.self_attn.o_proj": {
2965
+ "group_size": 64,
2966
+ "bits": 2
2967
+ },
2968
+ "model.layers.46.self_attn.rope": false,
2969
+ "model.layers.46.mlp.gate_proj": {
2970
+ "group_size": 64,
2971
+ "bits": 2
2972
+ },
2973
+ "model.layers.46.mlp.down_proj": {
2974
+ "group_size": 64,
2975
+ "bits": 4
2976
+ },
2977
+ "model.layers.46.mlp.up_proj": {
2978
+ "group_size": 64,
2979
+ "bits": 2
2980
+ },
2981
+ "model.layers.46.input_layernorm": false,
2982
+ "model.layers.46.post_attention_layernorm": false,
2983
+ "model.layers.47.self_attn.q_proj": {
2984
+ "group_size": 64,
2985
+ "bits": 2
2986
+ },
2987
+ "model.layers.47.self_attn.k_proj": {
2988
+ "group_size": 64,
2989
+ "bits": 2
2990
+ },
2991
+ "model.layers.47.self_attn.v_proj": {
2992
+ "group_size": 64,
2993
+ "bits": 4
2994
+ },
2995
+ "model.layers.47.self_attn.o_proj": {
2996
+ "group_size": 64,
2997
+ "bits": 2
2998
+ },
2999
+ "model.layers.47.self_attn.rope": false,
3000
+ "model.layers.47.mlp.gate_proj": {
3001
+ "group_size": 64,
3002
+ "bits": 2
3003
+ },
3004
+ "model.layers.47.mlp.down_proj": {
3005
+ "group_size": 64,
3006
+ "bits": 4
3007
+ },
3008
+ "model.layers.47.mlp.up_proj": {
3009
+ "group_size": 64,
3010
+ "bits": 2
3011
+ },
3012
+ "model.layers.47.input_layernorm": false,
3013
+ "model.layers.47.post_attention_layernorm": false,
3014
+ "model.norm": false,
3015
+ "lm_head": {
3016
+ "group_size": 64,
3017
+ "bits": 4
3018
+ }
3019
+ },
3020
+ "rms_norm_eps": 1e-05,
3021
+ "rope_theta": 1000000.0,
3022
+ "sliding_window": 131072,
3023
+ "tie_word_embeddings": false,
3024
+ "torch_dtype": "bfloat16",
3025
+ "transformers_version": "4.43.1",
3026
+ "use_cache": true,
3027
+ "use_sliding_window": false,
3028
+ "vocab_size": 152064
3029
+ }
DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8726362ae8224aeb9f3a381f02178df6a572c720ad9ab7ce890e40b19aea314d
3
+ size 5267968765
DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/model.safetensors.index.json ADDED
@@ -0,0 +1,1262 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 5267826688
4
+ },
5
+ "weight_map": {
6
+ "lm_head.biases": "model.safetensors",
7
+ "lm_head.scales": "model.safetensors",
8
+ "lm_head.weight": "model.safetensors",
9
+ "model.embed_tokens.biases": "model.safetensors",
10
+ "model.embed_tokens.scales": "model.safetensors",
11
+ "model.embed_tokens.weight": "model.safetensors",
12
+ "model.layers.0.input_layernorm.weight": "model.safetensors",
13
+ "model.layers.0.mlp.down_proj.biases": "model.safetensors",
14
+ "model.layers.0.mlp.down_proj.scales": "model.safetensors",
15
+ "model.layers.0.mlp.down_proj.weight": "model.safetensors",
16
+ "model.layers.0.mlp.gate_proj.biases": "model.safetensors",
17
+ "model.layers.0.mlp.gate_proj.scales": "model.safetensors",
18
+ "model.layers.0.mlp.gate_proj.weight": "model.safetensors",
19
+ "model.layers.0.mlp.up_proj.biases": "model.safetensors",
20
+ "model.layers.0.mlp.up_proj.scales": "model.safetensors",
21
+ "model.layers.0.mlp.up_proj.weight": "model.safetensors",
22
+ "model.layers.0.post_attention_layernorm.weight": "model.safetensors",
23
+ "model.layers.0.self_attn.k_proj.bias": "model.safetensors",
24
+ "model.layers.0.self_attn.k_proj.biases": "model.safetensors",
25
+ "model.layers.0.self_attn.k_proj.scales": "model.safetensors",
26
+ "model.layers.0.self_attn.k_proj.weight": "model.safetensors",
27
+ "model.layers.0.self_attn.o_proj.biases": "model.safetensors",
28
+ "model.layers.0.self_attn.o_proj.scales": "model.safetensors",
29
+ "model.layers.0.self_attn.o_proj.weight": "model.safetensors",
30
+ "model.layers.0.self_attn.q_proj.bias": "model.safetensors",
31
+ "model.layers.0.self_attn.q_proj.biases": "model.safetensors",
32
+ "model.layers.0.self_attn.q_proj.scales": "model.safetensors",
33
+ "model.layers.0.self_attn.q_proj.weight": "model.safetensors",
34
+ "model.layers.0.self_attn.v_proj.bias": "model.safetensors",
35
+ "model.layers.0.self_attn.v_proj.biases": "model.safetensors",
36
+ "model.layers.0.self_attn.v_proj.scales": "model.safetensors",
37
+ "model.layers.0.self_attn.v_proj.weight": "model.safetensors",
38
+ "model.layers.1.input_layernorm.weight": "model.safetensors",
39
+ "model.layers.1.mlp.down_proj.biases": "model.safetensors",
40
+ "model.layers.1.mlp.down_proj.scales": "model.safetensors",
41
+ "model.layers.1.mlp.down_proj.weight": "model.safetensors",
42
+ "model.layers.1.mlp.gate_proj.biases": "model.safetensors",
43
+ "model.layers.1.mlp.gate_proj.scales": "model.safetensors",
44
+ "model.layers.1.mlp.gate_proj.weight": "model.safetensors",
45
+ "model.layers.1.mlp.up_proj.biases": "model.safetensors",
46
+ "model.layers.1.mlp.up_proj.scales": "model.safetensors",
47
+ "model.layers.1.mlp.up_proj.weight": "model.safetensors",
48
+ "model.layers.1.post_attention_layernorm.weight": "model.safetensors",
49
+ "model.layers.1.self_attn.k_proj.bias": "model.safetensors",
50
+ "model.layers.1.self_attn.k_proj.biases": "model.safetensors",
51
+ "model.layers.1.self_attn.k_proj.scales": "model.safetensors",
52
+ "model.layers.1.self_attn.k_proj.weight": "model.safetensors",
53
+ "model.layers.1.self_attn.o_proj.biases": "model.safetensors",
54
+ "model.layers.1.self_attn.o_proj.scales": "model.safetensors",
55
+ "model.layers.1.self_attn.o_proj.weight": "model.safetensors",
56
+ "model.layers.1.self_attn.q_proj.bias": "model.safetensors",
57
+ "model.layers.1.self_attn.q_proj.biases": "model.safetensors",
58
+ "model.layers.1.self_attn.q_proj.scales": "model.safetensors",
59
+ "model.layers.1.self_attn.q_proj.weight": "model.safetensors",
60
+ "model.layers.1.self_attn.v_proj.bias": "model.safetensors",
61
+ "model.layers.1.self_attn.v_proj.biases": "model.safetensors",
62
+ "model.layers.1.self_attn.v_proj.scales": "model.safetensors",
63
+ "model.layers.1.self_attn.v_proj.weight": "model.safetensors",
64
+ "model.layers.10.input_layernorm.weight": "model.safetensors",
65
+ "model.layers.10.mlp.down_proj.biases": "model.safetensors",
66
+ "model.layers.10.mlp.down_proj.scales": "model.safetensors",
67
+ "model.layers.10.mlp.down_proj.weight": "model.safetensors",
68
+ "model.layers.10.mlp.gate_proj.biases": "model.safetensors",
69
+ "model.layers.10.mlp.gate_proj.scales": "model.safetensors",
70
+ "model.layers.10.mlp.gate_proj.weight": "model.safetensors",
71
+ "model.layers.10.mlp.up_proj.biases": "model.safetensors",
72
+ "model.layers.10.mlp.up_proj.scales": "model.safetensors",
73
+ "model.layers.10.mlp.up_proj.weight": "model.safetensors",
74
+ "model.layers.10.post_attention_layernorm.weight": "model.safetensors",
75
+ "model.layers.10.self_attn.k_proj.bias": "model.safetensors",
76
+ "model.layers.10.self_attn.k_proj.biases": "model.safetensors",
77
+ "model.layers.10.self_attn.k_proj.scales": "model.safetensors",
78
+ "model.layers.10.self_attn.k_proj.weight": "model.safetensors",
79
+ "model.layers.10.self_attn.o_proj.biases": "model.safetensors",
80
+ "model.layers.10.self_attn.o_proj.scales": "model.safetensors",
81
+ "model.layers.10.self_attn.o_proj.weight": "model.safetensors",
82
+ "model.layers.10.self_attn.q_proj.bias": "model.safetensors",
83
+ "model.layers.10.self_attn.q_proj.biases": "model.safetensors",
84
+ "model.layers.10.self_attn.q_proj.scales": "model.safetensors",
85
+ "model.layers.10.self_attn.q_proj.weight": "model.safetensors",
86
+ "model.layers.10.self_attn.v_proj.bias": "model.safetensors",
87
+ "model.layers.10.self_attn.v_proj.biases": "model.safetensors",
88
+ "model.layers.10.self_attn.v_proj.scales": "model.safetensors",
89
+ "model.layers.10.self_attn.v_proj.weight": "model.safetensors",
90
+ "model.layers.11.input_layernorm.weight": "model.safetensors",
91
+ "model.layers.11.mlp.down_proj.biases": "model.safetensors",
92
+ "model.layers.11.mlp.down_proj.scales": "model.safetensors",
93
+ "model.layers.11.mlp.down_proj.weight": "model.safetensors",
94
+ "model.layers.11.mlp.gate_proj.biases": "model.safetensors",
95
+ "model.layers.11.mlp.gate_proj.scales": "model.safetensors",
96
+ "model.layers.11.mlp.gate_proj.weight": "model.safetensors",
97
+ "model.layers.11.mlp.up_proj.biases": "model.safetensors",
98
+ "model.layers.11.mlp.up_proj.scales": "model.safetensors",
99
+ "model.layers.11.mlp.up_proj.weight": "model.safetensors",
100
+ "model.layers.11.post_attention_layernorm.weight": "model.safetensors",
101
+ "model.layers.11.self_attn.k_proj.bias": "model.safetensors",
102
+ "model.layers.11.self_attn.k_proj.biases": "model.safetensors",
103
+ "model.layers.11.self_attn.k_proj.scales": "model.safetensors",
104
+ "model.layers.11.self_attn.k_proj.weight": "model.safetensors",
105
+ "model.layers.11.self_attn.o_proj.biases": "model.safetensors",
106
+ "model.layers.11.self_attn.o_proj.scales": "model.safetensors",
107
+ "model.layers.11.self_attn.o_proj.weight": "model.safetensors",
108
+ "model.layers.11.self_attn.q_proj.bias": "model.safetensors",
109
+ "model.layers.11.self_attn.q_proj.biases": "model.safetensors",
110
+ "model.layers.11.self_attn.q_proj.scales": "model.safetensors",
111
+ "model.layers.11.self_attn.q_proj.weight": "model.safetensors",
112
+ "model.layers.11.self_attn.v_proj.bias": "model.safetensors",
113
+ "model.layers.11.self_attn.v_proj.biases": "model.safetensors",
114
+ "model.layers.11.self_attn.v_proj.scales": "model.safetensors",
115
+ "model.layers.11.self_attn.v_proj.weight": "model.safetensors",
116
+ "model.layers.12.input_layernorm.weight": "model.safetensors",
117
+ "model.layers.12.mlp.down_proj.biases": "model.safetensors",
118
+ "model.layers.12.mlp.down_proj.scales": "model.safetensors",
119
+ "model.layers.12.mlp.down_proj.weight": "model.safetensors",
120
+ "model.layers.12.mlp.gate_proj.biases": "model.safetensors",
121
+ "model.layers.12.mlp.gate_proj.scales": "model.safetensors",
122
+ "model.layers.12.mlp.gate_proj.weight": "model.safetensors",
123
+ "model.layers.12.mlp.up_proj.biases": "model.safetensors",
124
+ "model.layers.12.mlp.up_proj.scales": "model.safetensors",
125
+ "model.layers.12.mlp.up_proj.weight": "model.safetensors",
126
+ "model.layers.12.post_attention_layernorm.weight": "model.safetensors",
127
+ "model.layers.12.self_attn.k_proj.bias": "model.safetensors",
128
+ "model.layers.12.self_attn.k_proj.biases": "model.safetensors",
129
+ "model.layers.12.self_attn.k_proj.scales": "model.safetensors",
130
+ "model.layers.12.self_attn.k_proj.weight": "model.safetensors",
131
+ "model.layers.12.self_attn.o_proj.biases": "model.safetensors",
132
+ "model.layers.12.self_attn.o_proj.scales": "model.safetensors",
133
+ "model.layers.12.self_attn.o_proj.weight": "model.safetensors",
134
+ "model.layers.12.self_attn.q_proj.bias": "model.safetensors",
135
+ "model.layers.12.self_attn.q_proj.biases": "model.safetensors",
136
+ "model.layers.12.self_attn.q_proj.scales": "model.safetensors",
137
+ "model.layers.12.self_attn.q_proj.weight": "model.safetensors",
138
+ "model.layers.12.self_attn.v_proj.bias": "model.safetensors",
139
+ "model.layers.12.self_attn.v_proj.biases": "model.safetensors",
140
+ "model.layers.12.self_attn.v_proj.scales": "model.safetensors",
141
+ "model.layers.12.self_attn.v_proj.weight": "model.safetensors",
142
+ "model.layers.13.input_layernorm.weight": "model.safetensors",
143
+ "model.layers.13.mlp.down_proj.biases": "model.safetensors",
144
+ "model.layers.13.mlp.down_proj.scales": "model.safetensors",
145
+ "model.layers.13.mlp.down_proj.weight": "model.safetensors",
146
+ "model.layers.13.mlp.gate_proj.biases": "model.safetensors",
147
+ "model.layers.13.mlp.gate_proj.scales": "model.safetensors",
148
+ "model.layers.13.mlp.gate_proj.weight": "model.safetensors",
149
+ "model.layers.13.mlp.up_proj.biases": "model.safetensors",
150
+ "model.layers.13.mlp.up_proj.scales": "model.safetensors",
151
+ "model.layers.13.mlp.up_proj.weight": "model.safetensors",
152
+ "model.layers.13.post_attention_layernorm.weight": "model.safetensors",
153
+ "model.layers.13.self_attn.k_proj.bias": "model.safetensors",
154
+ "model.layers.13.self_attn.k_proj.biases": "model.safetensors",
155
+ "model.layers.13.self_attn.k_proj.scales": "model.safetensors",
156
+ "model.layers.13.self_attn.k_proj.weight": "model.safetensors",
157
+ "model.layers.13.self_attn.o_proj.biases": "model.safetensors",
158
+ "model.layers.13.self_attn.o_proj.scales": "model.safetensors",
159
+ "model.layers.13.self_attn.o_proj.weight": "model.safetensors",
160
+ "model.layers.13.self_attn.q_proj.bias": "model.safetensors",
161
+ "model.layers.13.self_attn.q_proj.biases": "model.safetensors",
162
+ "model.layers.13.self_attn.q_proj.scales": "model.safetensors",
163
+ "model.layers.13.self_attn.q_proj.weight": "model.safetensors",
164
+ "model.layers.13.self_attn.v_proj.bias": "model.safetensors",
165
+ "model.layers.13.self_attn.v_proj.biases": "model.safetensors",
166
+ "model.layers.13.self_attn.v_proj.scales": "model.safetensors",
167
+ "model.layers.13.self_attn.v_proj.weight": "model.safetensors",
168
+ "model.layers.14.input_layernorm.weight": "model.safetensors",
169
+ "model.layers.14.mlp.down_proj.biases": "model.safetensors",
170
+ "model.layers.14.mlp.down_proj.scales": "model.safetensors",
171
+ "model.layers.14.mlp.down_proj.weight": "model.safetensors",
172
+ "model.layers.14.mlp.gate_proj.biases": "model.safetensors",
173
+ "model.layers.14.mlp.gate_proj.scales": "model.safetensors",
174
+ "model.layers.14.mlp.gate_proj.weight": "model.safetensors",
175
+ "model.layers.14.mlp.up_proj.biases": "model.safetensors",
176
+ "model.layers.14.mlp.up_proj.scales": "model.safetensors",
177
+ "model.layers.14.mlp.up_proj.weight": "model.safetensors",
178
+ "model.layers.14.post_attention_layernorm.weight": "model.safetensors",
179
+ "model.layers.14.self_attn.k_proj.bias": "model.safetensors",
180
+ "model.layers.14.self_attn.k_proj.biases": "model.safetensors",
181
+ "model.layers.14.self_attn.k_proj.scales": "model.safetensors",
182
+ "model.layers.14.self_attn.k_proj.weight": "model.safetensors",
183
+ "model.layers.14.self_attn.o_proj.biases": "model.safetensors",
184
+ "model.layers.14.self_attn.o_proj.scales": "model.safetensors",
185
+ "model.layers.14.self_attn.o_proj.weight": "model.safetensors",
186
+ "model.layers.14.self_attn.q_proj.bias": "model.safetensors",
187
+ "model.layers.14.self_attn.q_proj.biases": "model.safetensors",
188
+ "model.layers.14.self_attn.q_proj.scales": "model.safetensors",
189
+ "model.layers.14.self_attn.q_proj.weight": "model.safetensors",
190
+ "model.layers.14.self_attn.v_proj.bias": "model.safetensors",
191
+ "model.layers.14.self_attn.v_proj.biases": "model.safetensors",
192
+ "model.layers.14.self_attn.v_proj.scales": "model.safetensors",
193
+ "model.layers.14.self_attn.v_proj.weight": "model.safetensors",
194
+ "model.layers.15.input_layernorm.weight": "model.safetensors",
195
+ "model.layers.15.mlp.down_proj.biases": "model.safetensors",
196
+ "model.layers.15.mlp.down_proj.scales": "model.safetensors",
197
+ "model.layers.15.mlp.down_proj.weight": "model.safetensors",
198
+ "model.layers.15.mlp.gate_proj.biases": "model.safetensors",
199
+ "model.layers.15.mlp.gate_proj.scales": "model.safetensors",
200
+ "model.layers.15.mlp.gate_proj.weight": "model.safetensors",
201
+ "model.layers.15.mlp.up_proj.biases": "model.safetensors",
202
+ "model.layers.15.mlp.up_proj.scales": "model.safetensors",
203
+ "model.layers.15.mlp.up_proj.weight": "model.safetensors",
204
+ "model.layers.15.post_attention_layernorm.weight": "model.safetensors",
205
+ "model.layers.15.self_attn.k_proj.bias": "model.safetensors",
206
+ "model.layers.15.self_attn.k_proj.biases": "model.safetensors",
207
+ "model.layers.15.self_attn.k_proj.scales": "model.safetensors",
208
+ "model.layers.15.self_attn.k_proj.weight": "model.safetensors",
209
+ "model.layers.15.self_attn.o_proj.biases": "model.safetensors",
210
+ "model.layers.15.self_attn.o_proj.scales": "model.safetensors",
211
+ "model.layers.15.self_attn.o_proj.weight": "model.safetensors",
212
+ "model.layers.15.self_attn.q_proj.bias": "model.safetensors",
213
+ "model.layers.15.self_attn.q_proj.biases": "model.safetensors",
214
+ "model.layers.15.self_attn.q_proj.scales": "model.safetensors",
215
+ "model.layers.15.self_attn.q_proj.weight": "model.safetensors",
216
+ "model.layers.15.self_attn.v_proj.bias": "model.safetensors",
217
+ "model.layers.15.self_attn.v_proj.biases": "model.safetensors",
218
+ "model.layers.15.self_attn.v_proj.scales": "model.safetensors",
219
+ "model.layers.15.self_attn.v_proj.weight": "model.safetensors",
220
+ "model.layers.16.input_layernorm.weight": "model.safetensors",
221
+ "model.layers.16.mlp.down_proj.biases": "model.safetensors",
222
+ "model.layers.16.mlp.down_proj.scales": "model.safetensors",
223
+ "model.layers.16.mlp.down_proj.weight": "model.safetensors",
224
+ "model.layers.16.mlp.gate_proj.biases": "model.safetensors",
225
+ "model.layers.16.mlp.gate_proj.scales": "model.safetensors",
226
+ "model.layers.16.mlp.gate_proj.weight": "model.safetensors",
227
+ "model.layers.16.mlp.up_proj.biases": "model.safetensors",
228
+ "model.layers.16.mlp.up_proj.scales": "model.safetensors",
229
+ "model.layers.16.mlp.up_proj.weight": "model.safetensors",
230
+ "model.layers.16.post_attention_layernorm.weight": "model.safetensors",
231
+ "model.layers.16.self_attn.k_proj.bias": "model.safetensors",
232
+ "model.layers.16.self_attn.k_proj.biases": "model.safetensors",
233
+ "model.layers.16.self_attn.k_proj.scales": "model.safetensors",
234
+ "model.layers.16.self_attn.k_proj.weight": "model.safetensors",
235
+ "model.layers.16.self_attn.o_proj.biases": "model.safetensors",
236
+ "model.layers.16.self_attn.o_proj.scales": "model.safetensors",
237
+ "model.layers.16.self_attn.o_proj.weight": "model.safetensors",
238
+ "model.layers.16.self_attn.q_proj.bias": "model.safetensors",
239
+ "model.layers.16.self_attn.q_proj.biases": "model.safetensors",
240
+ "model.layers.16.self_attn.q_proj.scales": "model.safetensors",
241
+ "model.layers.16.self_attn.q_proj.weight": "model.safetensors",
242
+ "model.layers.16.self_attn.v_proj.bias": "model.safetensors",
243
+ "model.layers.16.self_attn.v_proj.biases": "model.safetensors",
244
+ "model.layers.16.self_attn.v_proj.scales": "model.safetensors",
245
+ "model.layers.16.self_attn.v_proj.weight": "model.safetensors",
246
+ "model.layers.17.input_layernorm.weight": "model.safetensors",
247
+ "model.layers.17.mlp.down_proj.biases": "model.safetensors",
248
+ "model.layers.17.mlp.down_proj.scales": "model.safetensors",
249
+ "model.layers.17.mlp.down_proj.weight": "model.safetensors",
250
+ "model.layers.17.mlp.gate_proj.biases": "model.safetensors",
251
+ "model.layers.17.mlp.gate_proj.scales": "model.safetensors",
252
+ "model.layers.17.mlp.gate_proj.weight": "model.safetensors",
253
+ "model.layers.17.mlp.up_proj.biases": "model.safetensors",
254
+ "model.layers.17.mlp.up_proj.scales": "model.safetensors",
255
+ "model.layers.17.mlp.up_proj.weight": "model.safetensors",
256
+ "model.layers.17.post_attention_layernorm.weight": "model.safetensors",
257
+ "model.layers.17.self_attn.k_proj.bias": "model.safetensors",
258
+ "model.layers.17.self_attn.k_proj.biases": "model.safetensors",
259
+ "model.layers.17.self_attn.k_proj.scales": "model.safetensors",
260
+ "model.layers.17.self_attn.k_proj.weight": "model.safetensors",
261
+ "model.layers.17.self_attn.o_proj.biases": "model.safetensors",
262
+ "model.layers.17.self_attn.o_proj.scales": "model.safetensors",
263
+ "model.layers.17.self_attn.o_proj.weight": "model.safetensors",
264
+ "model.layers.17.self_attn.q_proj.bias": "model.safetensors",
265
+ "model.layers.17.self_attn.q_proj.biases": "model.safetensors",
266
+ "model.layers.17.self_attn.q_proj.scales": "model.safetensors",
267
+ "model.layers.17.self_attn.q_proj.weight": "model.safetensors",
268
+ "model.layers.17.self_attn.v_proj.bias": "model.safetensors",
269
+ "model.layers.17.self_attn.v_proj.biases": "model.safetensors",
270
+ "model.layers.17.self_attn.v_proj.scales": "model.safetensors",
271
+ "model.layers.17.self_attn.v_proj.weight": "model.safetensors",
272
+ "model.layers.18.input_layernorm.weight": "model.safetensors",
273
+ "model.layers.18.mlp.down_proj.biases": "model.safetensors",
274
+ "model.layers.18.mlp.down_proj.scales": "model.safetensors",
275
+ "model.layers.18.mlp.down_proj.weight": "model.safetensors",
276
+ "model.layers.18.mlp.gate_proj.biases": "model.safetensors",
277
+ "model.layers.18.mlp.gate_proj.scales": "model.safetensors",
278
+ "model.layers.18.mlp.gate_proj.weight": "model.safetensors",
279
+ "model.layers.18.mlp.up_proj.biases": "model.safetensors",
280
+ "model.layers.18.mlp.up_proj.scales": "model.safetensors",
281
+ "model.layers.18.mlp.up_proj.weight": "model.safetensors",
282
+ "model.layers.18.post_attention_layernorm.weight": "model.safetensors",
283
+ "model.layers.18.self_attn.k_proj.bias": "model.safetensors",
284
+ "model.layers.18.self_attn.k_proj.biases": "model.safetensors",
285
+ "model.layers.18.self_attn.k_proj.scales": "model.safetensors",
286
+ "model.layers.18.self_attn.k_proj.weight": "model.safetensors",
287
+ "model.layers.18.self_attn.o_proj.biases": "model.safetensors",
288
+ "model.layers.18.self_attn.o_proj.scales": "model.safetensors",
289
+ "model.layers.18.self_attn.o_proj.weight": "model.safetensors",
290
+ "model.layers.18.self_attn.q_proj.bias": "model.safetensors",
291
+ "model.layers.18.self_attn.q_proj.biases": "model.safetensors",
292
+ "model.layers.18.self_attn.q_proj.scales": "model.safetensors",
293
+ "model.layers.18.self_attn.q_proj.weight": "model.safetensors",
294
+ "model.layers.18.self_attn.v_proj.bias": "model.safetensors",
295
+ "model.layers.18.self_attn.v_proj.biases": "model.safetensors",
296
+ "model.layers.18.self_attn.v_proj.scales": "model.safetensors",
297
+ "model.layers.18.self_attn.v_proj.weight": "model.safetensors",
298
+ "model.layers.19.input_layernorm.weight": "model.safetensors",
299
+ "model.layers.19.mlp.down_proj.biases": "model.safetensors",
300
+ "model.layers.19.mlp.down_proj.scales": "model.safetensors",
301
+ "model.layers.19.mlp.down_proj.weight": "model.safetensors",
302
+ "model.layers.19.mlp.gate_proj.biases": "model.safetensors",
303
+ "model.layers.19.mlp.gate_proj.scales": "model.safetensors",
304
+ "model.layers.19.mlp.gate_proj.weight": "model.safetensors",
305
+ "model.layers.19.mlp.up_proj.biases": "model.safetensors",
306
+ "model.layers.19.mlp.up_proj.scales": "model.safetensors",
307
+ "model.layers.19.mlp.up_proj.weight": "model.safetensors",
308
+ "model.layers.19.post_attention_layernorm.weight": "model.safetensors",
309
+ "model.layers.19.self_attn.k_proj.bias": "model.safetensors",
310
+ "model.layers.19.self_attn.k_proj.biases": "model.safetensors",
311
+ "model.layers.19.self_attn.k_proj.scales": "model.safetensors",
312
+ "model.layers.19.self_attn.k_proj.weight": "model.safetensors",
313
+ "model.layers.19.self_attn.o_proj.biases": "model.safetensors",
314
+ "model.layers.19.self_attn.o_proj.scales": "model.safetensors",
315
+ "model.layers.19.self_attn.o_proj.weight": "model.safetensors",
316
+ "model.layers.19.self_attn.q_proj.bias": "model.safetensors",
317
+ "model.layers.19.self_attn.q_proj.biases": "model.safetensors",
318
+ "model.layers.19.self_attn.q_proj.scales": "model.safetensors",
319
+ "model.layers.19.self_attn.q_proj.weight": "model.safetensors",
320
+ "model.layers.19.self_attn.v_proj.bias": "model.safetensors",
321
+ "model.layers.19.self_attn.v_proj.biases": "model.safetensors",
322
+ "model.layers.19.self_attn.v_proj.scales": "model.safetensors",
323
+ "model.layers.19.self_attn.v_proj.weight": "model.safetensors",
324
+ "model.layers.2.input_layernorm.weight": "model.safetensors",
325
+ "model.layers.2.mlp.down_proj.biases": "model.safetensors",
326
+ "model.layers.2.mlp.down_proj.scales": "model.safetensors",
327
+ "model.layers.2.mlp.down_proj.weight": "model.safetensors",
328
+ "model.layers.2.mlp.gate_proj.biases": "model.safetensors",
329
+ "model.layers.2.mlp.gate_proj.scales": "model.safetensors",
330
+ "model.layers.2.mlp.gate_proj.weight": "model.safetensors",
331
+ "model.layers.2.mlp.up_proj.biases": "model.safetensors",
332
+ "model.layers.2.mlp.up_proj.scales": "model.safetensors",
333
+ "model.layers.2.mlp.up_proj.weight": "model.safetensors",
334
+ "model.layers.2.post_attention_layernorm.weight": "model.safetensors",
335
+ "model.layers.2.self_attn.k_proj.bias": "model.safetensors",
336
+ "model.layers.2.self_attn.k_proj.biases": "model.safetensors",
337
+ "model.layers.2.self_attn.k_proj.scales": "model.safetensors",
338
+ "model.layers.2.self_attn.k_proj.weight": "model.safetensors",
339
+ "model.layers.2.self_attn.o_proj.biases": "model.safetensors",
340
+ "model.layers.2.self_attn.o_proj.scales": "model.safetensors",
341
+ "model.layers.2.self_attn.o_proj.weight": "model.safetensors",
342
+ "model.layers.2.self_attn.q_proj.bias": "model.safetensors",
343
+ "model.layers.2.self_attn.q_proj.biases": "model.safetensors",
344
+ "model.layers.2.self_attn.q_proj.scales": "model.safetensors",
345
+ "model.layers.2.self_attn.q_proj.weight": "model.safetensors",
346
+ "model.layers.2.self_attn.v_proj.bias": "model.safetensors",
347
+ "model.layers.2.self_attn.v_proj.biases": "model.safetensors",
348
+ "model.layers.2.self_attn.v_proj.scales": "model.safetensors",
349
+ "model.layers.2.self_attn.v_proj.weight": "model.safetensors",
350
+ "model.layers.20.input_layernorm.weight": "model.safetensors",
351
+ "model.layers.20.mlp.down_proj.biases": "model.safetensors",
352
+ "model.layers.20.mlp.down_proj.scales": "model.safetensors",
353
+ "model.layers.20.mlp.down_proj.weight": "model.safetensors",
354
+ "model.layers.20.mlp.gate_proj.biases": "model.safetensors",
355
+ "model.layers.20.mlp.gate_proj.scales": "model.safetensors",
356
+ "model.layers.20.mlp.gate_proj.weight": "model.safetensors",
357
+ "model.layers.20.mlp.up_proj.biases": "model.safetensors",
358
+ "model.layers.20.mlp.up_proj.scales": "model.safetensors",
359
+ "model.layers.20.mlp.up_proj.weight": "model.safetensors",
360
+ "model.layers.20.post_attention_layernorm.weight": "model.safetensors",
361
+ "model.layers.20.self_attn.k_proj.bias": "model.safetensors",
362
+ "model.layers.20.self_attn.k_proj.biases": "model.safetensors",
363
+ "model.layers.20.self_attn.k_proj.scales": "model.safetensors",
364
+ "model.layers.20.self_attn.k_proj.weight": "model.safetensors",
365
+ "model.layers.20.self_attn.o_proj.biases": "model.safetensors",
366
+ "model.layers.20.self_attn.o_proj.scales": "model.safetensors",
367
+ "model.layers.20.self_attn.o_proj.weight": "model.safetensors",
368
+ "model.layers.20.self_attn.q_proj.bias": "model.safetensors",
369
+ "model.layers.20.self_attn.q_proj.biases": "model.safetensors",
370
+ "model.layers.20.self_attn.q_proj.scales": "model.safetensors",
371
+ "model.layers.20.self_attn.q_proj.weight": "model.safetensors",
372
+ "model.layers.20.self_attn.v_proj.bias": "model.safetensors",
373
+ "model.layers.20.self_attn.v_proj.biases": "model.safetensors",
374
+ "model.layers.20.self_attn.v_proj.scales": "model.safetensors",
375
+ "model.layers.20.self_attn.v_proj.weight": "model.safetensors",
376
+ "model.layers.21.input_layernorm.weight": "model.safetensors",
377
+ "model.layers.21.mlp.down_proj.biases": "model.safetensors",
378
+ "model.layers.21.mlp.down_proj.scales": "model.safetensors",
379
+ "model.layers.21.mlp.down_proj.weight": "model.safetensors",
380
+ "model.layers.21.mlp.gate_proj.biases": "model.safetensors",
381
+ "model.layers.21.mlp.gate_proj.scales": "model.safetensors",
382
+ "model.layers.21.mlp.gate_proj.weight": "model.safetensors",
383
+ "model.layers.21.mlp.up_proj.biases": "model.safetensors",
384
+ "model.layers.21.mlp.up_proj.scales": "model.safetensors",
385
+ "model.layers.21.mlp.up_proj.weight": "model.safetensors",
386
+ "model.layers.21.post_attention_layernorm.weight": "model.safetensors",
387
+ "model.layers.21.self_attn.k_proj.bias": "model.safetensors",
388
+ "model.layers.21.self_attn.k_proj.biases": "model.safetensors",
389
+ "model.layers.21.self_attn.k_proj.scales": "model.safetensors",
390
+ "model.layers.21.self_attn.k_proj.weight": "model.safetensors",
391
+ "model.layers.21.self_attn.o_proj.biases": "model.safetensors",
392
+ "model.layers.21.self_attn.o_proj.scales": "model.safetensors",
393
+ "model.layers.21.self_attn.o_proj.weight": "model.safetensors",
394
+ "model.layers.21.self_attn.q_proj.bias": "model.safetensors",
395
+ "model.layers.21.self_attn.q_proj.biases": "model.safetensors",
396
+ "model.layers.21.self_attn.q_proj.scales": "model.safetensors",
397
+ "model.layers.21.self_attn.q_proj.weight": "model.safetensors",
398
+ "model.layers.21.self_attn.v_proj.bias": "model.safetensors",
399
+ "model.layers.21.self_attn.v_proj.biases": "model.safetensors",
400
+ "model.layers.21.self_attn.v_proj.scales": "model.safetensors",
401
+ "model.layers.21.self_attn.v_proj.weight": "model.safetensors",
402
+ "model.layers.22.input_layernorm.weight": "model.safetensors",
403
+ "model.layers.22.mlp.down_proj.biases": "model.safetensors",
404
+ "model.layers.22.mlp.down_proj.scales": "model.safetensors",
405
+ "model.layers.22.mlp.down_proj.weight": "model.safetensors",
406
+ "model.layers.22.mlp.gate_proj.biases": "model.safetensors",
407
+ "model.layers.22.mlp.gate_proj.scales": "model.safetensors",
408
+ "model.layers.22.mlp.gate_proj.weight": "model.safetensors",
409
+ "model.layers.22.mlp.up_proj.biases": "model.safetensors",
410
+ "model.layers.22.mlp.up_proj.scales": "model.safetensors",
411
+ "model.layers.22.mlp.up_proj.weight": "model.safetensors",
412
+ "model.layers.22.post_attention_layernorm.weight": "model.safetensors",
413
+ "model.layers.22.self_attn.k_proj.bias": "model.safetensors",
414
+ "model.layers.22.self_attn.k_proj.biases": "model.safetensors",
415
+ "model.layers.22.self_attn.k_proj.scales": "model.safetensors",
416
+ "model.layers.22.self_attn.k_proj.weight": "model.safetensors",
417
+ "model.layers.22.self_attn.o_proj.biases": "model.safetensors",
418
+ "model.layers.22.self_attn.o_proj.scales": "model.safetensors",
419
+ "model.layers.22.self_attn.o_proj.weight": "model.safetensors",
420
+ "model.layers.22.self_attn.q_proj.bias": "model.safetensors",
421
+ "model.layers.22.self_attn.q_proj.biases": "model.safetensors",
422
+ "model.layers.22.self_attn.q_proj.scales": "model.safetensors",
423
+ "model.layers.22.self_attn.q_proj.weight": "model.safetensors",
424
+ "model.layers.22.self_attn.v_proj.bias": "model.safetensors",
425
+ "model.layers.22.self_attn.v_proj.biases": "model.safetensors",
426
+ "model.layers.22.self_attn.v_proj.scales": "model.safetensors",
427
+ "model.layers.22.self_attn.v_proj.weight": "model.safetensors",
428
+ "model.layers.23.input_layernorm.weight": "model.safetensors",
429
+ "model.layers.23.mlp.down_proj.biases": "model.safetensors",
430
+ "model.layers.23.mlp.down_proj.scales": "model.safetensors",
431
+ "model.layers.23.mlp.down_proj.weight": "model.safetensors",
432
+ "model.layers.23.mlp.gate_proj.biases": "model.safetensors",
433
+ "model.layers.23.mlp.gate_proj.scales": "model.safetensors",
434
+ "model.layers.23.mlp.gate_proj.weight": "model.safetensors",
435
+ "model.layers.23.mlp.up_proj.biases": "model.safetensors",
436
+ "model.layers.23.mlp.up_proj.scales": "model.safetensors",
437
+ "model.layers.23.mlp.up_proj.weight": "model.safetensors",
438
+ "model.layers.23.post_attention_layernorm.weight": "model.safetensors",
439
+ "model.layers.23.self_attn.k_proj.bias": "model.safetensors",
440
+ "model.layers.23.self_attn.k_proj.biases": "model.safetensors",
441
+ "model.layers.23.self_attn.k_proj.scales": "model.safetensors",
442
+ "model.layers.23.self_attn.k_proj.weight": "model.safetensors",
443
+ "model.layers.23.self_attn.o_proj.biases": "model.safetensors",
444
+ "model.layers.23.self_attn.o_proj.scales": "model.safetensors",
445
+ "model.layers.23.self_attn.o_proj.weight": "model.safetensors",
446
+ "model.layers.23.self_attn.q_proj.bias": "model.safetensors",
447
+ "model.layers.23.self_attn.q_proj.biases": "model.safetensors",
448
+ "model.layers.23.self_attn.q_proj.scales": "model.safetensors",
449
+ "model.layers.23.self_attn.q_proj.weight": "model.safetensors",
450
+ "model.layers.23.self_attn.v_proj.bias": "model.safetensors",
451
+ "model.layers.23.self_attn.v_proj.biases": "model.safetensors",
452
+ "model.layers.23.self_attn.v_proj.scales": "model.safetensors",
453
+ "model.layers.23.self_attn.v_proj.weight": "model.safetensors",
454
+ "model.layers.24.input_layernorm.weight": "model.safetensors",
455
+ "model.layers.24.mlp.down_proj.biases": "model.safetensors",
456
+ "model.layers.24.mlp.down_proj.scales": "model.safetensors",
457
+ "model.layers.24.mlp.down_proj.weight": "model.safetensors",
458
+ "model.layers.24.mlp.gate_proj.biases": "model.safetensors",
459
+ "model.layers.24.mlp.gate_proj.scales": "model.safetensors",
460
+ "model.layers.24.mlp.gate_proj.weight": "model.safetensors",
461
+ "model.layers.24.mlp.up_proj.biases": "model.safetensors",
462
+ "model.layers.24.mlp.up_proj.scales": "model.safetensors",
463
+ "model.layers.24.mlp.up_proj.weight": "model.safetensors",
464
+ "model.layers.24.post_attention_layernorm.weight": "model.safetensors",
465
+ "model.layers.24.self_attn.k_proj.bias": "model.safetensors",
466
+ "model.layers.24.self_attn.k_proj.biases": "model.safetensors",
467
+ "model.layers.24.self_attn.k_proj.scales": "model.safetensors",
468
+ "model.layers.24.self_attn.k_proj.weight": "model.safetensors",
469
+ "model.layers.24.self_attn.o_proj.biases": "model.safetensors",
470
+ "model.layers.24.self_attn.o_proj.scales": "model.safetensors",
471
+ "model.layers.24.self_attn.o_proj.weight": "model.safetensors",
472
+ "model.layers.24.self_attn.q_proj.bias": "model.safetensors",
473
+ "model.layers.24.self_attn.q_proj.biases": "model.safetensors",
474
+ "model.layers.24.self_attn.q_proj.scales": "model.safetensors",
475
+ "model.layers.24.self_attn.q_proj.weight": "model.safetensors",
476
+ "model.layers.24.self_attn.v_proj.bias": "model.safetensors",
477
+ "model.layers.24.self_attn.v_proj.biases": "model.safetensors",
478
+ "model.layers.24.self_attn.v_proj.scales": "model.safetensors",
479
+ "model.layers.24.self_attn.v_proj.weight": "model.safetensors",
480
+ "model.layers.25.input_layernorm.weight": "model.safetensors",
481
+ "model.layers.25.mlp.down_proj.biases": "model.safetensors",
482
+ "model.layers.25.mlp.down_proj.scales": "model.safetensors",
483
+ "model.layers.25.mlp.down_proj.weight": "model.safetensors",
484
+ "model.layers.25.mlp.gate_proj.biases": "model.safetensors",
485
+ "model.layers.25.mlp.gate_proj.scales": "model.safetensors",
486
+ "model.layers.25.mlp.gate_proj.weight": "model.safetensors",
487
+ "model.layers.25.mlp.up_proj.biases": "model.safetensors",
488
+ "model.layers.25.mlp.up_proj.scales": "model.safetensors",
489
+ "model.layers.25.mlp.up_proj.weight": "model.safetensors",
490
+ "model.layers.25.post_attention_layernorm.weight": "model.safetensors",
491
+ "model.layers.25.self_attn.k_proj.bias": "model.safetensors",
492
+ "model.layers.25.self_attn.k_proj.biases": "model.safetensors",
493
+ "model.layers.25.self_attn.k_proj.scales": "model.safetensors",
494
+ "model.layers.25.self_attn.k_proj.weight": "model.safetensors",
495
+ "model.layers.25.self_attn.o_proj.biases": "model.safetensors",
496
+ "model.layers.25.self_attn.o_proj.scales": "model.safetensors",
497
+ "model.layers.25.self_attn.o_proj.weight": "model.safetensors",
498
+ "model.layers.25.self_attn.q_proj.bias": "model.safetensors",
499
+ "model.layers.25.self_attn.q_proj.biases": "model.safetensors",
500
+ "model.layers.25.self_attn.q_proj.scales": "model.safetensors",
501
+ "model.layers.25.self_attn.q_proj.weight": "model.safetensors",
502
+ "model.layers.25.self_attn.v_proj.bias": "model.safetensors",
503
+ "model.layers.25.self_attn.v_proj.biases": "model.safetensors",
504
+ "model.layers.25.self_attn.v_proj.scales": "model.safetensors",
505
+ "model.layers.25.self_attn.v_proj.weight": "model.safetensors",
506
+ "model.layers.26.input_layernorm.weight": "model.safetensors",
507
+ "model.layers.26.mlp.down_proj.biases": "model.safetensors",
508
+ "model.layers.26.mlp.down_proj.scales": "model.safetensors",
509
+ "model.layers.26.mlp.down_proj.weight": "model.safetensors",
510
+ "model.layers.26.mlp.gate_proj.biases": "model.safetensors",
511
+ "model.layers.26.mlp.gate_proj.scales": "model.safetensors",
512
+ "model.layers.26.mlp.gate_proj.weight": "model.safetensors",
513
+ "model.layers.26.mlp.up_proj.biases": "model.safetensors",
514
+ "model.layers.26.mlp.up_proj.scales": "model.safetensors",
515
+ "model.layers.26.mlp.up_proj.weight": "model.safetensors",
516
+ "model.layers.26.post_attention_layernorm.weight": "model.safetensors",
517
+ "model.layers.26.self_attn.k_proj.bias": "model.safetensors",
518
+ "model.layers.26.self_attn.k_proj.biases": "model.safetensors",
519
+ "model.layers.26.self_attn.k_proj.scales": "model.safetensors",
520
+ "model.layers.26.self_attn.k_proj.weight": "model.safetensors",
521
+ "model.layers.26.self_attn.o_proj.biases": "model.safetensors",
522
+ "model.layers.26.self_attn.o_proj.scales": "model.safetensors",
523
+ "model.layers.26.self_attn.o_proj.weight": "model.safetensors",
524
+ "model.layers.26.self_attn.q_proj.bias": "model.safetensors",
525
+ "model.layers.26.self_attn.q_proj.biases": "model.safetensors",
526
+ "model.layers.26.self_attn.q_proj.scales": "model.safetensors",
527
+ "model.layers.26.self_attn.q_proj.weight": "model.safetensors",
528
+ "model.layers.26.self_attn.v_proj.bias": "model.safetensors",
529
+ "model.layers.26.self_attn.v_proj.biases": "model.safetensors",
530
+ "model.layers.26.self_attn.v_proj.scales": "model.safetensors",
531
+ "model.layers.26.self_attn.v_proj.weight": "model.safetensors",
532
+ "model.layers.27.input_layernorm.weight": "model.safetensors",
533
+ "model.layers.27.mlp.down_proj.biases": "model.safetensors",
534
+ "model.layers.27.mlp.down_proj.scales": "model.safetensors",
535
+ "model.layers.27.mlp.down_proj.weight": "model.safetensors",
536
+ "model.layers.27.mlp.gate_proj.biases": "model.safetensors",
537
+ "model.layers.27.mlp.gate_proj.scales": "model.safetensors",
538
+ "model.layers.27.mlp.gate_proj.weight": "model.safetensors",
539
+ "model.layers.27.mlp.up_proj.biases": "model.safetensors",
540
+ "model.layers.27.mlp.up_proj.scales": "model.safetensors",
541
+ "model.layers.27.mlp.up_proj.weight": "model.safetensors",
542
+ "model.layers.27.post_attention_layernorm.weight": "model.safetensors",
543
+ "model.layers.27.self_attn.k_proj.bias": "model.safetensors",
544
+ "model.layers.27.self_attn.k_proj.biases": "model.safetensors",
545
+ "model.layers.27.self_attn.k_proj.scales": "model.safetensors",
546
+ "model.layers.27.self_attn.k_proj.weight": "model.safetensors",
547
+ "model.layers.27.self_attn.o_proj.biases": "model.safetensors",
548
+ "model.layers.27.self_attn.o_proj.scales": "model.safetensors",
549
+ "model.layers.27.self_attn.o_proj.weight": "model.safetensors",
550
+ "model.layers.27.self_attn.q_proj.bias": "model.safetensors",
551
+ "model.layers.27.self_attn.q_proj.biases": "model.safetensors",
552
+ "model.layers.27.self_attn.q_proj.scales": "model.safetensors",
553
+ "model.layers.27.self_attn.q_proj.weight": "model.safetensors",
554
+ "model.layers.27.self_attn.v_proj.bias": "model.safetensors",
555
+ "model.layers.27.self_attn.v_proj.biases": "model.safetensors",
556
+ "model.layers.27.self_attn.v_proj.scales": "model.safetensors",
557
+ "model.layers.27.self_attn.v_proj.weight": "model.safetensors",
558
+ "model.layers.28.input_layernorm.weight": "model.safetensors",
559
+ "model.layers.28.mlp.down_proj.biases": "model.safetensors",
560
+ "model.layers.28.mlp.down_proj.scales": "model.safetensors",
561
+ "model.layers.28.mlp.down_proj.weight": "model.safetensors",
562
+ "model.layers.28.mlp.gate_proj.biases": "model.safetensors",
563
+ "model.layers.28.mlp.gate_proj.scales": "model.safetensors",
564
+ "model.layers.28.mlp.gate_proj.weight": "model.safetensors",
565
+ "model.layers.28.mlp.up_proj.biases": "model.safetensors",
566
+ "model.layers.28.mlp.up_proj.scales": "model.safetensors",
567
+ "model.layers.28.mlp.up_proj.weight": "model.safetensors",
568
+ "model.layers.28.post_attention_layernorm.weight": "model.safetensors",
569
+ "model.layers.28.self_attn.k_proj.bias": "model.safetensors",
570
+ "model.layers.28.self_attn.k_proj.biases": "model.safetensors",
571
+ "model.layers.28.self_attn.k_proj.scales": "model.safetensors",
572
+ "model.layers.28.self_attn.k_proj.weight": "model.safetensors",
573
+ "model.layers.28.self_attn.o_proj.biases": "model.safetensors",
574
+ "model.layers.28.self_attn.o_proj.scales": "model.safetensors",
575
+ "model.layers.28.self_attn.o_proj.weight": "model.safetensors",
576
+ "model.layers.28.self_attn.q_proj.bias": "model.safetensors",
577
+ "model.layers.28.self_attn.q_proj.biases": "model.safetensors",
578
+ "model.layers.28.self_attn.q_proj.scales": "model.safetensors",
579
+ "model.layers.28.self_attn.q_proj.weight": "model.safetensors",
580
+ "model.layers.28.self_attn.v_proj.bias": "model.safetensors",
581
+ "model.layers.28.self_attn.v_proj.biases": "model.safetensors",
582
+ "model.layers.28.self_attn.v_proj.scales": "model.safetensors",
583
+ "model.layers.28.self_attn.v_proj.weight": "model.safetensors",
584
+ "model.layers.29.input_layernorm.weight": "model.safetensors",
585
+ "model.layers.29.mlp.down_proj.biases": "model.safetensors",
586
+ "model.layers.29.mlp.down_proj.scales": "model.safetensors",
587
+ "model.layers.29.mlp.down_proj.weight": "model.safetensors",
588
+ "model.layers.29.mlp.gate_proj.biases": "model.safetensors",
589
+ "model.layers.29.mlp.gate_proj.scales": "model.safetensors",
590
+ "model.layers.29.mlp.gate_proj.weight": "model.safetensors",
591
+ "model.layers.29.mlp.up_proj.biases": "model.safetensors",
592
+ "model.layers.29.mlp.up_proj.scales": "model.safetensors",
593
+ "model.layers.29.mlp.up_proj.weight": "model.safetensors",
594
+ "model.layers.29.post_attention_layernorm.weight": "model.safetensors",
595
+ "model.layers.29.self_attn.k_proj.bias": "model.safetensors",
596
+ "model.layers.29.self_attn.k_proj.biases": "model.safetensors",
597
+ "model.layers.29.self_attn.k_proj.scales": "model.safetensors",
598
+ "model.layers.29.self_attn.k_proj.weight": "model.safetensors",
599
+ "model.layers.29.self_attn.o_proj.biases": "model.safetensors",
600
+ "model.layers.29.self_attn.o_proj.scales": "model.safetensors",
601
+ "model.layers.29.self_attn.o_proj.weight": "model.safetensors",
602
+ "model.layers.29.self_attn.q_proj.bias": "model.safetensors",
603
+ "model.layers.29.self_attn.q_proj.biases": "model.safetensors",
604
+ "model.layers.29.self_attn.q_proj.scales": "model.safetensors",
605
+ "model.layers.29.self_attn.q_proj.weight": "model.safetensors",
606
+ "model.layers.29.self_attn.v_proj.bias": "model.safetensors",
607
+ "model.layers.29.self_attn.v_proj.biases": "model.safetensors",
608
+ "model.layers.29.self_attn.v_proj.scales": "model.safetensors",
609
+ "model.layers.29.self_attn.v_proj.weight": "model.safetensors",
610
+ "model.layers.3.input_layernorm.weight": "model.safetensors",
611
+ "model.layers.3.mlp.down_proj.biases": "model.safetensors",
612
+ "model.layers.3.mlp.down_proj.scales": "model.safetensors",
613
+ "model.layers.3.mlp.down_proj.weight": "model.safetensors",
614
+ "model.layers.3.mlp.gate_proj.biases": "model.safetensors",
615
+ "model.layers.3.mlp.gate_proj.scales": "model.safetensors",
616
+ "model.layers.3.mlp.gate_proj.weight": "model.safetensors",
617
+ "model.layers.3.mlp.up_proj.biases": "model.safetensors",
618
+ "model.layers.3.mlp.up_proj.scales": "model.safetensors",
619
+ "model.layers.3.mlp.up_proj.weight": "model.safetensors",
620
+ "model.layers.3.post_attention_layernorm.weight": "model.safetensors",
621
+ "model.layers.3.self_attn.k_proj.bias": "model.safetensors",
622
+ "model.layers.3.self_attn.k_proj.biases": "model.safetensors",
623
+ "model.layers.3.self_attn.k_proj.scales": "model.safetensors",
624
+ "model.layers.3.self_attn.k_proj.weight": "model.safetensors",
625
+ "model.layers.3.self_attn.o_proj.biases": "model.safetensors",
626
+ "model.layers.3.self_attn.o_proj.scales": "model.safetensors",
627
+ "model.layers.3.self_attn.o_proj.weight": "model.safetensors",
628
+ "model.layers.3.self_attn.q_proj.bias": "model.safetensors",
629
+ "model.layers.3.self_attn.q_proj.biases": "model.safetensors",
630
+ "model.layers.3.self_attn.q_proj.scales": "model.safetensors",
631
+ "model.layers.3.self_attn.q_proj.weight": "model.safetensors",
632
+ "model.layers.3.self_attn.v_proj.bias": "model.safetensors",
633
+ "model.layers.3.self_attn.v_proj.biases": "model.safetensors",
634
+ "model.layers.3.self_attn.v_proj.scales": "model.safetensors",
635
+ "model.layers.3.self_attn.v_proj.weight": "model.safetensors",
636
+ "model.layers.30.input_layernorm.weight": "model.safetensors",
637
+ "model.layers.30.mlp.down_proj.biases": "model.safetensors",
638
+ "model.layers.30.mlp.down_proj.scales": "model.safetensors",
639
+ "model.layers.30.mlp.down_proj.weight": "model.safetensors",
640
+ "model.layers.30.mlp.gate_proj.biases": "model.safetensors",
641
+ "model.layers.30.mlp.gate_proj.scales": "model.safetensors",
642
+ "model.layers.30.mlp.gate_proj.weight": "model.safetensors",
643
+ "model.layers.30.mlp.up_proj.biases": "model.safetensors",
644
+ "model.layers.30.mlp.up_proj.scales": "model.safetensors",
645
+ "model.layers.30.mlp.up_proj.weight": "model.safetensors",
646
+ "model.layers.30.post_attention_layernorm.weight": "model.safetensors",
647
+ "model.layers.30.self_attn.k_proj.bias": "model.safetensors",
648
+ "model.layers.30.self_attn.k_proj.biases": "model.safetensors",
649
+ "model.layers.30.self_attn.k_proj.scales": "model.safetensors",
650
+ "model.layers.30.self_attn.k_proj.weight": "model.safetensors",
651
+ "model.layers.30.self_attn.o_proj.biases": "model.safetensors",
652
+ "model.layers.30.self_attn.o_proj.scales": "model.safetensors",
653
+ "model.layers.30.self_attn.o_proj.weight": "model.safetensors",
654
+ "model.layers.30.self_attn.q_proj.bias": "model.safetensors",
655
+ "model.layers.30.self_attn.q_proj.biases": "model.safetensors",
656
+ "model.layers.30.self_attn.q_proj.scales": "model.safetensors",
657
+ "model.layers.30.self_attn.q_proj.weight": "model.safetensors",
658
+ "model.layers.30.self_attn.v_proj.bias": "model.safetensors",
659
+ "model.layers.30.self_attn.v_proj.biases": "model.safetensors",
660
+ "model.layers.30.self_attn.v_proj.scales": "model.safetensors",
661
+ "model.layers.30.self_attn.v_proj.weight": "model.safetensors",
662
+ "model.layers.31.input_layernorm.weight": "model.safetensors",
663
+ "model.layers.31.mlp.down_proj.biases": "model.safetensors",
664
+ "model.layers.31.mlp.down_proj.scales": "model.safetensors",
665
+ "model.layers.31.mlp.down_proj.weight": "model.safetensors",
666
+ "model.layers.31.mlp.gate_proj.biases": "model.safetensors",
667
+ "model.layers.31.mlp.gate_proj.scales": "model.safetensors",
668
+ "model.layers.31.mlp.gate_proj.weight": "model.safetensors",
669
+ "model.layers.31.mlp.up_proj.biases": "model.safetensors",
670
+ "model.layers.31.mlp.up_proj.scales": "model.safetensors",
671
+ "model.layers.31.mlp.up_proj.weight": "model.safetensors",
672
+ "model.layers.31.post_attention_layernorm.weight": "model.safetensors",
673
+ "model.layers.31.self_attn.k_proj.bias": "model.safetensors",
674
+ "model.layers.31.self_attn.k_proj.biases": "model.safetensors",
675
+ "model.layers.31.self_attn.k_proj.scales": "model.safetensors",
676
+ "model.layers.31.self_attn.k_proj.weight": "model.safetensors",
677
+ "model.layers.31.self_attn.o_proj.biases": "model.safetensors",
678
+ "model.layers.31.self_attn.o_proj.scales": "model.safetensors",
679
+ "model.layers.31.self_attn.o_proj.weight": "model.safetensors",
680
+ "model.layers.31.self_attn.q_proj.bias": "model.safetensors",
681
+ "model.layers.31.self_attn.q_proj.biases": "model.safetensors",
682
+ "model.layers.31.self_attn.q_proj.scales": "model.safetensors",
683
+ "model.layers.31.self_attn.q_proj.weight": "model.safetensors",
684
+ "model.layers.31.self_attn.v_proj.bias": "model.safetensors",
685
+ "model.layers.31.self_attn.v_proj.biases": "model.safetensors",
686
+ "model.layers.31.self_attn.v_proj.scales": "model.safetensors",
687
+ "model.layers.31.self_attn.v_proj.weight": "model.safetensors",
688
+ "model.layers.32.input_layernorm.weight": "model.safetensors",
689
+ "model.layers.32.mlp.down_proj.biases": "model.safetensors",
690
+ "model.layers.32.mlp.down_proj.scales": "model.safetensors",
691
+ "model.layers.32.mlp.down_proj.weight": "model.safetensors",
692
+ "model.layers.32.mlp.gate_proj.biases": "model.safetensors",
693
+ "model.layers.32.mlp.gate_proj.scales": "model.safetensors",
694
+ "model.layers.32.mlp.gate_proj.weight": "model.safetensors",
695
+ "model.layers.32.mlp.up_proj.biases": "model.safetensors",
696
+ "model.layers.32.mlp.up_proj.scales": "model.safetensors",
697
+ "model.layers.32.mlp.up_proj.weight": "model.safetensors",
698
+ "model.layers.32.post_attention_layernorm.weight": "model.safetensors",
699
+ "model.layers.32.self_attn.k_proj.bias": "model.safetensors",
700
+ "model.layers.32.self_attn.k_proj.biases": "model.safetensors",
701
+ "model.layers.32.self_attn.k_proj.scales": "model.safetensors",
702
+ "model.layers.32.self_attn.k_proj.weight": "model.safetensors",
703
+ "model.layers.32.self_attn.o_proj.biases": "model.safetensors",
704
+ "model.layers.32.self_attn.o_proj.scales": "model.safetensors",
705
+ "model.layers.32.self_attn.o_proj.weight": "model.safetensors",
706
+ "model.layers.32.self_attn.q_proj.bias": "model.safetensors",
707
+ "model.layers.32.self_attn.q_proj.biases": "model.safetensors",
708
+ "model.layers.32.self_attn.q_proj.scales": "model.safetensors",
709
+ "model.layers.32.self_attn.q_proj.weight": "model.safetensors",
710
+ "model.layers.32.self_attn.v_proj.bias": "model.safetensors",
711
+ "model.layers.32.self_attn.v_proj.biases": "model.safetensors",
712
+ "model.layers.32.self_attn.v_proj.scales": "model.safetensors",
713
+ "model.layers.32.self_attn.v_proj.weight": "model.safetensors",
714
+ "model.layers.33.input_layernorm.weight": "model.safetensors",
715
+ "model.layers.33.mlp.down_proj.biases": "model.safetensors",
716
+ "model.layers.33.mlp.down_proj.scales": "model.safetensors",
717
+ "model.layers.33.mlp.down_proj.weight": "model.safetensors",
718
+ "model.layers.33.mlp.gate_proj.biases": "model.safetensors",
719
+ "model.layers.33.mlp.gate_proj.scales": "model.safetensors",
720
+ "model.layers.33.mlp.gate_proj.weight": "model.safetensors",
721
+ "model.layers.33.mlp.up_proj.biases": "model.safetensors",
722
+ "model.layers.33.mlp.up_proj.scales": "model.safetensors",
723
+ "model.layers.33.mlp.up_proj.weight": "model.safetensors",
724
+ "model.layers.33.post_attention_layernorm.weight": "model.safetensors",
725
+ "model.layers.33.self_attn.k_proj.bias": "model.safetensors",
726
+ "model.layers.33.self_attn.k_proj.biases": "model.safetensors",
727
+ "model.layers.33.self_attn.k_proj.scales": "model.safetensors",
728
+ "model.layers.33.self_attn.k_proj.weight": "model.safetensors",
729
+ "model.layers.33.self_attn.o_proj.biases": "model.safetensors",
730
+ "model.layers.33.self_attn.o_proj.scales": "model.safetensors",
731
+ "model.layers.33.self_attn.o_proj.weight": "model.safetensors",
732
+ "model.layers.33.self_attn.q_proj.bias": "model.safetensors",
733
+ "model.layers.33.self_attn.q_proj.biases": "model.safetensors",
734
+ "model.layers.33.self_attn.q_proj.scales": "model.safetensors",
735
+ "model.layers.33.self_attn.q_proj.weight": "model.safetensors",
736
+ "model.layers.33.self_attn.v_proj.bias": "model.safetensors",
737
+ "model.layers.33.self_attn.v_proj.biases": "model.safetensors",
738
+ "model.layers.33.self_attn.v_proj.scales": "model.safetensors",
739
+ "model.layers.33.self_attn.v_proj.weight": "model.safetensors",
740
+ "model.layers.34.input_layernorm.weight": "model.safetensors",
741
+ "model.layers.34.mlp.down_proj.biases": "model.safetensors",
742
+ "model.layers.34.mlp.down_proj.scales": "model.safetensors",
743
+ "model.layers.34.mlp.down_proj.weight": "model.safetensors",
744
+ "model.layers.34.mlp.gate_proj.biases": "model.safetensors",
745
+ "model.layers.34.mlp.gate_proj.scales": "model.safetensors",
746
+ "model.layers.34.mlp.gate_proj.weight": "model.safetensors",
747
+ "model.layers.34.mlp.up_proj.biases": "model.safetensors",
748
+ "model.layers.34.mlp.up_proj.scales": "model.safetensors",
749
+ "model.layers.34.mlp.up_proj.weight": "model.safetensors",
750
+ "model.layers.34.post_attention_layernorm.weight": "model.safetensors",
751
+ "model.layers.34.self_attn.k_proj.bias": "model.safetensors",
752
+ "model.layers.34.self_attn.k_proj.biases": "model.safetensors",
753
+ "model.layers.34.self_attn.k_proj.scales": "model.safetensors",
754
+ "model.layers.34.self_attn.k_proj.weight": "model.safetensors",
755
+ "model.layers.34.self_attn.o_proj.biases": "model.safetensors",
756
+ "model.layers.34.self_attn.o_proj.scales": "model.safetensors",
757
+ "model.layers.34.self_attn.o_proj.weight": "model.safetensors",
758
+ "model.layers.34.self_attn.q_proj.bias": "model.safetensors",
759
+ "model.layers.34.self_attn.q_proj.biases": "model.safetensors",
760
+ "model.layers.34.self_attn.q_proj.scales": "model.safetensors",
761
+ "model.layers.34.self_attn.q_proj.weight": "model.safetensors",
762
+ "model.layers.34.self_attn.v_proj.bias": "model.safetensors",
763
+ "model.layers.34.self_attn.v_proj.biases": "model.safetensors",
764
+ "model.layers.34.self_attn.v_proj.scales": "model.safetensors",
765
+ "model.layers.34.self_attn.v_proj.weight": "model.safetensors",
766
+ "model.layers.35.input_layernorm.weight": "model.safetensors",
767
+ "model.layers.35.mlp.down_proj.biases": "model.safetensors",
768
+ "model.layers.35.mlp.down_proj.scales": "model.safetensors",
769
+ "model.layers.35.mlp.down_proj.weight": "model.safetensors",
770
+ "model.layers.35.mlp.gate_proj.biases": "model.safetensors",
771
+ "model.layers.35.mlp.gate_proj.scales": "model.safetensors",
772
+ "model.layers.35.mlp.gate_proj.weight": "model.safetensors",
773
+ "model.layers.35.mlp.up_proj.biases": "model.safetensors",
774
+ "model.layers.35.mlp.up_proj.scales": "model.safetensors",
775
+ "model.layers.35.mlp.up_proj.weight": "model.safetensors",
776
+ "model.layers.35.post_attention_layernorm.weight": "model.safetensors",
777
+ "model.layers.35.self_attn.k_proj.bias": "model.safetensors",
778
+ "model.layers.35.self_attn.k_proj.biases": "model.safetensors",
779
+ "model.layers.35.self_attn.k_proj.scales": "model.safetensors",
780
+ "model.layers.35.self_attn.k_proj.weight": "model.safetensors",
781
+ "model.layers.35.self_attn.o_proj.biases": "model.safetensors",
782
+ "model.layers.35.self_attn.o_proj.scales": "model.safetensors",
783
+ "model.layers.35.self_attn.o_proj.weight": "model.safetensors",
784
+ "model.layers.35.self_attn.q_proj.bias": "model.safetensors",
785
+ "model.layers.35.self_attn.q_proj.biases": "model.safetensors",
786
+ "model.layers.35.self_attn.q_proj.scales": "model.safetensors",
787
+ "model.layers.35.self_attn.q_proj.weight": "model.safetensors",
788
+ "model.layers.35.self_attn.v_proj.bias": "model.safetensors",
789
+ "model.layers.35.self_attn.v_proj.biases": "model.safetensors",
790
+ "model.layers.35.self_attn.v_proj.scales": "model.safetensors",
791
+ "model.layers.35.self_attn.v_proj.weight": "model.safetensors",
792
+ "model.layers.36.input_layernorm.weight": "model.safetensors",
793
+ "model.layers.36.mlp.down_proj.biases": "model.safetensors",
794
+ "model.layers.36.mlp.down_proj.scales": "model.safetensors",
795
+ "model.layers.36.mlp.down_proj.weight": "model.safetensors",
796
+ "model.layers.36.mlp.gate_proj.biases": "model.safetensors",
797
+ "model.layers.36.mlp.gate_proj.scales": "model.safetensors",
798
+ "model.layers.36.mlp.gate_proj.weight": "model.safetensors",
799
+ "model.layers.36.mlp.up_proj.biases": "model.safetensors",
800
+ "model.layers.36.mlp.up_proj.scales": "model.safetensors",
801
+ "model.layers.36.mlp.up_proj.weight": "model.safetensors",
802
+ "model.layers.36.post_attention_layernorm.weight": "model.safetensors",
803
+ "model.layers.36.self_attn.k_proj.bias": "model.safetensors",
804
+ "model.layers.36.self_attn.k_proj.biases": "model.safetensors",
805
+ "model.layers.36.self_attn.k_proj.scales": "model.safetensors",
806
+ "model.layers.36.self_attn.k_proj.weight": "model.safetensors",
807
+ "model.layers.36.self_attn.o_proj.biases": "model.safetensors",
808
+ "model.layers.36.self_attn.o_proj.scales": "model.safetensors",
809
+ "model.layers.36.self_attn.o_proj.weight": "model.safetensors",
810
+ "model.layers.36.self_attn.q_proj.bias": "model.safetensors",
811
+ "model.layers.36.self_attn.q_proj.biases": "model.safetensors",
812
+ "model.layers.36.self_attn.q_proj.scales": "model.safetensors",
813
+ "model.layers.36.self_attn.q_proj.weight": "model.safetensors",
814
+ "model.layers.36.self_attn.v_proj.bias": "model.safetensors",
815
+ "model.layers.36.self_attn.v_proj.biases": "model.safetensors",
816
+ "model.layers.36.self_attn.v_proj.scales": "model.safetensors",
817
+ "model.layers.36.self_attn.v_proj.weight": "model.safetensors",
818
+ "model.layers.37.input_layernorm.weight": "model.safetensors",
819
+ "model.layers.37.mlp.down_proj.biases": "model.safetensors",
820
+ "model.layers.37.mlp.down_proj.scales": "model.safetensors",
821
+ "model.layers.37.mlp.down_proj.weight": "model.safetensors",
822
+ "model.layers.37.mlp.gate_proj.biases": "model.safetensors",
823
+ "model.layers.37.mlp.gate_proj.scales": "model.safetensors",
824
+ "model.layers.37.mlp.gate_proj.weight": "model.safetensors",
825
+ "model.layers.37.mlp.up_proj.biases": "model.safetensors",
826
+ "model.layers.37.mlp.up_proj.scales": "model.safetensors",
827
+ "model.layers.37.mlp.up_proj.weight": "model.safetensors",
828
+ "model.layers.37.post_attention_layernorm.weight": "model.safetensors",
829
+ "model.layers.37.self_attn.k_proj.bias": "model.safetensors",
830
+ "model.layers.37.self_attn.k_proj.biases": "model.safetensors",
831
+ "model.layers.37.self_attn.k_proj.scales": "model.safetensors",
832
+ "model.layers.37.self_attn.k_proj.weight": "model.safetensors",
833
+ "model.layers.37.self_attn.o_proj.biases": "model.safetensors",
834
+ "model.layers.37.self_attn.o_proj.scales": "model.safetensors",
835
+ "model.layers.37.self_attn.o_proj.weight": "model.safetensors",
836
+ "model.layers.37.self_attn.q_proj.bias": "model.safetensors",
837
+ "model.layers.37.self_attn.q_proj.biases": "model.safetensors",
838
+ "model.layers.37.self_attn.q_proj.scales": "model.safetensors",
839
+ "model.layers.37.self_attn.q_proj.weight": "model.safetensors",
840
+ "model.layers.37.self_attn.v_proj.bias": "model.safetensors",
841
+ "model.layers.37.self_attn.v_proj.biases": "model.safetensors",
842
+ "model.layers.37.self_attn.v_proj.scales": "model.safetensors",
843
+ "model.layers.37.self_attn.v_proj.weight": "model.safetensors",
844
+ "model.layers.38.input_layernorm.weight": "model.safetensors",
845
+ "model.layers.38.mlp.down_proj.biases": "model.safetensors",
846
+ "model.layers.38.mlp.down_proj.scales": "model.safetensors",
847
+ "model.layers.38.mlp.down_proj.weight": "model.safetensors",
848
+ "model.layers.38.mlp.gate_proj.biases": "model.safetensors",
849
+ "model.layers.38.mlp.gate_proj.scales": "model.safetensors",
850
+ "model.layers.38.mlp.gate_proj.weight": "model.safetensors",
851
+ "model.layers.38.mlp.up_proj.biases": "model.safetensors",
852
+ "model.layers.38.mlp.up_proj.scales": "model.safetensors",
853
+ "model.layers.38.mlp.up_proj.weight": "model.safetensors",
854
+ "model.layers.38.post_attention_layernorm.weight": "model.safetensors",
855
+ "model.layers.38.self_attn.k_proj.bias": "model.safetensors",
856
+ "model.layers.38.self_attn.k_proj.biases": "model.safetensors",
857
+ "model.layers.38.self_attn.k_proj.scales": "model.safetensors",
858
+ "model.layers.38.self_attn.k_proj.weight": "model.safetensors",
859
+ "model.layers.38.self_attn.o_proj.biases": "model.safetensors",
860
+ "model.layers.38.self_attn.o_proj.scales": "model.safetensors",
861
+ "model.layers.38.self_attn.o_proj.weight": "model.safetensors",
862
+ "model.layers.38.self_attn.q_proj.bias": "model.safetensors",
863
+ "model.layers.38.self_attn.q_proj.biases": "model.safetensors",
864
+ "model.layers.38.self_attn.q_proj.scales": "model.safetensors",
865
+ "model.layers.38.self_attn.q_proj.weight": "model.safetensors",
866
+ "model.layers.38.self_attn.v_proj.bias": "model.safetensors",
867
+ "model.layers.38.self_attn.v_proj.biases": "model.safetensors",
868
+ "model.layers.38.self_attn.v_proj.scales": "model.safetensors",
869
+ "model.layers.38.self_attn.v_proj.weight": "model.safetensors",
870
+ "model.layers.39.input_layernorm.weight": "model.safetensors",
871
+ "model.layers.39.mlp.down_proj.biases": "model.safetensors",
872
+ "model.layers.39.mlp.down_proj.scales": "model.safetensors",
873
+ "model.layers.39.mlp.down_proj.weight": "model.safetensors",
874
+ "model.layers.39.mlp.gate_proj.biases": "model.safetensors",
875
+ "model.layers.39.mlp.gate_proj.scales": "model.safetensors",
876
+ "model.layers.39.mlp.gate_proj.weight": "model.safetensors",
877
+ "model.layers.39.mlp.up_proj.biases": "model.safetensors",
878
+ "model.layers.39.mlp.up_proj.scales": "model.safetensors",
879
+ "model.layers.39.mlp.up_proj.weight": "model.safetensors",
880
+ "model.layers.39.post_attention_layernorm.weight": "model.safetensors",
881
+ "model.layers.39.self_attn.k_proj.bias": "model.safetensors",
882
+ "model.layers.39.self_attn.k_proj.biases": "model.safetensors",
883
+ "model.layers.39.self_attn.k_proj.scales": "model.safetensors",
884
+ "model.layers.39.self_attn.k_proj.weight": "model.safetensors",
885
+ "model.layers.39.self_attn.o_proj.biases": "model.safetensors",
886
+ "model.layers.39.self_attn.o_proj.scales": "model.safetensors",
887
+ "model.layers.39.self_attn.o_proj.weight": "model.safetensors",
888
+ "model.layers.39.self_attn.q_proj.bias": "model.safetensors",
889
+ "model.layers.39.self_attn.q_proj.biases": "model.safetensors",
890
+ "model.layers.39.self_attn.q_proj.scales": "model.safetensors",
891
+ "model.layers.39.self_attn.q_proj.weight": "model.safetensors",
892
+ "model.layers.39.self_attn.v_proj.bias": "model.safetensors",
893
+ "model.layers.39.self_attn.v_proj.biases": "model.safetensors",
894
+ "model.layers.39.self_attn.v_proj.scales": "model.safetensors",
895
+ "model.layers.39.self_attn.v_proj.weight": "model.safetensors",
896
+ "model.layers.4.input_layernorm.weight": "model.safetensors",
897
+ "model.layers.4.mlp.down_proj.biases": "model.safetensors",
898
+ "model.layers.4.mlp.down_proj.scales": "model.safetensors",
899
+ "model.layers.4.mlp.down_proj.weight": "model.safetensors",
900
+ "model.layers.4.mlp.gate_proj.biases": "model.safetensors",
901
+ "model.layers.4.mlp.gate_proj.scales": "model.safetensors",
902
+ "model.layers.4.mlp.gate_proj.weight": "model.safetensors",
903
+ "model.layers.4.mlp.up_proj.biases": "model.safetensors",
904
+ "model.layers.4.mlp.up_proj.scales": "model.safetensors",
905
+ "model.layers.4.mlp.up_proj.weight": "model.safetensors",
906
+ "model.layers.4.post_attention_layernorm.weight": "model.safetensors",
907
+ "model.layers.4.self_attn.k_proj.bias": "model.safetensors",
908
+ "model.layers.4.self_attn.k_proj.biases": "model.safetensors",
909
+ "model.layers.4.self_attn.k_proj.scales": "model.safetensors",
910
+ "model.layers.4.self_attn.k_proj.weight": "model.safetensors",
911
+ "model.layers.4.self_attn.o_proj.biases": "model.safetensors",
912
+ "model.layers.4.self_attn.o_proj.scales": "model.safetensors",
913
+ "model.layers.4.self_attn.o_proj.weight": "model.safetensors",
914
+ "model.layers.4.self_attn.q_proj.bias": "model.safetensors",
915
+ "model.layers.4.self_attn.q_proj.biases": "model.safetensors",
916
+ "model.layers.4.self_attn.q_proj.scales": "model.safetensors",
917
+ "model.layers.4.self_attn.q_proj.weight": "model.safetensors",
918
+ "model.layers.4.self_attn.v_proj.bias": "model.safetensors",
919
+ "model.layers.4.self_attn.v_proj.biases": "model.safetensors",
920
+ "model.layers.4.self_attn.v_proj.scales": "model.safetensors",
921
+ "model.layers.4.self_attn.v_proj.weight": "model.safetensors",
922
+ "model.layers.40.input_layernorm.weight": "model.safetensors",
923
+ "model.layers.40.mlp.down_proj.biases": "model.safetensors",
924
+ "model.layers.40.mlp.down_proj.scales": "model.safetensors",
925
+ "model.layers.40.mlp.down_proj.weight": "model.safetensors",
926
+ "model.layers.40.mlp.gate_proj.biases": "model.safetensors",
927
+ "model.layers.40.mlp.gate_proj.scales": "model.safetensors",
928
+ "model.layers.40.mlp.gate_proj.weight": "model.safetensors",
929
+ "model.layers.40.mlp.up_proj.biases": "model.safetensors",
930
+ "model.layers.40.mlp.up_proj.scales": "model.safetensors",
931
+ "model.layers.40.mlp.up_proj.weight": "model.safetensors",
932
+ "model.layers.40.post_attention_layernorm.weight": "model.safetensors",
933
+ "model.layers.40.self_attn.k_proj.bias": "model.safetensors",
934
+ "model.layers.40.self_attn.k_proj.biases": "model.safetensors",
935
+ "model.layers.40.self_attn.k_proj.scales": "model.safetensors",
936
+ "model.layers.40.self_attn.k_proj.weight": "model.safetensors",
937
+ "model.layers.40.self_attn.o_proj.biases": "model.safetensors",
938
+ "model.layers.40.self_attn.o_proj.scales": "model.safetensors",
939
+ "model.layers.40.self_attn.o_proj.weight": "model.safetensors",
940
+ "model.layers.40.self_attn.q_proj.bias": "model.safetensors",
941
+ "model.layers.40.self_attn.q_proj.biases": "model.safetensors",
942
+ "model.layers.40.self_attn.q_proj.scales": "model.safetensors",
943
+ "model.layers.40.self_attn.q_proj.weight": "model.safetensors",
944
+ "model.layers.40.self_attn.v_proj.bias": "model.safetensors",
945
+ "model.layers.40.self_attn.v_proj.biases": "model.safetensors",
946
+ "model.layers.40.self_attn.v_proj.scales": "model.safetensors",
947
+ "model.layers.40.self_attn.v_proj.weight": "model.safetensors",
948
+ "model.layers.41.input_layernorm.weight": "model.safetensors",
949
+ "model.layers.41.mlp.down_proj.biases": "model.safetensors",
950
+ "model.layers.41.mlp.down_proj.scales": "model.safetensors",
951
+ "model.layers.41.mlp.down_proj.weight": "model.safetensors",
952
+ "model.layers.41.mlp.gate_proj.biases": "model.safetensors",
953
+ "model.layers.41.mlp.gate_proj.scales": "model.safetensors",
954
+ "model.layers.41.mlp.gate_proj.weight": "model.safetensors",
955
+ "model.layers.41.mlp.up_proj.biases": "model.safetensors",
956
+ "model.layers.41.mlp.up_proj.scales": "model.safetensors",
957
+ "model.layers.41.mlp.up_proj.weight": "model.safetensors",
958
+ "model.layers.41.post_attention_layernorm.weight": "model.safetensors",
959
+ "model.layers.41.self_attn.k_proj.bias": "model.safetensors",
960
+ "model.layers.41.self_attn.k_proj.biases": "model.safetensors",
961
+ "model.layers.41.self_attn.k_proj.scales": "model.safetensors",
962
+ "model.layers.41.self_attn.k_proj.weight": "model.safetensors",
963
+ "model.layers.41.self_attn.o_proj.biases": "model.safetensors",
964
+ "model.layers.41.self_attn.o_proj.scales": "model.safetensors",
965
+ "model.layers.41.self_attn.o_proj.weight": "model.safetensors",
966
+ "model.layers.41.self_attn.q_proj.bias": "model.safetensors",
967
+ "model.layers.41.self_attn.q_proj.biases": "model.safetensors",
968
+ "model.layers.41.self_attn.q_proj.scales": "model.safetensors",
969
+ "model.layers.41.self_attn.q_proj.weight": "model.safetensors",
970
+ "model.layers.41.self_attn.v_proj.bias": "model.safetensors",
971
+ "model.layers.41.self_attn.v_proj.biases": "model.safetensors",
972
+ "model.layers.41.self_attn.v_proj.scales": "model.safetensors",
973
+ "model.layers.41.self_attn.v_proj.weight": "model.safetensors",
974
+ "model.layers.42.input_layernorm.weight": "model.safetensors",
975
+ "model.layers.42.mlp.down_proj.biases": "model.safetensors",
976
+ "model.layers.42.mlp.down_proj.scales": "model.safetensors",
977
+ "model.layers.42.mlp.down_proj.weight": "model.safetensors",
978
+ "model.layers.42.mlp.gate_proj.biases": "model.safetensors",
979
+ "model.layers.42.mlp.gate_proj.scales": "model.safetensors",
980
+ "model.layers.42.mlp.gate_proj.weight": "model.safetensors",
981
+ "model.layers.42.mlp.up_proj.biases": "model.safetensors",
982
+ "model.layers.42.mlp.up_proj.scales": "model.safetensors",
983
+ "model.layers.42.mlp.up_proj.weight": "model.safetensors",
984
+ "model.layers.42.post_attention_layernorm.weight": "model.safetensors",
985
+ "model.layers.42.self_attn.k_proj.bias": "model.safetensors",
986
+ "model.layers.42.self_attn.k_proj.biases": "model.safetensors",
987
+ "model.layers.42.self_attn.k_proj.scales": "model.safetensors",
988
+ "model.layers.42.self_attn.k_proj.weight": "model.safetensors",
989
+ "model.layers.42.self_attn.o_proj.biases": "model.safetensors",
990
+ "model.layers.42.self_attn.o_proj.scales": "model.safetensors",
991
+ "model.layers.42.self_attn.o_proj.weight": "model.safetensors",
992
+ "model.layers.42.self_attn.q_proj.bias": "model.safetensors",
993
+ "model.layers.42.self_attn.q_proj.biases": "model.safetensors",
994
+ "model.layers.42.self_attn.q_proj.scales": "model.safetensors",
995
+ "model.layers.42.self_attn.q_proj.weight": "model.safetensors",
996
+ "model.layers.42.self_attn.v_proj.bias": "model.safetensors",
997
+ "model.layers.42.self_attn.v_proj.biases": "model.safetensors",
998
+ "model.layers.42.self_attn.v_proj.scales": "model.safetensors",
999
+ "model.layers.42.self_attn.v_proj.weight": "model.safetensors",
1000
+ "model.layers.43.input_layernorm.weight": "model.safetensors",
1001
+ "model.layers.43.mlp.down_proj.biases": "model.safetensors",
1002
+ "model.layers.43.mlp.down_proj.scales": "model.safetensors",
1003
+ "model.layers.43.mlp.down_proj.weight": "model.safetensors",
1004
+ "model.layers.43.mlp.gate_proj.biases": "model.safetensors",
1005
+ "model.layers.43.mlp.gate_proj.scales": "model.safetensors",
1006
+ "model.layers.43.mlp.gate_proj.weight": "model.safetensors",
1007
+ "model.layers.43.mlp.up_proj.biases": "model.safetensors",
1008
+ "model.layers.43.mlp.up_proj.scales": "model.safetensors",
1009
+ "model.layers.43.mlp.up_proj.weight": "model.safetensors",
1010
+ "model.layers.43.post_attention_layernorm.weight": "model.safetensors",
1011
+ "model.layers.43.self_attn.k_proj.bias": "model.safetensors",
1012
+ "model.layers.43.self_attn.k_proj.biases": "model.safetensors",
1013
+ "model.layers.43.self_attn.k_proj.scales": "model.safetensors",
1014
+ "model.layers.43.self_attn.k_proj.weight": "model.safetensors",
1015
+ "model.layers.43.self_attn.o_proj.biases": "model.safetensors",
1016
+ "model.layers.43.self_attn.o_proj.scales": "model.safetensors",
1017
+ "model.layers.43.self_attn.o_proj.weight": "model.safetensors",
1018
+ "model.layers.43.self_attn.q_proj.bias": "model.safetensors",
1019
+ "model.layers.43.self_attn.q_proj.biases": "model.safetensors",
1020
+ "model.layers.43.self_attn.q_proj.scales": "model.safetensors",
1021
+ "model.layers.43.self_attn.q_proj.weight": "model.safetensors",
1022
+ "model.layers.43.self_attn.v_proj.bias": "model.safetensors",
1023
+ "model.layers.43.self_attn.v_proj.biases": "model.safetensors",
1024
+ "model.layers.43.self_attn.v_proj.scales": "model.safetensors",
1025
+ "model.layers.43.self_attn.v_proj.weight": "model.safetensors",
1026
+ "model.layers.44.input_layernorm.weight": "model.safetensors",
1027
+ "model.layers.44.mlp.down_proj.biases": "model.safetensors",
1028
+ "model.layers.44.mlp.down_proj.scales": "model.safetensors",
1029
+ "model.layers.44.mlp.down_proj.weight": "model.safetensors",
1030
+ "model.layers.44.mlp.gate_proj.biases": "model.safetensors",
1031
+ "model.layers.44.mlp.gate_proj.scales": "model.safetensors",
1032
+ "model.layers.44.mlp.gate_proj.weight": "model.safetensors",
1033
+ "model.layers.44.mlp.up_proj.biases": "model.safetensors",
1034
+ "model.layers.44.mlp.up_proj.scales": "model.safetensors",
1035
+ "model.layers.44.mlp.up_proj.weight": "model.safetensors",
1036
+ "model.layers.44.post_attention_layernorm.weight": "model.safetensors",
1037
+ "model.layers.44.self_attn.k_proj.bias": "model.safetensors",
1038
+ "model.layers.44.self_attn.k_proj.biases": "model.safetensors",
1039
+ "model.layers.44.self_attn.k_proj.scales": "model.safetensors",
1040
+ "model.layers.44.self_attn.k_proj.weight": "model.safetensors",
1041
+ "model.layers.44.self_attn.o_proj.biases": "model.safetensors",
1042
+ "model.layers.44.self_attn.o_proj.scales": "model.safetensors",
1043
+ "model.layers.44.self_attn.o_proj.weight": "model.safetensors",
1044
+ "model.layers.44.self_attn.q_proj.bias": "model.safetensors",
1045
+ "model.layers.44.self_attn.q_proj.biases": "model.safetensors",
1046
+ "model.layers.44.self_attn.q_proj.scales": "model.safetensors",
1047
+ "model.layers.44.self_attn.q_proj.weight": "model.safetensors",
1048
+ "model.layers.44.self_attn.v_proj.bias": "model.safetensors",
1049
+ "model.layers.44.self_attn.v_proj.biases": "model.safetensors",
1050
+ "model.layers.44.self_attn.v_proj.scales": "model.safetensors",
1051
+ "model.layers.44.self_attn.v_proj.weight": "model.safetensors",
1052
+ "model.layers.45.input_layernorm.weight": "model.safetensors",
1053
+ "model.layers.45.mlp.down_proj.biases": "model.safetensors",
1054
+ "model.layers.45.mlp.down_proj.scales": "model.safetensors",
1055
+ "model.layers.45.mlp.down_proj.weight": "model.safetensors",
1056
+ "model.layers.45.mlp.gate_proj.biases": "model.safetensors",
1057
+ "model.layers.45.mlp.gate_proj.scales": "model.safetensors",
1058
+ "model.layers.45.mlp.gate_proj.weight": "model.safetensors",
1059
+ "model.layers.45.mlp.up_proj.biases": "model.safetensors",
1060
+ "model.layers.45.mlp.up_proj.scales": "model.safetensors",
1061
+ "model.layers.45.mlp.up_proj.weight": "model.safetensors",
1062
+ "model.layers.45.post_attention_layernorm.weight": "model.safetensors",
1063
+ "model.layers.45.self_attn.k_proj.bias": "model.safetensors",
1064
+ "model.layers.45.self_attn.k_proj.biases": "model.safetensors",
1065
+ "model.layers.45.self_attn.k_proj.scales": "model.safetensors",
1066
+ "model.layers.45.self_attn.k_proj.weight": "model.safetensors",
1067
+ "model.layers.45.self_attn.o_proj.biases": "model.safetensors",
1068
+ "model.layers.45.self_attn.o_proj.scales": "model.safetensors",
1069
+ "model.layers.45.self_attn.o_proj.weight": "model.safetensors",
1070
+ "model.layers.45.self_attn.q_proj.bias": "model.safetensors",
1071
+ "model.layers.45.self_attn.q_proj.biases": "model.safetensors",
1072
+ "model.layers.45.self_attn.q_proj.scales": "model.safetensors",
1073
+ "model.layers.45.self_attn.q_proj.weight": "model.safetensors",
1074
+ "model.layers.45.self_attn.v_proj.bias": "model.safetensors",
1075
+ "model.layers.45.self_attn.v_proj.biases": "model.safetensors",
1076
+ "model.layers.45.self_attn.v_proj.scales": "model.safetensors",
1077
+ "model.layers.45.self_attn.v_proj.weight": "model.safetensors",
1078
+ "model.layers.46.input_layernorm.weight": "model.safetensors",
1079
+ "model.layers.46.mlp.down_proj.biases": "model.safetensors",
1080
+ "model.layers.46.mlp.down_proj.scales": "model.safetensors",
1081
+ "model.layers.46.mlp.down_proj.weight": "model.safetensors",
1082
+ "model.layers.46.mlp.gate_proj.biases": "model.safetensors",
1083
+ "model.layers.46.mlp.gate_proj.scales": "model.safetensors",
1084
+ "model.layers.46.mlp.gate_proj.weight": "model.safetensors",
1085
+ "model.layers.46.mlp.up_proj.biases": "model.safetensors",
1086
+ "model.layers.46.mlp.up_proj.scales": "model.safetensors",
1087
+ "model.layers.46.mlp.up_proj.weight": "model.safetensors",
1088
+ "model.layers.46.post_attention_layernorm.weight": "model.safetensors",
1089
+ "model.layers.46.self_attn.k_proj.bias": "model.safetensors",
1090
+ "model.layers.46.self_attn.k_proj.biases": "model.safetensors",
1091
+ "model.layers.46.self_attn.k_proj.scales": "model.safetensors",
1092
+ "model.layers.46.self_attn.k_proj.weight": "model.safetensors",
1093
+ "model.layers.46.self_attn.o_proj.biases": "model.safetensors",
1094
+ "model.layers.46.self_attn.o_proj.scales": "model.safetensors",
1095
+ "model.layers.46.self_attn.o_proj.weight": "model.safetensors",
1096
+ "model.layers.46.self_attn.q_proj.bias": "model.safetensors",
1097
+ "model.layers.46.self_attn.q_proj.biases": "model.safetensors",
1098
+ "model.layers.46.self_attn.q_proj.scales": "model.safetensors",
1099
+ "model.layers.46.self_attn.q_proj.weight": "model.safetensors",
1100
+ "model.layers.46.self_attn.v_proj.bias": "model.safetensors",
1101
+ "model.layers.46.self_attn.v_proj.biases": "model.safetensors",
1102
+ "model.layers.46.self_attn.v_proj.scales": "model.safetensors",
1103
+ "model.layers.46.self_attn.v_proj.weight": "model.safetensors",
1104
+ "model.layers.47.input_layernorm.weight": "model.safetensors",
1105
+ "model.layers.47.mlp.down_proj.biases": "model.safetensors",
1106
+ "model.layers.47.mlp.down_proj.scales": "model.safetensors",
1107
+ "model.layers.47.mlp.down_proj.weight": "model.safetensors",
1108
+ "model.layers.47.mlp.gate_proj.biases": "model.safetensors",
1109
+ "model.layers.47.mlp.gate_proj.scales": "model.safetensors",
1110
+ "model.layers.47.mlp.gate_proj.weight": "model.safetensors",
1111
+ "model.layers.47.mlp.up_proj.biases": "model.safetensors",
1112
+ "model.layers.47.mlp.up_proj.scales": "model.safetensors",
1113
+ "model.layers.47.mlp.up_proj.weight": "model.safetensors",
1114
+ "model.layers.47.post_attention_layernorm.weight": "model.safetensors",
1115
+ "model.layers.47.self_attn.k_proj.bias": "model.safetensors",
1116
+ "model.layers.47.self_attn.k_proj.biases": "model.safetensors",
1117
+ "model.layers.47.self_attn.k_proj.scales": "model.safetensors",
1118
+ "model.layers.47.self_attn.k_proj.weight": "model.safetensors",
1119
+ "model.layers.47.self_attn.o_proj.biases": "model.safetensors",
1120
+ "model.layers.47.self_attn.o_proj.scales": "model.safetensors",
1121
+ "model.layers.47.self_attn.o_proj.weight": "model.safetensors",
1122
+ "model.layers.47.self_attn.q_proj.bias": "model.safetensors",
1123
+ "model.layers.47.self_attn.q_proj.biases": "model.safetensors",
1124
+ "model.layers.47.self_attn.q_proj.scales": "model.safetensors",
1125
+ "model.layers.47.self_attn.q_proj.weight": "model.safetensors",
1126
+ "model.layers.47.self_attn.v_proj.bias": "model.safetensors",
1127
+ "model.layers.47.self_attn.v_proj.biases": "model.safetensors",
1128
+ "model.layers.47.self_attn.v_proj.scales": "model.safetensors",
1129
+ "model.layers.47.self_attn.v_proj.weight": "model.safetensors",
1130
+ "model.layers.5.input_layernorm.weight": "model.safetensors",
1131
+ "model.layers.5.mlp.down_proj.biases": "model.safetensors",
1132
+ "model.layers.5.mlp.down_proj.scales": "model.safetensors",
1133
+ "model.layers.5.mlp.down_proj.weight": "model.safetensors",
1134
+ "model.layers.5.mlp.gate_proj.biases": "model.safetensors",
1135
+ "model.layers.5.mlp.gate_proj.scales": "model.safetensors",
1136
+ "model.layers.5.mlp.gate_proj.weight": "model.safetensors",
1137
+ "model.layers.5.mlp.up_proj.biases": "model.safetensors",
1138
+ "model.layers.5.mlp.up_proj.scales": "model.safetensors",
1139
+ "model.layers.5.mlp.up_proj.weight": "model.safetensors",
1140
+ "model.layers.5.post_attention_layernorm.weight": "model.safetensors",
1141
+ "model.layers.5.self_attn.k_proj.bias": "model.safetensors",
1142
+ "model.layers.5.self_attn.k_proj.biases": "model.safetensors",
1143
+ "model.layers.5.self_attn.k_proj.scales": "model.safetensors",
1144
+ "model.layers.5.self_attn.k_proj.weight": "model.safetensors",
1145
+ "model.layers.5.self_attn.o_proj.biases": "model.safetensors",
1146
+ "model.layers.5.self_attn.o_proj.scales": "model.safetensors",
1147
+ "model.layers.5.self_attn.o_proj.weight": "model.safetensors",
1148
+ "model.layers.5.self_attn.q_proj.bias": "model.safetensors",
1149
+ "model.layers.5.self_attn.q_proj.biases": "model.safetensors",
1150
+ "model.layers.5.self_attn.q_proj.scales": "model.safetensors",
1151
+ "model.layers.5.self_attn.q_proj.weight": "model.safetensors",
1152
+ "model.layers.5.self_attn.v_proj.bias": "model.safetensors",
1153
+ "model.layers.5.self_attn.v_proj.biases": "model.safetensors",
1154
+ "model.layers.5.self_attn.v_proj.scales": "model.safetensors",
1155
+ "model.layers.5.self_attn.v_proj.weight": "model.safetensors",
1156
+ "model.layers.6.input_layernorm.weight": "model.safetensors",
1157
+ "model.layers.6.mlp.down_proj.biases": "model.safetensors",
1158
+ "model.layers.6.mlp.down_proj.scales": "model.safetensors",
1159
+ "model.layers.6.mlp.down_proj.weight": "model.safetensors",
1160
+ "model.layers.6.mlp.gate_proj.biases": "model.safetensors",
1161
+ "model.layers.6.mlp.gate_proj.scales": "model.safetensors",
1162
+ "model.layers.6.mlp.gate_proj.weight": "model.safetensors",
1163
+ "model.layers.6.mlp.up_proj.biases": "model.safetensors",
1164
+ "model.layers.6.mlp.up_proj.scales": "model.safetensors",
1165
+ "model.layers.6.mlp.up_proj.weight": "model.safetensors",
1166
+ "model.layers.6.post_attention_layernorm.weight": "model.safetensors",
1167
+ "model.layers.6.self_attn.k_proj.bias": "model.safetensors",
1168
+ "model.layers.6.self_attn.k_proj.biases": "model.safetensors",
1169
+ "model.layers.6.self_attn.k_proj.scales": "model.safetensors",
1170
+ "model.layers.6.self_attn.k_proj.weight": "model.safetensors",
1171
+ "model.layers.6.self_attn.o_proj.biases": "model.safetensors",
1172
+ "model.layers.6.self_attn.o_proj.scales": "model.safetensors",
1173
+ "model.layers.6.self_attn.o_proj.weight": "model.safetensors",
1174
+ "model.layers.6.self_attn.q_proj.bias": "model.safetensors",
1175
+ "model.layers.6.self_attn.q_proj.biases": "model.safetensors",
1176
+ "model.layers.6.self_attn.q_proj.scales": "model.safetensors",
1177
+ "model.layers.6.self_attn.q_proj.weight": "model.safetensors",
1178
+ "model.layers.6.self_attn.v_proj.bias": "model.safetensors",
1179
+ "model.layers.6.self_attn.v_proj.biases": "model.safetensors",
1180
+ "model.layers.6.self_attn.v_proj.scales": "model.safetensors",
1181
+ "model.layers.6.self_attn.v_proj.weight": "model.safetensors",
1182
+ "model.layers.7.input_layernorm.weight": "model.safetensors",
1183
+ "model.layers.7.mlp.down_proj.biases": "model.safetensors",
1184
+ "model.layers.7.mlp.down_proj.scales": "model.safetensors",
1185
+ "model.layers.7.mlp.down_proj.weight": "model.safetensors",
1186
+ "model.layers.7.mlp.gate_proj.biases": "model.safetensors",
1187
+ "model.layers.7.mlp.gate_proj.scales": "model.safetensors",
1188
+ "model.layers.7.mlp.gate_proj.weight": "model.safetensors",
1189
+ "model.layers.7.mlp.up_proj.biases": "model.safetensors",
1190
+ "model.layers.7.mlp.up_proj.scales": "model.safetensors",
1191
+ "model.layers.7.mlp.up_proj.weight": "model.safetensors",
1192
+ "model.layers.7.post_attention_layernorm.weight": "model.safetensors",
1193
+ "model.layers.7.self_attn.k_proj.bias": "model.safetensors",
1194
+ "model.layers.7.self_attn.k_proj.biases": "model.safetensors",
1195
+ "model.layers.7.self_attn.k_proj.scales": "model.safetensors",
1196
+ "model.layers.7.self_attn.k_proj.weight": "model.safetensors",
1197
+ "model.layers.7.self_attn.o_proj.biases": "model.safetensors",
1198
+ "model.layers.7.self_attn.o_proj.scales": "model.safetensors",
1199
+ "model.layers.7.self_attn.o_proj.weight": "model.safetensors",
1200
+ "model.layers.7.self_attn.q_proj.bias": "model.safetensors",
1201
+ "model.layers.7.self_attn.q_proj.biases": "model.safetensors",
1202
+ "model.layers.7.self_attn.q_proj.scales": "model.safetensors",
1203
+ "model.layers.7.self_attn.q_proj.weight": "model.safetensors",
1204
+ "model.layers.7.self_attn.v_proj.bias": "model.safetensors",
1205
+ "model.layers.7.self_attn.v_proj.biases": "model.safetensors",
1206
+ "model.layers.7.self_attn.v_proj.scales": "model.safetensors",
1207
+ "model.layers.7.self_attn.v_proj.weight": "model.safetensors",
1208
+ "model.layers.8.input_layernorm.weight": "model.safetensors",
1209
+ "model.layers.8.mlp.down_proj.biases": "model.safetensors",
1210
+ "model.layers.8.mlp.down_proj.scales": "model.safetensors",
1211
+ "model.layers.8.mlp.down_proj.weight": "model.safetensors",
1212
+ "model.layers.8.mlp.gate_proj.biases": "model.safetensors",
1213
+ "model.layers.8.mlp.gate_proj.scales": "model.safetensors",
1214
+ "model.layers.8.mlp.gate_proj.weight": "model.safetensors",
1215
+ "model.layers.8.mlp.up_proj.biases": "model.safetensors",
1216
+ "model.layers.8.mlp.up_proj.scales": "model.safetensors",
1217
+ "model.layers.8.mlp.up_proj.weight": "model.safetensors",
1218
+ "model.layers.8.post_attention_layernorm.weight": "model.safetensors",
1219
+ "model.layers.8.self_attn.k_proj.bias": "model.safetensors",
1220
+ "model.layers.8.self_attn.k_proj.biases": "model.safetensors",
1221
+ "model.layers.8.self_attn.k_proj.scales": "model.safetensors",
1222
+ "model.layers.8.self_attn.k_proj.weight": "model.safetensors",
1223
+ "model.layers.8.self_attn.o_proj.biases": "model.safetensors",
1224
+ "model.layers.8.self_attn.o_proj.scales": "model.safetensors",
1225
+ "model.layers.8.self_attn.o_proj.weight": "model.safetensors",
1226
+ "model.layers.8.self_attn.q_proj.bias": "model.safetensors",
1227
+ "model.layers.8.self_attn.q_proj.biases": "model.safetensors",
1228
+ "model.layers.8.self_attn.q_proj.scales": "model.safetensors",
1229
+ "model.layers.8.self_attn.q_proj.weight": "model.safetensors",
1230
+ "model.layers.8.self_attn.v_proj.bias": "model.safetensors",
1231
+ "model.layers.8.self_attn.v_proj.biases": "model.safetensors",
1232
+ "model.layers.8.self_attn.v_proj.scales": "model.safetensors",
1233
+ "model.layers.8.self_attn.v_proj.weight": "model.safetensors",
1234
+ "model.layers.9.input_layernorm.weight": "model.safetensors",
1235
+ "model.layers.9.mlp.down_proj.biases": "model.safetensors",
1236
+ "model.layers.9.mlp.down_proj.scales": "model.safetensors",
1237
+ "model.layers.9.mlp.down_proj.weight": "model.safetensors",
1238
+ "model.layers.9.mlp.gate_proj.biases": "model.safetensors",
1239
+ "model.layers.9.mlp.gate_proj.scales": "model.safetensors",
1240
+ "model.layers.9.mlp.gate_proj.weight": "model.safetensors",
1241
+ "model.layers.9.mlp.up_proj.biases": "model.safetensors",
1242
+ "model.layers.9.mlp.up_proj.scales": "model.safetensors",
1243
+ "model.layers.9.mlp.up_proj.weight": "model.safetensors",
1244
+ "model.layers.9.post_attention_layernorm.weight": "model.safetensors",
1245
+ "model.layers.9.self_attn.k_proj.bias": "model.safetensors",
1246
+ "model.layers.9.self_attn.k_proj.biases": "model.safetensors",
1247
+ "model.layers.9.self_attn.k_proj.scales": "model.safetensors",
1248
+ "model.layers.9.self_attn.k_proj.weight": "model.safetensors",
1249
+ "model.layers.9.self_attn.o_proj.biases": "model.safetensors",
1250
+ "model.layers.9.self_attn.o_proj.scales": "model.safetensors",
1251
+ "model.layers.9.self_attn.o_proj.weight": "model.safetensors",
1252
+ "model.layers.9.self_attn.q_proj.bias": "model.safetensors",
1253
+ "model.layers.9.self_attn.q_proj.biases": "model.safetensors",
1254
+ "model.layers.9.self_attn.q_proj.scales": "model.safetensors",
1255
+ "model.layers.9.self_attn.q_proj.weight": "model.safetensors",
1256
+ "model.layers.9.self_attn.v_proj.bias": "model.safetensors",
1257
+ "model.layers.9.self_attn.v_proj.biases": "model.safetensors",
1258
+ "model.layers.9.self_attn.v_proj.scales": "model.safetensors",
1259
+ "model.layers.9.self_attn.v_proj.weight": "model.safetensors",
1260
+ "model.norm.weight": "model.safetensors"
1261
+ }
1262
+ }
DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/special_tokens_map.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<｜begin▁of▁sentence｜>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<｜end▁of▁sentence｜>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<｜end▁of▁sentence｜>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ }
23
+ }
DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e20ddafc659ba90242154b55275402edeca0715e5dbb30f56815a4ce081f4893
3
+ size 11422778
DeepSeek-R1-Distill-Qwen-14B-2,6_mixed/tokenizer_config.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "151643": {
7
+ "content": "<｜end▁of▁sentence｜>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "151644": {
15
+ "content": "<｜User｜>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": false
21
+ },
22
+ "151645": {
23
+ "content": "<｜Assistant｜>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": false
29
+ },
30
+ "151646": {
31
+ "content": "<｜begin▁of▁sentence｜>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false,
36
+ "special": true
37
+ },
38
+ "151647": {
39
+ "content": "<|EOT|>",
40
+ "lstrip": false,
41
+ "normalized": false,
42
+ "rstrip": false,
43
+ "single_word": false,
44
+ "special": false
45
+ },
46
+ "151648": {
47
+ "content": "<think>",
48
+ "lstrip": false,
49
+ "normalized": false,
50
+ "rstrip": false,
51
+ "single_word": false,
52
+ "special": false
53
+ },
54
+ "151649": {
55
+ "content": "</think>",
56
+ "lstrip": false,
57
+ "normalized": false,
58
+ "rstrip": false,
59
+ "single_word": false,
60
+ "special": false
61
+ },
62
+ "151650": {
63
+ "content": "<|quad_start|>",
64
+ "lstrip": false,
65
+ "normalized": false,
66
+ "rstrip": false,
67
+ "single_word": false,
68
+ "special": true
69
+ },
70
+ "151651": {
71
+ "content": "<|quad_end|>",
72
+ "lstrip": false,
73
+ "normalized": false,
74
+ "rstrip": false,
75
+ "single_word": false,
76
+ "special": true
77
+ },
78
+ "151652": {
79
+ "content": "<|vision_start|>",
80
+ "lstrip": false,
81
+ "normalized": false,
82
+ "rstrip": false,
83
+ "single_word": false,
84
+ "special": true
85
+ },
86
+ "151653": {
87
+ "content": "<|vision_end|>",
88
+ "lstrip": false,
89
+ "normalized": false,
90
+ "rstrip": false,
91
+ "single_word": false,
92
+ "special": true
93
+ },
94
+ "151654": {
95
+ "content": "<|vision_pad|>",
96
+ "lstrip": false,
97
+ "normalized": false,
98
+ "rstrip": false,
99
+ "single_word": false,
100
+ "special": true
101
+ },
102
+ "151655": {
103
+ "content": "<|image_pad|>",
104
+ "lstrip": false,
105
+ "normalized": false,
106
+ "rstrip": false,
107
+ "single_word": false,
108
+ "special": true
109
+ },
110
+ "151656": {
111
+ "content": "<|video_pad|>",
112
+ "lstrip": false,
113
+ "normalized": false,
114
+ "rstrip": false,
115
+ "single_word": false,
116
+ "special": true
117
+ },
118
+ "151657": {
119
+ "content": "<tool_call>",
120
+ "lstrip": false,
121
+ "normalized": false,
122
+ "rstrip": false,
123
+ "single_word": false,
124
+ "special": false
125
+ },
126
+ "151658": {
127
+ "content": "</tool_call>",
128
+ "lstrip": false,
129
+ "normalized": false,
130
+ "rstrip": false,
131
+ "single_word": false,
132
+ "special": false
133
+ },
134
+ "151659": {
135
+ "content": "<|fim_prefix|>",
136
+ "lstrip": false,
137
+ "normalized": false,
138
+ "rstrip": false,
139
+ "single_word": false,
140
+ "special": false
141
+ },
142
+ "151660": {
143
+ "content": "<|fim_middle|>",
144
+ "lstrip": false,
145
+ "normalized": false,
146
+ "rstrip": false,
147
+ "single_word": false,
148
+ "special": false
149
+ },
150
+ "151661": {
151
+ "content": "<|fim_suffix|>",
152
+ "lstrip": false,
153
+ "normalized": false,
154
+ "rstrip": false,
155
+ "single_word": false,
156
+ "special": false
157
+ },
158
+ "151662": {
159
+ "content": "<|fim_pad|>",
160
+ "lstrip": false,
161
+ "normalized": false,
162
+ "rstrip": false,
163
+ "single_word": false,
164
+ "special": false
165
+ },
166
+ "151663": {
167
+ "content": "<|repo_name|>",
168
+ "lstrip": false,
169
+ "normalized": false,
170
+ "rstrip": false,
171
+ "single_word": false,
172
+ "special": false
173
+ },
174
+ "151664": {
175
+ "content": "<|file_sep|>",
176
+ "lstrip": false,
177
+ "normalized": false,
178
+ "rstrip": false,
179
+ "single_word": false,
180
+ "special": false
181
+ }
182
+ },
183
+ "bos_token": "<｜begin▁of▁sentence｜>",
184
+ "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='') %}{%- for message in messages %}{%- if message['role'] == 'system' %}{% set ns.system_prompt = message['content'] %}{%- endif %}{%- endfor %}{{bos_token}}{{ns.system_prompt}}{%- for message in messages %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{{'<｜User｜>' + message['content']}}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is none %}{%- set ns.is_tool = false -%}{%- for tool in message['tool_calls']%}{%- if not ns.is_first %}{{'<｜Assistant｜><｜tool▁calls▁begin｜><｜tool▁call▁begin��>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{%- set ns.is_first = true -%}{%- else %}{{'\\n' + '<｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{{'<｜tool▁calls▁end｜><｜end▁of▁sentence｜>'}}{%- endif %}{%- endfor %}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is not none %}{%- if ns.is_tool %}{{'<｜tool▁outputs▁end｜>' + message['content'] + '<｜end▁of▁sentence｜>'}}{%- set ns.is_tool = false -%}{%- else %}{% set content = message['content'] %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{% endif %}{{'<｜Assistant｜>' + content + '<｜end▁of▁sentence｜>'}}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<｜tool▁outputs▁begin｜><｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- set ns.is_output_first = false %}{%- else %}{{'\\n<｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<｜tool▁outputs▁end｜>'}}{% endif %}{% if add_generation_prompt and not ns.is_tool %}{{'<｜Assistant｜><think>\\n'}}{% endif %}",
185
+ "clean_up_tokenization_spaces": false,
186
+ "eos_token": "<｜end▁of▁sentence｜>",
187
+ "extra_special_tokens": {},
188
+ "legacy": true,
189
+ "model_max_length": 16384,
190
+ "pad_token": "<｜end▁of▁sentence｜>",
191
+ "sp_model_kwargs": {},
192
+ "tokenizer_class": "LlamaTokenizerFast",
193
+ "unk_token": null,
194
+ "use_default_system_prompt": false
195
+ }
DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/config.json ADDED
@@ -0,0 +1,3029 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151643,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 5120,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 13824,
12
+ "max_position_embeddings": 131072,
13
+ "max_window_layers": 48,
14
+ "model_type": "qwen2",
15
+ "num_attention_heads": 40,
16
+ "num_hidden_layers": 48,
17
+ "num_key_value_heads": 8,
18
+ "quantization": {
19
+ "group_size": 64,
20
+ "bits": null,
21
+ "model.embed_tokens": {
22
+ "group_size": 64,
23
+ "bits": 3
24
+ },
25
+ "model.layers.0.self_attn.q_proj": {
26
+ "group_size": 64,
27
+ "bits": 3
28
+ },
29
+ "model.layers.0.self_attn.k_proj": {
30
+ "group_size": 64,
31
+ "bits": 3
32
+ },
33
+ "model.layers.0.self_attn.v_proj": {
34
+ "group_size": 64,
35
+ "bits": 4
36
+ },
37
+ "model.layers.0.self_attn.o_proj": {
38
+ "group_size": 64,
39
+ "bits": 3
40
+ },
41
+ "model.layers.0.self_attn.rope": false,
42
+ "model.layers.0.mlp.gate_proj": {
43
+ "group_size": 64,
44
+ "bits": 3
45
+ },
46
+ "model.layers.0.mlp.down_proj": {
47
+ "group_size": 64,
48
+ "bits": 4
49
+ },
50
+ "model.layers.0.mlp.up_proj": {
51
+ "group_size": 64,
52
+ "bits": 3
53
+ },
54
+ "model.layers.0.input_layernorm": false,
55
+ "model.layers.0.post_attention_layernorm": false,
56
+ "model.layers.1.self_attn.q_proj": {
57
+ "group_size": 64,
58
+ "bits": 3
59
+ },
60
+ "model.layers.1.self_attn.k_proj": {
61
+ "group_size": 64,
62
+ "bits": 3
63
+ },
64
+ "model.layers.1.self_attn.v_proj": {
65
+ "group_size": 64,
66
+ "bits": 4
67
+ },
68
+ "model.layers.1.self_attn.o_proj": {
69
+ "group_size": 64,
70
+ "bits": 3
71
+ },
72
+ "model.layers.1.self_attn.rope": false,
73
+ "model.layers.1.mlp.gate_proj": {
74
+ "group_size": 64,
75
+ "bits": 3
76
+ },
77
+ "model.layers.1.mlp.down_proj": {
78
+ "group_size": 64,
79
+ "bits": 4
80
+ },
81
+ "model.layers.1.mlp.up_proj": {
82
+ "group_size": 64,
83
+ "bits": 3
84
+ },
85
+ "model.layers.1.input_layernorm": false,
86
+ "model.layers.1.post_attention_layernorm": false,
87
+ "model.layers.2.self_attn.q_proj": {
88
+ "group_size": 64,
89
+ "bits": 3
90
+ },
91
+ "model.layers.2.self_attn.k_proj": {
92
+ "group_size": 64,
93
+ "bits": 3
94
+ },
95
+ "model.layers.2.self_attn.v_proj": {
96
+ "group_size": 64,
97
+ "bits": 4
98
+ },
99
+ "model.layers.2.self_attn.o_proj": {
100
+ "group_size": 64,
101
+ "bits": 3
102
+ },
103
+ "model.layers.2.self_attn.rope": false,
104
+ "model.layers.2.mlp.gate_proj": {
105
+ "group_size": 64,
106
+ "bits": 3
107
+ },
108
+ "model.layers.2.mlp.down_proj": {
109
+ "group_size": 64,
110
+ "bits": 4
111
+ },
112
+ "model.layers.2.mlp.up_proj": {
113
+ "group_size": 64,
114
+ "bits": 3
115
+ },
116
+ "model.layers.2.input_layernorm": false,
117
+ "model.layers.2.post_attention_layernorm": false,
118
+ "model.layers.3.self_attn.q_proj": {
119
+ "group_size": 64,
120
+ "bits": 3
121
+ },
122
+ "model.layers.3.self_attn.k_proj": {
123
+ "group_size": 64,
124
+ "bits": 3
125
+ },
126
+ "model.layers.3.self_attn.v_proj": {
127
+ "group_size": 64,
128
+ "bits": 4
129
+ },
130
+ "model.layers.3.self_attn.o_proj": {
131
+ "group_size": 64,
132
+ "bits": 3
133
+ },
134
+ "model.layers.3.self_attn.rope": false,
135
+ "model.layers.3.mlp.gate_proj": {
136
+ "group_size": 64,
137
+ "bits": 3
138
+ },
139
+ "model.layers.3.mlp.down_proj": {
140
+ "group_size": 64,
141
+ "bits": 4
142
+ },
143
+ "model.layers.3.mlp.up_proj": {
144
+ "group_size": 64,
145
+ "bits": 3
146
+ },
147
+ "model.layers.3.input_layernorm": false,
148
+ "model.layers.3.post_attention_layernorm": false,
149
+ "model.layers.4.self_attn.q_proj": {
150
+ "group_size": 64,
151
+ "bits": 3
152
+ },
153
+ "model.layers.4.self_attn.k_proj": {
154
+ "group_size": 64,
155
+ "bits": 3
156
+ },
157
+ "model.layers.4.self_attn.v_proj": {
158
+ "group_size": 64,
159
+ "bits": 4
160
+ },
161
+ "model.layers.4.self_attn.o_proj": {
162
+ "group_size": 64,
163
+ "bits": 3
164
+ },
165
+ "model.layers.4.self_attn.rope": false,
166
+ "model.layers.4.mlp.gate_proj": {
167
+ "group_size": 64,
168
+ "bits": 3
169
+ },
170
+ "model.layers.4.mlp.down_proj": {
171
+ "group_size": 64,
172
+ "bits": 4
173
+ },
174
+ "model.layers.4.mlp.up_proj": {
175
+ "group_size": 64,
176
+ "bits": 3
177
+ },
178
+ "model.layers.4.input_layernorm": false,
179
+ "model.layers.4.post_attention_layernorm": false,
180
+ "model.layers.5.self_attn.q_proj": {
181
+ "group_size": 64,
182
+ "bits": 3
183
+ },
184
+ "model.layers.5.self_attn.k_proj": {
185
+ "group_size": 64,
186
+ "bits": 3
187
+ },
188
+ "model.layers.5.self_attn.v_proj": {
189
+ "group_size": 64,
190
+ "bits": 4
191
+ },
192
+ "model.layers.5.self_attn.o_proj": {
193
+ "group_size": 64,
194
+ "bits": 3
195
+ },
196
+ "model.layers.5.self_attn.rope": false,
197
+ "model.layers.5.mlp.gate_proj": {
198
+ "group_size": 64,
199
+ "bits": 3
200
+ },
201
+ "model.layers.5.mlp.down_proj": {
202
+ "group_size": 64,
203
+ "bits": 4
204
+ },
205
+ "model.layers.5.mlp.up_proj": {
206
+ "group_size": 64,
207
+ "bits": 3
208
+ },
209
+ "model.layers.5.input_layernorm": false,
210
+ "model.layers.5.post_attention_layernorm": false,
211
+ "model.layers.6.self_attn.q_proj": {
212
+ "group_size": 64,
213
+ "bits": 3
214
+ },
215
+ "model.layers.6.self_attn.k_proj": {
216
+ "group_size": 64,
217
+ "bits": 3
218
+ },
219
+ "model.layers.6.self_attn.v_proj": {
220
+ "group_size": 64,
221
+ "bits": 3
222
+ },
223
+ "model.layers.6.self_attn.o_proj": {
224
+ "group_size": 64,
225
+ "bits": 3
226
+ },
227
+ "model.layers.6.self_attn.rope": false,
228
+ "model.layers.6.mlp.gate_proj": {
229
+ "group_size": 64,
230
+ "bits": 3
231
+ },
232
+ "model.layers.6.mlp.down_proj": {
233
+ "group_size": 64,
234
+ "bits": 3
235
+ },
236
+ "model.layers.6.mlp.up_proj": {
237
+ "group_size": 64,
238
+ "bits": 3
239
+ },
240
+ "model.layers.6.input_layernorm": false,
241
+ "model.layers.6.post_attention_layernorm": false,
242
+ "model.layers.7.self_attn.q_proj": {
243
+ "group_size": 64,
244
+ "bits": 3
245
+ },
246
+ "model.layers.7.self_attn.k_proj": {
247
+ "group_size": 64,
248
+ "bits": 3
249
+ },
250
+ "model.layers.7.self_attn.v_proj": {
251
+ "group_size": 64,
252
+ "bits": 3
253
+ },
254
+ "model.layers.7.self_attn.o_proj": {
255
+ "group_size": 64,
256
+ "bits": 3
257
+ },
258
+ "model.layers.7.self_attn.rope": false,
259
+ "model.layers.7.mlp.gate_proj": {
260
+ "group_size": 64,
261
+ "bits": 3
262
+ },
263
+ "model.layers.7.mlp.down_proj": {
264
+ "group_size": 64,
265
+ "bits": 3
266
+ },
267
+ "model.layers.7.mlp.up_proj": {
268
+ "group_size": 64,
269
+ "bits": 3
270
+ },
271
+ "model.layers.7.input_layernorm": false,
272
+ "model.layers.7.post_attention_layernorm": false,
273
+ "model.layers.8.self_attn.q_proj": {
274
+ "group_size": 64,
275
+ "bits": 3
276
+ },
277
+ "model.layers.8.self_attn.k_proj": {
278
+ "group_size": 64,
279
+ "bits": 3
280
+ },
281
+ "model.layers.8.self_attn.v_proj": {
282
+ "group_size": 64,
283
+ "bits": 4
284
+ },
285
+ "model.layers.8.self_attn.o_proj": {
286
+ "group_size": 64,
287
+ "bits": 3
288
+ },
289
+ "model.layers.8.self_attn.rope": false,
290
+ "model.layers.8.mlp.gate_proj": {
291
+ "group_size": 64,
292
+ "bits": 3
293
+ },
294
+ "model.layers.8.mlp.down_proj": {
295
+ "group_size": 64,
296
+ "bits": 4
297
+ },
298
+ "model.layers.8.mlp.up_proj": {
299
+ "group_size": 64,
300
+ "bits": 3
301
+ },
302
+ "model.layers.8.input_layernorm": false,
303
+ "model.layers.8.post_attention_layernorm": false,
304
+ "model.layers.9.self_attn.q_proj": {
305
+ "group_size": 64,
306
+ "bits": 3
307
+ },
308
+ "model.layers.9.self_attn.k_proj": {
309
+ "group_size": 64,
310
+ "bits": 3
311
+ },
312
+ "model.layers.9.self_attn.v_proj": {
313
+ "group_size": 64,
314
+ "bits": 3
315
+ },
316
+ "model.layers.9.self_attn.o_proj": {
317
+ "group_size": 64,
318
+ "bits": 3
319
+ },
320
+ "model.layers.9.self_attn.rope": false,
321
+ "model.layers.9.mlp.gate_proj": {
322
+ "group_size": 64,
323
+ "bits": 3
324
+ },
325
+ "model.layers.9.mlp.down_proj": {
326
+ "group_size": 64,
327
+ "bits": 3
328
+ },
329
+ "model.layers.9.mlp.up_proj": {
330
+ "group_size": 64,
331
+ "bits": 3
332
+ },
333
+ "model.layers.9.input_layernorm": false,
334
+ "model.layers.9.post_attention_layernorm": false,
335
+ "model.layers.10.self_attn.q_proj": {
336
+ "group_size": 64,
337
+ "bits": 3
338
+ },
339
+ "model.layers.10.self_attn.k_proj": {
340
+ "group_size": 64,
341
+ "bits": 3
342
+ },
343
+ "model.layers.10.self_attn.v_proj": {
344
+ "group_size": 64,
345
+ "bits": 3
346
+ },
347
+ "model.layers.10.self_attn.o_proj": {
348
+ "group_size": 64,
349
+ "bits": 3
350
+ },
351
+ "model.layers.10.self_attn.rope": false,
352
+ "model.layers.10.mlp.gate_proj": {
353
+ "group_size": 64,
354
+ "bits": 3
355
+ },
356
+ "model.layers.10.mlp.down_proj": {
357
+ "group_size": 64,
358
+ "bits": 3
359
+ },
360
+ "model.layers.10.mlp.up_proj": {
361
+ "group_size": 64,
362
+ "bits": 3
363
+ },
364
+ "model.layers.10.input_layernorm": false,
365
+ "model.layers.10.post_attention_layernorm": false,
366
+ "model.layers.11.self_attn.q_proj": {
367
+ "group_size": 64,
368
+ "bits": 3
369
+ },
370
+ "model.layers.11.self_attn.k_proj": {
371
+ "group_size": 64,
372
+ "bits": 3
373
+ },
374
+ "model.layers.11.self_attn.v_proj": {
375
+ "group_size": 64,
376
+ "bits": 4
377
+ },
378
+ "model.layers.11.self_attn.o_proj": {
379
+ "group_size": 64,
380
+ "bits": 3
381
+ },
382
+ "model.layers.11.self_attn.rope": false,
383
+ "model.layers.11.mlp.gate_proj": {
384
+ "group_size": 64,
385
+ "bits": 3
386
+ },
387
+ "model.layers.11.mlp.down_proj": {
388
+ "group_size": 64,
389
+ "bits": 4
390
+ },
391
+ "model.layers.11.mlp.up_proj": {
392
+ "group_size": 64,
393
+ "bits": 3
394
+ },
395
+ "model.layers.11.input_layernorm": false,
396
+ "model.layers.11.post_attention_layernorm": false,
397
+ "model.layers.12.self_attn.q_proj": {
398
+ "group_size": 64,
399
+ "bits": 3
400
+ },
401
+ "model.layers.12.self_attn.k_proj": {
402
+ "group_size": 64,
403
+ "bits": 3
404
+ },
405
+ "model.layers.12.self_attn.v_proj": {
406
+ "group_size": 64,
407
+ "bits": 3
408
+ },
409
+ "model.layers.12.self_attn.o_proj": {
410
+ "group_size": 64,
411
+ "bits": 3
412
+ },
413
+ "model.layers.12.self_attn.rope": false,
414
+ "model.layers.12.mlp.gate_proj": {
415
+ "group_size": 64,
416
+ "bits": 3
417
+ },
418
+ "model.layers.12.mlp.down_proj": {
419
+ "group_size": 64,
420
+ "bits": 3
421
+ },
422
+ "model.layers.12.mlp.up_proj": {
423
+ "group_size": 64,
424
+ "bits": 3
425
+ },
426
+ "model.layers.12.input_layernorm": false,
427
+ "model.layers.12.post_attention_layernorm": false,
428
+ "model.layers.13.self_attn.q_proj": {
429
+ "group_size": 64,
430
+ "bits": 3
431
+ },
432
+ "model.layers.13.self_attn.k_proj": {
433
+ "group_size": 64,
434
+ "bits": 3
435
+ },
436
+ "model.layers.13.self_attn.v_proj": {
437
+ "group_size": 64,
438
+ "bits": 3
439
+ },
440
+ "model.layers.13.self_attn.o_proj": {
441
+ "group_size": 64,
442
+ "bits": 3
443
+ },
444
+ "model.layers.13.self_attn.rope": false,
445
+ "model.layers.13.mlp.gate_proj": {
446
+ "group_size": 64,
447
+ "bits": 3
448
+ },
449
+ "model.layers.13.mlp.down_proj": {
450
+ "group_size": 64,
451
+ "bits": 3
452
+ },
453
+ "model.layers.13.mlp.up_proj": {
454
+ "group_size": 64,
455
+ "bits": 3
456
+ },
457
+ "model.layers.13.input_layernorm": false,
458
+ "model.layers.13.post_attention_layernorm": false,
459
+ "model.layers.14.self_attn.q_proj": {
460
+ "group_size": 64,
461
+ "bits": 3
462
+ },
463
+ "model.layers.14.self_attn.k_proj": {
464
+ "group_size": 64,
465
+ "bits": 3
466
+ },
467
+ "model.layers.14.self_attn.v_proj": {
468
+ "group_size": 64,
469
+ "bits": 4
470
+ },
471
+ "model.layers.14.self_attn.o_proj": {
472
+ "group_size": 64,
473
+ "bits": 3
474
+ },
475
+ "model.layers.14.self_attn.rope": false,
476
+ "model.layers.14.mlp.gate_proj": {
477
+ "group_size": 64,
478
+ "bits": 3
479
+ },
480
+ "model.layers.14.mlp.down_proj": {
481
+ "group_size": 64,
482
+ "bits": 4
483
+ },
484
+ "model.layers.14.mlp.up_proj": {
485
+ "group_size": 64,
486
+ "bits": 3
487
+ },
488
+ "model.layers.14.input_layernorm": false,
489
+ "model.layers.14.post_attention_layernorm": false,
490
+ "model.layers.15.self_attn.q_proj": {
491
+ "group_size": 64,
492
+ "bits": 3
493
+ },
494
+ "model.layers.15.self_attn.k_proj": {
495
+ "group_size": 64,
496
+ "bits": 3
497
+ },
498
+ "model.layers.15.self_attn.v_proj": {
499
+ "group_size": 64,
500
+ "bits": 3
501
+ },
502
+ "model.layers.15.self_attn.o_proj": {
503
+ "group_size": 64,
504
+ "bits": 3
505
+ },
506
+ "model.layers.15.self_attn.rope": false,
507
+ "model.layers.15.mlp.gate_proj": {
508
+ "group_size": 64,
509
+ "bits": 3
510
+ },
511
+ "model.layers.15.mlp.down_proj": {
512
+ "group_size": 64,
513
+ "bits": 3
514
+ },
515
+ "model.layers.15.mlp.up_proj": {
516
+ "group_size": 64,
517
+ "bits": 3
518
+ },
519
+ "model.layers.15.input_layernorm": false,
520
+ "model.layers.15.post_attention_layernorm": false,
521
+ "model.layers.16.self_attn.q_proj": {
522
+ "group_size": 64,
523
+ "bits": 3
524
+ },
525
+ "model.layers.16.self_attn.k_proj": {
526
+ "group_size": 64,
527
+ "bits": 3
528
+ },
529
+ "model.layers.16.self_attn.v_proj": {
530
+ "group_size": 64,
531
+ "bits": 3
532
+ },
533
+ "model.layers.16.self_attn.o_proj": {
534
+ "group_size": 64,
535
+ "bits": 3
536
+ },
537
+ "model.layers.16.self_attn.rope": false,
538
+ "model.layers.16.mlp.gate_proj": {
539
+ "group_size": 64,
540
+ "bits": 3
541
+ },
542
+ "model.layers.16.mlp.down_proj": {
543
+ "group_size": 64,
544
+ "bits": 3
545
+ },
546
+ "model.layers.16.mlp.up_proj": {
547
+ "group_size": 64,
548
+ "bits": 3
549
+ },
550
+ "model.layers.16.input_layernorm": false,
551
+ "model.layers.16.post_attention_layernorm": false,
552
+ "model.layers.17.self_attn.q_proj": {
553
+ "group_size": 64,
554
+ "bits": 3
555
+ },
556
+ "model.layers.17.self_attn.k_proj": {
557
+ "group_size": 64,
558
+ "bits": 3
559
+ },
560
+ "model.layers.17.self_attn.v_proj": {
561
+ "group_size": 64,
562
+ "bits": 4
563
+ },
564
+ "model.layers.17.self_attn.o_proj": {
565
+ "group_size": 64,
566
+ "bits": 3
567
+ },
568
+ "model.layers.17.self_attn.rope": false,
569
+ "model.layers.17.mlp.gate_proj": {
570
+ "group_size": 64,
571
+ "bits": 3
572
+ },
573
+ "model.layers.17.mlp.down_proj": {
574
+ "group_size": 64,
575
+ "bits": 4
576
+ },
577
+ "model.layers.17.mlp.up_proj": {
578
+ "group_size": 64,
579
+ "bits": 3
580
+ },
581
+ "model.layers.17.input_layernorm": false,
582
+ "model.layers.17.post_attention_layernorm": false,
583
+ "model.layers.18.self_attn.q_proj": {
584
+ "group_size": 64,
585
+ "bits": 3
586
+ },
587
+ "model.layers.18.self_attn.k_proj": {
588
+ "group_size": 64,
589
+ "bits": 3
590
+ },
591
+ "model.layers.18.self_attn.v_proj": {
592
+ "group_size": 64,
593
+ "bits": 3
594
+ },
595
+ "model.layers.18.self_attn.o_proj": {
596
+ "group_size": 64,
597
+ "bits": 3
598
+ },
599
+ "model.layers.18.self_attn.rope": false,
600
+ "model.layers.18.mlp.gate_proj": {
601
+ "group_size": 64,
602
+ "bits": 3
603
+ },
604
+ "model.layers.18.mlp.down_proj": {
605
+ "group_size": 64,
606
+ "bits": 3
607
+ },
608
+ "model.layers.18.mlp.up_proj": {
609
+ "group_size": 64,
610
+ "bits": 3
611
+ },
612
+ "model.layers.18.input_layernorm": false,
613
+ "model.layers.18.post_attention_layernorm": false,
614
+ "model.layers.19.self_attn.q_proj": {
615
+ "group_size": 64,
616
+ "bits": 3
617
+ },
618
+ "model.layers.19.self_attn.k_proj": {
619
+ "group_size": 64,
620
+ "bits": 3
621
+ },
622
+ "model.layers.19.self_attn.v_proj": {
623
+ "group_size": 64,
624
+ "bits": 3
625
+ },
626
+ "model.layers.19.self_attn.o_proj": {
627
+ "group_size": 64,
628
+ "bits": 3
629
+ },
630
+ "model.layers.19.self_attn.rope": false,
631
+ "model.layers.19.mlp.gate_proj": {
632
+ "group_size": 64,
633
+ "bits": 3
634
+ },
635
+ "model.layers.19.mlp.down_proj": {
636
+ "group_size": 64,
637
+ "bits": 3
638
+ },
639
+ "model.layers.19.mlp.up_proj": {
640
+ "group_size": 64,
641
+ "bits": 3
642
+ },
643
+ "model.layers.19.input_layernorm": false,
644
+ "model.layers.19.post_attention_layernorm": false,
645
+ "model.layers.20.self_attn.q_proj": {
646
+ "group_size": 64,
647
+ "bits": 3
648
+ },
649
+ "model.layers.20.self_attn.k_proj": {
650
+ "group_size": 64,
651
+ "bits": 3
652
+ },
653
+ "model.layers.20.self_attn.v_proj": {
654
+ "group_size": 64,
655
+ "bits": 4
656
+ },
657
+ "model.layers.20.self_attn.o_proj": {
658
+ "group_size": 64,
659
+ "bits": 3
660
+ },
661
+ "model.layers.20.self_attn.rope": false,
662
+ "model.layers.20.mlp.gate_proj": {
663
+ "group_size": 64,
664
+ "bits": 3
665
+ },
666
+ "model.layers.20.mlp.down_proj": {
667
+ "group_size": 64,
668
+ "bits": 4
669
+ },
670
+ "model.layers.20.mlp.up_proj": {
671
+ "group_size": 64,
672
+ "bits": 3
673
+ },
674
+ "model.layers.20.input_layernorm": false,
675
+ "model.layers.20.post_attention_layernorm": false,
676
+ "model.layers.21.self_attn.q_proj": {
677
+ "group_size": 64,
678
+ "bits": 3
679
+ },
680
+ "model.layers.21.self_attn.k_proj": {
681
+ "group_size": 64,
682
+ "bits": 3
683
+ },
684
+ "model.layers.21.self_attn.v_proj": {
685
+ "group_size": 64,
686
+ "bits": 3
687
+ },
688
+ "model.layers.21.self_attn.o_proj": {
689
+ "group_size": 64,
690
+ "bits": 3
691
+ },
692
+ "model.layers.21.self_attn.rope": false,
693
+ "model.layers.21.mlp.gate_proj": {
694
+ "group_size": 64,
695
+ "bits": 3
696
+ },
697
+ "model.layers.21.mlp.down_proj": {
698
+ "group_size": 64,
699
+ "bits": 3
700
+ },
701
+ "model.layers.21.mlp.up_proj": {
702
+ "group_size": 64,
703
+ "bits": 3
704
+ },
705
+ "model.layers.21.input_layernorm": false,
706
+ "model.layers.21.post_attention_layernorm": false,
707
+ "model.layers.22.self_attn.q_proj": {
708
+ "group_size": 64,
709
+ "bits": 3
710
+ },
711
+ "model.layers.22.self_attn.k_proj": {
712
+ "group_size": 64,
713
+ "bits": 3
714
+ },
715
+ "model.layers.22.self_attn.v_proj": {
716
+ "group_size": 64,
717
+ "bits": 3
718
+ },
719
+ "model.layers.22.self_attn.o_proj": {
720
+ "group_size": 64,
721
+ "bits": 3
722
+ },
723
+ "model.layers.22.self_attn.rope": false,
724
+ "model.layers.22.mlp.gate_proj": {
725
+ "group_size": 64,
726
+ "bits": 3
727
+ },
728
+ "model.layers.22.mlp.down_proj": {
729
+ "group_size": 64,
730
+ "bits": 3
731
+ },
732
+ "model.layers.22.mlp.up_proj": {
733
+ "group_size": 64,
734
+ "bits": 3
735
+ },
736
+ "model.layers.22.input_layernorm": false,
737
+ "model.layers.22.post_attention_layernorm": false,
738
+ "model.layers.23.self_attn.q_proj": {
739
+ "group_size": 64,
740
+ "bits": 3
741
+ },
742
+ "model.layers.23.self_attn.k_proj": {
743
+ "group_size": 64,
744
+ "bits": 3
745
+ },
746
+ "model.layers.23.self_attn.v_proj": {
747
+ "group_size": 64,
748
+ "bits": 4
749
+ },
750
+ "model.layers.23.self_attn.o_proj": {
751
+ "group_size": 64,
752
+ "bits": 3
753
+ },
754
+ "model.layers.23.self_attn.rope": false,
755
+ "model.layers.23.mlp.gate_proj": {
756
+ "group_size": 64,
757
+ "bits": 3
758
+ },
759
+ "model.layers.23.mlp.down_proj": {
760
+ "group_size": 64,
761
+ "bits": 4
762
+ },
763
+ "model.layers.23.mlp.up_proj": {
764
+ "group_size": 64,
765
+ "bits": 3
766
+ },
767
+ "model.layers.23.input_layernorm": false,
768
+ "model.layers.23.post_attention_layernorm": false,
769
+ "model.layers.24.self_attn.q_proj": {
770
+ "group_size": 64,
771
+ "bits": 3
772
+ },
773
+ "model.layers.24.self_attn.k_proj": {
774
+ "group_size": 64,
775
+ "bits": 3
776
+ },
777
+ "model.layers.24.self_attn.v_proj": {
778
+ "group_size": 64,
779
+ "bits": 3
780
+ },
781
+ "model.layers.24.self_attn.o_proj": {
782
+ "group_size": 64,
783
+ "bits": 3
784
+ },
785
+ "model.layers.24.self_attn.rope": false,
786
+ "model.layers.24.mlp.gate_proj": {
787
+ "group_size": 64,
788
+ "bits": 3
789
+ },
790
+ "model.layers.24.mlp.down_proj": {
791
+ "group_size": 64,
792
+ "bits": 3
793
+ },
794
+ "model.layers.24.mlp.up_proj": {
795
+ "group_size": 64,
796
+ "bits": 3
797
+ },
798
+ "model.layers.24.input_layernorm": false,
799
+ "model.layers.24.post_attention_layernorm": false,
800
+ "model.layers.25.self_attn.q_proj": {
801
+ "group_size": 64,
802
+ "bits": 3
803
+ },
804
+ "model.layers.25.self_attn.k_proj": {
805
+ "group_size": 64,
806
+ "bits": 3
807
+ },
808
+ "model.layers.25.self_attn.v_proj": {
809
+ "group_size": 64,
810
+ "bits": 3
811
+ },
812
+ "model.layers.25.self_attn.o_proj": {
813
+ "group_size": 64,
814
+ "bits": 3
815
+ },
816
+ "model.layers.25.self_attn.rope": false,
817
+ "model.layers.25.mlp.gate_proj": {
818
+ "group_size": 64,
819
+ "bits": 3
820
+ },
821
+ "model.layers.25.mlp.down_proj": {
822
+ "group_size": 64,
823
+ "bits": 3
824
+ },
825
+ "model.layers.25.mlp.up_proj": {
826
+ "group_size": 64,
827
+ "bits": 3
828
+ },
829
+ "model.layers.25.input_layernorm": false,
830
+ "model.layers.25.post_attention_layernorm": false,
831
+ "model.layers.26.self_attn.q_proj": {
832
+ "group_size": 64,
833
+ "bits": 3
834
+ },
835
+ "model.layers.26.self_attn.k_proj": {
836
+ "group_size": 64,
837
+ "bits": 3
838
+ },
839
+ "model.layers.26.self_attn.v_proj": {
840
+ "group_size": 64,
841
+ "bits": 4
842
+ },
843
+ "model.layers.26.self_attn.o_proj": {
844
+ "group_size": 64,
845
+ "bits": 3
846
+ },
847
+ "model.layers.26.self_attn.rope": false,
848
+ "model.layers.26.mlp.gate_proj": {
849
+ "group_size": 64,
850
+ "bits": 3
851
+ },
852
+ "model.layers.26.mlp.down_proj": {
853
+ "group_size": 64,
854
+ "bits": 4
855
+ },
856
+ "model.layers.26.mlp.up_proj": {
857
+ "group_size": 64,
858
+ "bits": 3
859
+ },
860
+ "model.layers.26.input_layernorm": false,
861
+ "model.layers.26.post_attention_layernorm": false,
862
+ "model.layers.27.self_attn.q_proj": {
863
+ "group_size": 64,
864
+ "bits": 3
865
+ },
866
+ "model.layers.27.self_attn.k_proj": {
867
+ "group_size": 64,
868
+ "bits": 3
869
+ },
870
+ "model.layers.27.self_attn.v_proj": {
871
+ "group_size": 64,
872
+ "bits": 3
873
+ },
874
+ "model.layers.27.self_attn.o_proj": {
875
+ "group_size": 64,
876
+ "bits": 3
877
+ },
878
+ "model.layers.27.self_attn.rope": false,
879
+ "model.layers.27.mlp.gate_proj": {
880
+ "group_size": 64,
881
+ "bits": 3
882
+ },
883
+ "model.layers.27.mlp.down_proj": {
884
+ "group_size": 64,
885
+ "bits": 3
886
+ },
887
+ "model.layers.27.mlp.up_proj": {
888
+ "group_size": 64,
889
+ "bits": 3
890
+ },
891
+ "model.layers.27.input_layernorm": false,
892
+ "model.layers.27.post_attention_layernorm": false,
893
+ "model.layers.28.self_attn.q_proj": {
894
+ "group_size": 64,
895
+ "bits": 3
896
+ },
897
+ "model.layers.28.self_attn.k_proj": {
898
+ "group_size": 64,
899
+ "bits": 3
900
+ },
901
+ "model.layers.28.self_attn.v_proj": {
902
+ "group_size": 64,
903
+ "bits": 3
904
+ },
905
+ "model.layers.28.self_attn.o_proj": {
906
+ "group_size": 64,
907
+ "bits": 3
908
+ },
909
+ "model.layers.28.self_attn.rope": false,
910
+ "model.layers.28.mlp.gate_proj": {
911
+ "group_size": 64,
912
+ "bits": 3
913
+ },
914
+ "model.layers.28.mlp.down_proj": {
915
+ "group_size": 64,
916
+ "bits": 3
917
+ },
918
+ "model.layers.28.mlp.up_proj": {
919
+ "group_size": 64,
920
+ "bits": 3
921
+ },
922
+ "model.layers.28.input_layernorm": false,
923
+ "model.layers.28.post_attention_layernorm": false,
924
+ "model.layers.29.self_attn.q_proj": {
925
+ "group_size": 64,
926
+ "bits": 3
927
+ },
928
+ "model.layers.29.self_attn.k_proj": {
929
+ "group_size": 64,
930
+ "bits": 3
931
+ },
932
+ "model.layers.29.self_attn.v_proj": {
933
+ "group_size": 64,
934
+ "bits": 4
935
+ },
936
+ "model.layers.29.self_attn.o_proj": {
937
+ "group_size": 64,
938
+ "bits": 3
939
+ },
940
+ "model.layers.29.self_attn.rope": false,
941
+ "model.layers.29.mlp.gate_proj": {
942
+ "group_size": 64,
943
+ "bits": 3
944
+ },
945
+ "model.layers.29.mlp.down_proj": {
946
+ "group_size": 64,
947
+ "bits": 4
948
+ },
949
+ "model.layers.29.mlp.up_proj": {
950
+ "group_size": 64,
951
+ "bits": 3
952
+ },
953
+ "model.layers.29.input_layernorm": false,
954
+ "model.layers.29.post_attention_layernorm": false,
955
+ "model.layers.30.self_attn.q_proj": {
956
+ "group_size": 64,
957
+ "bits": 3
958
+ },
959
+ "model.layers.30.self_attn.k_proj": {
960
+ "group_size": 64,
961
+ "bits": 3
962
+ },
963
+ "model.layers.30.self_attn.v_proj": {
964
+ "group_size": 64,
965
+ "bits": 3
966
+ },
967
+ "model.layers.30.self_attn.o_proj": {
968
+ "group_size": 64,
969
+ "bits": 3
970
+ },
971
+ "model.layers.30.self_attn.rope": false,
972
+ "model.layers.30.mlp.gate_proj": {
973
+ "group_size": 64,
974
+ "bits": 3
975
+ },
976
+ "model.layers.30.mlp.down_proj": {
977
+ "group_size": 64,
978
+ "bits": 3
979
+ },
980
+ "model.layers.30.mlp.up_proj": {
981
+ "group_size": 64,
982
+ "bits": 3
983
+ },
984
+ "model.layers.30.input_layernorm": false,
985
+ "model.layers.30.post_attention_layernorm": false,
986
+ "model.layers.31.self_attn.q_proj": {
987
+ "group_size": 64,
988
+ "bits": 3
989
+ },
990
+ "model.layers.31.self_attn.k_proj": {
991
+ "group_size": 64,
992
+ "bits": 3
993
+ },
994
+ "model.layers.31.self_attn.v_proj": {
995
+ "group_size": 64,
996
+ "bits": 3
997
+ },
998
+ "model.layers.31.self_attn.o_proj": {
999
+ "group_size": 64,
1000
+ "bits": 3
1001
+ },
1002
+ "model.layers.31.self_attn.rope": false,
1003
+ "model.layers.31.mlp.gate_proj": {
1004
+ "group_size": 64,
1005
+ "bits": 3
1006
+ },
1007
+ "model.layers.31.mlp.down_proj": {
1008
+ "group_size": 64,
1009
+ "bits": 3
1010
+ },
1011
+ "model.layers.31.mlp.up_proj": {
1012
+ "group_size": 64,
1013
+ "bits": 3
1014
+ },
1015
+ "model.layers.31.input_layernorm": false,
1016
+ "model.layers.31.post_attention_layernorm": false,
1017
+ "model.layers.32.self_attn.q_proj": {
1018
+ "group_size": 64,
1019
+ "bits": 3
1020
+ },
1021
+ "model.layers.32.self_attn.k_proj": {
1022
+ "group_size": 64,
1023
+ "bits": 3
1024
+ },
1025
+ "model.layers.32.self_attn.v_proj": {
1026
+ "group_size": 64,
1027
+ "bits": 4
1028
+ },
1029
+ "model.layers.32.self_attn.o_proj": {
1030
+ "group_size": 64,
1031
+ "bits": 3
1032
+ },
1033
+ "model.layers.32.self_attn.rope": false,
1034
+ "model.layers.32.mlp.gate_proj": {
1035
+ "group_size": 64,
1036
+ "bits": 3
1037
+ },
1038
+ "model.layers.32.mlp.down_proj": {
1039
+ "group_size": 64,
1040
+ "bits": 4
1041
+ },
1042
+ "model.layers.32.mlp.up_proj": {
1043
+ "group_size": 64,
1044
+ "bits": 3
1045
+ },
1046
+ "model.layers.32.input_layernorm": false,
1047
+ "model.layers.32.post_attention_layernorm": false,
1048
+ "model.layers.33.self_attn.q_proj": {
1049
+ "group_size": 64,
1050
+ "bits": 3
1051
+ },
1052
+ "model.layers.33.self_attn.k_proj": {
1053
+ "group_size": 64,
1054
+ "bits": 3
1055
+ },
1056
+ "model.layers.33.self_attn.v_proj": {
1057
+ "group_size": 64,
1058
+ "bits": 3
1059
+ },
1060
+ "model.layers.33.self_attn.o_proj": {
1061
+ "group_size": 64,
1062
+ "bits": 3
1063
+ },
1064
+ "model.layers.33.self_attn.rope": false,
1065
+ "model.layers.33.mlp.gate_proj": {
1066
+ "group_size": 64,
1067
+ "bits": 3
1068
+ },
1069
+ "model.layers.33.mlp.down_proj": {
1070
+ "group_size": 64,
1071
+ "bits": 3
1072
+ },
1073
+ "model.layers.33.mlp.up_proj": {
1074
+ "group_size": 64,
1075
+ "bits": 3
1076
+ },
1077
+ "model.layers.33.input_layernorm": false,
1078
+ "model.layers.33.post_attention_layernorm": false,
1079
+ "model.layers.34.self_attn.q_proj": {
1080
+ "group_size": 64,
1081
+ "bits": 3
1082
+ },
1083
+ "model.layers.34.self_attn.k_proj": {
1084
+ "group_size": 64,
1085
+ "bits": 3
1086
+ },
1087
+ "model.layers.34.self_attn.v_proj": {
1088
+ "group_size": 64,
1089
+ "bits": 3
1090
+ },
1091
+ "model.layers.34.self_attn.o_proj": {
1092
+ "group_size": 64,
1093
+ "bits": 3
1094
+ },
1095
+ "model.layers.34.self_attn.rope": false,
1096
+ "model.layers.34.mlp.gate_proj": {
1097
+ "group_size": 64,
1098
+ "bits": 3
1099
+ },
1100
+ "model.layers.34.mlp.down_proj": {
1101
+ "group_size": 64,
1102
+ "bits": 3
1103
+ },
1104
+ "model.layers.34.mlp.up_proj": {
1105
+ "group_size": 64,
1106
+ "bits": 3
1107
+ },
1108
+ "model.layers.34.input_layernorm": false,
1109
+ "model.layers.34.post_attention_layernorm": false,
1110
+ "model.layers.35.self_attn.q_proj": {
1111
+ "group_size": 64,
1112
+ "bits": 3
1113
+ },
1114
+ "model.layers.35.self_attn.k_proj": {
1115
+ "group_size": 64,
1116
+ "bits": 3
1117
+ },
1118
+ "model.layers.35.self_attn.v_proj": {
1119
+ "group_size": 64,
1120
+ "bits": 4
1121
+ },
1122
+ "model.layers.35.self_attn.o_proj": {
1123
+ "group_size": 64,
1124
+ "bits": 3
1125
+ },
1126
+ "model.layers.35.self_attn.rope": false,
1127
+ "model.layers.35.mlp.gate_proj": {
1128
+ "group_size": 64,
1129
+ "bits": 3
1130
+ },
1131
+ "model.layers.35.mlp.down_proj": {
1132
+ "group_size": 64,
1133
+ "bits": 4
1134
+ },
1135
+ "model.layers.35.mlp.up_proj": {
1136
+ "group_size": 64,
1137
+ "bits": 3
1138
+ },
1139
+ "model.layers.35.input_layernorm": false,
1140
+ "model.layers.35.post_attention_layernorm": false,
1141
+ "model.layers.36.self_attn.q_proj": {
1142
+ "group_size": 64,
1143
+ "bits": 3
1144
+ },
1145
+ "model.layers.36.self_attn.k_proj": {
1146
+ "group_size": 64,
1147
+ "bits": 3
1148
+ },
1149
+ "model.layers.36.self_attn.v_proj": {
1150
+ "group_size": 64,
1151
+ "bits": 3
1152
+ },
1153
+ "model.layers.36.self_attn.o_proj": {
1154
+ "group_size": 64,
1155
+ "bits": 3
1156
+ },
1157
+ "model.layers.36.self_attn.rope": false,
1158
+ "model.layers.36.mlp.gate_proj": {
1159
+ "group_size": 64,
1160
+ "bits": 3
1161
+ },
1162
+ "model.layers.36.mlp.down_proj": {
1163
+ "group_size": 64,
1164
+ "bits": 3
1165
+ },
1166
+ "model.layers.36.mlp.up_proj": {
1167
+ "group_size": 64,
1168
+ "bits": 3
1169
+ },
1170
+ "model.layers.36.input_layernorm": false,
1171
+ "model.layers.36.post_attention_layernorm": false,
1172
+ "model.layers.37.self_attn.q_proj": {
1173
+ "group_size": 64,
1174
+ "bits": 3
1175
+ },
1176
+ "model.layers.37.self_attn.k_proj": {
1177
+ "group_size": 64,
1178
+ "bits": 3
1179
+ },
1180
+ "model.layers.37.self_attn.v_proj": {
1181
+ "group_size": 64,
1182
+ "bits": 3
1183
+ },
1184
+ "model.layers.37.self_attn.o_proj": {
1185
+ "group_size": 64,
1186
+ "bits": 3
1187
+ },
1188
+ "model.layers.37.self_attn.rope": false,
1189
+ "model.layers.37.mlp.gate_proj": {
1190
+ "group_size": 64,
1191
+ "bits": 3
1192
+ },
1193
+ "model.layers.37.mlp.down_proj": {
1194
+ "group_size": 64,
1195
+ "bits": 3
1196
+ },
1197
+ "model.layers.37.mlp.up_proj": {
1198
+ "group_size": 64,
1199
+ "bits": 3
1200
+ },
1201
+ "model.layers.37.input_layernorm": false,
1202
+ "model.layers.37.post_attention_layernorm": false,
1203
+ "model.layers.38.self_attn.q_proj": {
1204
+ "group_size": 64,
1205
+ "bits": 3
1206
+ },
1207
+ "model.layers.38.self_attn.k_proj": {
1208
+ "group_size": 64,
1209
+ "bits": 3
1210
+ },
1211
+ "model.layers.38.self_attn.v_proj": {
1212
+ "group_size": 64,
1213
+ "bits": 4
1214
+ },
1215
+ "model.layers.38.self_attn.o_proj": {
1216
+ "group_size": 64,
1217
+ "bits": 3
1218
+ },
1219
+ "model.layers.38.self_attn.rope": false,
1220
+ "model.layers.38.mlp.gate_proj": {
1221
+ "group_size": 64,
1222
+ "bits": 3
1223
+ },
1224
+ "model.layers.38.mlp.down_proj": {
1225
+ "group_size": 64,
1226
+ "bits": 4
1227
+ },
1228
+ "model.layers.38.mlp.up_proj": {
1229
+ "group_size": 64,
1230
+ "bits": 3
1231
+ },
1232
+ "model.layers.38.input_layernorm": false,
1233
+ "model.layers.38.post_attention_layernorm": false,
1234
+ "model.layers.39.self_attn.q_proj": {
1235
+ "group_size": 64,
1236
+ "bits": 3
1237
+ },
1238
+ "model.layers.39.self_attn.k_proj": {
1239
+ "group_size": 64,
1240
+ "bits": 3
1241
+ },
1242
+ "model.layers.39.self_attn.v_proj": {
1243
+ "group_size": 64,
1244
+ "bits": 3
1245
+ },
1246
+ "model.layers.39.self_attn.o_proj": {
1247
+ "group_size": 64,
1248
+ "bits": 3
1249
+ },
1250
+ "model.layers.39.self_attn.rope": false,
1251
+ "model.layers.39.mlp.gate_proj": {
1252
+ "group_size": 64,
1253
+ "bits": 3
1254
+ },
1255
+ "model.layers.39.mlp.down_proj": {
1256
+ "group_size": 64,
1257
+ "bits": 3
1258
+ },
1259
+ "model.layers.39.mlp.up_proj": {
1260
+ "group_size": 64,
1261
+ "bits": 3
1262
+ },
1263
+ "model.layers.39.input_layernorm": false,
1264
+ "model.layers.39.post_attention_layernorm": false,
1265
+ "model.layers.40.self_attn.q_proj": {
1266
+ "group_size": 64,
1267
+ "bits": 3
1268
+ },
1269
+ "model.layers.40.self_attn.k_proj": {
1270
+ "group_size": 64,
1271
+ "bits": 3
1272
+ },
1273
+ "model.layers.40.self_attn.v_proj": {
1274
+ "group_size": 64,
1275
+ "bits": 3
1276
+ },
1277
+ "model.layers.40.self_attn.o_proj": {
1278
+ "group_size": 64,
1279
+ "bits": 3
1280
+ },
1281
+ "model.layers.40.self_attn.rope": false,
1282
+ "model.layers.40.mlp.gate_proj": {
1283
+ "group_size": 64,
1284
+ "bits": 3
1285
+ },
1286
+ "model.layers.40.mlp.down_proj": {
1287
+ "group_size": 64,
1288
+ "bits": 3
1289
+ },
1290
+ "model.layers.40.mlp.up_proj": {
1291
+ "group_size": 64,
1292
+ "bits": 3
1293
+ },
1294
+ "model.layers.40.input_layernorm": false,
1295
+ "model.layers.40.post_attention_layernorm": false,
1296
+ "model.layers.41.self_attn.q_proj": {
1297
+ "group_size": 64,
1298
+ "bits": 3
1299
+ },
1300
+ "model.layers.41.self_attn.k_proj": {
1301
+ "group_size": 64,
1302
+ "bits": 3
1303
+ },
1304
+ "model.layers.41.self_attn.v_proj": {
1305
+ "group_size": 64,
1306
+ "bits": 4
1307
+ },
1308
+ "model.layers.41.self_attn.o_proj": {
1309
+ "group_size": 64,
1310
+ "bits": 3
1311
+ },
1312
+ "model.layers.41.self_attn.rope": false,
1313
+ "model.layers.41.mlp.gate_proj": {
1314
+ "group_size": 64,
1315
+ "bits": 3
1316
+ },
1317
+ "model.layers.41.mlp.down_proj": {
1318
+ "group_size": 64,
1319
+ "bits": 4
1320
+ },
1321
+ "model.layers.41.mlp.up_proj": {
1322
+ "group_size": 64,
1323
+ "bits": 3
1324
+ },
1325
+ "model.layers.41.input_layernorm": false,
1326
+ "model.layers.41.post_attention_layernorm": false,
1327
+ "model.layers.42.self_attn.q_proj": {
1328
+ "group_size": 64,
1329
+ "bits": 3
1330
+ },
1331
+ "model.layers.42.self_attn.k_proj": {
1332
+ "group_size": 64,
1333
+ "bits": 3
1334
+ },
1335
+ "model.layers.42.self_attn.v_proj": {
1336
+ "group_size": 64,
1337
+ "bits": 4
1338
+ },
1339
+ "model.layers.42.self_attn.o_proj": {
1340
+ "group_size": 64,
1341
+ "bits": 3
1342
+ },
1343
+ "model.layers.42.self_attn.rope": false,
1344
+ "model.layers.42.mlp.gate_proj": {
1345
+ "group_size": 64,
1346
+ "bits": 3
1347
+ },
1348
+ "model.layers.42.mlp.down_proj": {
1349
+ "group_size": 64,
1350
+ "bits": 4
1351
+ },
1352
+ "model.layers.42.mlp.up_proj": {
1353
+ "group_size": 64,
1354
+ "bits": 3
1355
+ },
1356
+ "model.layers.42.input_layernorm": false,
1357
+ "model.layers.42.post_attention_layernorm": false,
1358
+ "model.layers.43.self_attn.q_proj": {
1359
+ "group_size": 64,
1360
+ "bits": 3
1361
+ },
1362
+ "model.layers.43.self_attn.k_proj": {
1363
+ "group_size": 64,
1364
+ "bits": 3
1365
+ },
1366
+ "model.layers.43.self_attn.v_proj": {
1367
+ "group_size": 64,
1368
+ "bits": 4
1369
+ },
1370
+ "model.layers.43.self_attn.o_proj": {
1371
+ "group_size": 64,
1372
+ "bits": 3
1373
+ },
1374
+ "model.layers.43.self_attn.rope": false,
1375
+ "model.layers.43.mlp.gate_proj": {
1376
+ "group_size": 64,
1377
+ "bits": 3
1378
+ },
1379
+ "model.layers.43.mlp.down_proj": {
1380
+ "group_size": 64,
1381
+ "bits": 4
1382
+ },
1383
+ "model.layers.43.mlp.up_proj": {
1384
+ "group_size": 64,
1385
+ "bits": 3
1386
+ },
1387
+ "model.layers.43.input_layernorm": false,
1388
+ "model.layers.43.post_attention_layernorm": false,
1389
+ "model.layers.44.self_attn.q_proj": {
1390
+ "group_size": 64,
1391
+ "bits": 3
1392
+ },
1393
+ "model.layers.44.self_attn.k_proj": {
1394
+ "group_size": 64,
1395
+ "bits": 3
1396
+ },
1397
+ "model.layers.44.self_attn.v_proj": {
1398
+ "group_size": 64,
1399
+ "bits": 4
1400
+ },
1401
+ "model.layers.44.self_attn.o_proj": {
1402
+ "group_size": 64,
1403
+ "bits": 3
1404
+ },
1405
+ "model.layers.44.self_attn.rope": false,
1406
+ "model.layers.44.mlp.gate_proj": {
1407
+ "group_size": 64,
1408
+ "bits": 3
1409
+ },
1410
+ "model.layers.44.mlp.down_proj": {
1411
+ "group_size": 64,
1412
+ "bits": 4
1413
+ },
1414
+ "model.layers.44.mlp.up_proj": {
1415
+ "group_size": 64,
1416
+ "bits": 3
1417
+ },
1418
+ "model.layers.44.input_layernorm": false,
1419
+ "model.layers.44.post_attention_layernorm": false,
1420
+ "model.layers.45.self_attn.q_proj": {
1421
+ "group_size": 64,
1422
+ "bits": 3
1423
+ },
1424
+ "model.layers.45.self_attn.k_proj": {
1425
+ "group_size": 64,
1426
+ "bits": 3
1427
+ },
1428
+ "model.layers.45.self_attn.v_proj": {
1429
+ "group_size": 64,
1430
+ "bits": 4
1431
+ },
1432
+ "model.layers.45.self_attn.o_proj": {
1433
+ "group_size": 64,
1434
+ "bits": 3
1435
+ },
1436
+ "model.layers.45.self_attn.rope": false,
1437
+ "model.layers.45.mlp.gate_proj": {
1438
+ "group_size": 64,
1439
+ "bits": 3
1440
+ },
1441
+ "model.layers.45.mlp.down_proj": {
1442
+ "group_size": 64,
1443
+ "bits": 4
1444
+ },
1445
+ "model.layers.45.mlp.up_proj": {
1446
+ "group_size": 64,
1447
+ "bits": 3
1448
+ },
1449
+ "model.layers.45.input_layernorm": false,
1450
+ "model.layers.45.post_attention_layernorm": false,
1451
+ "model.layers.46.self_attn.q_proj": {
1452
+ "group_size": 64,
1453
+ "bits": 3
1454
+ },
1455
+ "model.layers.46.self_attn.k_proj": {
1456
+ "group_size": 64,
1457
+ "bits": 3
1458
+ },
1459
+ "model.layers.46.self_attn.v_proj": {
1460
+ "group_size": 64,
1461
+ "bits": 4
1462
+ },
1463
+ "model.layers.46.self_attn.o_proj": {
1464
+ "group_size": 64,
1465
+ "bits": 3
1466
+ },
1467
+ "model.layers.46.self_attn.rope": false,
1468
+ "model.layers.46.mlp.gate_proj": {
1469
+ "group_size": 64,
1470
+ "bits": 3
1471
+ },
1472
+ "model.layers.46.mlp.down_proj": {
1473
+ "group_size": 64,
1474
+ "bits": 4
1475
+ },
1476
+ "model.layers.46.mlp.up_proj": {
1477
+ "group_size": 64,
1478
+ "bits": 3
1479
+ },
1480
+ "model.layers.46.input_layernorm": false,
1481
+ "model.layers.46.post_attention_layernorm": false,
1482
+ "model.layers.47.self_attn.q_proj": {
1483
+ "group_size": 64,
1484
+ "bits": 3
1485
+ },
1486
+ "model.layers.47.self_attn.k_proj": {
1487
+ "group_size": 64,
1488
+ "bits": 3
1489
+ },
1490
+ "model.layers.47.self_attn.v_proj": {
1491
+ "group_size": 64,
1492
+ "bits": 4
1493
+ },
1494
+ "model.layers.47.self_attn.o_proj": {
1495
+ "group_size": 64,
1496
+ "bits": 3
1497
+ },
1498
+ "model.layers.47.self_attn.rope": false,
1499
+ "model.layers.47.mlp.gate_proj": {
1500
+ "group_size": 64,
1501
+ "bits": 3
1502
+ },
1503
+ "model.layers.47.mlp.down_proj": {
1504
+ "group_size": 64,
1505
+ "bits": 4
1506
+ },
1507
+ "model.layers.47.mlp.up_proj": {
1508
+ "group_size": 64,
1509
+ "bits": 3
1510
+ },
1511
+ "model.layers.47.input_layernorm": false,
1512
+ "model.layers.47.post_attention_layernorm": false,
1513
+ "model.norm": false,
1514
+ "lm_head": {
1515
+ "group_size": 64,
1516
+ "bits": 4
1517
+ }
1518
+ },
1519
+ "quantization_config": {
1520
+ "group_size": 64,
1521
+ "bits": null,
1522
+ "model.embed_tokens": {
1523
+ "group_size": 64,
1524
+ "bits": 3
1525
+ },
1526
+ "model.layers.0.self_attn.q_proj": {
1527
+ "group_size": 64,
1528
+ "bits": 3
1529
+ },
1530
+ "model.layers.0.self_attn.k_proj": {
1531
+ "group_size": 64,
1532
+ "bits": 3
1533
+ },
1534
+ "model.layers.0.self_attn.v_proj": {
1535
+ "group_size": 64,
1536
+ "bits": 4
1537
+ },
1538
+ "model.layers.0.self_attn.o_proj": {
1539
+ "group_size": 64,
1540
+ "bits": 3
1541
+ },
1542
+ "model.layers.0.self_attn.rope": false,
1543
+ "model.layers.0.mlp.gate_proj": {
1544
+ "group_size": 64,
1545
+ "bits": 3
1546
+ },
1547
+ "model.layers.0.mlp.down_proj": {
1548
+ "group_size": 64,
1549
+ "bits": 4
1550
+ },
1551
+ "model.layers.0.mlp.up_proj": {
1552
+ "group_size": 64,
1553
+ "bits": 3
1554
+ },
1555
+ "model.layers.0.input_layernorm": false,
1556
+ "model.layers.0.post_attention_layernorm": false,
1557
+ "model.layers.1.self_attn.q_proj": {
1558
+ "group_size": 64,
1559
+ "bits": 3
1560
+ },
1561
+ "model.layers.1.self_attn.k_proj": {
1562
+ "group_size": 64,
1563
+ "bits": 3
1564
+ },
1565
+ "model.layers.1.self_attn.v_proj": {
1566
+ "group_size": 64,
1567
+ "bits": 4
1568
+ },
1569
+ "model.layers.1.self_attn.o_proj": {
1570
+ "group_size": 64,
1571
+ "bits": 3
1572
+ },
1573
+ "model.layers.1.self_attn.rope": false,
1574
+ "model.layers.1.mlp.gate_proj": {
1575
+ "group_size": 64,
1576
+ "bits": 3
1577
+ },
1578
+ "model.layers.1.mlp.down_proj": {
1579
+ "group_size": 64,
1580
+ "bits": 4
1581
+ },
1582
+ "model.layers.1.mlp.up_proj": {
1583
+ "group_size": 64,
1584
+ "bits": 3
1585
+ },
1586
+ "model.layers.1.input_layernorm": false,
1587
+ "model.layers.1.post_attention_layernorm": false,
1588
+ "model.layers.2.self_attn.q_proj": {
1589
+ "group_size": 64,
1590
+ "bits": 3
1591
+ },
1592
+ "model.layers.2.self_attn.k_proj": {
1593
+ "group_size": 64,
1594
+ "bits": 3
1595
+ },
1596
+ "model.layers.2.self_attn.v_proj": {
1597
+ "group_size": 64,
1598
+ "bits": 4
1599
+ },
1600
+ "model.layers.2.self_attn.o_proj": {
1601
+ "group_size": 64,
1602
+ "bits": 3
1603
+ },
1604
+ "model.layers.2.self_attn.rope": false,
1605
+ "model.layers.2.mlp.gate_proj": {
1606
+ "group_size": 64,
1607
+ "bits": 3
1608
+ },
1609
+ "model.layers.2.mlp.down_proj": {
1610
+ "group_size": 64,
1611
+ "bits": 4
1612
+ },
1613
+ "model.layers.2.mlp.up_proj": {
1614
+ "group_size": 64,
1615
+ "bits": 3
1616
+ },
1617
+ "model.layers.2.input_layernorm": false,
1618
+ "model.layers.2.post_attention_layernorm": false,
1619
+ "model.layers.3.self_attn.q_proj": {
1620
+ "group_size": 64,
1621
+ "bits": 3
1622
+ },
1623
+ "model.layers.3.self_attn.k_proj": {
1624
+ "group_size": 64,
1625
+ "bits": 3
1626
+ },
1627
+ "model.layers.3.self_attn.v_proj": {
1628
+ "group_size": 64,
1629
+ "bits": 4
1630
+ },
1631
+ "model.layers.3.self_attn.o_proj": {
1632
+ "group_size": 64,
1633
+ "bits": 3
1634
+ },
1635
+ "model.layers.3.self_attn.rope": false,
1636
+ "model.layers.3.mlp.gate_proj": {
1637
+ "group_size": 64,
1638
+ "bits": 3
1639
+ },
1640
+ "model.layers.3.mlp.down_proj": {
1641
+ "group_size": 64,
1642
+ "bits": 4
1643
+ },
1644
+ "model.layers.3.mlp.up_proj": {
1645
+ "group_size": 64,
1646
+ "bits": 3
1647
+ },
1648
+ "model.layers.3.input_layernorm": false,
1649
+ "model.layers.3.post_attention_layernorm": false,
1650
+ "model.layers.4.self_attn.q_proj": {
1651
+ "group_size": 64,
1652
+ "bits": 3
1653
+ },
1654
+ "model.layers.4.self_attn.k_proj": {
1655
+ "group_size": 64,
1656
+ "bits": 3
1657
+ },
1658
+ "model.layers.4.self_attn.v_proj": {
1659
+ "group_size": 64,
1660
+ "bits": 4
1661
+ },
1662
+ "model.layers.4.self_attn.o_proj": {
1663
+ "group_size": 64,
1664
+ "bits": 3
1665
+ },
1666
+ "model.layers.4.self_attn.rope": false,
1667
+ "model.layers.4.mlp.gate_proj": {
1668
+ "group_size": 64,
1669
+ "bits": 3
1670
+ },
1671
+ "model.layers.4.mlp.down_proj": {
1672
+ "group_size": 64,
1673
+ "bits": 4
1674
+ },
1675
+ "model.layers.4.mlp.up_proj": {
1676
+ "group_size": 64,
1677
+ "bits": 3
1678
+ },
1679
+ "model.layers.4.input_layernorm": false,
1680
+ "model.layers.4.post_attention_layernorm": false,
1681
+ "model.layers.5.self_attn.q_proj": {
1682
+ "group_size": 64,
1683
+ "bits": 3
1684
+ },
1685
+ "model.layers.5.self_attn.k_proj": {
1686
+ "group_size": 64,
1687
+ "bits": 3
1688
+ },
1689
+ "model.layers.5.self_attn.v_proj": {
1690
+ "group_size": 64,
1691
+ "bits": 4
1692
+ },
1693
+ "model.layers.5.self_attn.o_proj": {
1694
+ "group_size": 64,
1695
+ "bits": 3
1696
+ },
1697
+ "model.layers.5.self_attn.rope": false,
1698
+ "model.layers.5.mlp.gate_proj": {
1699
+ "group_size": 64,
1700
+ "bits": 3
1701
+ },
1702
+ "model.layers.5.mlp.down_proj": {
1703
+ "group_size": 64,
1704
+ "bits": 4
1705
+ },
1706
+ "model.layers.5.mlp.up_proj": {
1707
+ "group_size": 64,
1708
+ "bits": 3
1709
+ },
1710
+ "model.layers.5.input_layernorm": false,
1711
+ "model.layers.5.post_attention_layernorm": false,
1712
+ "model.layers.6.self_attn.q_proj": {
1713
+ "group_size": 64,
1714
+ "bits": 3
1715
+ },
1716
+ "model.layers.6.self_attn.k_proj": {
1717
+ "group_size": 64,
1718
+ "bits": 3
1719
+ },
1720
+ "model.layers.6.self_attn.v_proj": {
1721
+ "group_size": 64,
1722
+ "bits": 3
1723
+ },
1724
+ "model.layers.6.self_attn.o_proj": {
1725
+ "group_size": 64,
1726
+ "bits": 3
1727
+ },
1728
+ "model.layers.6.self_attn.rope": false,
1729
+ "model.layers.6.mlp.gate_proj": {
1730
+ "group_size": 64,
1731
+ "bits": 3
1732
+ },
1733
+ "model.layers.6.mlp.down_proj": {
1734
+ "group_size": 64,
1735
+ "bits": 3
1736
+ },
1737
+ "model.layers.6.mlp.up_proj": {
1738
+ "group_size": 64,
1739
+ "bits": 3
1740
+ },
1741
+ "model.layers.6.input_layernorm": false,
1742
+ "model.layers.6.post_attention_layernorm": false,
1743
+ "model.layers.7.self_attn.q_proj": {
1744
+ "group_size": 64,
1745
+ "bits": 3
1746
+ },
1747
+ "model.layers.7.self_attn.k_proj": {
1748
+ "group_size": 64,
1749
+ "bits": 3
1750
+ },
1751
+ "model.layers.7.self_attn.v_proj": {
1752
+ "group_size": 64,
1753
+ "bits": 3
1754
+ },
1755
+ "model.layers.7.self_attn.o_proj": {
1756
+ "group_size": 64,
1757
+ "bits": 3
1758
+ },
1759
+ "model.layers.7.self_attn.rope": false,
1760
+ "model.layers.7.mlp.gate_proj": {
1761
+ "group_size": 64,
1762
+ "bits": 3
1763
+ },
1764
+ "model.layers.7.mlp.down_proj": {
1765
+ "group_size": 64,
1766
+ "bits": 3
1767
+ },
1768
+ "model.layers.7.mlp.up_proj": {
1769
+ "group_size": 64,
1770
+ "bits": 3
1771
+ },
1772
+ "model.layers.7.input_layernorm": false,
1773
+ "model.layers.7.post_attention_layernorm": false,
1774
+ "model.layers.8.self_attn.q_proj": {
1775
+ "group_size": 64,
1776
+ "bits": 3
1777
+ },
1778
+ "model.layers.8.self_attn.k_proj": {
1779
+ "group_size": 64,
1780
+ "bits": 3
1781
+ },
1782
+ "model.layers.8.self_attn.v_proj": {
1783
+ "group_size": 64,
1784
+ "bits": 4
1785
+ },
1786
+ "model.layers.8.self_attn.o_proj": {
1787
+ "group_size": 64,
1788
+ "bits": 3
1789
+ },
1790
+ "model.layers.8.self_attn.rope": false,
1791
+ "model.layers.8.mlp.gate_proj": {
1792
+ "group_size": 64,
1793
+ "bits": 3
1794
+ },
1795
+ "model.layers.8.mlp.down_proj": {
1796
+ "group_size": 64,
1797
+ "bits": 4
1798
+ },
1799
+ "model.layers.8.mlp.up_proj": {
1800
+ "group_size": 64,
1801
+ "bits": 3
1802
+ },
1803
+ "model.layers.8.input_layernorm": false,
1804
+ "model.layers.8.post_attention_layernorm": false,
1805
+ "model.layers.9.self_attn.q_proj": {
1806
+ "group_size": 64,
1807
+ "bits": 3
1808
+ },
1809
+ "model.layers.9.self_attn.k_proj": {
1810
+ "group_size": 64,
1811
+ "bits": 3
1812
+ },
1813
+ "model.layers.9.self_attn.v_proj": {
1814
+ "group_size": 64,
1815
+ "bits": 3
1816
+ },
1817
+ "model.layers.9.self_attn.o_proj": {
1818
+ "group_size": 64,
1819
+ "bits": 3
1820
+ },
1821
+ "model.layers.9.self_attn.rope": false,
1822
+ "model.layers.9.mlp.gate_proj": {
1823
+ "group_size": 64,
1824
+ "bits": 3
1825
+ },
1826
+ "model.layers.9.mlp.down_proj": {
1827
+ "group_size": 64,
1828
+ "bits": 3
1829
+ },
1830
+ "model.layers.9.mlp.up_proj": {
1831
+ "group_size": 64,
1832
+ "bits": 3
1833
+ },
1834
+ "model.layers.9.input_layernorm": false,
1835
+ "model.layers.9.post_attention_layernorm": false,
1836
+ "model.layers.10.self_attn.q_proj": {
1837
+ "group_size": 64,
1838
+ "bits": 3
1839
+ },
1840
+ "model.layers.10.self_attn.k_proj": {
1841
+ "group_size": 64,
1842
+ "bits": 3
1843
+ },
1844
+ "model.layers.10.self_attn.v_proj": {
1845
+ "group_size": 64,
1846
+ "bits": 3
1847
+ },
1848
+ "model.layers.10.self_attn.o_proj": {
1849
+ "group_size": 64,
1850
+ "bits": 3
1851
+ },
1852
+ "model.layers.10.self_attn.rope": false,
1853
+ "model.layers.10.mlp.gate_proj": {
1854
+ "group_size": 64,
1855
+ "bits": 3
1856
+ },
1857
+ "model.layers.10.mlp.down_proj": {
1858
+ "group_size": 64,
1859
+ "bits": 3
1860
+ },
1861
+ "model.layers.10.mlp.up_proj": {
1862
+ "group_size": 64,
1863
+ "bits": 3
1864
+ },
1865
+ "model.layers.10.input_layernorm": false,
1866
+ "model.layers.10.post_attention_layernorm": false,
1867
+ "model.layers.11.self_attn.q_proj": {
1868
+ "group_size": 64,
1869
+ "bits": 3
1870
+ },
1871
+ "model.layers.11.self_attn.k_proj": {
1872
+ "group_size": 64,
1873
+ "bits": 3
1874
+ },
1875
+ "model.layers.11.self_attn.v_proj": {
1876
+ "group_size": 64,
1877
+ "bits": 4
1878
+ },
1879
+ "model.layers.11.self_attn.o_proj": {
1880
+ "group_size": 64,
1881
+ "bits": 3
1882
+ },
1883
+ "model.layers.11.self_attn.rope": false,
1884
+ "model.layers.11.mlp.gate_proj": {
1885
+ "group_size": 64,
1886
+ "bits": 3
1887
+ },
1888
+ "model.layers.11.mlp.down_proj": {
1889
+ "group_size": 64,
1890
+ "bits": 4
1891
+ },
1892
+ "model.layers.11.mlp.up_proj": {
1893
+ "group_size": 64,
1894
+ "bits": 3
1895
+ },
1896
+ "model.layers.11.input_layernorm": false,
1897
+ "model.layers.11.post_attention_layernorm": false,
1898
+ "model.layers.12.self_attn.q_proj": {
1899
+ "group_size": 64,
1900
+ "bits": 3
1901
+ },
1902
+ "model.layers.12.self_attn.k_proj": {
1903
+ "group_size": 64,
1904
+ "bits": 3
1905
+ },
1906
+ "model.layers.12.self_attn.v_proj": {
1907
+ "group_size": 64,
1908
+ "bits": 3
1909
+ },
1910
+ "model.layers.12.self_attn.o_proj": {
1911
+ "group_size": 64,
1912
+ "bits": 3
1913
+ },
1914
+ "model.layers.12.self_attn.rope": false,
1915
+ "model.layers.12.mlp.gate_proj": {
1916
+ "group_size": 64,
1917
+ "bits": 3
1918
+ },
1919
+ "model.layers.12.mlp.down_proj": {
1920
+ "group_size": 64,
1921
+ "bits": 3
1922
+ },
1923
+ "model.layers.12.mlp.up_proj": {
1924
+ "group_size": 64,
1925
+ "bits": 3
1926
+ },
1927
+ "model.layers.12.input_layernorm": false,
1928
+ "model.layers.12.post_attention_layernorm": false,
1929
+ "model.layers.13.self_attn.q_proj": {
1930
+ "group_size": 64,
1931
+ "bits": 3
1932
+ },
1933
+ "model.layers.13.self_attn.k_proj": {
1934
+ "group_size": 64,
1935
+ "bits": 3
1936
+ },
1937
+ "model.layers.13.self_attn.v_proj": {
1938
+ "group_size": 64,
1939
+ "bits": 3
1940
+ },
1941
+ "model.layers.13.self_attn.o_proj": {
1942
+ "group_size": 64,
1943
+ "bits": 3
1944
+ },
1945
+ "model.layers.13.self_attn.rope": false,
1946
+ "model.layers.13.mlp.gate_proj": {
1947
+ "group_size": 64,
1948
+ "bits": 3
1949
+ },
1950
+ "model.layers.13.mlp.down_proj": {
1951
+ "group_size": 64,
1952
+ "bits": 3
1953
+ },
1954
+ "model.layers.13.mlp.up_proj": {
1955
+ "group_size": 64,
1956
+ "bits": 3
1957
+ },
1958
+ "model.layers.13.input_layernorm": false,
1959
+ "model.layers.13.post_attention_layernorm": false,
1960
+ "model.layers.14.self_attn.q_proj": {
1961
+ "group_size": 64,
1962
+ "bits": 3
1963
+ },
1964
+ "model.layers.14.self_attn.k_proj": {
1965
+ "group_size": 64,
1966
+ "bits": 3
1967
+ },
1968
+ "model.layers.14.self_attn.v_proj": {
1969
+ "group_size": 64,
1970
+ "bits": 4
1971
+ },
1972
+ "model.layers.14.self_attn.o_proj": {
1973
+ "group_size": 64,
1974
+ "bits": 3
1975
+ },
1976
+ "model.layers.14.self_attn.rope": false,
1977
+ "model.layers.14.mlp.gate_proj": {
1978
+ "group_size": 64,
1979
+ "bits": 3
1980
+ },
1981
+ "model.layers.14.mlp.down_proj": {
1982
+ "group_size": 64,
1983
+ "bits": 4
1984
+ },
1985
+ "model.layers.14.mlp.up_proj": {
1986
+ "group_size": 64,
1987
+ "bits": 3
1988
+ },
1989
+ "model.layers.14.input_layernorm": false,
1990
+ "model.layers.14.post_attention_layernorm": false,
1991
+ "model.layers.15.self_attn.q_proj": {
1992
+ "group_size": 64,
1993
+ "bits": 3
1994
+ },
1995
+ "model.layers.15.self_attn.k_proj": {
1996
+ "group_size": 64,
1997
+ "bits": 3
1998
+ },
1999
+ "model.layers.15.self_attn.v_proj": {
2000
+ "group_size": 64,
2001
+ "bits": 3
2002
+ },
2003
+ "model.layers.15.self_attn.o_proj": {
2004
+ "group_size": 64,
2005
+ "bits": 3
2006
+ },
2007
+ "model.layers.15.self_attn.rope": false,
2008
+ "model.layers.15.mlp.gate_proj": {
2009
+ "group_size": 64,
2010
+ "bits": 3
2011
+ },
2012
+ "model.layers.15.mlp.down_proj": {
2013
+ "group_size": 64,
2014
+ "bits": 3
2015
+ },
2016
+ "model.layers.15.mlp.up_proj": {
2017
+ "group_size": 64,
2018
+ "bits": 3
2019
+ },
2020
+ "model.layers.15.input_layernorm": false,
2021
+ "model.layers.15.post_attention_layernorm": false,
2022
+ "model.layers.16.self_attn.q_proj": {
2023
+ "group_size": 64,
2024
+ "bits": 3
2025
+ },
2026
+ "model.layers.16.self_attn.k_proj": {
2027
+ "group_size": 64,
2028
+ "bits": 3
2029
+ },
2030
+ "model.layers.16.self_attn.v_proj": {
2031
+ "group_size": 64,
2032
+ "bits": 3
2033
+ },
2034
+ "model.layers.16.self_attn.o_proj": {
2035
+ "group_size": 64,
2036
+ "bits": 3
2037
+ },
2038
+ "model.layers.16.self_attn.rope": false,
2039
+ "model.layers.16.mlp.gate_proj": {
2040
+ "group_size": 64,
2041
+ "bits": 3
2042
+ },
2043
+ "model.layers.16.mlp.down_proj": {
2044
+ "group_size": 64,
2045
+ "bits": 3
2046
+ },
2047
+ "model.layers.16.mlp.up_proj": {
2048
+ "group_size": 64,
2049
+ "bits": 3
2050
+ },
2051
+ "model.layers.16.input_layernorm": false,
2052
+ "model.layers.16.post_attention_layernorm": false,
2053
+ "model.layers.17.self_attn.q_proj": {
2054
+ "group_size": 64,
2055
+ "bits": 3
2056
+ },
2057
+ "model.layers.17.self_attn.k_proj": {
2058
+ "group_size": 64,
2059
+ "bits": 3
2060
+ },
2061
+ "model.layers.17.self_attn.v_proj": {
2062
+ "group_size": 64,
2063
+ "bits": 4
2064
+ },
2065
+ "model.layers.17.self_attn.o_proj": {
2066
+ "group_size": 64,
2067
+ "bits": 3
2068
+ },
2069
+ "model.layers.17.self_attn.rope": false,
2070
+ "model.layers.17.mlp.gate_proj": {
2071
+ "group_size": 64,
2072
+ "bits": 3
2073
+ },
2074
+ "model.layers.17.mlp.down_proj": {
2075
+ "group_size": 64,
2076
+ "bits": 4
2077
+ },
2078
+ "model.layers.17.mlp.up_proj": {
2079
+ "group_size": 64,
2080
+ "bits": 3
2081
+ },
2082
+ "model.layers.17.input_layernorm": false,
2083
+ "model.layers.17.post_attention_layernorm": false,
2084
+ "model.layers.18.self_attn.q_proj": {
2085
+ "group_size": 64,
2086
+ "bits": 3
2087
+ },
2088
+ "model.layers.18.self_attn.k_proj": {
2089
+ "group_size": 64,
2090
+ "bits": 3
2091
+ },
2092
+ "model.layers.18.self_attn.v_proj": {
2093
+ "group_size": 64,
2094
+ "bits": 3
2095
+ },
2096
+ "model.layers.18.self_attn.o_proj": {
2097
+ "group_size": 64,
2098
+ "bits": 3
2099
+ },
2100
+ "model.layers.18.self_attn.rope": false,
2101
+ "model.layers.18.mlp.gate_proj": {
2102
+ "group_size": 64,
2103
+ "bits": 3
2104
+ },
2105
+ "model.layers.18.mlp.down_proj": {
2106
+ "group_size": 64,
2107
+ "bits": 3
2108
+ },
2109
+ "model.layers.18.mlp.up_proj": {
2110
+ "group_size": 64,
2111
+ "bits": 3
2112
+ },
2113
+ "model.layers.18.input_layernorm": false,
2114
+ "model.layers.18.post_attention_layernorm": false,
2115
+ "model.layers.19.self_attn.q_proj": {
2116
+ "group_size": 64,
2117
+ "bits": 3
2118
+ },
2119
+ "model.layers.19.self_attn.k_proj": {
2120
+ "group_size": 64,
2121
+ "bits": 3
2122
+ },
2123
+ "model.layers.19.self_attn.v_proj": {
2124
+ "group_size": 64,
2125
+ "bits": 3
2126
+ },
2127
+ "model.layers.19.self_attn.o_proj": {
2128
+ "group_size": 64,
2129
+ "bits": 3
2130
+ },
2131
+ "model.layers.19.self_attn.rope": false,
2132
+ "model.layers.19.mlp.gate_proj": {
2133
+ "group_size": 64,
2134
+ "bits": 3
2135
+ },
2136
+ "model.layers.19.mlp.down_proj": {
2137
+ "group_size": 64,
2138
+ "bits": 3
2139
+ },
2140
+ "model.layers.19.mlp.up_proj": {
2141
+ "group_size": 64,
2142
+ "bits": 3
2143
+ },
2144
+ "model.layers.19.input_layernorm": false,
2145
+ "model.layers.19.post_attention_layernorm": false,
2146
+ "model.layers.20.self_attn.q_proj": {
2147
+ "group_size": 64,
2148
+ "bits": 3
2149
+ },
2150
+ "model.layers.20.self_attn.k_proj": {
2151
+ "group_size": 64,
2152
+ "bits": 3
2153
+ },
2154
+ "model.layers.20.self_attn.v_proj": {
2155
+ "group_size": 64,
2156
+ "bits": 4
2157
+ },
2158
+ "model.layers.20.self_attn.o_proj": {
2159
+ "group_size": 64,
2160
+ "bits": 3
2161
+ },
2162
+ "model.layers.20.self_attn.rope": false,
2163
+ "model.layers.20.mlp.gate_proj": {
2164
+ "group_size": 64,
2165
+ "bits": 3
2166
+ },
2167
+ "model.layers.20.mlp.down_proj": {
2168
+ "group_size": 64,
2169
+ "bits": 4
2170
+ },
2171
+ "model.layers.20.mlp.up_proj": {
2172
+ "group_size": 64,
2173
+ "bits": 3
2174
+ },
2175
+ "model.layers.20.input_layernorm": false,
2176
+ "model.layers.20.post_attention_layernorm": false,
2177
+ "model.layers.21.self_attn.q_proj": {
2178
+ "group_size": 64,
2179
+ "bits": 3
2180
+ },
2181
+ "model.layers.21.self_attn.k_proj": {
2182
+ "group_size": 64,
2183
+ "bits": 3
2184
+ },
2185
+ "model.layers.21.self_attn.v_proj": {
2186
+ "group_size": 64,
2187
+ "bits": 3
2188
+ },
2189
+ "model.layers.21.self_attn.o_proj": {
2190
+ "group_size": 64,
2191
+ "bits": 3
2192
+ },
2193
+ "model.layers.21.self_attn.rope": false,
2194
+ "model.layers.21.mlp.gate_proj": {
2195
+ "group_size": 64,
2196
+ "bits": 3
2197
+ },
2198
+ "model.layers.21.mlp.down_proj": {
2199
+ "group_size": 64,
2200
+ "bits": 3
2201
+ },
2202
+ "model.layers.21.mlp.up_proj": {
2203
+ "group_size": 64,
2204
+ "bits": 3
2205
+ },
2206
+ "model.layers.21.input_layernorm": false,
2207
+ "model.layers.21.post_attention_layernorm": false,
2208
+ "model.layers.22.self_attn.q_proj": {
2209
+ "group_size": 64,
2210
+ "bits": 3
2211
+ },
2212
+ "model.layers.22.self_attn.k_proj": {
2213
+ "group_size": 64,
2214
+ "bits": 3
2215
+ },
2216
+ "model.layers.22.self_attn.v_proj": {
2217
+ "group_size": 64,
2218
+ "bits": 3
2219
+ },
2220
+ "model.layers.22.self_attn.o_proj": {
2221
+ "group_size": 64,
2222
+ "bits": 3
2223
+ },
2224
+ "model.layers.22.self_attn.rope": false,
2225
+ "model.layers.22.mlp.gate_proj": {
2226
+ "group_size": 64,
2227
+ "bits": 3
2228
+ },
2229
+ "model.layers.22.mlp.down_proj": {
2230
+ "group_size": 64,
2231
+ "bits": 3
2232
+ },
2233
+ "model.layers.22.mlp.up_proj": {
2234
+ "group_size": 64,
2235
+ "bits": 3
2236
+ },
2237
+ "model.layers.22.input_layernorm": false,
2238
+ "model.layers.22.post_attention_layernorm": false,
2239
+ "model.layers.23.self_attn.q_proj": {
2240
+ "group_size": 64,
2241
+ "bits": 3
2242
+ },
2243
+ "model.layers.23.self_attn.k_proj": {
2244
+ "group_size": 64,
2245
+ "bits": 3
2246
+ },
2247
+ "model.layers.23.self_attn.v_proj": {
2248
+ "group_size": 64,
2249
+ "bits": 4
2250
+ },
2251
+ "model.layers.23.self_attn.o_proj": {
2252
+ "group_size": 64,
2253
+ "bits": 3
2254
+ },
2255
+ "model.layers.23.self_attn.rope": false,
2256
+ "model.layers.23.mlp.gate_proj": {
2257
+ "group_size": 64,
2258
+ "bits": 3
2259
+ },
2260
+ "model.layers.23.mlp.down_proj": {
2261
+ "group_size": 64,
2262
+ "bits": 4
2263
+ },
2264
+ "model.layers.23.mlp.up_proj": {
2265
+ "group_size": 64,
2266
+ "bits": 3
2267
+ },
2268
+ "model.layers.23.input_layernorm": false,
2269
+ "model.layers.23.post_attention_layernorm": false,
2270
+ "model.layers.24.self_attn.q_proj": {
2271
+ "group_size": 64,
2272
+ "bits": 3
2273
+ },
2274
+ "model.layers.24.self_attn.k_proj": {
2275
+ "group_size": 64,
2276
+ "bits": 3
2277
+ },
2278
+ "model.layers.24.self_attn.v_proj": {
2279
+ "group_size": 64,
2280
+ "bits": 3
2281
+ },
2282
+ "model.layers.24.self_attn.o_proj": {
2283
+ "group_size": 64,
2284
+ "bits": 3
2285
+ },
2286
+ "model.layers.24.self_attn.rope": false,
2287
+ "model.layers.24.mlp.gate_proj": {
2288
+ "group_size": 64,
2289
+ "bits": 3
2290
+ },
2291
+ "model.layers.24.mlp.down_proj": {
2292
+ "group_size": 64,
2293
+ "bits": 3
2294
+ },
2295
+ "model.layers.24.mlp.up_proj": {
2296
+ "group_size": 64,
2297
+ "bits": 3
2298
+ },
2299
+ "model.layers.24.input_layernorm": false,
2300
+ "model.layers.24.post_attention_layernorm": false,
2301
+ "model.layers.25.self_attn.q_proj": {
2302
+ "group_size": 64,
2303
+ "bits": 3
2304
+ },
2305
+ "model.layers.25.self_attn.k_proj": {
2306
+ "group_size": 64,
2307
+ "bits": 3
2308
+ },
2309
+ "model.layers.25.self_attn.v_proj": {
2310
+ "group_size": 64,
2311
+ "bits": 3
2312
+ },
2313
+ "model.layers.25.self_attn.o_proj": {
2314
+ "group_size": 64,
2315
+ "bits": 3
2316
+ },
2317
+ "model.layers.25.self_attn.rope": false,
2318
+ "model.layers.25.mlp.gate_proj": {
2319
+ "group_size": 64,
2320
+ "bits": 3
2321
+ },
2322
+ "model.layers.25.mlp.down_proj": {
2323
+ "group_size": 64,
2324
+ "bits": 3
2325
+ },
2326
+ "model.layers.25.mlp.up_proj": {
2327
+ "group_size": 64,
2328
+ "bits": 3
2329
+ },
2330
+ "model.layers.25.input_layernorm": false,
2331
+ "model.layers.25.post_attention_layernorm": false,
2332
+ "model.layers.26.self_attn.q_proj": {
2333
+ "group_size": 64,
2334
+ "bits": 3
2335
+ },
2336
+ "model.layers.26.self_attn.k_proj": {
2337
+ "group_size": 64,
2338
+ "bits": 3
2339
+ },
2340
+ "model.layers.26.self_attn.v_proj": {
2341
+ "group_size": 64,
2342
+ "bits": 4
2343
+ },
2344
+ "model.layers.26.self_attn.o_proj": {
2345
+ "group_size": 64,
2346
+ "bits": 3
2347
+ },
2348
+ "model.layers.26.self_attn.rope": false,
2349
+ "model.layers.26.mlp.gate_proj": {
2350
+ "group_size": 64,
2351
+ "bits": 3
2352
+ },
2353
+ "model.layers.26.mlp.down_proj": {
2354
+ "group_size": 64,
2355
+ "bits": 4
2356
+ },
2357
+ "model.layers.26.mlp.up_proj": {
2358
+ "group_size": 64,
2359
+ "bits": 3
2360
+ },
2361
+ "model.layers.26.input_layernorm": false,
2362
+ "model.layers.26.post_attention_layernorm": false,
2363
+ "model.layers.27.self_attn.q_proj": {
2364
+ "group_size": 64,
2365
+ "bits": 3
2366
+ },
2367
+ "model.layers.27.self_attn.k_proj": {
2368
+ "group_size": 64,
2369
+ "bits": 3
2370
+ },
2371
+ "model.layers.27.self_attn.v_proj": {
2372
+ "group_size": 64,
2373
+ "bits": 3
2374
+ },
2375
+ "model.layers.27.self_attn.o_proj": {
2376
+ "group_size": 64,
2377
+ "bits": 3
2378
+ },
2379
+ "model.layers.27.self_attn.rope": false,
2380
+ "model.layers.27.mlp.gate_proj": {
2381
+ "group_size": 64,
2382
+ "bits": 3
2383
+ },
2384
+ "model.layers.27.mlp.down_proj": {
2385
+ "group_size": 64,
2386
+ "bits": 3
2387
+ },
2388
+ "model.layers.27.mlp.up_proj": {
2389
+ "group_size": 64,
2390
+ "bits": 3
2391
+ },
2392
+ "model.layers.27.input_layernorm": false,
2393
+ "model.layers.27.post_attention_layernorm": false,
2394
+ "model.layers.28.self_attn.q_proj": {
2395
+ "group_size": 64,
2396
+ "bits": 3
2397
+ },
2398
+ "model.layers.28.self_attn.k_proj": {
2399
+ "group_size": 64,
2400
+ "bits": 3
2401
+ },
2402
+ "model.layers.28.self_attn.v_proj": {
2403
+ "group_size": 64,
2404
+ "bits": 3
2405
+ },
2406
+ "model.layers.28.self_attn.o_proj": {
2407
+ "group_size": 64,
2408
+ "bits": 3
2409
+ },
2410
+ "model.layers.28.self_attn.rope": false,
2411
+ "model.layers.28.mlp.gate_proj": {
2412
+ "group_size": 64,
2413
+ "bits": 3
2414
+ },
2415
+ "model.layers.28.mlp.down_proj": {
2416
+ "group_size": 64,
2417
+ "bits": 3
2418
+ },
2419
+ "model.layers.28.mlp.up_proj": {
2420
+ "group_size": 64,
2421
+ "bits": 3
2422
+ },
2423
+ "model.layers.28.input_layernorm": false,
2424
+ "model.layers.28.post_attention_layernorm": false,
2425
+ "model.layers.29.self_attn.q_proj": {
2426
+ "group_size": 64,
2427
+ "bits": 3
2428
+ },
2429
+ "model.layers.29.self_attn.k_proj": {
2430
+ "group_size": 64,
2431
+ "bits": 3
2432
+ },
2433
+ "model.layers.29.self_attn.v_proj": {
2434
+ "group_size": 64,
2435
+ "bits": 4
2436
+ },
2437
+ "model.layers.29.self_attn.o_proj": {
2438
+ "group_size": 64,
2439
+ "bits": 3
2440
+ },
2441
+ "model.layers.29.self_attn.rope": false,
2442
+ "model.layers.29.mlp.gate_proj": {
2443
+ "group_size": 64,
2444
+ "bits": 3
2445
+ },
2446
+ "model.layers.29.mlp.down_proj": {
2447
+ "group_size": 64,
2448
+ "bits": 4
2449
+ },
2450
+ "model.layers.29.mlp.up_proj": {
2451
+ "group_size": 64,
2452
+ "bits": 3
2453
+ },
2454
+ "model.layers.29.input_layernorm": false,
2455
+ "model.layers.29.post_attention_layernorm": false,
2456
+ "model.layers.30.self_attn.q_proj": {
2457
+ "group_size": 64,
2458
+ "bits": 3
2459
+ },
2460
+ "model.layers.30.self_attn.k_proj": {
2461
+ "group_size": 64,
2462
+ "bits": 3
2463
+ },
2464
+ "model.layers.30.self_attn.v_proj": {
2465
+ "group_size": 64,
2466
+ "bits": 3
2467
+ },
2468
+ "model.layers.30.self_attn.o_proj": {
2469
+ "group_size": 64,
2470
+ "bits": 3
2471
+ },
2472
+ "model.layers.30.self_attn.rope": false,
2473
+ "model.layers.30.mlp.gate_proj": {
2474
+ "group_size": 64,
2475
+ "bits": 3
2476
+ },
2477
+ "model.layers.30.mlp.down_proj": {
2478
+ "group_size": 64,
2479
+ "bits": 3
2480
+ },
2481
+ "model.layers.30.mlp.up_proj": {
2482
+ "group_size": 64,
2483
+ "bits": 3
2484
+ },
2485
+ "model.layers.30.input_layernorm": false,
2486
+ "model.layers.30.post_attention_layernorm": false,
2487
+ "model.layers.31.self_attn.q_proj": {
2488
+ "group_size": 64,
2489
+ "bits": 3
2490
+ },
2491
+ "model.layers.31.self_attn.k_proj": {
2492
+ "group_size": 64,
2493
+ "bits": 3
2494
+ },
2495
+ "model.layers.31.self_attn.v_proj": {
2496
+ "group_size": 64,
2497
+ "bits": 3
2498
+ },
2499
+ "model.layers.31.self_attn.o_proj": {
2500
+ "group_size": 64,
2501
+ "bits": 3
2502
+ },
2503
+ "model.layers.31.self_attn.rope": false,
2504
+ "model.layers.31.mlp.gate_proj": {
2505
+ "group_size": 64,
2506
+ "bits": 3
2507
+ },
2508
+ "model.layers.31.mlp.down_proj": {
2509
+ "group_size": 64,
2510
+ "bits": 3
2511
+ },
2512
+ "model.layers.31.mlp.up_proj": {
2513
+ "group_size": 64,
2514
+ "bits": 3
2515
+ },
2516
+ "model.layers.31.input_layernorm": false,
2517
+ "model.layers.31.post_attention_layernorm": false,
2518
+ "model.layers.32.self_attn.q_proj": {
2519
+ "group_size": 64,
2520
+ "bits": 3
2521
+ },
2522
+ "model.layers.32.self_attn.k_proj": {
2523
+ "group_size": 64,
2524
+ "bits": 3
2525
+ },
2526
+ "model.layers.32.self_attn.v_proj": {
2527
+ "group_size": 64,
2528
+ "bits": 4
2529
+ },
2530
+ "model.layers.32.self_attn.o_proj": {
2531
+ "group_size": 64,
2532
+ "bits": 3
2533
+ },
2534
+ "model.layers.32.self_attn.rope": false,
2535
+ "model.layers.32.mlp.gate_proj": {
2536
+ "group_size": 64,
2537
+ "bits": 3
2538
+ },
2539
+ "model.layers.32.mlp.down_proj": {
2540
+ "group_size": 64,
2541
+ "bits": 4
2542
+ },
2543
+ "model.layers.32.mlp.up_proj": {
2544
+ "group_size": 64,
2545
+ "bits": 3
2546
+ },
2547
+ "model.layers.32.input_layernorm": false,
2548
+ "model.layers.32.post_attention_layernorm": false,
2549
+ "model.layers.33.self_attn.q_proj": {
2550
+ "group_size": 64,
2551
+ "bits": 3
2552
+ },
2553
+ "model.layers.33.self_attn.k_proj": {
2554
+ "group_size": 64,
2555
+ "bits": 3
2556
+ },
2557
+ "model.layers.33.self_attn.v_proj": {
2558
+ "group_size": 64,
2559
+ "bits": 3
2560
+ },
2561
+ "model.layers.33.self_attn.o_proj": {
2562
+ "group_size": 64,
2563
+ "bits": 3
2564
+ },
2565
+ "model.layers.33.self_attn.rope": false,
2566
+ "model.layers.33.mlp.gate_proj": {
2567
+ "group_size": 64,
2568
+ "bits": 3
2569
+ },
2570
+ "model.layers.33.mlp.down_proj": {
2571
+ "group_size": 64,
2572
+ "bits": 3
2573
+ },
2574
+ "model.layers.33.mlp.up_proj": {
2575
+ "group_size": 64,
2576
+ "bits": 3
2577
+ },
2578
+ "model.layers.33.input_layernorm": false,
2579
+ "model.layers.33.post_attention_layernorm": false,
2580
+ "model.layers.34.self_attn.q_proj": {
2581
+ "group_size": 64,
2582
+ "bits": 3
2583
+ },
2584
+ "model.layers.34.self_attn.k_proj": {
2585
+ "group_size": 64,
2586
+ "bits": 3
2587
+ },
2588
+ "model.layers.34.self_attn.v_proj": {
2589
+ "group_size": 64,
2590
+ "bits": 3
2591
+ },
2592
+ "model.layers.34.self_attn.o_proj": {
2593
+ "group_size": 64,
2594
+ "bits": 3
2595
+ },
2596
+ "model.layers.34.self_attn.rope": false,
2597
+ "model.layers.34.mlp.gate_proj": {
2598
+ "group_size": 64,
2599
+ "bits": 3
2600
+ },
2601
+ "model.layers.34.mlp.down_proj": {
2602
+ "group_size": 64,
2603
+ "bits": 3
2604
+ },
2605
+ "model.layers.34.mlp.up_proj": {
2606
+ "group_size": 64,
2607
+ "bits": 3
2608
+ },
2609
+ "model.layers.34.input_layernorm": false,
2610
+ "model.layers.34.post_attention_layernorm": false,
2611
+ "model.layers.35.self_attn.q_proj": {
2612
+ "group_size": 64,
2613
+ "bits": 3
2614
+ },
2615
+ "model.layers.35.self_attn.k_proj": {
2616
+ "group_size": 64,
2617
+ "bits": 3
2618
+ },
2619
+ "model.layers.35.self_attn.v_proj": {
2620
+ "group_size": 64,
2621
+ "bits": 4
2622
+ },
2623
+ "model.layers.35.self_attn.o_proj": {
2624
+ "group_size": 64,
2625
+ "bits": 3
2626
+ },
2627
+ "model.layers.35.self_attn.rope": false,
2628
+ "model.layers.35.mlp.gate_proj": {
2629
+ "group_size": 64,
2630
+ "bits": 3
2631
+ },
2632
+ "model.layers.35.mlp.down_proj": {
2633
+ "group_size": 64,
2634
+ "bits": 4
2635
+ },
2636
+ "model.layers.35.mlp.up_proj": {
2637
+ "group_size": 64,
2638
+ "bits": 3
2639
+ },
2640
+ "model.layers.35.input_layernorm": false,
2641
+ "model.layers.35.post_attention_layernorm": false,
2642
+ "model.layers.36.self_attn.q_proj": {
2643
+ "group_size": 64,
2644
+ "bits": 3
2645
+ },
2646
+ "model.layers.36.self_attn.k_proj": {
2647
+ "group_size": 64,
2648
+ "bits": 3
2649
+ },
2650
+ "model.layers.36.self_attn.v_proj": {
2651
+ "group_size": 64,
2652
+ "bits": 3
2653
+ },
2654
+ "model.layers.36.self_attn.o_proj": {
2655
+ "group_size": 64,
2656
+ "bits": 3
2657
+ },
2658
+ "model.layers.36.self_attn.rope": false,
2659
+ "model.layers.36.mlp.gate_proj": {
2660
+ "group_size": 64,
2661
+ "bits": 3
2662
+ },
2663
+ "model.layers.36.mlp.down_proj": {
2664
+ "group_size": 64,
2665
+ "bits": 3
2666
+ },
2667
+ "model.layers.36.mlp.up_proj": {
2668
+ "group_size": 64,
2669
+ "bits": 3
2670
+ },
2671
+ "model.layers.36.input_layernorm": false,
2672
+ "model.layers.36.post_attention_layernorm": false,
2673
+ "model.layers.37.self_attn.q_proj": {
2674
+ "group_size": 64,
2675
+ "bits": 3
2676
+ },
2677
+ "model.layers.37.self_attn.k_proj": {
2678
+ "group_size": 64,
2679
+ "bits": 3
2680
+ },
2681
+ "model.layers.37.self_attn.v_proj": {
2682
+ "group_size": 64,
2683
+ "bits": 3
2684
+ },
2685
+ "model.layers.37.self_attn.o_proj": {
2686
+ "group_size": 64,
2687
+ "bits": 3
2688
+ },
2689
+ "model.layers.37.self_attn.rope": false,
2690
+ "model.layers.37.mlp.gate_proj": {
2691
+ "group_size": 64,
2692
+ "bits": 3
2693
+ },
2694
+ "model.layers.37.mlp.down_proj": {
2695
+ "group_size": 64,
2696
+ "bits": 3
2697
+ },
2698
+ "model.layers.37.mlp.up_proj": {
2699
+ "group_size": 64,
2700
+ "bits": 3
2701
+ },
2702
+ "model.layers.37.input_layernorm": false,
2703
+ "model.layers.37.post_attention_layernorm": false,
2704
+ "model.layers.38.self_attn.q_proj": {
2705
+ "group_size": 64,
2706
+ "bits": 3
2707
+ },
2708
+ "model.layers.38.self_attn.k_proj": {
2709
+ "group_size": 64,
2710
+ "bits": 3
2711
+ },
2712
+ "model.layers.38.self_attn.v_proj": {
2713
+ "group_size": 64,
2714
+ "bits": 4
2715
+ },
2716
+ "model.layers.38.self_attn.o_proj": {
2717
+ "group_size": 64,
2718
+ "bits": 3
2719
+ },
2720
+ "model.layers.38.self_attn.rope": false,
2721
+ "model.layers.38.mlp.gate_proj": {
2722
+ "group_size": 64,
2723
+ "bits": 3
2724
+ },
2725
+ "model.layers.38.mlp.down_proj": {
2726
+ "group_size": 64,
2727
+ "bits": 4
2728
+ },
2729
+ "model.layers.38.mlp.up_proj": {
2730
+ "group_size": 64,
2731
+ "bits": 3
2732
+ },
2733
+ "model.layers.38.input_layernorm": false,
2734
+ "model.layers.38.post_attention_layernorm": false,
2735
+ "model.layers.39.self_attn.q_proj": {
2736
+ "group_size": 64,
2737
+ "bits": 3
2738
+ },
2739
+ "model.layers.39.self_attn.k_proj": {
2740
+ "group_size": 64,
2741
+ "bits": 3
2742
+ },
2743
+ "model.layers.39.self_attn.v_proj": {
2744
+ "group_size": 64,
2745
+ "bits": 3
2746
+ },
2747
+ "model.layers.39.self_attn.o_proj": {
2748
+ "group_size": 64,
2749
+ "bits": 3
2750
+ },
2751
+ "model.layers.39.self_attn.rope": false,
2752
+ "model.layers.39.mlp.gate_proj": {
2753
+ "group_size": 64,
2754
+ "bits": 3
2755
+ },
2756
+ "model.layers.39.mlp.down_proj": {
2757
+ "group_size": 64,
2758
+ "bits": 3
2759
+ },
2760
+ "model.layers.39.mlp.up_proj": {
2761
+ "group_size": 64,
2762
+ "bits": 3
2763
+ },
2764
+ "model.layers.39.input_layernorm": false,
2765
+ "model.layers.39.post_attention_layernorm": false,
2766
+ "model.layers.40.self_attn.q_proj": {
2767
+ "group_size": 64,
2768
+ "bits": 3
2769
+ },
2770
+ "model.layers.40.self_attn.k_proj": {
2771
+ "group_size": 64,
2772
+ "bits": 3
2773
+ },
2774
+ "model.layers.40.self_attn.v_proj": {
2775
+ "group_size": 64,
2776
+ "bits": 3
2777
+ },
2778
+ "model.layers.40.self_attn.o_proj": {
2779
+ "group_size": 64,
2780
+ "bits": 3
2781
+ },
2782
+ "model.layers.40.self_attn.rope": false,
2783
+ "model.layers.40.mlp.gate_proj": {
2784
+ "group_size": 64,
2785
+ "bits": 3
2786
+ },
2787
+ "model.layers.40.mlp.down_proj": {
2788
+ "group_size": 64,
2789
+ "bits": 3
2790
+ },
2791
+ "model.layers.40.mlp.up_proj": {
2792
+ "group_size": 64,
2793
+ "bits": 3
2794
+ },
2795
+ "model.layers.40.input_layernorm": false,
2796
+ "model.layers.40.post_attention_layernorm": false,
2797
+ "model.layers.41.self_attn.q_proj": {
2798
+ "group_size": 64,
2799
+ "bits": 3
2800
+ },
2801
+ "model.layers.41.self_attn.k_proj": {
2802
+ "group_size": 64,
2803
+ "bits": 3
2804
+ },
2805
+ "model.layers.41.self_attn.v_proj": {
2806
+ "group_size": 64,
2807
+ "bits": 4
2808
+ },
2809
+ "model.layers.41.self_attn.o_proj": {
2810
+ "group_size": 64,
2811
+ "bits": 3
2812
+ },
2813
+ "model.layers.41.self_attn.rope": false,
2814
+ "model.layers.41.mlp.gate_proj": {
2815
+ "group_size": 64,
2816
+ "bits": 3
2817
+ },
2818
+ "model.layers.41.mlp.down_proj": {
2819
+ "group_size": 64,
2820
+ "bits": 4
2821
+ },
2822
+ "model.layers.41.mlp.up_proj": {
2823
+ "group_size": 64,
2824
+ "bits": 3
2825
+ },
2826
+ "model.layers.41.input_layernorm": false,
2827
+ "model.layers.41.post_attention_layernorm": false,
2828
+ "model.layers.42.self_attn.q_proj": {
2829
+ "group_size": 64,
2830
+ "bits": 3
2831
+ },
2832
+ "model.layers.42.self_attn.k_proj": {
2833
+ "group_size": 64,
2834
+ "bits": 3
2835
+ },
2836
+ "model.layers.42.self_attn.v_proj": {
2837
+ "group_size": 64,
2838
+ "bits": 4
2839
+ },
2840
+ "model.layers.42.self_attn.o_proj": {
2841
+ "group_size": 64,
2842
+ "bits": 3
2843
+ },
2844
+ "model.layers.42.self_attn.rope": false,
2845
+ "model.layers.42.mlp.gate_proj": {
2846
+ "group_size": 64,
2847
+ "bits": 3
2848
+ },
2849
+ "model.layers.42.mlp.down_proj": {
2850
+ "group_size": 64,
2851
+ "bits": 4
2852
+ },
2853
+ "model.layers.42.mlp.up_proj": {
2854
+ "group_size": 64,
2855
+ "bits": 3
2856
+ },
2857
+ "model.layers.42.input_layernorm": false,
2858
+ "model.layers.42.post_attention_layernorm": false,
2859
+ "model.layers.43.self_attn.q_proj": {
2860
+ "group_size": 64,
2861
+ "bits": 3
2862
+ },
2863
+ "model.layers.43.self_attn.k_proj": {
2864
+ "group_size": 64,
2865
+ "bits": 3
2866
+ },
2867
+ "model.layers.43.self_attn.v_proj": {
2868
+ "group_size": 64,
2869
+ "bits": 4
2870
+ },
2871
+ "model.layers.43.self_attn.o_proj": {
2872
+ "group_size": 64,
2873
+ "bits": 3
2874
+ },
2875
+ "model.layers.43.self_attn.rope": false,
2876
+ "model.layers.43.mlp.gate_proj": {
2877
+ "group_size": 64,
2878
+ "bits": 3
2879
+ },
2880
+ "model.layers.43.mlp.down_proj": {
2881
+ "group_size": 64,
2882
+ "bits": 4
2883
+ },
2884
+ "model.layers.43.mlp.up_proj": {
2885
+ "group_size": 64,
2886
+ "bits": 3
2887
+ },
2888
+ "model.layers.43.input_layernorm": false,
2889
+ "model.layers.43.post_attention_layernorm": false,
2890
+ "model.layers.44.self_attn.q_proj": {
2891
+ "group_size": 64,
2892
+ "bits": 3
2893
+ },
2894
+ "model.layers.44.self_attn.k_proj": {
2895
+ "group_size": 64,
2896
+ "bits": 3
2897
+ },
2898
+ "model.layers.44.self_attn.v_proj": {
2899
+ "group_size": 64,
2900
+ "bits": 4
2901
+ },
2902
+ "model.layers.44.self_attn.o_proj": {
2903
+ "group_size": 64,
2904
+ "bits": 3
2905
+ },
2906
+ "model.layers.44.self_attn.rope": false,
2907
+ "model.layers.44.mlp.gate_proj": {
2908
+ "group_size": 64,
2909
+ "bits": 3
2910
+ },
2911
+ "model.layers.44.mlp.down_proj": {
2912
+ "group_size": 64,
2913
+ "bits": 4
2914
+ },
2915
+ "model.layers.44.mlp.up_proj": {
2916
+ "group_size": 64,
2917
+ "bits": 3
2918
+ },
2919
+ "model.layers.44.input_layernorm": false,
2920
+ "model.layers.44.post_attention_layernorm": false,
2921
+ "model.layers.45.self_attn.q_proj": {
2922
+ "group_size": 64,
2923
+ "bits": 3
2924
+ },
2925
+ "model.layers.45.self_attn.k_proj": {
2926
+ "group_size": 64,
2927
+ "bits": 3
2928
+ },
2929
+ "model.layers.45.self_attn.v_proj": {
2930
+ "group_size": 64,
2931
+ "bits": 4
2932
+ },
2933
+ "model.layers.45.self_attn.o_proj": {
2934
+ "group_size": 64,
2935
+ "bits": 3
2936
+ },
2937
+ "model.layers.45.self_attn.rope": false,
2938
+ "model.layers.45.mlp.gate_proj": {
2939
+ "group_size": 64,
2940
+ "bits": 3
2941
+ },
2942
+ "model.layers.45.mlp.down_proj": {
2943
+ "group_size": 64,
2944
+ "bits": 4
2945
+ },
2946
+ "model.layers.45.mlp.up_proj": {
2947
+ "group_size": 64,
2948
+ "bits": 3
2949
+ },
2950
+ "model.layers.45.input_layernorm": false,
2951
+ "model.layers.45.post_attention_layernorm": false,
2952
+ "model.layers.46.self_attn.q_proj": {
2953
+ "group_size": 64,
2954
+ "bits": 3
2955
+ },
2956
+ "model.layers.46.self_attn.k_proj": {
2957
+ "group_size": 64,
2958
+ "bits": 3
2959
+ },
2960
+ "model.layers.46.self_attn.v_proj": {
2961
+ "group_size": 64,
2962
+ "bits": 4
2963
+ },
2964
+ "model.layers.46.self_attn.o_proj": {
2965
+ "group_size": 64,
2966
+ "bits": 3
2967
+ },
2968
+ "model.layers.46.self_attn.rope": false,
2969
+ "model.layers.46.mlp.gate_proj": {
2970
+ "group_size": 64,
2971
+ "bits": 3
2972
+ },
2973
+ "model.layers.46.mlp.down_proj": {
2974
+ "group_size": 64,
2975
+ "bits": 4
2976
+ },
2977
+ "model.layers.46.mlp.up_proj": {
2978
+ "group_size": 64,
2979
+ "bits": 3
2980
+ },
2981
+ "model.layers.46.input_layernorm": false,
2982
+ "model.layers.46.post_attention_layernorm": false,
2983
+ "model.layers.47.self_attn.q_proj": {
2984
+ "group_size": 64,
2985
+ "bits": 3
2986
+ },
2987
+ "model.layers.47.self_attn.k_proj": {
2988
+ "group_size": 64,
2989
+ "bits": 3
2990
+ },
2991
+ "model.layers.47.self_attn.v_proj": {
2992
+ "group_size": 64,
2993
+ "bits": 4
2994
+ },
2995
+ "model.layers.47.self_attn.o_proj": {
2996
+ "group_size": 64,
2997
+ "bits": 3
2998
+ },
2999
+ "model.layers.47.self_attn.rope": false,
3000
+ "model.layers.47.mlp.gate_proj": {
3001
+ "group_size": 64,
3002
+ "bits": 3
3003
+ },
3004
+ "model.layers.47.mlp.down_proj": {
3005
+ "group_size": 64,
3006
+ "bits": 4
3007
+ },
3008
+ "model.layers.47.mlp.up_proj": {
3009
+ "group_size": 64,
3010
+ "bits": 3
3011
+ },
3012
+ "model.layers.47.input_layernorm": false,
3013
+ "model.layers.47.post_attention_layernorm": false,
3014
+ "model.norm": false,
3015
+ "lm_head": {
3016
+ "group_size": 64,
3017
+ "bits": 4
3018
+ }
3019
+ },
3020
+ "rms_norm_eps": 1e-05,
3021
+ "rope_theta": 1000000.0,
3022
+ "sliding_window": 131072,
3023
+ "tie_word_embeddings": false,
3024
+ "torch_dtype": "bfloat16",
3025
+ "transformers_version": "4.43.1",
3026
+ "use_cache": true,
3027
+ "use_sliding_window": false,
3028
+ "vocab_size": 152064
3029
+ }
DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ccdd1c75fe542d1c9b5ded243fd1ae6c627c852e07f466d67181991d448555ea
3
+ size 5348117358
DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:744229f4c2a22a1301aa5a3d582efa7ba228c9b460eaf47b2dd85b8e7a57fb24
3
+ size 1440613893
DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/special_tokens_map.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<｜begin▁of▁sentence｜>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<｜end▁of▁sentence｜>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<｜end▁of▁sentence｜>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ }
23
+ }
DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e20ddafc659ba90242154b55275402edeca0715e5dbb30f56815a4ce081f4893
3
+ size 11422778
DeepSeek-R1-Distill-Qwen-14B-3,4_mixed/tokenizer_config.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "151643": {
7
+ "content": "<｜end▁of▁sentence｜>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "151644": {
15
+ "content": "<｜User｜>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": false
21
+ },
22
+ "151645": {
23
+ "content": "<｜Assistant｜>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": false
29
+ },
30
+ "151646": {
31
+ "content": "<｜begin▁of▁sentence｜>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false,
36
+ "special": true
37
+ },
38
+ "151647": {
39
+ "content": "<|EOT|>",
40
+ "lstrip": false,
41
+ "normalized": false,
42
+ "rstrip": false,
43
+ "single_word": false,
44
+ "special": false
45
+ },
46
+ "151648": {
47
+ "content": "<think>",
48
+ "lstrip": false,
49
+ "normalized": false,
50
+ "rstrip": false,
51
+ "single_word": false,
52
+ "special": false
53
+ },
54
+ "151649": {
55
+ "content": "</think>",
56
+ "lstrip": false,
57
+ "normalized": false,
58
+ "rstrip": false,
59
+ "single_word": false,
60
+ "special": false
61
+ },
62
+ "151650": {
63
+ "content": "<|quad_start|>",
64
+ "lstrip": false,
65
+ "normalized": false,
66
+ "rstrip": false,
67
+ "single_word": false,
68
+ "special": true
69
+ },
70
+ "151651": {
71
+ "content": "<|quad_end|>",
72
+ "lstrip": false,
73
+ "normalized": false,
74
+ "rstrip": false,
75
+ "single_word": false,
76
+ "special": true
77
+ },
78
+ "151652": {
79
+ "content": "<|vision_start|>",
80
+ "lstrip": false,
81
+ "normalized": false,
82
+ "rstrip": false,
83
+ "single_word": false,
84
+ "special": true
85
+ },
86
+ "151653": {
87
+ "content": "<|vision_end|>",
88
+ "lstrip": false,
89
+ "normalized": false,
90
+ "rstrip": false,
91
+ "single_word": false,
92
+ "special": true
93
+ },
94
+ "151654": {
95
+ "content": "<|vision_pad|>",
96
+ "lstrip": false,
97
+ "normalized": false,
98
+ "rstrip": false,
99
+ "single_word": false,
100
+ "special": true
101
+ },
102
+ "151655": {
103
+ "content": "<|image_pad|>",
104
+ "lstrip": false,
105
+ "normalized": false,
106
+ "rstrip": false,
107
+ "single_word": false,
108
+ "special": true
109
+ },
110
+ "151656": {
111
+ "content": "<|video_pad|>",
112
+ "lstrip": false,
113
+ "normalized": false,
114
+ "rstrip": false,
115
+ "single_word": false,
116
+ "special": true
117
+ },
118
+ "151657": {
119
+ "content": "<tool_call>",
120
+ "lstrip": false,
121
+ "normalized": false,
122
+ "rstrip": false,
123
+ "single_word": false,
124
+ "special": false
125
+ },
126
+ "151658": {
127
+ "content": "</tool_call>",
128
+ "lstrip": false,
129
+ "normalized": false,
130
+ "rstrip": false,
131
+ "single_word": false,
132
+ "special": false
133
+ },
134
+ "151659": {
135
+ "content": "<|fim_prefix|>",
136
+ "lstrip": false,
137
+ "normalized": false,
138
+ "rstrip": false,
139
+ "single_word": false,
140
+ "special": false
141
+ },
142
+ "151660": {
143
+ "content": "<|fim_middle|>",
144
+ "lstrip": false,
145
+ "normalized": false,
146
+ "rstrip": false,
147
+ "single_word": false,
148
+ "special": false
149
+ },
150
+ "151661": {
151
+ "content": "<|fim_suffix|>",
152
+ "lstrip": false,
153
+ "normalized": false,
154
+ "rstrip": false,
155
+ "single_word": false,
156
+ "special": false
157
+ },
158
+ "151662": {
159
+ "content": "<|fim_pad|>",
160
+ "lstrip": false,
161
+ "normalized": false,
162
+ "rstrip": false,
163
+ "single_word": false,
164
+ "special": false
165
+ },
166
+ "151663": {
167
+ "content": "<|repo_name|>",
168
+ "lstrip": false,
169
+ "normalized": false,
170
+ "rstrip": false,
171
+ "single_word": false,
172
+ "special": false
173
+ },
174
+ "151664": {
175
+ "content": "<|file_sep|>",
176
+ "lstrip": false,
177
+ "normalized": false,
178
+ "rstrip": false,
179
+ "single_word": false,
180
+ "special": false
181
+ }
182
+ },
183
+ "bos_token": "<｜begin▁of▁sentence｜>",
184
+ "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='') %}{%- for message in messages %}{%- if message['role'] == 'system' %}{% set ns.system_prompt = message['content'] %}{%- endif %}{%- endfor %}{{bos_token}}{{ns.system_prompt}}{%- for message in messages %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{{'<｜User｜>' + message['content']}}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is none %}{%- set ns.is_tool = false -%}{%- for tool in message['tool_calls']%}{%- if not ns.is_first %}{{'<｜Assistant｜><｜tool▁calls▁begin｜><｜tool▁call▁begin��>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{%- set ns.is_first = true -%}{%- else %}{{'\\n' + '<｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{{'<｜tool▁calls▁end｜><｜end▁of▁sentence｜>'}}{%- endif %}{%- endfor %}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is not none %}{%- if ns.is_tool %}{{'<｜tool▁outputs▁end｜>' + message['content'] + '<｜end▁of▁sentence｜>'}}{%- set ns.is_tool = false -%}{%- else %}{% set content = message['content'] %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{% endif %}{{'<｜Assistant｜>' + content + '<｜end▁of▁sentence｜>'}}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<｜tool▁outputs▁begin｜><｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- set ns.is_output_first = false %}{%- else %}{{'\\n<｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<｜tool▁outputs▁end｜>'}}{% endif %}{% if add_generation_prompt and not ns.is_tool %}{{'<｜Assistant｜><think>\\n'}}{% endif %}",
185
+ "clean_up_tokenization_spaces": false,
186
+ "eos_token": "<｜end▁of▁sentence｜>",
187
+ "extra_special_tokens": {},
188
+ "legacy": true,
189
+ "model_max_length": 16384,
190
+ "pad_token": "<｜end▁of▁sentence｜>",
191
+ "sp_model_kwargs": {},
192
+ "tokenizer_class": "LlamaTokenizerFast",
193
+ "unk_token": null,
194
+ "use_default_system_prompt": false
195
+ }
DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/config.json ADDED
@@ -0,0 +1,3029 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151643,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 5120,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 13824,
12
+ "max_position_embeddings": 131072,
13
+ "max_window_layers": 48,
14
+ "model_type": "qwen2",
15
+ "num_attention_heads": 40,
16
+ "num_hidden_layers": 48,
17
+ "num_key_value_heads": 8,
18
+ "quantization": {
19
+ "group_size": 64,
20
+ "bits": null,
21
+ "model.embed_tokens": {
22
+ "group_size": 64,
23
+ "bits": 3
24
+ },
25
+ "model.layers.0.self_attn.q_proj": {
26
+ "group_size": 64,
27
+ "bits": 3
28
+ },
29
+ "model.layers.0.self_attn.k_proj": {
30
+ "group_size": 64,
31
+ "bits": 3
32
+ },
33
+ "model.layers.0.self_attn.v_proj": {
34
+ "group_size": 64,
35
+ "bits": 4
36
+ },
37
+ "model.layers.0.self_attn.o_proj": {
38
+ "group_size": 64,
39
+ "bits": 3
40
+ },
41
+ "model.layers.0.self_attn.rope": false,
42
+ "model.layers.0.mlp.gate_proj": {
43
+ "group_size": 64,
44
+ "bits": 3
45
+ },
46
+ "model.layers.0.mlp.down_proj": {
47
+ "group_size": 64,
48
+ "bits": 4
49
+ },
50
+ "model.layers.0.mlp.up_proj": {
51
+ "group_size": 64,
52
+ "bits": 3
53
+ },
54
+ "model.layers.0.input_layernorm": false,
55
+ "model.layers.0.post_attention_layernorm": false,
56
+ "model.layers.1.self_attn.q_proj": {
57
+ "group_size": 64,
58
+ "bits": 3
59
+ },
60
+ "model.layers.1.self_attn.k_proj": {
61
+ "group_size": 64,
62
+ "bits": 3
63
+ },
64
+ "model.layers.1.self_attn.v_proj": {
65
+ "group_size": 64,
66
+ "bits": 4
67
+ },
68
+ "model.layers.1.self_attn.o_proj": {
69
+ "group_size": 64,
70
+ "bits": 3
71
+ },
72
+ "model.layers.1.self_attn.rope": false,
73
+ "model.layers.1.mlp.gate_proj": {
74
+ "group_size": 64,
75
+ "bits": 3
76
+ },
77
+ "model.layers.1.mlp.down_proj": {
78
+ "group_size": 64,
79
+ "bits": 4
80
+ },
81
+ "model.layers.1.mlp.up_proj": {
82
+ "group_size": 64,
83
+ "bits": 3
84
+ },
85
+ "model.layers.1.input_layernorm": false,
86
+ "model.layers.1.post_attention_layernorm": false,
87
+ "model.layers.2.self_attn.q_proj": {
88
+ "group_size": 64,
89
+ "bits": 3
90
+ },
91
+ "model.layers.2.self_attn.k_proj": {
92
+ "group_size": 64,
93
+ "bits": 3
94
+ },
95
+ "model.layers.2.self_attn.v_proj": {
96
+ "group_size": 64,
97
+ "bits": 4
98
+ },
99
+ "model.layers.2.self_attn.o_proj": {
100
+ "group_size": 64,
101
+ "bits": 3
102
+ },
103
+ "model.layers.2.self_attn.rope": false,
104
+ "model.layers.2.mlp.gate_proj": {
105
+ "group_size": 64,
106
+ "bits": 3
107
+ },
108
+ "model.layers.2.mlp.down_proj": {
109
+ "group_size": 64,
110
+ "bits": 4
111
+ },
112
+ "model.layers.2.mlp.up_proj": {
113
+ "group_size": 64,
114
+ "bits": 3
115
+ },
116
+ "model.layers.2.input_layernorm": false,
117
+ "model.layers.2.post_attention_layernorm": false,
118
+ "model.layers.3.self_attn.q_proj": {
119
+ "group_size": 64,
120
+ "bits": 3
121
+ },
122
+ "model.layers.3.self_attn.k_proj": {
123
+ "group_size": 64,
124
+ "bits": 3
125
+ },
126
+ "model.layers.3.self_attn.v_proj": {
127
+ "group_size": 64,
128
+ "bits": 4
129
+ },
130
+ "model.layers.3.self_attn.o_proj": {
131
+ "group_size": 64,
132
+ "bits": 3
133
+ },
134
+ "model.layers.3.self_attn.rope": false,
135
+ "model.layers.3.mlp.gate_proj": {
136
+ "group_size": 64,
137
+ "bits": 3
138
+ },
139
+ "model.layers.3.mlp.down_proj": {
140
+ "group_size": 64,
141
+ "bits": 4
142
+ },
143
+ "model.layers.3.mlp.up_proj": {
144
+ "group_size": 64,
145
+ "bits": 3
146
+ },
147
+ "model.layers.3.input_layernorm": false,
148
+ "model.layers.3.post_attention_layernorm": false,
149
+ "model.layers.4.self_attn.q_proj": {
150
+ "group_size": 64,
151
+ "bits": 3
152
+ },
153
+ "model.layers.4.self_attn.k_proj": {
154
+ "group_size": 64,
155
+ "bits": 3
156
+ },
157
+ "model.layers.4.self_attn.v_proj": {
158
+ "group_size": 64,
159
+ "bits": 4
160
+ },
161
+ "model.layers.4.self_attn.o_proj": {
162
+ "group_size": 64,
163
+ "bits": 3
164
+ },
165
+ "model.layers.4.self_attn.rope": false,
166
+ "model.layers.4.mlp.gate_proj": {
167
+ "group_size": 64,
168
+ "bits": 3
169
+ },
170
+ "model.layers.4.mlp.down_proj": {
171
+ "group_size": 64,
172
+ "bits": 4
173
+ },
174
+ "model.layers.4.mlp.up_proj": {
175
+ "group_size": 64,
176
+ "bits": 3
177
+ },
178
+ "model.layers.4.input_layernorm": false,
179
+ "model.layers.4.post_attention_layernorm": false,
180
+ "model.layers.5.self_attn.q_proj": {
181
+ "group_size": 64,
182
+ "bits": 3
183
+ },
184
+ "model.layers.5.self_attn.k_proj": {
185
+ "group_size": 64,
186
+ "bits": 3
187
+ },
188
+ "model.layers.5.self_attn.v_proj": {
189
+ "group_size": 64,
190
+ "bits": 4
191
+ },
192
+ "model.layers.5.self_attn.o_proj": {
193
+ "group_size": 64,
194
+ "bits": 3
195
+ },
196
+ "model.layers.5.self_attn.rope": false,
197
+ "model.layers.5.mlp.gate_proj": {
198
+ "group_size": 64,
199
+ "bits": 3
200
+ },
201
+ "model.layers.5.mlp.down_proj": {
202
+ "group_size": 64,
203
+ "bits": 4
204
+ },
205
+ "model.layers.5.mlp.up_proj": {
206
+ "group_size": 64,
207
+ "bits": 3
208
+ },
209
+ "model.layers.5.input_layernorm": false,
210
+ "model.layers.5.post_attention_layernorm": false,
211
+ "model.layers.6.self_attn.q_proj": {
212
+ "group_size": 64,
213
+ "bits": 3
214
+ },
215
+ "model.layers.6.self_attn.k_proj": {
216
+ "group_size": 64,
217
+ "bits": 3
218
+ },
219
+ "model.layers.6.self_attn.v_proj": {
220
+ "group_size": 64,
221
+ "bits": 3
222
+ },
223
+ "model.layers.6.self_attn.o_proj": {
224
+ "group_size": 64,
225
+ "bits": 3
226
+ },
227
+ "model.layers.6.self_attn.rope": false,
228
+ "model.layers.6.mlp.gate_proj": {
229
+ "group_size": 64,
230
+ "bits": 3
231
+ },
232
+ "model.layers.6.mlp.down_proj": {
233
+ "group_size": 64,
234
+ "bits": 3
235
+ },
236
+ "model.layers.6.mlp.up_proj": {
237
+ "group_size": 64,
238
+ "bits": 3
239
+ },
240
+ "model.layers.6.input_layernorm": false,
241
+ "model.layers.6.post_attention_layernorm": false,
242
+ "model.layers.7.self_attn.q_proj": {
243
+ "group_size": 64,
244
+ "bits": 3
245
+ },
246
+ "model.layers.7.self_attn.k_proj": {
247
+ "group_size": 64,
248
+ "bits": 3
249
+ },
250
+ "model.layers.7.self_attn.v_proj": {
251
+ "group_size": 64,
252
+ "bits": 3
253
+ },
254
+ "model.layers.7.self_attn.o_proj": {
255
+ "group_size": 64,
256
+ "bits": 3
257
+ },
258
+ "model.layers.7.self_attn.rope": false,
259
+ "model.layers.7.mlp.gate_proj": {
260
+ "group_size": 64,
261
+ "bits": 3
262
+ },
263
+ "model.layers.7.mlp.down_proj": {
264
+ "group_size": 64,
265
+ "bits": 3
266
+ },
267
+ "model.layers.7.mlp.up_proj": {
268
+ "group_size": 64,
269
+ "bits": 3
270
+ },
271
+ "model.layers.7.input_layernorm": false,
272
+ "model.layers.7.post_attention_layernorm": false,
273
+ "model.layers.8.self_attn.q_proj": {
274
+ "group_size": 64,
275
+ "bits": 3
276
+ },
277
+ "model.layers.8.self_attn.k_proj": {
278
+ "group_size": 64,
279
+ "bits": 3
280
+ },
281
+ "model.layers.8.self_attn.v_proj": {
282
+ "group_size": 64,
283
+ "bits": 4
284
+ },
285
+ "model.layers.8.self_attn.o_proj": {
286
+ "group_size": 64,
287
+ "bits": 3
288
+ },
289
+ "model.layers.8.self_attn.rope": false,
290
+ "model.layers.8.mlp.gate_proj": {
291
+ "group_size": 64,
292
+ "bits": 3
293
+ },
294
+ "model.layers.8.mlp.down_proj": {
295
+ "group_size": 64,
296
+ "bits": 4
297
+ },
298
+ "model.layers.8.mlp.up_proj": {
299
+ "group_size": 64,
300
+ "bits": 3
301
+ },
302
+ "model.layers.8.input_layernorm": false,
303
+ "model.layers.8.post_attention_layernorm": false,
304
+ "model.layers.9.self_attn.q_proj": {
305
+ "group_size": 64,
306
+ "bits": 3
307
+ },
308
+ "model.layers.9.self_attn.k_proj": {
309
+ "group_size": 64,
310
+ "bits": 3
311
+ },
312
+ "model.layers.9.self_attn.v_proj": {
313
+ "group_size": 64,
314
+ "bits": 3
315
+ },
316
+ "model.layers.9.self_attn.o_proj": {
317
+ "group_size": 64,
318
+ "bits": 3
319
+ },
320
+ "model.layers.9.self_attn.rope": false,
321
+ "model.layers.9.mlp.gate_proj": {
322
+ "group_size": 64,
323
+ "bits": 3
324
+ },
325
+ "model.layers.9.mlp.down_proj": {
326
+ "group_size": 64,
327
+ "bits": 3
328
+ },
329
+ "model.layers.9.mlp.up_proj": {
330
+ "group_size": 64,
331
+ "bits": 3
332
+ },
333
+ "model.layers.9.input_layernorm": false,
334
+ "model.layers.9.post_attention_layernorm": false,
335
+ "model.layers.10.self_attn.q_proj": {
336
+ "group_size": 64,
337
+ "bits": 3
338
+ },
339
+ "model.layers.10.self_attn.k_proj": {
340
+ "group_size": 64,
341
+ "bits": 3
342
+ },
343
+ "model.layers.10.self_attn.v_proj": {
344
+ "group_size": 64,
345
+ "bits": 3
346
+ },
347
+ "model.layers.10.self_attn.o_proj": {
348
+ "group_size": 64,
349
+ "bits": 3
350
+ },
351
+ "model.layers.10.self_attn.rope": false,
352
+ "model.layers.10.mlp.gate_proj": {
353
+ "group_size": 64,
354
+ "bits": 3
355
+ },
356
+ "model.layers.10.mlp.down_proj": {
357
+ "group_size": 64,
358
+ "bits": 3
359
+ },
360
+ "model.layers.10.mlp.up_proj": {
361
+ "group_size": 64,
362
+ "bits": 3
363
+ },
364
+ "model.layers.10.input_layernorm": false,
365
+ "model.layers.10.post_attention_layernorm": false,
366
+ "model.layers.11.self_attn.q_proj": {
367
+ "group_size": 64,
368
+ "bits": 3
369
+ },
370
+ "model.layers.11.self_attn.k_proj": {
371
+ "group_size": 64,
372
+ "bits": 3
373
+ },
374
+ "model.layers.11.self_attn.v_proj": {
375
+ "group_size": 64,
376
+ "bits": 4
377
+ },
378
+ "model.layers.11.self_attn.o_proj": {
379
+ "group_size": 64,
380
+ "bits": 3
381
+ },
382
+ "model.layers.11.self_attn.rope": false,
383
+ "model.layers.11.mlp.gate_proj": {
384
+ "group_size": 64,
385
+ "bits": 3
386
+ },
387
+ "model.layers.11.mlp.down_proj": {
388
+ "group_size": 64,
389
+ "bits": 4
390
+ },
391
+ "model.layers.11.mlp.up_proj": {
392
+ "group_size": 64,
393
+ "bits": 3
394
+ },
395
+ "model.layers.11.input_layernorm": false,
396
+ "model.layers.11.post_attention_layernorm": false,
397
+ "model.layers.12.self_attn.q_proj": {
398
+ "group_size": 64,
399
+ "bits": 3
400
+ },
401
+ "model.layers.12.self_attn.k_proj": {
402
+ "group_size": 64,
403
+ "bits": 3
404
+ },
405
+ "model.layers.12.self_attn.v_proj": {
406
+ "group_size": 64,
407
+ "bits": 3
408
+ },
409
+ "model.layers.12.self_attn.o_proj": {
410
+ "group_size": 64,
411
+ "bits": 3
412
+ },
413
+ "model.layers.12.self_attn.rope": false,
414
+ "model.layers.12.mlp.gate_proj": {
415
+ "group_size": 64,
416
+ "bits": 3
417
+ },
418
+ "model.layers.12.mlp.down_proj": {
419
+ "group_size": 64,
420
+ "bits": 3
421
+ },
422
+ "model.layers.12.mlp.up_proj": {
423
+ "group_size": 64,
424
+ "bits": 3
425
+ },
426
+ "model.layers.12.input_layernorm": false,
427
+ "model.layers.12.post_attention_layernorm": false,
428
+ "model.layers.13.self_attn.q_proj": {
429
+ "group_size": 64,
430
+ "bits": 3
431
+ },
432
+ "model.layers.13.self_attn.k_proj": {
433
+ "group_size": 64,
434
+ "bits": 3
435
+ },
436
+ "model.layers.13.self_attn.v_proj": {
437
+ "group_size": 64,
438
+ "bits": 3
439
+ },
440
+ "model.layers.13.self_attn.o_proj": {
441
+ "group_size": 64,
442
+ "bits": 3
443
+ },
444
+ "model.layers.13.self_attn.rope": false,
445
+ "model.layers.13.mlp.gate_proj": {
446
+ "group_size": 64,
447
+ "bits": 3
448
+ },
449
+ "model.layers.13.mlp.down_proj": {
450
+ "group_size": 64,
451
+ "bits": 3
452
+ },
453
+ "model.layers.13.mlp.up_proj": {
454
+ "group_size": 64,
455
+ "bits": 3
456
+ },
457
+ "model.layers.13.input_layernorm": false,
458
+ "model.layers.13.post_attention_layernorm": false,
459
+ "model.layers.14.self_attn.q_proj": {
460
+ "group_size": 64,
461
+ "bits": 3
462
+ },
463
+ "model.layers.14.self_attn.k_proj": {
464
+ "group_size": 64,
465
+ "bits": 3
466
+ },
467
+ "model.layers.14.self_attn.v_proj": {
468
+ "group_size": 64,
469
+ "bits": 4
470
+ },
471
+ "model.layers.14.self_attn.o_proj": {
472
+ "group_size": 64,
473
+ "bits": 3
474
+ },
475
+ "model.layers.14.self_attn.rope": false,
476
+ "model.layers.14.mlp.gate_proj": {
477
+ "group_size": 64,
478
+ "bits": 3
479
+ },
480
+ "model.layers.14.mlp.down_proj": {
481
+ "group_size": 64,
482
+ "bits": 4
483
+ },
484
+ "model.layers.14.mlp.up_proj": {
485
+ "group_size": 64,
486
+ "bits": 3
487
+ },
488
+ "model.layers.14.input_layernorm": false,
489
+ "model.layers.14.post_attention_layernorm": false,
490
+ "model.layers.15.self_attn.q_proj": {
491
+ "group_size": 64,
492
+ "bits": 3
493
+ },
494
+ "model.layers.15.self_attn.k_proj": {
495
+ "group_size": 64,
496
+ "bits": 3
497
+ },
498
+ "model.layers.15.self_attn.v_proj": {
499
+ "group_size": 64,
500
+ "bits": 3
501
+ },
502
+ "model.layers.15.self_attn.o_proj": {
503
+ "group_size": 64,
504
+ "bits": 3
505
+ },
506
+ "model.layers.15.self_attn.rope": false,
507
+ "model.layers.15.mlp.gate_proj": {
508
+ "group_size": 64,
509
+ "bits": 3
510
+ },
511
+ "model.layers.15.mlp.down_proj": {
512
+ "group_size": 64,
513
+ "bits": 3
514
+ },
515
+ "model.layers.15.mlp.up_proj": {
516
+ "group_size": 64,
517
+ "bits": 3
518
+ },
519
+ "model.layers.15.input_layernorm": false,
520
+ "model.layers.15.post_attention_layernorm": false,
521
+ "model.layers.16.self_attn.q_proj": {
522
+ "group_size": 64,
523
+ "bits": 3
524
+ },
525
+ "model.layers.16.self_attn.k_proj": {
526
+ "group_size": 64,
527
+ "bits": 3
528
+ },
529
+ "model.layers.16.self_attn.v_proj": {
530
+ "group_size": 64,
531
+ "bits": 3
532
+ },
533
+ "model.layers.16.self_attn.o_proj": {
534
+ "group_size": 64,
535
+ "bits": 3
536
+ },
537
+ "model.layers.16.self_attn.rope": false,
538
+ "model.layers.16.mlp.gate_proj": {
539
+ "group_size": 64,
540
+ "bits": 3
541
+ },
542
+ "model.layers.16.mlp.down_proj": {
543
+ "group_size": 64,
544
+ "bits": 3
545
+ },
546
+ "model.layers.16.mlp.up_proj": {
547
+ "group_size": 64,
548
+ "bits": 3
549
+ },
550
+ "model.layers.16.input_layernorm": false,
551
+ "model.layers.16.post_attention_layernorm": false,
552
+ "model.layers.17.self_attn.q_proj": {
553
+ "group_size": 64,
554
+ "bits": 3
555
+ },
556
+ "model.layers.17.self_attn.k_proj": {
557
+ "group_size": 64,
558
+ "bits": 3
559
+ },
560
+ "model.layers.17.self_attn.v_proj": {
561
+ "group_size": 64,
562
+ "bits": 4
563
+ },
564
+ "model.layers.17.self_attn.o_proj": {
565
+ "group_size": 64,
566
+ "bits": 3
567
+ },
568
+ "model.layers.17.self_attn.rope": false,
569
+ "model.layers.17.mlp.gate_proj": {
570
+ "group_size": 64,
571
+ "bits": 3
572
+ },
573
+ "model.layers.17.mlp.down_proj": {
574
+ "group_size": 64,
575
+ "bits": 4
576
+ },
577
+ "model.layers.17.mlp.up_proj": {
578
+ "group_size": 64,
579
+ "bits": 3
580
+ },
581
+ "model.layers.17.input_layernorm": false,
582
+ "model.layers.17.post_attention_layernorm": false,
583
+ "model.layers.18.self_attn.q_proj": {
584
+ "group_size": 64,
585
+ "bits": 3
586
+ },
587
+ "model.layers.18.self_attn.k_proj": {
588
+ "group_size": 64,
589
+ "bits": 3
590
+ },
591
+ "model.layers.18.self_attn.v_proj": {
592
+ "group_size": 64,
593
+ "bits": 3
594
+ },
595
+ "model.layers.18.self_attn.o_proj": {
596
+ "group_size": 64,
597
+ "bits": 3
598
+ },
599
+ "model.layers.18.self_attn.rope": false,
600
+ "model.layers.18.mlp.gate_proj": {
601
+ "group_size": 64,
602
+ "bits": 3
603
+ },
604
+ "model.layers.18.mlp.down_proj": {
605
+ "group_size": 64,
606
+ "bits": 3
607
+ },
608
+ "model.layers.18.mlp.up_proj": {
609
+ "group_size": 64,
610
+ "bits": 3
611
+ },
612
+ "model.layers.18.input_layernorm": false,
613
+ "model.layers.18.post_attention_layernorm": false,
614
+ "model.layers.19.self_attn.q_proj": {
615
+ "group_size": 64,
616
+ "bits": 3
617
+ },
618
+ "model.layers.19.self_attn.k_proj": {
619
+ "group_size": 64,
620
+ "bits": 3
621
+ },
622
+ "model.layers.19.self_attn.v_proj": {
623
+ "group_size": 64,
624
+ "bits": 3
625
+ },
626
+ "model.layers.19.self_attn.o_proj": {
627
+ "group_size": 64,
628
+ "bits": 3
629
+ },
630
+ "model.layers.19.self_attn.rope": false,
631
+ "model.layers.19.mlp.gate_proj": {
632
+ "group_size": 64,
633
+ "bits": 3
634
+ },
635
+ "model.layers.19.mlp.down_proj": {
636
+ "group_size": 64,
637
+ "bits": 3
638
+ },
639
+ "model.layers.19.mlp.up_proj": {
640
+ "group_size": 64,
641
+ "bits": 3
642
+ },
643
+ "model.layers.19.input_layernorm": false,
644
+ "model.layers.19.post_attention_layernorm": false,
645
+ "model.layers.20.self_attn.q_proj": {
646
+ "group_size": 64,
647
+ "bits": 3
648
+ },
649
+ "model.layers.20.self_attn.k_proj": {
650
+ "group_size": 64,
651
+ "bits": 3
652
+ },
653
+ "model.layers.20.self_attn.v_proj": {
654
+ "group_size": 64,
655
+ "bits": 4
656
+ },
657
+ "model.layers.20.self_attn.o_proj": {
658
+ "group_size": 64,
659
+ "bits": 3
660
+ },
661
+ "model.layers.20.self_attn.rope": false,
662
+ "model.layers.20.mlp.gate_proj": {
663
+ "group_size": 64,
664
+ "bits": 3
665
+ },
666
+ "model.layers.20.mlp.down_proj": {
667
+ "group_size": 64,
668
+ "bits": 4
669
+ },
670
+ "model.layers.20.mlp.up_proj": {
671
+ "group_size": 64,
672
+ "bits": 3
673
+ },
674
+ "model.layers.20.input_layernorm": false,
675
+ "model.layers.20.post_attention_layernorm": false,
676
+ "model.layers.21.self_attn.q_proj": {
677
+ "group_size": 64,
678
+ "bits": 3
679
+ },
680
+ "model.layers.21.self_attn.k_proj": {
681
+ "group_size": 64,
682
+ "bits": 3
683
+ },
684
+ "model.layers.21.self_attn.v_proj": {
685
+ "group_size": 64,
686
+ "bits": 3
687
+ },
688
+ "model.layers.21.self_attn.o_proj": {
689
+ "group_size": 64,
690
+ "bits": 3
691
+ },
692
+ "model.layers.21.self_attn.rope": false,
693
+ "model.layers.21.mlp.gate_proj": {
694
+ "group_size": 64,
695
+ "bits": 3
696
+ },
697
+ "model.layers.21.mlp.down_proj": {
698
+ "group_size": 64,
699
+ "bits": 3
700
+ },
701
+ "model.layers.21.mlp.up_proj": {
702
+ "group_size": 64,
703
+ "bits": 3
704
+ },
705
+ "model.layers.21.input_layernorm": false,
706
+ "model.layers.21.post_attention_layernorm": false,
707
+ "model.layers.22.self_attn.q_proj": {
708
+ "group_size": 64,
709
+ "bits": 3
710
+ },
711
+ "model.layers.22.self_attn.k_proj": {
712
+ "group_size": 64,
713
+ "bits": 3
714
+ },
715
+ "model.layers.22.self_attn.v_proj": {
716
+ "group_size": 64,
717
+ "bits": 3
718
+ },
719
+ "model.layers.22.self_attn.o_proj": {
720
+ "group_size": 64,
721
+ "bits": 3
722
+ },
723
+ "model.layers.22.self_attn.rope": false,
724
+ "model.layers.22.mlp.gate_proj": {
725
+ "group_size": 64,
726
+ "bits": 3
727
+ },
728
+ "model.layers.22.mlp.down_proj": {
729
+ "group_size": 64,
730
+ "bits": 3
731
+ },
732
+ "model.layers.22.mlp.up_proj": {
733
+ "group_size": 64,
734
+ "bits": 3
735
+ },
736
+ "model.layers.22.input_layernorm": false,
737
+ "model.layers.22.post_attention_layernorm": false,
738
+ "model.layers.23.self_attn.q_proj": {
739
+ "group_size": 64,
740
+ "bits": 3
741
+ },
742
+ "model.layers.23.self_attn.k_proj": {
743
+ "group_size": 64,
744
+ "bits": 3
745
+ },
746
+ "model.layers.23.self_attn.v_proj": {
747
+ "group_size": 64,
748
+ "bits": 4
749
+ },
750
+ "model.layers.23.self_attn.o_proj": {
751
+ "group_size": 64,
752
+ "bits": 3
753
+ },
754
+ "model.layers.23.self_attn.rope": false,
755
+ "model.layers.23.mlp.gate_proj": {
756
+ "group_size": 64,
757
+ "bits": 3
758
+ },
759
+ "model.layers.23.mlp.down_proj": {
760
+ "group_size": 64,
761
+ "bits": 4
762
+ },
763
+ "model.layers.23.mlp.up_proj": {
764
+ "group_size": 64,
765
+ "bits": 3
766
+ },
767
+ "model.layers.23.input_layernorm": false,
768
+ "model.layers.23.post_attention_layernorm": false,
769
+ "model.layers.24.self_attn.q_proj": {
770
+ "group_size": 64,
771
+ "bits": 3
772
+ },
773
+ "model.layers.24.self_attn.k_proj": {
774
+ "group_size": 64,
775
+ "bits": 3
776
+ },
777
+ "model.layers.24.self_attn.v_proj": {
778
+ "group_size": 64,
779
+ "bits": 3
780
+ },
781
+ "model.layers.24.self_attn.o_proj": {
782
+ "group_size": 64,
783
+ "bits": 3
784
+ },
785
+ "model.layers.24.self_attn.rope": false,
786
+ "model.layers.24.mlp.gate_proj": {
787
+ "group_size": 64,
788
+ "bits": 3
789
+ },
790
+ "model.layers.24.mlp.down_proj": {
791
+ "group_size": 64,
792
+ "bits": 3
793
+ },
794
+ "model.layers.24.mlp.up_proj": {
795
+ "group_size": 64,
796
+ "bits": 3
797
+ },
798
+ "model.layers.24.input_layernorm": false,
799
+ "model.layers.24.post_attention_layernorm": false,
800
+ "model.layers.25.self_attn.q_proj": {
801
+ "group_size": 64,
802
+ "bits": 3
803
+ },
804
+ "model.layers.25.self_attn.k_proj": {
805
+ "group_size": 64,
806
+ "bits": 3
807
+ },
808
+ "model.layers.25.self_attn.v_proj": {
809
+ "group_size": 64,
810
+ "bits": 3
811
+ },
812
+ "model.layers.25.self_attn.o_proj": {
813
+ "group_size": 64,
814
+ "bits": 3
815
+ },
816
+ "model.layers.25.self_attn.rope": false,
817
+ "model.layers.25.mlp.gate_proj": {
818
+ "group_size": 64,
819
+ "bits": 3
820
+ },
821
+ "model.layers.25.mlp.down_proj": {
822
+ "group_size": 64,
823
+ "bits": 3
824
+ },
825
+ "model.layers.25.mlp.up_proj": {
826
+ "group_size": 64,
827
+ "bits": 3
828
+ },
829
+ "model.layers.25.input_layernorm": false,
830
+ "model.layers.25.post_attention_layernorm": false,
831
+ "model.layers.26.self_attn.q_proj": {
832
+ "group_size": 64,
833
+ "bits": 3
834
+ },
835
+ "model.layers.26.self_attn.k_proj": {
836
+ "group_size": 64,
837
+ "bits": 3
838
+ },
839
+ "model.layers.26.self_attn.v_proj": {
840
+ "group_size": 64,
841
+ "bits": 4
842
+ },
843
+ "model.layers.26.self_attn.o_proj": {
844
+ "group_size": 64,
845
+ "bits": 3
846
+ },
847
+ "model.layers.26.self_attn.rope": false,
848
+ "model.layers.26.mlp.gate_proj": {
849
+ "group_size": 64,
850
+ "bits": 3
851
+ },
852
+ "model.layers.26.mlp.down_proj": {
853
+ "group_size": 64,
854
+ "bits": 4
855
+ },
856
+ "model.layers.26.mlp.up_proj": {
857
+ "group_size": 64,
858
+ "bits": 3
859
+ },
860
+ "model.layers.26.input_layernorm": false,
861
+ "model.layers.26.post_attention_layernorm": false,
862
+ "model.layers.27.self_attn.q_proj": {
863
+ "group_size": 64,
864
+ "bits": 3
865
+ },
866
+ "model.layers.27.self_attn.k_proj": {
867
+ "group_size": 64,
868
+ "bits": 3
869
+ },
870
+ "model.layers.27.self_attn.v_proj": {
871
+ "group_size": 64,
872
+ "bits": 3
873
+ },
874
+ "model.layers.27.self_attn.o_proj": {
875
+ "group_size": 64,
876
+ "bits": 3
877
+ },
878
+ "model.layers.27.self_attn.rope": false,
879
+ "model.layers.27.mlp.gate_proj": {
880
+ "group_size": 64,
881
+ "bits": 3
882
+ },
883
+ "model.layers.27.mlp.down_proj": {
884
+ "group_size": 64,
885
+ "bits": 3
886
+ },
887
+ "model.layers.27.mlp.up_proj": {
888
+ "group_size": 64,
889
+ "bits": 3
890
+ },
891
+ "model.layers.27.input_layernorm": false,
892
+ "model.layers.27.post_attention_layernorm": false,
893
+ "model.layers.28.self_attn.q_proj": {
894
+ "group_size": 64,
895
+ "bits": 3
896
+ },
897
+ "model.layers.28.self_attn.k_proj": {
898
+ "group_size": 64,
899
+ "bits": 3
900
+ },
901
+ "model.layers.28.self_attn.v_proj": {
902
+ "group_size": 64,
903
+ "bits": 3
904
+ },
905
+ "model.layers.28.self_attn.o_proj": {
906
+ "group_size": 64,
907
+ "bits": 3
908
+ },
909
+ "model.layers.28.self_attn.rope": false,
910
+ "model.layers.28.mlp.gate_proj": {
911
+ "group_size": 64,
912
+ "bits": 3
913
+ },
914
+ "model.layers.28.mlp.down_proj": {
915
+ "group_size": 64,
916
+ "bits": 3
917
+ },
918
+ "model.layers.28.mlp.up_proj": {
919
+ "group_size": 64,
920
+ "bits": 3
921
+ },
922
+ "model.layers.28.input_layernorm": false,
923
+ "model.layers.28.post_attention_layernorm": false,
924
+ "model.layers.29.self_attn.q_proj": {
925
+ "group_size": 64,
926
+ "bits": 3
927
+ },
928
+ "model.layers.29.self_attn.k_proj": {
929
+ "group_size": 64,
930
+ "bits": 3
931
+ },
932
+ "model.layers.29.self_attn.v_proj": {
933
+ "group_size": 64,
934
+ "bits": 4
935
+ },
936
+ "model.layers.29.self_attn.o_proj": {
937
+ "group_size": 64,
938
+ "bits": 3
939
+ },
940
+ "model.layers.29.self_attn.rope": false,
941
+ "model.layers.29.mlp.gate_proj": {
942
+ "group_size": 64,
943
+ "bits": 3
944
+ },
945
+ "model.layers.29.mlp.down_proj": {
946
+ "group_size": 64,
947
+ "bits": 4
948
+ },
949
+ "model.layers.29.mlp.up_proj": {
950
+ "group_size": 64,
951
+ "bits": 3
952
+ },
953
+ "model.layers.29.input_layernorm": false,
954
+ "model.layers.29.post_attention_layernorm": false,
955
+ "model.layers.30.self_attn.q_proj": {
956
+ "group_size": 64,
957
+ "bits": 3
958
+ },
959
+ "model.layers.30.self_attn.k_proj": {
960
+ "group_size": 64,
961
+ "bits": 3
962
+ },
963
+ "model.layers.30.self_attn.v_proj": {
964
+ "group_size": 64,
965
+ "bits": 3
966
+ },
967
+ "model.layers.30.self_attn.o_proj": {
968
+ "group_size": 64,
969
+ "bits": 3
970
+ },
971
+ "model.layers.30.self_attn.rope": false,
972
+ "model.layers.30.mlp.gate_proj": {
973
+ "group_size": 64,
974
+ "bits": 3
975
+ },
976
+ "model.layers.30.mlp.down_proj": {
977
+ "group_size": 64,
978
+ "bits": 3
979
+ },
980
+ "model.layers.30.mlp.up_proj": {
981
+ "group_size": 64,
982
+ "bits": 3
983
+ },
984
+ "model.layers.30.input_layernorm": false,
985
+ "model.layers.30.post_attention_layernorm": false,
986
+ "model.layers.31.self_attn.q_proj": {
987
+ "group_size": 64,
988
+ "bits": 3
989
+ },
990
+ "model.layers.31.self_attn.k_proj": {
991
+ "group_size": 64,
992
+ "bits": 3
993
+ },
994
+ "model.layers.31.self_attn.v_proj": {
995
+ "group_size": 64,
996
+ "bits": 3
997
+ },
998
+ "model.layers.31.self_attn.o_proj": {
999
+ "group_size": 64,
1000
+ "bits": 3
1001
+ },
1002
+ "model.layers.31.self_attn.rope": false,
1003
+ "model.layers.31.mlp.gate_proj": {
1004
+ "group_size": 64,
1005
+ "bits": 3
1006
+ },
1007
+ "model.layers.31.mlp.down_proj": {
1008
+ "group_size": 64,
1009
+ "bits": 3
1010
+ },
1011
+ "model.layers.31.mlp.up_proj": {
1012
+ "group_size": 64,
1013
+ "bits": 3
1014
+ },
1015
+ "model.layers.31.input_layernorm": false,
1016
+ "model.layers.31.post_attention_layernorm": false,
1017
+ "model.layers.32.self_attn.q_proj": {
1018
+ "group_size": 64,
1019
+ "bits": 3
1020
+ },
1021
+ "model.layers.32.self_attn.k_proj": {
1022
+ "group_size": 64,
1023
+ "bits": 3
1024
+ },
1025
+ "model.layers.32.self_attn.v_proj": {
1026
+ "group_size": 64,
1027
+ "bits": 4
1028
+ },
1029
+ "model.layers.32.self_attn.o_proj": {
1030
+ "group_size": 64,
1031
+ "bits": 3
1032
+ },
1033
+ "model.layers.32.self_attn.rope": false,
1034
+ "model.layers.32.mlp.gate_proj": {
1035
+ "group_size": 64,
1036
+ "bits": 3
1037
+ },
1038
+ "model.layers.32.mlp.down_proj": {
1039
+ "group_size": 64,
1040
+ "bits": 4
1041
+ },
1042
+ "model.layers.32.mlp.up_proj": {
1043
+ "group_size": 64,
1044
+ "bits": 3
1045
+ },
1046
+ "model.layers.32.input_layernorm": false,
1047
+ "model.layers.32.post_attention_layernorm": false,
1048
+ "model.layers.33.self_attn.q_proj": {
1049
+ "group_size": 64,
1050
+ "bits": 3
1051
+ },
1052
+ "model.layers.33.self_attn.k_proj": {
1053
+ "group_size": 64,
1054
+ "bits": 3
1055
+ },
1056
+ "model.layers.33.self_attn.v_proj": {
1057
+ "group_size": 64,
1058
+ "bits": 3
1059
+ },
1060
+ "model.layers.33.self_attn.o_proj": {
1061
+ "group_size": 64,
1062
+ "bits": 3
1063
+ },
1064
+ "model.layers.33.self_attn.rope": false,
1065
+ "model.layers.33.mlp.gate_proj": {
1066
+ "group_size": 64,
1067
+ "bits": 3
1068
+ },
1069
+ "model.layers.33.mlp.down_proj": {
1070
+ "group_size": 64,
1071
+ "bits": 3
1072
+ },
1073
+ "model.layers.33.mlp.up_proj": {
1074
+ "group_size": 64,
1075
+ "bits": 3
1076
+ },
1077
+ "model.layers.33.input_layernorm": false,
1078
+ "model.layers.33.post_attention_layernorm": false,
1079
+ "model.layers.34.self_attn.q_proj": {
1080
+ "group_size": 64,
1081
+ "bits": 3
1082
+ },
1083
+ "model.layers.34.self_attn.k_proj": {
1084
+ "group_size": 64,
1085
+ "bits": 3
1086
+ },
1087
+ "model.layers.34.self_attn.v_proj": {
1088
+ "group_size": 64,
1089
+ "bits": 3
1090
+ },
1091
+ "model.layers.34.self_attn.o_proj": {
1092
+ "group_size": 64,
1093
+ "bits": 3
1094
+ },
1095
+ "model.layers.34.self_attn.rope": false,
1096
+ "model.layers.34.mlp.gate_proj": {
1097
+ "group_size": 64,
1098
+ "bits": 3
1099
+ },
1100
+ "model.layers.34.mlp.down_proj": {
1101
+ "group_size": 64,
1102
+ "bits": 3
1103
+ },
1104
+ "model.layers.34.mlp.up_proj": {
1105
+ "group_size": 64,
1106
+ "bits": 3
1107
+ },
1108
+ "model.layers.34.input_layernorm": false,
1109
+ "model.layers.34.post_attention_layernorm": false,
1110
+ "model.layers.35.self_attn.q_proj": {
1111
+ "group_size": 64,
1112
+ "bits": 3
1113
+ },
1114
+ "model.layers.35.self_attn.k_proj": {
1115
+ "group_size": 64,
1116
+ "bits": 3
1117
+ },
1118
+ "model.layers.35.self_attn.v_proj": {
1119
+ "group_size": 64,
1120
+ "bits": 4
1121
+ },
1122
+ "model.layers.35.self_attn.o_proj": {
1123
+ "group_size": 64,
1124
+ "bits": 3
1125
+ },
1126
+ "model.layers.35.self_attn.rope": false,
1127
+ "model.layers.35.mlp.gate_proj": {
1128
+ "group_size": 64,
1129
+ "bits": 3
1130
+ },
1131
+ "model.layers.35.mlp.down_proj": {
1132
+ "group_size": 64,
1133
+ "bits": 4
1134
+ },
1135
+ "model.layers.35.mlp.up_proj": {
1136
+ "group_size": 64,
1137
+ "bits": 3
1138
+ },
1139
+ "model.layers.35.input_layernorm": false,
1140
+ "model.layers.35.post_attention_layernorm": false,
1141
+ "model.layers.36.self_attn.q_proj": {
1142
+ "group_size": 64,
1143
+ "bits": 3
1144
+ },
1145
+ "model.layers.36.self_attn.k_proj": {
1146
+ "group_size": 64,
1147
+ "bits": 3
1148
+ },
1149
+ "model.layers.36.self_attn.v_proj": {
1150
+ "group_size": 64,
1151
+ "bits": 3
1152
+ },
1153
+ "model.layers.36.self_attn.o_proj": {
1154
+ "group_size": 64,
1155
+ "bits": 3
1156
+ },
1157
+ "model.layers.36.self_attn.rope": false,
1158
+ "model.layers.36.mlp.gate_proj": {
1159
+ "group_size": 64,
1160
+ "bits": 3
1161
+ },
1162
+ "model.layers.36.mlp.down_proj": {
1163
+ "group_size": 64,
1164
+ "bits": 3
1165
+ },
1166
+ "model.layers.36.mlp.up_proj": {
1167
+ "group_size": 64,
1168
+ "bits": 3
1169
+ },
1170
+ "model.layers.36.input_layernorm": false,
1171
+ "model.layers.36.post_attention_layernorm": false,
1172
+ "model.layers.37.self_attn.q_proj": {
1173
+ "group_size": 64,
1174
+ "bits": 3
1175
+ },
1176
+ "model.layers.37.self_attn.k_proj": {
1177
+ "group_size": 64,
1178
+ "bits": 3
1179
+ },
1180
+ "model.layers.37.self_attn.v_proj": {
1181
+ "group_size": 64,
1182
+ "bits": 3
1183
+ },
1184
+ "model.layers.37.self_attn.o_proj": {
1185
+ "group_size": 64,
1186
+ "bits": 3
1187
+ },
1188
+ "model.layers.37.self_attn.rope": false,
1189
+ "model.layers.37.mlp.gate_proj": {
1190
+ "group_size": 64,
1191
+ "bits": 3
1192
+ },
1193
+ "model.layers.37.mlp.down_proj": {
1194
+ "group_size": 64,
1195
+ "bits": 3
1196
+ },
1197
+ "model.layers.37.mlp.up_proj": {
1198
+ "group_size": 64,
1199
+ "bits": 3
1200
+ },
1201
+ "model.layers.37.input_layernorm": false,
1202
+ "model.layers.37.post_attention_layernorm": false,
1203
+ "model.layers.38.self_attn.q_proj": {
1204
+ "group_size": 64,
1205
+ "bits": 3
1206
+ },
1207
+ "model.layers.38.self_attn.k_proj": {
1208
+ "group_size": 64,
1209
+ "bits": 3
1210
+ },
1211
+ "model.layers.38.self_attn.v_proj": {
1212
+ "group_size": 64,
1213
+ "bits": 4
1214
+ },
1215
+ "model.layers.38.self_attn.o_proj": {
1216
+ "group_size": 64,
1217
+ "bits": 3
1218
+ },
1219
+ "model.layers.38.self_attn.rope": false,
1220
+ "model.layers.38.mlp.gate_proj": {
1221
+ "group_size": 64,
1222
+ "bits": 3
1223
+ },
1224
+ "model.layers.38.mlp.down_proj": {
1225
+ "group_size": 64,
1226
+ "bits": 4
1227
+ },
1228
+ "model.layers.38.mlp.up_proj": {
1229
+ "group_size": 64,
1230
+ "bits": 3
1231
+ },
1232
+ "model.layers.38.input_layernorm": false,
1233
+ "model.layers.38.post_attention_layernorm": false,
1234
+ "model.layers.39.self_attn.q_proj": {
1235
+ "group_size": 64,
1236
+ "bits": 3
1237
+ },
1238
+ "model.layers.39.self_attn.k_proj": {
1239
+ "group_size": 64,
1240
+ "bits": 3
1241
+ },
1242
+ "model.layers.39.self_attn.v_proj": {
1243
+ "group_size": 64,
1244
+ "bits": 3
1245
+ },
1246
+ "model.layers.39.self_attn.o_proj": {
1247
+ "group_size": 64,
1248
+ "bits": 3
1249
+ },
1250
+ "model.layers.39.self_attn.rope": false,
1251
+ "model.layers.39.mlp.gate_proj": {
1252
+ "group_size": 64,
1253
+ "bits": 3
1254
+ },
1255
+ "model.layers.39.mlp.down_proj": {
1256
+ "group_size": 64,
1257
+ "bits": 3
1258
+ },
1259
+ "model.layers.39.mlp.up_proj": {
1260
+ "group_size": 64,
1261
+ "bits": 3
1262
+ },
1263
+ "model.layers.39.input_layernorm": false,
1264
+ "model.layers.39.post_attention_layernorm": false,
1265
+ "model.layers.40.self_attn.q_proj": {
1266
+ "group_size": 64,
1267
+ "bits": 3
1268
+ },
1269
+ "model.layers.40.self_attn.k_proj": {
1270
+ "group_size": 64,
1271
+ "bits": 3
1272
+ },
1273
+ "model.layers.40.self_attn.v_proj": {
1274
+ "group_size": 64,
1275
+ "bits": 3
1276
+ },
1277
+ "model.layers.40.self_attn.o_proj": {
1278
+ "group_size": 64,
1279
+ "bits": 3
1280
+ },
1281
+ "model.layers.40.self_attn.rope": false,
1282
+ "model.layers.40.mlp.gate_proj": {
1283
+ "group_size": 64,
1284
+ "bits": 3
1285
+ },
1286
+ "model.layers.40.mlp.down_proj": {
1287
+ "group_size": 64,
1288
+ "bits": 3
1289
+ },
1290
+ "model.layers.40.mlp.up_proj": {
1291
+ "group_size": 64,
1292
+ "bits": 3
1293
+ },
1294
+ "model.layers.40.input_layernorm": false,
1295
+ "model.layers.40.post_attention_layernorm": false,
1296
+ "model.layers.41.self_attn.q_proj": {
1297
+ "group_size": 64,
1298
+ "bits": 3
1299
+ },
1300
+ "model.layers.41.self_attn.k_proj": {
1301
+ "group_size": 64,
1302
+ "bits": 3
1303
+ },
1304
+ "model.layers.41.self_attn.v_proj": {
1305
+ "group_size": 64,
1306
+ "bits": 4
1307
+ },
1308
+ "model.layers.41.self_attn.o_proj": {
1309
+ "group_size": 64,
1310
+ "bits": 3
1311
+ },
1312
+ "model.layers.41.self_attn.rope": false,
1313
+ "model.layers.41.mlp.gate_proj": {
1314
+ "group_size": 64,
1315
+ "bits": 3
1316
+ },
1317
+ "model.layers.41.mlp.down_proj": {
1318
+ "group_size": 64,
1319
+ "bits": 4
1320
+ },
1321
+ "model.layers.41.mlp.up_proj": {
1322
+ "group_size": 64,
1323
+ "bits": 3
1324
+ },
1325
+ "model.layers.41.input_layernorm": false,
1326
+ "model.layers.41.post_attention_layernorm": false,
1327
+ "model.layers.42.self_attn.q_proj": {
1328
+ "group_size": 64,
1329
+ "bits": 3
1330
+ },
1331
+ "model.layers.42.self_attn.k_proj": {
1332
+ "group_size": 64,
1333
+ "bits": 3
1334
+ },
1335
+ "model.layers.42.self_attn.v_proj": {
1336
+ "group_size": 64,
1337
+ "bits": 4
1338
+ },
1339
+ "model.layers.42.self_attn.o_proj": {
1340
+ "group_size": 64,
1341
+ "bits": 3
1342
+ },
1343
+ "model.layers.42.self_attn.rope": false,
1344
+ "model.layers.42.mlp.gate_proj": {
1345
+ "group_size": 64,
1346
+ "bits": 3
1347
+ },
1348
+ "model.layers.42.mlp.down_proj": {
1349
+ "group_size": 64,
1350
+ "bits": 4
1351
+ },
1352
+ "model.layers.42.mlp.up_proj": {
1353
+ "group_size": 64,
1354
+ "bits": 3
1355
+ },
1356
+ "model.layers.42.input_layernorm": false,
1357
+ "model.layers.42.post_attention_layernorm": false,
1358
+ "model.layers.43.self_attn.q_proj": {
1359
+ "group_size": 64,
1360
+ "bits": 3
1361
+ },
1362
+ "model.layers.43.self_attn.k_proj": {
1363
+ "group_size": 64,
1364
+ "bits": 3
1365
+ },
1366
+ "model.layers.43.self_attn.v_proj": {
1367
+ "group_size": 64,
1368
+ "bits": 4
1369
+ },
1370
+ "model.layers.43.self_attn.o_proj": {
1371
+ "group_size": 64,
1372
+ "bits": 3
1373
+ },
1374
+ "model.layers.43.self_attn.rope": false,
1375
+ "model.layers.43.mlp.gate_proj": {
1376
+ "group_size": 64,
1377
+ "bits": 3
1378
+ },
1379
+ "model.layers.43.mlp.down_proj": {
1380
+ "group_size": 64,
1381
+ "bits": 4
1382
+ },
1383
+ "model.layers.43.mlp.up_proj": {
1384
+ "group_size": 64,
1385
+ "bits": 3
1386
+ },
1387
+ "model.layers.43.input_layernorm": false,
1388
+ "model.layers.43.post_attention_layernorm": false,
1389
+ "model.layers.44.self_attn.q_proj": {
1390
+ "group_size": 64,
1391
+ "bits": 3
1392
+ },
1393
+ "model.layers.44.self_attn.k_proj": {
1394
+ "group_size": 64,
1395
+ "bits": 3
1396
+ },
1397
+ "model.layers.44.self_attn.v_proj": {
1398
+ "group_size": 64,
1399
+ "bits": 4
1400
+ },
1401
+ "model.layers.44.self_attn.o_proj": {
1402
+ "group_size": 64,
1403
+ "bits": 3
1404
+ },
1405
+ "model.layers.44.self_attn.rope": false,
1406
+ "model.layers.44.mlp.gate_proj": {
1407
+ "group_size": 64,
1408
+ "bits": 3
1409
+ },
1410
+ "model.layers.44.mlp.down_proj": {
1411
+ "group_size": 64,
1412
+ "bits": 4
1413
+ },
1414
+ "model.layers.44.mlp.up_proj": {
1415
+ "group_size": 64,
1416
+ "bits": 3
1417
+ },
1418
+ "model.layers.44.input_layernorm": false,
1419
+ "model.layers.44.post_attention_layernorm": false,
1420
+ "model.layers.45.self_attn.q_proj": {
1421
+ "group_size": 64,
1422
+ "bits": 3
1423
+ },
1424
+ "model.layers.45.self_attn.k_proj": {
1425
+ "group_size": 64,
1426
+ "bits": 3
1427
+ },
1428
+ "model.layers.45.self_attn.v_proj": {
1429
+ "group_size": 64,
1430
+ "bits": 4
1431
+ },
1432
+ "model.layers.45.self_attn.o_proj": {
1433
+ "group_size": 64,
1434
+ "bits": 3
1435
+ },
1436
+ "model.layers.45.self_attn.rope": false,
1437
+ "model.layers.45.mlp.gate_proj": {
1438
+ "group_size": 64,
1439
+ "bits": 3
1440
+ },
1441
+ "model.layers.45.mlp.down_proj": {
1442
+ "group_size": 64,
1443
+ "bits": 4
1444
+ },
1445
+ "model.layers.45.mlp.up_proj": {
1446
+ "group_size": 64,
1447
+ "bits": 3
1448
+ },
1449
+ "model.layers.45.input_layernorm": false,
1450
+ "model.layers.45.post_attention_layernorm": false,
1451
+ "model.layers.46.self_attn.q_proj": {
1452
+ "group_size": 64,
1453
+ "bits": 3
1454
+ },
1455
+ "model.layers.46.self_attn.k_proj": {
1456
+ "group_size": 64,
1457
+ "bits": 3
1458
+ },
1459
+ "model.layers.46.self_attn.v_proj": {
1460
+ "group_size": 64,
1461
+ "bits": 4
1462
+ },
1463
+ "model.layers.46.self_attn.o_proj": {
1464
+ "group_size": 64,
1465
+ "bits": 3
1466
+ },
1467
+ "model.layers.46.self_attn.rope": false,
1468
+ "model.layers.46.mlp.gate_proj": {
1469
+ "group_size": 64,
1470
+ "bits": 3
1471
+ },
1472
+ "model.layers.46.mlp.down_proj": {
1473
+ "group_size": 64,
1474
+ "bits": 4
1475
+ },
1476
+ "model.layers.46.mlp.up_proj": {
1477
+ "group_size": 64,
1478
+ "bits": 3
1479
+ },
1480
+ "model.layers.46.input_layernorm": false,
1481
+ "model.layers.46.post_attention_layernorm": false,
1482
+ "model.layers.47.self_attn.q_proj": {
1483
+ "group_size": 64,
1484
+ "bits": 3
1485
+ },
1486
+ "model.layers.47.self_attn.k_proj": {
1487
+ "group_size": 64,
1488
+ "bits": 3
1489
+ },
1490
+ "model.layers.47.self_attn.v_proj": {
1491
+ "group_size": 64,
1492
+ "bits": 4
1493
+ },
1494
+ "model.layers.47.self_attn.o_proj": {
1495
+ "group_size": 64,
1496
+ "bits": 3
1497
+ },
1498
+ "model.layers.47.self_attn.rope": false,
1499
+ "model.layers.47.mlp.gate_proj": {
1500
+ "group_size": 64,
1501
+ "bits": 3
1502
+ },
1503
+ "model.layers.47.mlp.down_proj": {
1504
+ "group_size": 64,
1505
+ "bits": 4
1506
+ },
1507
+ "model.layers.47.mlp.up_proj": {
1508
+ "group_size": 64,
1509
+ "bits": 3
1510
+ },
1511
+ "model.layers.47.input_layernorm": false,
1512
+ "model.layers.47.post_attention_layernorm": false,
1513
+ "model.norm": false,
1514
+ "lm_head": {
1515
+ "group_size": 64,
1516
+ "bits": 4
1517
+ }
1518
+ },
1519
+ "quantization_config": {
1520
+ "group_size": 64,
1521
+ "bits": null,
1522
+ "model.embed_tokens": {
1523
+ "group_size": 64,
1524
+ "bits": 3
1525
+ },
1526
+ "model.layers.0.self_attn.q_proj": {
1527
+ "group_size": 64,
1528
+ "bits": 3
1529
+ },
1530
+ "model.layers.0.self_attn.k_proj": {
1531
+ "group_size": 64,
1532
+ "bits": 3
1533
+ },
1534
+ "model.layers.0.self_attn.v_proj": {
1535
+ "group_size": 64,
1536
+ "bits": 4
1537
+ },
1538
+ "model.layers.0.self_attn.o_proj": {
1539
+ "group_size": 64,
1540
+ "bits": 3
1541
+ },
1542
+ "model.layers.0.self_attn.rope": false,
1543
+ "model.layers.0.mlp.gate_proj": {
1544
+ "group_size": 64,
1545
+ "bits": 3
1546
+ },
1547
+ "model.layers.0.mlp.down_proj": {
1548
+ "group_size": 64,
1549
+ "bits": 4
1550
+ },
1551
+ "model.layers.0.mlp.up_proj": {
1552
+ "group_size": 64,
1553
+ "bits": 3
1554
+ },
1555
+ "model.layers.0.input_layernorm": false,
1556
+ "model.layers.0.post_attention_layernorm": false,
1557
+ "model.layers.1.self_attn.q_proj": {
1558
+ "group_size": 64,
1559
+ "bits": 3
1560
+ },
1561
+ "model.layers.1.self_attn.k_proj": {
1562
+ "group_size": 64,
1563
+ "bits": 3
1564
+ },
1565
+ "model.layers.1.self_attn.v_proj": {
1566
+ "group_size": 64,
1567
+ "bits": 4
1568
+ },
1569
+ "model.layers.1.self_attn.o_proj": {
1570
+ "group_size": 64,
1571
+ "bits": 3
1572
+ },
1573
+ "model.layers.1.self_attn.rope": false,
1574
+ "model.layers.1.mlp.gate_proj": {
1575
+ "group_size": 64,
1576
+ "bits": 3
1577
+ },
1578
+ "model.layers.1.mlp.down_proj": {
1579
+ "group_size": 64,
1580
+ "bits": 4
1581
+ },
1582
+ "model.layers.1.mlp.up_proj": {
1583
+ "group_size": 64,
1584
+ "bits": 3
1585
+ },
1586
+ "model.layers.1.input_layernorm": false,
1587
+ "model.layers.1.post_attention_layernorm": false,
1588
+ "model.layers.2.self_attn.q_proj": {
1589
+ "group_size": 64,
1590
+ "bits": 3
1591
+ },
1592
+ "model.layers.2.self_attn.k_proj": {
1593
+ "group_size": 64,
1594
+ "bits": 3
1595
+ },
1596
+ "model.layers.2.self_attn.v_proj": {
1597
+ "group_size": 64,
1598
+ "bits": 4
1599
+ },
1600
+ "model.layers.2.self_attn.o_proj": {
1601
+ "group_size": 64,
1602
+ "bits": 3
1603
+ },
1604
+ "model.layers.2.self_attn.rope": false,
1605
+ "model.layers.2.mlp.gate_proj": {
1606
+ "group_size": 64,
1607
+ "bits": 3
1608
+ },
1609
+ "model.layers.2.mlp.down_proj": {
1610
+ "group_size": 64,
1611
+ "bits": 4
1612
+ },
1613
+ "model.layers.2.mlp.up_proj": {
1614
+ "group_size": 64,
1615
+ "bits": 3
1616
+ },
1617
+ "model.layers.2.input_layernorm": false,
1618
+ "model.layers.2.post_attention_layernorm": false,
1619
+ "model.layers.3.self_attn.q_proj": {
1620
+ "group_size": 64,
1621
+ "bits": 3
1622
+ },
1623
+ "model.layers.3.self_attn.k_proj": {
1624
+ "group_size": 64,
1625
+ "bits": 3
1626
+ },
1627
+ "model.layers.3.self_attn.v_proj": {
1628
+ "group_size": 64,
1629
+ "bits": 4
1630
+ },
1631
+ "model.layers.3.self_attn.o_proj": {
1632
+ "group_size": 64,
1633
+ "bits": 3
1634
+ },
1635
+ "model.layers.3.self_attn.rope": false,
1636
+ "model.layers.3.mlp.gate_proj": {
1637
+ "group_size": 64,
1638
+ "bits": 3
1639
+ },
1640
+ "model.layers.3.mlp.down_proj": {
1641
+ "group_size": 64,
1642
+ "bits": 4
1643
+ },
1644
+ "model.layers.3.mlp.up_proj": {
1645
+ "group_size": 64,
1646
+ "bits": 3
1647
+ },
1648
+ "model.layers.3.input_layernorm": false,
1649
+ "model.layers.3.post_attention_layernorm": false,
1650
+ "model.layers.4.self_attn.q_proj": {
1651
+ "group_size": 64,
1652
+ "bits": 3
1653
+ },
1654
+ "model.layers.4.self_attn.k_proj": {
1655
+ "group_size": 64,
1656
+ "bits": 3
1657
+ },
1658
+ "model.layers.4.self_attn.v_proj": {
1659
+ "group_size": 64,
1660
+ "bits": 4
1661
+ },
1662
+ "model.layers.4.self_attn.o_proj": {
1663
+ "group_size": 64,
1664
+ "bits": 3
1665
+ },
1666
+ "model.layers.4.self_attn.rope": false,
1667
+ "model.layers.4.mlp.gate_proj": {
1668
+ "group_size": 64,
1669
+ "bits": 3
1670
+ },
1671
+ "model.layers.4.mlp.down_proj": {
1672
+ "group_size": 64,
1673
+ "bits": 4
1674
+ },
1675
+ "model.layers.4.mlp.up_proj": {
1676
+ "group_size": 64,
1677
+ "bits": 3
1678
+ },
1679
+ "model.layers.4.input_layernorm": false,
1680
+ "model.layers.4.post_attention_layernorm": false,
1681
+ "model.layers.5.self_attn.q_proj": {
1682
+ "group_size": 64,
1683
+ "bits": 3
1684
+ },
1685
+ "model.layers.5.self_attn.k_proj": {
1686
+ "group_size": 64,
1687
+ "bits": 3
1688
+ },
1689
+ "model.layers.5.self_attn.v_proj": {
1690
+ "group_size": 64,
1691
+ "bits": 4
1692
+ },
1693
+ "model.layers.5.self_attn.o_proj": {
1694
+ "group_size": 64,
1695
+ "bits": 3
1696
+ },
1697
+ "model.layers.5.self_attn.rope": false,
1698
+ "model.layers.5.mlp.gate_proj": {
1699
+ "group_size": 64,
1700
+ "bits": 3
1701
+ },
1702
+ "model.layers.5.mlp.down_proj": {
1703
+ "group_size": 64,
1704
+ "bits": 4
1705
+ },
1706
+ "model.layers.5.mlp.up_proj": {
1707
+ "group_size": 64,
1708
+ "bits": 3
1709
+ },
1710
+ "model.layers.5.input_layernorm": false,
1711
+ "model.layers.5.post_attention_layernorm": false,
1712
+ "model.layers.6.self_attn.q_proj": {
1713
+ "group_size": 64,
1714
+ "bits": 3
1715
+ },
1716
+ "model.layers.6.self_attn.k_proj": {
1717
+ "group_size": 64,
1718
+ "bits": 3
1719
+ },
1720
+ "model.layers.6.self_attn.v_proj": {
1721
+ "group_size": 64,
1722
+ "bits": 3
1723
+ },
1724
+ "model.layers.6.self_attn.o_proj": {
1725
+ "group_size": 64,
1726
+ "bits": 3
1727
+ },
1728
+ "model.layers.6.self_attn.rope": false,
1729
+ "model.layers.6.mlp.gate_proj": {
1730
+ "group_size": 64,
1731
+ "bits": 3
1732
+ },
1733
+ "model.layers.6.mlp.down_proj": {
1734
+ "group_size": 64,
1735
+ "bits": 3
1736
+ },
1737
+ "model.layers.6.mlp.up_proj": {
1738
+ "group_size": 64,
1739
+ "bits": 3
1740
+ },
1741
+ "model.layers.6.input_layernorm": false,
1742
+ "model.layers.6.post_attention_layernorm": false,
1743
+ "model.layers.7.self_attn.q_proj": {
1744
+ "group_size": 64,
1745
+ "bits": 3
1746
+ },
1747
+ "model.layers.7.self_attn.k_proj": {
1748
+ "group_size": 64,
1749
+ "bits": 3
1750
+ },
1751
+ "model.layers.7.self_attn.v_proj": {
1752
+ "group_size": 64,
1753
+ "bits": 3
1754
+ },
1755
+ "model.layers.7.self_attn.o_proj": {
1756
+ "group_size": 64,
1757
+ "bits": 3
1758
+ },
1759
+ "model.layers.7.self_attn.rope": false,
1760
+ "model.layers.7.mlp.gate_proj": {
1761
+ "group_size": 64,
1762
+ "bits": 3
1763
+ },
1764
+ "model.layers.7.mlp.down_proj": {
1765
+ "group_size": 64,
1766
+ "bits": 3
1767
+ },
1768
+ "model.layers.7.mlp.up_proj": {
1769
+ "group_size": 64,
1770
+ "bits": 3
1771
+ },
1772
+ "model.layers.7.input_layernorm": false,
1773
+ "model.layers.7.post_attention_layernorm": false,
1774
+ "model.layers.8.self_attn.q_proj": {
1775
+ "group_size": 64,
1776
+ "bits": 3
1777
+ },
1778
+ "model.layers.8.self_attn.k_proj": {
1779
+ "group_size": 64,
1780
+ "bits": 3
1781
+ },
1782
+ "model.layers.8.self_attn.v_proj": {
1783
+ "group_size": 64,
1784
+ "bits": 4
1785
+ },
1786
+ "model.layers.8.self_attn.o_proj": {
1787
+ "group_size": 64,
1788
+ "bits": 3
1789
+ },
1790
+ "model.layers.8.self_attn.rope": false,
1791
+ "model.layers.8.mlp.gate_proj": {
1792
+ "group_size": 64,
1793
+ "bits": 3
1794
+ },
1795
+ "model.layers.8.mlp.down_proj": {
1796
+ "group_size": 64,
1797
+ "bits": 4
1798
+ },
1799
+ "model.layers.8.mlp.up_proj": {
1800
+ "group_size": 64,
1801
+ "bits": 3
1802
+ },
1803
+ "model.layers.8.input_layernorm": false,
1804
+ "model.layers.8.post_attention_layernorm": false,
1805
+ "model.layers.9.self_attn.q_proj": {
1806
+ "group_size": 64,
1807
+ "bits": 3
1808
+ },
1809
+ "model.layers.9.self_attn.k_proj": {
1810
+ "group_size": 64,
1811
+ "bits": 3
1812
+ },
1813
+ "model.layers.9.self_attn.v_proj": {
1814
+ "group_size": 64,
1815
+ "bits": 3
1816
+ },
1817
+ "model.layers.9.self_attn.o_proj": {
1818
+ "group_size": 64,
1819
+ "bits": 3
1820
+ },
1821
+ "model.layers.9.self_attn.rope": false,
1822
+ "model.layers.9.mlp.gate_proj": {
1823
+ "group_size": 64,
1824
+ "bits": 3
1825
+ },
1826
+ "model.layers.9.mlp.down_proj": {
1827
+ "group_size": 64,
1828
+ "bits": 3
1829
+ },
1830
+ "model.layers.9.mlp.up_proj": {
1831
+ "group_size": 64,
1832
+ "bits": 3
1833
+ },
1834
+ "model.layers.9.input_layernorm": false,
1835
+ "model.layers.9.post_attention_layernorm": false,
1836
+ "model.layers.10.self_attn.q_proj": {
1837
+ "group_size": 64,
1838
+ "bits": 3
1839
+ },
1840
+ "model.layers.10.self_attn.k_proj": {
1841
+ "group_size": 64,
1842
+ "bits": 3
1843
+ },
1844
+ "model.layers.10.self_attn.v_proj": {
1845
+ "group_size": 64,
1846
+ "bits": 3
1847
+ },
1848
+ "model.layers.10.self_attn.o_proj": {
1849
+ "group_size": 64,
1850
+ "bits": 3
1851
+ },
1852
+ "model.layers.10.self_attn.rope": false,
1853
+ "model.layers.10.mlp.gate_proj": {
1854
+ "group_size": 64,
1855
+ "bits": 3
1856
+ },
1857
+ "model.layers.10.mlp.down_proj": {
1858
+ "group_size": 64,
1859
+ "bits": 3
1860
+ },
1861
+ "model.layers.10.mlp.up_proj": {
1862
+ "group_size": 64,
1863
+ "bits": 3
1864
+ },
1865
+ "model.layers.10.input_layernorm": false,
1866
+ "model.layers.10.post_attention_layernorm": false,
1867
+ "model.layers.11.self_attn.q_proj": {
1868
+ "group_size": 64,
1869
+ "bits": 3
1870
+ },
1871
+ "model.layers.11.self_attn.k_proj": {
1872
+ "group_size": 64,
1873
+ "bits": 3
1874
+ },
1875
+ "model.layers.11.self_attn.v_proj": {
1876
+ "group_size": 64,
1877
+ "bits": 4
1878
+ },
1879
+ "model.layers.11.self_attn.o_proj": {
1880
+ "group_size": 64,
1881
+ "bits": 3
1882
+ },
1883
+ "model.layers.11.self_attn.rope": false,
1884
+ "model.layers.11.mlp.gate_proj": {
1885
+ "group_size": 64,
1886
+ "bits": 3
1887
+ },
1888
+ "model.layers.11.mlp.down_proj": {
1889
+ "group_size": 64,
1890
+ "bits": 4
1891
+ },
1892
+ "model.layers.11.mlp.up_proj": {
1893
+ "group_size": 64,
1894
+ "bits": 3
1895
+ },
1896
+ "model.layers.11.input_layernorm": false,
1897
+ "model.layers.11.post_attention_layernorm": false,
1898
+ "model.layers.12.self_attn.q_proj": {
1899
+ "group_size": 64,
1900
+ "bits": 3
1901
+ },
1902
+ "model.layers.12.self_attn.k_proj": {
1903
+ "group_size": 64,
1904
+ "bits": 3
1905
+ },
1906
+ "model.layers.12.self_attn.v_proj": {
1907
+ "group_size": 64,
1908
+ "bits": 3
1909
+ },
1910
+ "model.layers.12.self_attn.o_proj": {
1911
+ "group_size": 64,
1912
+ "bits": 3
1913
+ },
1914
+ "model.layers.12.self_attn.rope": false,
1915
+ "model.layers.12.mlp.gate_proj": {
1916
+ "group_size": 64,
1917
+ "bits": 3
1918
+ },
1919
+ "model.layers.12.mlp.down_proj": {
1920
+ "group_size": 64,
1921
+ "bits": 3
1922
+ },
1923
+ "model.layers.12.mlp.up_proj": {
1924
+ "group_size": 64,
1925
+ "bits": 3
1926
+ },
1927
+ "model.layers.12.input_layernorm": false,
1928
+ "model.layers.12.post_attention_layernorm": false,
1929
+ "model.layers.13.self_attn.q_proj": {
1930
+ "group_size": 64,
1931
+ "bits": 3
1932
+ },
1933
+ "model.layers.13.self_attn.k_proj": {
1934
+ "group_size": 64,
1935
+ "bits": 3
1936
+ },
1937
+ "model.layers.13.self_attn.v_proj": {
1938
+ "group_size": 64,
1939
+ "bits": 3
1940
+ },
1941
+ "model.layers.13.self_attn.o_proj": {
1942
+ "group_size": 64,
1943
+ "bits": 3
1944
+ },
1945
+ "model.layers.13.self_attn.rope": false,
1946
+ "model.layers.13.mlp.gate_proj": {
1947
+ "group_size": 64,
1948
+ "bits": 3
1949
+ },
1950
+ "model.layers.13.mlp.down_proj": {
1951
+ "group_size": 64,
1952
+ "bits": 3
1953
+ },
1954
+ "model.layers.13.mlp.up_proj": {
1955
+ "group_size": 64,
1956
+ "bits": 3
1957
+ },
1958
+ "model.layers.13.input_layernorm": false,
1959
+ "model.layers.13.post_attention_layernorm": false,
1960
+ "model.layers.14.self_attn.q_proj": {
1961
+ "group_size": 64,
1962
+ "bits": 3
1963
+ },
1964
+ "model.layers.14.self_attn.k_proj": {
1965
+ "group_size": 64,
1966
+ "bits": 3
1967
+ },
1968
+ "model.layers.14.self_attn.v_proj": {
1969
+ "group_size": 64,
1970
+ "bits": 4
1971
+ },
1972
+ "model.layers.14.self_attn.o_proj": {
1973
+ "group_size": 64,
1974
+ "bits": 3
1975
+ },
1976
+ "model.layers.14.self_attn.rope": false,
1977
+ "model.layers.14.mlp.gate_proj": {
1978
+ "group_size": 64,
1979
+ "bits": 3
1980
+ },
1981
+ "model.layers.14.mlp.down_proj": {
1982
+ "group_size": 64,
1983
+ "bits": 4
1984
+ },
1985
+ "model.layers.14.mlp.up_proj": {
1986
+ "group_size": 64,
1987
+ "bits": 3
1988
+ },
1989
+ "model.layers.14.input_layernorm": false,
1990
+ "model.layers.14.post_attention_layernorm": false,
1991
+ "model.layers.15.self_attn.q_proj": {
1992
+ "group_size": 64,
1993
+ "bits": 3
1994
+ },
1995
+ "model.layers.15.self_attn.k_proj": {
1996
+ "group_size": 64,
1997
+ "bits": 3
1998
+ },
1999
+ "model.layers.15.self_attn.v_proj": {
2000
+ "group_size": 64,
2001
+ "bits": 3
2002
+ },
2003
+ "model.layers.15.self_attn.o_proj": {
2004
+ "group_size": 64,
2005
+ "bits": 3
2006
+ },
2007
+ "model.layers.15.self_attn.rope": false,
2008
+ "model.layers.15.mlp.gate_proj": {
2009
+ "group_size": 64,
2010
+ "bits": 3
2011
+ },
2012
+ "model.layers.15.mlp.down_proj": {
2013
+ "group_size": 64,
2014
+ "bits": 3
2015
+ },
2016
+ "model.layers.15.mlp.up_proj": {
2017
+ "group_size": 64,
2018
+ "bits": 3
2019
+ },
2020
+ "model.layers.15.input_layernorm": false,
2021
+ "model.layers.15.post_attention_layernorm": false,
2022
+ "model.layers.16.self_attn.q_proj": {
2023
+ "group_size": 64,
2024
+ "bits": 3
2025
+ },
2026
+ "model.layers.16.self_attn.k_proj": {
2027
+ "group_size": 64,
2028
+ "bits": 3
2029
+ },
2030
+ "model.layers.16.self_attn.v_proj": {
2031
+ "group_size": 64,
2032
+ "bits": 3
2033
+ },
2034
+ "model.layers.16.self_attn.o_proj": {
2035
+ "group_size": 64,
2036
+ "bits": 3
2037
+ },
2038
+ "model.layers.16.self_attn.rope": false,
2039
+ "model.layers.16.mlp.gate_proj": {
2040
+ "group_size": 64,
2041
+ "bits": 3
2042
+ },
2043
+ "model.layers.16.mlp.down_proj": {
2044
+ "group_size": 64,
2045
+ "bits": 3
2046
+ },
2047
+ "model.layers.16.mlp.up_proj": {
2048
+ "group_size": 64,
2049
+ "bits": 3
2050
+ },
2051
+ "model.layers.16.input_layernorm": false,
2052
+ "model.layers.16.post_attention_layernorm": false,
2053
+ "model.layers.17.self_attn.q_proj": {
2054
+ "group_size": 64,
2055
+ "bits": 3
2056
+ },
2057
+ "model.layers.17.self_attn.k_proj": {
2058
+ "group_size": 64,
2059
+ "bits": 3
2060
+ },
2061
+ "model.layers.17.self_attn.v_proj": {
2062
+ "group_size": 64,
2063
+ "bits": 4
2064
+ },
2065
+ "model.layers.17.self_attn.o_proj": {
2066
+ "group_size": 64,
2067
+ "bits": 3
2068
+ },
2069
+ "model.layers.17.self_attn.rope": false,
2070
+ "model.layers.17.mlp.gate_proj": {
2071
+ "group_size": 64,
2072
+ "bits": 3
2073
+ },
2074
+ "model.layers.17.mlp.down_proj": {
2075
+ "group_size": 64,
2076
+ "bits": 4
2077
+ },
2078
+ "model.layers.17.mlp.up_proj": {
2079
+ "group_size": 64,
2080
+ "bits": 3
2081
+ },
2082
+ "model.layers.17.input_layernorm": false,
2083
+ "model.layers.17.post_attention_layernorm": false,
2084
+ "model.layers.18.self_attn.q_proj": {
2085
+ "group_size": 64,
2086
+ "bits": 3
2087
+ },
2088
+ "model.layers.18.self_attn.k_proj": {
2089
+ "group_size": 64,
2090
+ "bits": 3
2091
+ },
2092
+ "model.layers.18.self_attn.v_proj": {
2093
+ "group_size": 64,
2094
+ "bits": 3
2095
+ },
2096
+ "model.layers.18.self_attn.o_proj": {
2097
+ "group_size": 64,
2098
+ "bits": 3
2099
+ },
2100
+ "model.layers.18.self_attn.rope": false,
2101
+ "model.layers.18.mlp.gate_proj": {
2102
+ "group_size": 64,
2103
+ "bits": 3
2104
+ },
2105
+ "model.layers.18.mlp.down_proj": {
2106
+ "group_size": 64,
2107
+ "bits": 3
2108
+ },
2109
+ "model.layers.18.mlp.up_proj": {
2110
+ "group_size": 64,
2111
+ "bits": 3
2112
+ },
2113
+ "model.layers.18.input_layernorm": false,
2114
+ "model.layers.18.post_attention_layernorm": false,
2115
+ "model.layers.19.self_attn.q_proj": {
2116
+ "group_size": 64,
2117
+ "bits": 3
2118
+ },
2119
+ "model.layers.19.self_attn.k_proj": {
2120
+ "group_size": 64,
2121
+ "bits": 3
2122
+ },
2123
+ "model.layers.19.self_attn.v_proj": {
2124
+ "group_size": 64,
2125
+ "bits": 3
2126
+ },
2127
+ "model.layers.19.self_attn.o_proj": {
2128
+ "group_size": 64,
2129
+ "bits": 3
2130
+ },
2131
+ "model.layers.19.self_attn.rope": false,
2132
+ "model.layers.19.mlp.gate_proj": {
2133
+ "group_size": 64,
2134
+ "bits": 3
2135
+ },
2136
+ "model.layers.19.mlp.down_proj": {
2137
+ "group_size": 64,
2138
+ "bits": 3
2139
+ },
2140
+ "model.layers.19.mlp.up_proj": {
2141
+ "group_size": 64,
2142
+ "bits": 3
2143
+ },
2144
+ "model.layers.19.input_layernorm": false,
2145
+ "model.layers.19.post_attention_layernorm": false,
2146
+ "model.layers.20.self_attn.q_proj": {
2147
+ "group_size": 64,
2148
+ "bits": 3
2149
+ },
2150
+ "model.layers.20.self_attn.k_proj": {
2151
+ "group_size": 64,
2152
+ "bits": 3
2153
+ },
2154
+ "model.layers.20.self_attn.v_proj": {
2155
+ "group_size": 64,
2156
+ "bits": 4
2157
+ },
2158
+ "model.layers.20.self_attn.o_proj": {
2159
+ "group_size": 64,
2160
+ "bits": 3
2161
+ },
2162
+ "model.layers.20.self_attn.rope": false,
2163
+ "model.layers.20.mlp.gate_proj": {
2164
+ "group_size": 64,
2165
+ "bits": 3
2166
+ },
2167
+ "model.layers.20.mlp.down_proj": {
2168
+ "group_size": 64,
2169
+ "bits": 4
2170
+ },
2171
+ "model.layers.20.mlp.up_proj": {
2172
+ "group_size": 64,
2173
+ "bits": 3
2174
+ },
2175
+ "model.layers.20.input_layernorm": false,
2176
+ "model.layers.20.post_attention_layernorm": false,
2177
+ "model.layers.21.self_attn.q_proj": {
2178
+ "group_size": 64,
2179
+ "bits": 3
2180
+ },
2181
+ "model.layers.21.self_attn.k_proj": {
2182
+ "group_size": 64,
2183
+ "bits": 3
2184
+ },
2185
+ "model.layers.21.self_attn.v_proj": {
2186
+ "group_size": 64,
2187
+ "bits": 3
2188
+ },
2189
+ "model.layers.21.self_attn.o_proj": {
2190
+ "group_size": 64,
2191
+ "bits": 3
2192
+ },
2193
+ "model.layers.21.self_attn.rope": false,
2194
+ "model.layers.21.mlp.gate_proj": {
2195
+ "group_size": 64,
2196
+ "bits": 3
2197
+ },
2198
+ "model.layers.21.mlp.down_proj": {
2199
+ "group_size": 64,
2200
+ "bits": 3
2201
+ },
2202
+ "model.layers.21.mlp.up_proj": {
2203
+ "group_size": 64,
2204
+ "bits": 3
2205
+ },
2206
+ "model.layers.21.input_layernorm": false,
2207
+ "model.layers.21.post_attention_layernorm": false,
2208
+ "model.layers.22.self_attn.q_proj": {
2209
+ "group_size": 64,
2210
+ "bits": 3
2211
+ },
2212
+ "model.layers.22.self_attn.k_proj": {
2213
+ "group_size": 64,
2214
+ "bits": 3
2215
+ },
2216
+ "model.layers.22.self_attn.v_proj": {
2217
+ "group_size": 64,
2218
+ "bits": 3
2219
+ },
2220
+ "model.layers.22.self_attn.o_proj": {
2221
+ "group_size": 64,
2222
+ "bits": 3
2223
+ },
2224
+ "model.layers.22.self_attn.rope": false,
2225
+ "model.layers.22.mlp.gate_proj": {
2226
+ "group_size": 64,
2227
+ "bits": 3
2228
+ },
2229
+ "model.layers.22.mlp.down_proj": {
2230
+ "group_size": 64,
2231
+ "bits": 3
2232
+ },
2233
+ "model.layers.22.mlp.up_proj": {
2234
+ "group_size": 64,
2235
+ "bits": 3
2236
+ },
2237
+ "model.layers.22.input_layernorm": false,
2238
+ "model.layers.22.post_attention_layernorm": false,
2239
+ "model.layers.23.self_attn.q_proj": {
2240
+ "group_size": 64,
2241
+ "bits": 3
2242
+ },
2243
+ "model.layers.23.self_attn.k_proj": {
2244
+ "group_size": 64,
2245
+ "bits": 3
2246
+ },
2247
+ "model.layers.23.self_attn.v_proj": {
2248
+ "group_size": 64,
2249
+ "bits": 4
2250
+ },
2251
+ "model.layers.23.self_attn.o_proj": {
2252
+ "group_size": 64,
2253
+ "bits": 3
2254
+ },
2255
+ "model.layers.23.self_attn.rope": false,
2256
+ "model.layers.23.mlp.gate_proj": {
2257
+ "group_size": 64,
2258
+ "bits": 3
2259
+ },
2260
+ "model.layers.23.mlp.down_proj": {
2261
+ "group_size": 64,
2262
+ "bits": 4
2263
+ },
2264
+ "model.layers.23.mlp.up_proj": {
2265
+ "group_size": 64,
2266
+ "bits": 3
2267
+ },
2268
+ "model.layers.23.input_layernorm": false,
2269
+ "model.layers.23.post_attention_layernorm": false,
2270
+ "model.layers.24.self_attn.q_proj": {
2271
+ "group_size": 64,
2272
+ "bits": 3
2273
+ },
2274
+ "model.layers.24.self_attn.k_proj": {
2275
+ "group_size": 64,
2276
+ "bits": 3
2277
+ },
2278
+ "model.layers.24.self_attn.v_proj": {
2279
+ "group_size": 64,
2280
+ "bits": 3
2281
+ },
2282
+ "model.layers.24.self_attn.o_proj": {
2283
+ "group_size": 64,
2284
+ "bits": 3
2285
+ },
2286
+ "model.layers.24.self_attn.rope": false,
2287
+ "model.layers.24.mlp.gate_proj": {
2288
+ "group_size": 64,
2289
+ "bits": 3
2290
+ },
2291
+ "model.layers.24.mlp.down_proj": {
2292
+ "group_size": 64,
2293
+ "bits": 3
2294
+ },
2295
+ "model.layers.24.mlp.up_proj": {
2296
+ "group_size": 64,
2297
+ "bits": 3
2298
+ },
2299
+ "model.layers.24.input_layernorm": false,
2300
+ "model.layers.24.post_attention_layernorm": false,
2301
+ "model.layers.25.self_attn.q_proj": {
2302
+ "group_size": 64,
2303
+ "bits": 3
2304
+ },
2305
+ "model.layers.25.self_attn.k_proj": {
2306
+ "group_size": 64,
2307
+ "bits": 3
2308
+ },
2309
+ "model.layers.25.self_attn.v_proj": {
2310
+ "group_size": 64,
2311
+ "bits": 3
2312
+ },
2313
+ "model.layers.25.self_attn.o_proj": {
2314
+ "group_size": 64,
2315
+ "bits": 3
2316
+ },
2317
+ "model.layers.25.self_attn.rope": false,
2318
+ "model.layers.25.mlp.gate_proj": {
2319
+ "group_size": 64,
2320
+ "bits": 3
2321
+ },
2322
+ "model.layers.25.mlp.down_proj": {
2323
+ "group_size": 64,
2324
+ "bits": 3
2325
+ },
2326
+ "model.layers.25.mlp.up_proj": {
2327
+ "group_size": 64,
2328
+ "bits": 3
2329
+ },
2330
+ "model.layers.25.input_layernorm": false,
2331
+ "model.layers.25.post_attention_layernorm": false,
2332
+ "model.layers.26.self_attn.q_proj": {
2333
+ "group_size": 64,
2334
+ "bits": 3
2335
+ },
2336
+ "model.layers.26.self_attn.k_proj": {
2337
+ "group_size": 64,
2338
+ "bits": 3
2339
+ },
2340
+ "model.layers.26.self_attn.v_proj": {
2341
+ "group_size": 64,
2342
+ "bits": 4
2343
+ },
2344
+ "model.layers.26.self_attn.o_proj": {
2345
+ "group_size": 64,
2346
+ "bits": 3
2347
+ },
2348
+ "model.layers.26.self_attn.rope": false,
2349
+ "model.layers.26.mlp.gate_proj": {
2350
+ "group_size": 64,
2351
+ "bits": 3
2352
+ },
2353
+ "model.layers.26.mlp.down_proj": {
2354
+ "group_size": 64,
2355
+ "bits": 4
2356
+ },
2357
+ "model.layers.26.mlp.up_proj": {
2358
+ "group_size": 64,
2359
+ "bits": 3
2360
+ },
2361
+ "model.layers.26.input_layernorm": false,
2362
+ "model.layers.26.post_attention_layernorm": false,
2363
+ "model.layers.27.self_attn.q_proj": {
2364
+ "group_size": 64,
2365
+ "bits": 3
2366
+ },
2367
+ "model.layers.27.self_attn.k_proj": {
2368
+ "group_size": 64,
2369
+ "bits": 3
2370
+ },
2371
+ "model.layers.27.self_attn.v_proj": {
2372
+ "group_size": 64,
2373
+ "bits": 3
2374
+ },
2375
+ "model.layers.27.self_attn.o_proj": {
2376
+ "group_size": 64,
2377
+ "bits": 3
2378
+ },
2379
+ "model.layers.27.self_attn.rope": false,
2380
+ "model.layers.27.mlp.gate_proj": {
2381
+ "group_size": 64,
2382
+ "bits": 3
2383
+ },
2384
+ "model.layers.27.mlp.down_proj": {
2385
+ "group_size": 64,
2386
+ "bits": 3
2387
+ },
2388
+ "model.layers.27.mlp.up_proj": {
2389
+ "group_size": 64,
2390
+ "bits": 3
2391
+ },
2392
+ "model.layers.27.input_layernorm": false,
2393
+ "model.layers.27.post_attention_layernorm": false,
2394
+ "model.layers.28.self_attn.q_proj": {
2395
+ "group_size": 64,
2396
+ "bits": 3
2397
+ },
2398
+ "model.layers.28.self_attn.k_proj": {
2399
+ "group_size": 64,
2400
+ "bits": 3
2401
+ },
2402
+ "model.layers.28.self_attn.v_proj": {
2403
+ "group_size": 64,
2404
+ "bits": 3
2405
+ },
2406
+ "model.layers.28.self_attn.o_proj": {
2407
+ "group_size": 64,
2408
+ "bits": 3
2409
+ },
2410
+ "model.layers.28.self_attn.rope": false,
2411
+ "model.layers.28.mlp.gate_proj": {
2412
+ "group_size": 64,
2413
+ "bits": 3
2414
+ },
2415
+ "model.layers.28.mlp.down_proj": {
2416
+ "group_size": 64,
2417
+ "bits": 3
2418
+ },
2419
+ "model.layers.28.mlp.up_proj": {
2420
+ "group_size": 64,
2421
+ "bits": 3
2422
+ },
2423
+ "model.layers.28.input_layernorm": false,
2424
+ "model.layers.28.post_attention_layernorm": false,
2425
+ "model.layers.29.self_attn.q_proj": {
2426
+ "group_size": 64,
2427
+ "bits": 3
2428
+ },
2429
+ "model.layers.29.self_attn.k_proj": {
2430
+ "group_size": 64,
2431
+ "bits": 3
2432
+ },
2433
+ "model.layers.29.self_attn.v_proj": {
2434
+ "group_size": 64,
2435
+ "bits": 4
2436
+ },
2437
+ "model.layers.29.self_attn.o_proj": {
2438
+ "group_size": 64,
2439
+ "bits": 3
2440
+ },
2441
+ "model.layers.29.self_attn.rope": false,
2442
+ "model.layers.29.mlp.gate_proj": {
2443
+ "group_size": 64,
2444
+ "bits": 3
2445
+ },
2446
+ "model.layers.29.mlp.down_proj": {
2447
+ "group_size": 64,
2448
+ "bits": 4
2449
+ },
2450
+ "model.layers.29.mlp.up_proj": {
2451
+ "group_size": 64,
2452
+ "bits": 3
2453
+ },
2454
+ "model.layers.29.input_layernorm": false,
2455
+ "model.layers.29.post_attention_layernorm": false,
2456
+ "model.layers.30.self_attn.q_proj": {
2457
+ "group_size": 64,
2458
+ "bits": 3
2459
+ },
2460
+ "model.layers.30.self_attn.k_proj": {
2461
+ "group_size": 64,
2462
+ "bits": 3
2463
+ },
2464
+ "model.layers.30.self_attn.v_proj": {
2465
+ "group_size": 64,
2466
+ "bits": 3
2467
+ },
2468
+ "model.layers.30.self_attn.o_proj": {
2469
+ "group_size": 64,
2470
+ "bits": 3
2471
+ },
2472
+ "model.layers.30.self_attn.rope": false,
2473
+ "model.layers.30.mlp.gate_proj": {
2474
+ "group_size": 64,
2475
+ "bits": 3
2476
+ },
2477
+ "model.layers.30.mlp.down_proj": {
2478
+ "group_size": 64,
2479
+ "bits": 3
2480
+ },
2481
+ "model.layers.30.mlp.up_proj": {
2482
+ "group_size": 64,
2483
+ "bits": 3
2484
+ },
2485
+ "model.layers.30.input_layernorm": false,
2486
+ "model.layers.30.post_attention_layernorm": false,
2487
+ "model.layers.31.self_attn.q_proj": {
2488
+ "group_size": 64,
2489
+ "bits": 3
2490
+ },
2491
+ "model.layers.31.self_attn.k_proj": {
2492
+ "group_size": 64,
2493
+ "bits": 3
2494
+ },
2495
+ "model.layers.31.self_attn.v_proj": {
2496
+ "group_size": 64,
2497
+ "bits": 3
2498
+ },
2499
+ "model.layers.31.self_attn.o_proj": {
2500
+ "group_size": 64,
2501
+ "bits": 3
2502
+ },
2503
+ "model.layers.31.self_attn.rope": false,
2504
+ "model.layers.31.mlp.gate_proj": {
2505
+ "group_size": 64,
2506
+ "bits": 3
2507
+ },
2508
+ "model.layers.31.mlp.down_proj": {
2509
+ "group_size": 64,
2510
+ "bits": 3
2511
+ },
2512
+ "model.layers.31.mlp.up_proj": {
2513
+ "group_size": 64,
2514
+ "bits": 3
2515
+ },
2516
+ "model.layers.31.input_layernorm": false,
2517
+ "model.layers.31.post_attention_layernorm": false,
2518
+ "model.layers.32.self_attn.q_proj": {
2519
+ "group_size": 64,
2520
+ "bits": 3
2521
+ },
2522
+ "model.layers.32.self_attn.k_proj": {
2523
+ "group_size": 64,
2524
+ "bits": 3
2525
+ },
2526
+ "model.layers.32.self_attn.v_proj": {
2527
+ "group_size": 64,
2528
+ "bits": 4
2529
+ },
2530
+ "model.layers.32.self_attn.o_proj": {
2531
+ "group_size": 64,
2532
+ "bits": 3
2533
+ },
2534
+ "model.layers.32.self_attn.rope": false,
2535
+ "model.layers.32.mlp.gate_proj": {
2536
+ "group_size": 64,
2537
+ "bits": 3
2538
+ },
2539
+ "model.layers.32.mlp.down_proj": {
2540
+ "group_size": 64,
2541
+ "bits": 4
2542
+ },
2543
+ "model.layers.32.mlp.up_proj": {
2544
+ "group_size": 64,
2545
+ "bits": 3
2546
+ },
2547
+ "model.layers.32.input_layernorm": false,
2548
+ "model.layers.32.post_attention_layernorm": false,
2549
+ "model.layers.33.self_attn.q_proj": {
2550
+ "group_size": 64,
2551
+ "bits": 3
2552
+ },
2553
+ "model.layers.33.self_attn.k_proj": {
2554
+ "group_size": 64,
2555
+ "bits": 3
2556
+ },
2557
+ "model.layers.33.self_attn.v_proj": {
2558
+ "group_size": 64,
2559
+ "bits": 3
2560
+ },
2561
+ "model.layers.33.self_attn.o_proj": {
2562
+ "group_size": 64,
2563
+ "bits": 3
2564
+ },
2565
+ "model.layers.33.self_attn.rope": false,
2566
+ "model.layers.33.mlp.gate_proj": {
2567
+ "group_size": 64,
2568
+ "bits": 3
2569
+ },
2570
+ "model.layers.33.mlp.down_proj": {
2571
+ "group_size": 64,
2572
+ "bits": 3
2573
+ },
2574
+ "model.layers.33.mlp.up_proj": {
2575
+ "group_size": 64,
2576
+ "bits": 3
2577
+ },
2578
+ "model.layers.33.input_layernorm": false,
2579
+ "model.layers.33.post_attention_layernorm": false,
2580
+ "model.layers.34.self_attn.q_proj": {
2581
+ "group_size": 64,
2582
+ "bits": 3
2583
+ },
2584
+ "model.layers.34.self_attn.k_proj": {
2585
+ "group_size": 64,
2586
+ "bits": 3
2587
+ },
2588
+ "model.layers.34.self_attn.v_proj": {
2589
+ "group_size": 64,
2590
+ "bits": 3
2591
+ },
2592
+ "model.layers.34.self_attn.o_proj": {
2593
+ "group_size": 64,
2594
+ "bits": 3
2595
+ },
2596
+ "model.layers.34.self_attn.rope": false,
2597
+ "model.layers.34.mlp.gate_proj": {
2598
+ "group_size": 64,
2599
+ "bits": 3
2600
+ },
2601
+ "model.layers.34.mlp.down_proj": {
2602
+ "group_size": 64,
2603
+ "bits": 3
2604
+ },
2605
+ "model.layers.34.mlp.up_proj": {
2606
+ "group_size": 64,
2607
+ "bits": 3
2608
+ },
2609
+ "model.layers.34.input_layernorm": false,
2610
+ "model.layers.34.post_attention_layernorm": false,
2611
+ "model.layers.35.self_attn.q_proj": {
2612
+ "group_size": 64,
2613
+ "bits": 3
2614
+ },
2615
+ "model.layers.35.self_attn.k_proj": {
2616
+ "group_size": 64,
2617
+ "bits": 3
2618
+ },
2619
+ "model.layers.35.self_attn.v_proj": {
2620
+ "group_size": 64,
2621
+ "bits": 4
2622
+ },
2623
+ "model.layers.35.self_attn.o_proj": {
2624
+ "group_size": 64,
2625
+ "bits": 3
2626
+ },
2627
+ "model.layers.35.self_attn.rope": false,
2628
+ "model.layers.35.mlp.gate_proj": {
2629
+ "group_size": 64,
2630
+ "bits": 3
2631
+ },
2632
+ "model.layers.35.mlp.down_proj": {
2633
+ "group_size": 64,
2634
+ "bits": 4
2635
+ },
2636
+ "model.layers.35.mlp.up_proj": {
2637
+ "group_size": 64,
2638
+ "bits": 3
2639
+ },
2640
+ "model.layers.35.input_layernorm": false,
2641
+ "model.layers.35.post_attention_layernorm": false,
2642
+ "model.layers.36.self_attn.q_proj": {
2643
+ "group_size": 64,
2644
+ "bits": 3
2645
+ },
2646
+ "model.layers.36.self_attn.k_proj": {
2647
+ "group_size": 64,
2648
+ "bits": 3
2649
+ },
2650
+ "model.layers.36.self_attn.v_proj": {
2651
+ "group_size": 64,
2652
+ "bits": 3
2653
+ },
2654
+ "model.layers.36.self_attn.o_proj": {
2655
+ "group_size": 64,
2656
+ "bits": 3
2657
+ },
2658
+ "model.layers.36.self_attn.rope": false,
2659
+ "model.layers.36.mlp.gate_proj": {
2660
+ "group_size": 64,
2661
+ "bits": 3
2662
+ },
2663
+ "model.layers.36.mlp.down_proj": {
2664
+ "group_size": 64,
2665
+ "bits": 3
2666
+ },
2667
+ "model.layers.36.mlp.up_proj": {
2668
+ "group_size": 64,
2669
+ "bits": 3
2670
+ },
2671
+ "model.layers.36.input_layernorm": false,
2672
+ "model.layers.36.post_attention_layernorm": false,
2673
+ "model.layers.37.self_attn.q_proj": {
2674
+ "group_size": 64,
2675
+ "bits": 3
2676
+ },
2677
+ "model.layers.37.self_attn.k_proj": {
2678
+ "group_size": 64,
2679
+ "bits": 3
2680
+ },
2681
+ "model.layers.37.self_attn.v_proj": {
2682
+ "group_size": 64,
2683
+ "bits": 3
2684
+ },
2685
+ "model.layers.37.self_attn.o_proj": {
2686
+ "group_size": 64,
2687
+ "bits": 3
2688
+ },
2689
+ "model.layers.37.self_attn.rope": false,
2690
+ "model.layers.37.mlp.gate_proj": {
2691
+ "group_size": 64,
2692
+ "bits": 3
2693
+ },
2694
+ "model.layers.37.mlp.down_proj": {
2695
+ "group_size": 64,
2696
+ "bits": 3
2697
+ },
2698
+ "model.layers.37.mlp.up_proj": {
2699
+ "group_size": 64,
2700
+ "bits": 3
2701
+ },
2702
+ "model.layers.37.input_layernorm": false,
2703
+ "model.layers.37.post_attention_layernorm": false,
2704
+ "model.layers.38.self_attn.q_proj": {
2705
+ "group_size": 64,
2706
+ "bits": 3
2707
+ },
2708
+ "model.layers.38.self_attn.k_proj": {
2709
+ "group_size": 64,
2710
+ "bits": 3
2711
+ },
2712
+ "model.layers.38.self_attn.v_proj": {
2713
+ "group_size": 64,
2714
+ "bits": 4
2715
+ },
2716
+ "model.layers.38.self_attn.o_proj": {
2717
+ "group_size": 64,
2718
+ "bits": 3
2719
+ },
2720
+ "model.layers.38.self_attn.rope": false,
2721
+ "model.layers.38.mlp.gate_proj": {
2722
+ "group_size": 64,
2723
+ "bits": 3
2724
+ },
2725
+ "model.layers.38.mlp.down_proj": {
2726
+ "group_size": 64,
2727
+ "bits": 4
2728
+ },
2729
+ "model.layers.38.mlp.up_proj": {
2730
+ "group_size": 64,
2731
+ "bits": 3
2732
+ },
2733
+ "model.layers.38.input_layernorm": false,
2734
+ "model.layers.38.post_attention_layernorm": false,
2735
+ "model.layers.39.self_attn.q_proj": {
2736
+ "group_size": 64,
2737
+ "bits": 3
2738
+ },
2739
+ "model.layers.39.self_attn.k_proj": {
2740
+ "group_size": 64,
2741
+ "bits": 3
2742
+ },
2743
+ "model.layers.39.self_attn.v_proj": {
2744
+ "group_size": 64,
2745
+ "bits": 3
2746
+ },
2747
+ "model.layers.39.self_attn.o_proj": {
2748
+ "group_size": 64,
2749
+ "bits": 3
2750
+ },
2751
+ "model.layers.39.self_attn.rope": false,
2752
+ "model.layers.39.mlp.gate_proj": {
2753
+ "group_size": 64,
2754
+ "bits": 3
2755
+ },
2756
+ "model.layers.39.mlp.down_proj": {
2757
+ "group_size": 64,
2758
+ "bits": 3
2759
+ },
2760
+ "model.layers.39.mlp.up_proj": {
2761
+ "group_size": 64,
2762
+ "bits": 3
2763
+ },
2764
+ "model.layers.39.input_layernorm": false,
2765
+ "model.layers.39.post_attention_layernorm": false,
2766
+ "model.layers.40.self_attn.q_proj": {
2767
+ "group_size": 64,
2768
+ "bits": 3
2769
+ },
2770
+ "model.layers.40.self_attn.k_proj": {
2771
+ "group_size": 64,
2772
+ "bits": 3
2773
+ },
2774
+ "model.layers.40.self_attn.v_proj": {
2775
+ "group_size": 64,
2776
+ "bits": 3
2777
+ },
2778
+ "model.layers.40.self_attn.o_proj": {
2779
+ "group_size": 64,
2780
+ "bits": 3
2781
+ },
2782
+ "model.layers.40.self_attn.rope": false,
2783
+ "model.layers.40.mlp.gate_proj": {
2784
+ "group_size": 64,
2785
+ "bits": 3
2786
+ },
2787
+ "model.layers.40.mlp.down_proj": {
2788
+ "group_size": 64,
2789
+ "bits": 3
2790
+ },
2791
+ "model.layers.40.mlp.up_proj": {
2792
+ "group_size": 64,
2793
+ "bits": 3
2794
+ },
2795
+ "model.layers.40.input_layernorm": false,
2796
+ "model.layers.40.post_attention_layernorm": false,
2797
+ "model.layers.41.self_attn.q_proj": {
2798
+ "group_size": 64,
2799
+ "bits": 3
2800
+ },
2801
+ "model.layers.41.self_attn.k_proj": {
2802
+ "group_size": 64,
2803
+ "bits": 3
2804
+ },
2805
+ "model.layers.41.self_attn.v_proj": {
2806
+ "group_size": 64,
2807
+ "bits": 4
2808
+ },
2809
+ "model.layers.41.self_attn.o_proj": {
2810
+ "group_size": 64,
2811
+ "bits": 3
2812
+ },
2813
+ "model.layers.41.self_attn.rope": false,
2814
+ "model.layers.41.mlp.gate_proj": {
2815
+ "group_size": 64,
2816
+ "bits": 3
2817
+ },
2818
+ "model.layers.41.mlp.down_proj": {
2819
+ "group_size": 64,
2820
+ "bits": 4
2821
+ },
2822
+ "model.layers.41.mlp.up_proj": {
2823
+ "group_size": 64,
2824
+ "bits": 3
2825
+ },
2826
+ "model.layers.41.input_layernorm": false,
2827
+ "model.layers.41.post_attention_layernorm": false,
2828
+ "model.layers.42.self_attn.q_proj": {
2829
+ "group_size": 64,
2830
+ "bits": 3
2831
+ },
2832
+ "model.layers.42.self_attn.k_proj": {
2833
+ "group_size": 64,
2834
+ "bits": 3
2835
+ },
2836
+ "model.layers.42.self_attn.v_proj": {
2837
+ "group_size": 64,
2838
+ "bits": 4
2839
+ },
2840
+ "model.layers.42.self_attn.o_proj": {
2841
+ "group_size": 64,
2842
+ "bits": 3
2843
+ },
2844
+ "model.layers.42.self_attn.rope": false,
2845
+ "model.layers.42.mlp.gate_proj": {
2846
+ "group_size": 64,
2847
+ "bits": 3
2848
+ },
2849
+ "model.layers.42.mlp.down_proj": {
2850
+ "group_size": 64,
2851
+ "bits": 4
2852
+ },
2853
+ "model.layers.42.mlp.up_proj": {
2854
+ "group_size": 64,
2855
+ "bits": 3
2856
+ },
2857
+ "model.layers.42.input_layernorm": false,
2858
+ "model.layers.42.post_attention_layernorm": false,
2859
+ "model.layers.43.self_attn.q_proj": {
2860
+ "group_size": 64,
2861
+ "bits": 3
2862
+ },
2863
+ "model.layers.43.self_attn.k_proj": {
2864
+ "group_size": 64,
2865
+ "bits": 3
2866
+ },
2867
+ "model.layers.43.self_attn.v_proj": {
2868
+ "group_size": 64,
2869
+ "bits": 4
2870
+ },
2871
+ "model.layers.43.self_attn.o_proj": {
2872
+ "group_size": 64,
2873
+ "bits": 3
2874
+ },
2875
+ "model.layers.43.self_attn.rope": false,
2876
+ "model.layers.43.mlp.gate_proj": {
2877
+ "group_size": 64,
2878
+ "bits": 3
2879
+ },
2880
+ "model.layers.43.mlp.down_proj": {
2881
+ "group_size": 64,
2882
+ "bits": 4
2883
+ },
2884
+ "model.layers.43.mlp.up_proj": {
2885
+ "group_size": 64,
2886
+ "bits": 3
2887
+ },
2888
+ "model.layers.43.input_layernorm": false,
2889
+ "model.layers.43.post_attention_layernorm": false,
2890
+ "model.layers.44.self_attn.q_proj": {
2891
+ "group_size": 64,
2892
+ "bits": 3
2893
+ },
2894
+ "model.layers.44.self_attn.k_proj": {
2895
+ "group_size": 64,
2896
+ "bits": 3
2897
+ },
2898
+ "model.layers.44.self_attn.v_proj": {
2899
+ "group_size": 64,
2900
+ "bits": 4
2901
+ },
2902
+ "model.layers.44.self_attn.o_proj": {
2903
+ "group_size": 64,
2904
+ "bits": 3
2905
+ },
2906
+ "model.layers.44.self_attn.rope": false,
2907
+ "model.layers.44.mlp.gate_proj": {
2908
+ "group_size": 64,
2909
+ "bits": 3
2910
+ },
2911
+ "model.layers.44.mlp.down_proj": {
2912
+ "group_size": 64,
2913
+ "bits": 4
2914
+ },
2915
+ "model.layers.44.mlp.up_proj": {
2916
+ "group_size": 64,
2917
+ "bits": 3
2918
+ },
2919
+ "model.layers.44.input_layernorm": false,
2920
+ "model.layers.44.post_attention_layernorm": false,
2921
+ "model.layers.45.self_attn.q_proj": {
2922
+ "group_size": 64,
2923
+ "bits": 3
2924
+ },
2925
+ "model.layers.45.self_attn.k_proj": {
2926
+ "group_size": 64,
2927
+ "bits": 3
2928
+ },
2929
+ "model.layers.45.self_attn.v_proj": {
2930
+ "group_size": 64,
2931
+ "bits": 4
2932
+ },
2933
+ "model.layers.45.self_attn.o_proj": {
2934
+ "group_size": 64,
2935
+ "bits": 3
2936
+ },
2937
+ "model.layers.45.self_attn.rope": false,
2938
+ "model.layers.45.mlp.gate_proj": {
2939
+ "group_size": 64,
2940
+ "bits": 3
2941
+ },
2942
+ "model.layers.45.mlp.down_proj": {
2943
+ "group_size": 64,
2944
+ "bits": 4
2945
+ },
2946
+ "model.layers.45.mlp.up_proj": {
2947
+ "group_size": 64,
2948
+ "bits": 3
2949
+ },
2950
+ "model.layers.45.input_layernorm": false,
2951
+ "model.layers.45.post_attention_layernorm": false,
2952
+ "model.layers.46.self_attn.q_proj": {
2953
+ "group_size": 64,
2954
+ "bits": 3
2955
+ },
2956
+ "model.layers.46.self_attn.k_proj": {
2957
+ "group_size": 64,
2958
+ "bits": 3
2959
+ },
2960
+ "model.layers.46.self_attn.v_proj": {
2961
+ "group_size": 64,
2962
+ "bits": 4
2963
+ },
2964
+ "model.layers.46.self_attn.o_proj": {
2965
+ "group_size": 64,
2966
+ "bits": 3
2967
+ },
2968
+ "model.layers.46.self_attn.rope": false,
2969
+ "model.layers.46.mlp.gate_proj": {
2970
+ "group_size": 64,
2971
+ "bits": 3
2972
+ },
2973
+ "model.layers.46.mlp.down_proj": {
2974
+ "group_size": 64,
2975
+ "bits": 4
2976
+ },
2977
+ "model.layers.46.mlp.up_proj": {
2978
+ "group_size": 64,
2979
+ "bits": 3
2980
+ },
2981
+ "model.layers.46.input_layernorm": false,
2982
+ "model.layers.46.post_attention_layernorm": false,
2983
+ "model.layers.47.self_attn.q_proj": {
2984
+ "group_size": 64,
2985
+ "bits": 3
2986
+ },
2987
+ "model.layers.47.self_attn.k_proj": {
2988
+ "group_size": 64,
2989
+ "bits": 3
2990
+ },
2991
+ "model.layers.47.self_attn.v_proj": {
2992
+ "group_size": 64,
2993
+ "bits": 4
2994
+ },
2995
+ "model.layers.47.self_attn.o_proj": {
2996
+ "group_size": 64,
2997
+ "bits": 3
2998
+ },
2999
+ "model.layers.47.self_attn.rope": false,
3000
+ "model.layers.47.mlp.gate_proj": {
3001
+ "group_size": 64,
3002
+ "bits": 3
3003
+ },
3004
+ "model.layers.47.mlp.down_proj": {
3005
+ "group_size": 64,
3006
+ "bits": 4
3007
+ },
3008
+ "model.layers.47.mlp.up_proj": {
3009
+ "group_size": 64,
3010
+ "bits": 3
3011
+ },
3012
+ "model.layers.47.input_layernorm": false,
3013
+ "model.layers.47.post_attention_layernorm": false,
3014
+ "model.norm": false,
3015
+ "lm_head": {
3016
+ "group_size": 64,
3017
+ "bits": 4
3018
+ }
3019
+ },
3020
+ "rms_norm_eps": 1e-05,
3021
+ "rope_theta": 1000000.0,
3022
+ "sliding_window": 131072,
3023
+ "tie_word_embeddings": false,
3024
+ "torch_dtype": "bfloat16",
3025
+ "transformers_version": "4.43.1",
3026
+ "use_cache": true,
3027
+ "use_sliding_window": false,
3028
+ "vocab_size": 152064
3029
+ }
DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ccdd1c75fe542d1c9b5ded243fd1ae6c627c852e07f466d67181991d448555ea
3
+ size 5348117358
DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:744229f4c2a22a1301aa5a3d582efa7ba228c9b460eaf47b2dd85b8e7a57fb24
3
+ size 1440613893
DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/special_tokens_map.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<｜begin▁of▁sentence｜>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<｜end▁of▁sentence｜>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<｜end▁of▁sentence｜>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ }
23
+ }
DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e20ddafc659ba90242154b55275402edeca0715e5dbb30f56815a4ce081f4893
3
+ size 11422778
DeepSeek-R1-Distill-Qwen-14B-3,6_mixed/tokenizer_config.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "151643": {
7
+ "content": "<｜end▁of▁sentence｜>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "151644": {
15
+ "content": "<｜User｜>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": false
21
+ },
22
+ "151645": {
23
+ "content": "<｜Assistant｜>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": false
29
+ },
30
+ "151646": {
31
+ "content": "<｜begin▁of▁sentence｜>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false,
36
+ "special": true
37
+ },
38
+ "151647": {
39
+ "content": "<|EOT|>",
40
+ "lstrip": false,
41
+ "normalized": false,
42
+ "rstrip": false,
43
+ "single_word": false,
44
+ "special": false
45
+ },
46
+ "151648": {
47
+ "content": "<think>",
48
+ "lstrip": false,
49
+ "normalized": false,
50
+ "rstrip": false,
51
+ "single_word": false,
52
+ "special": false
53
+ },
54
+ "151649": {
55
+ "content": "</think>",
56
+ "lstrip": false,
57
+ "normalized": false,
58
+ "rstrip": false,
59
+ "single_word": false,
60
+ "special": false
61
+ },
62
+ "151650": {
63
+ "content": "<|quad_start|>",
64
+ "lstrip": false,
65
+ "normalized": false,
66
+ "rstrip": false,
67
+ "single_word": false,
68
+ "special": true
69
+ },
70
+ "151651": {
71
+ "content": "<|quad_end|>",
72
+ "lstrip": false,
73
+ "normalized": false,
74
+ "rstrip": false,
75
+ "single_word": false,
76
+ "special": true
77
+ },
78
+ "151652": {
79
+ "content": "<|vision_start|>",
80
+ "lstrip": false,
81
+ "normalized": false,
82
+ "rstrip": false,
83
+ "single_word": false,
84
+ "special": true
85
+ },
86
+ "151653": {
87
+ "content": "<|vision_end|>",
88
+ "lstrip": false,
89
+ "normalized": false,
90
+ "rstrip": false,
91
+ "single_word": false,
92
+ "special": true
93
+ },
94
+ "151654": {
95
+ "content": "<|vision_pad|>",
96
+ "lstrip": false,
97
+ "normalized": false,
98
+ "rstrip": false,
99
+ "single_word": false,
100
+ "special": true
101
+ },
102
+ "151655": {
103
+ "content": "<|image_pad|>",
104
+ "lstrip": false,
105
+ "normalized": false,
106
+ "rstrip": false,
107
+ "single_word": false,
108
+ "special": true
109
+ },
110
+ "151656": {
111
+ "content": "<|video_pad|>",
112
+ "lstrip": false,
113
+ "normalized": false,
114
+ "rstrip": false,
115
+ "single_word": false,
116
+ "special": true
117
+ },
118
+ "151657": {
119
+ "content": "<tool_call>",
120
+ "lstrip": false,
121
+ "normalized": false,
122
+ "rstrip": false,
123
+ "single_word": false,
124
+ "special": false
125
+ },
126
+ "151658": {
127
+ "content": "</tool_call>",
128
+ "lstrip": false,
129
+ "normalized": false,
130
+ "rstrip": false,
131
+ "single_word": false,
132
+ "special": false
133
+ },
134
+ "151659": {
135
+ "content": "<|fim_prefix|>",
136
+ "lstrip": false,
137
+ "normalized": false,
138
+ "rstrip": false,
139
+ "single_word": false,
140
+ "special": false
141
+ },
142
+ "151660": {
143
+ "content": "<|fim_middle|>",
144
+ "lstrip": false,
145
+ "normalized": false,
146
+ "rstrip": false,
147
+ "single_word": false,
148
+ "special": false
149
+ },
150
+ "151661": {
151
+ "content": "<|fim_suffix|>",
152
+ "lstrip": false,
153
+ "normalized": false,
154
+ "rstrip": false,
155
+ "single_word": false,
156
+ "special": false
157
+ },
158
+ "151662": {
159
+ "content": "<|fim_pad|>",
160
+ "lstrip": false,
161
+ "normalized": false,
162
+ "rstrip": false,
163
+ "single_word": false,
164
+ "special": false
165
+ },
166
+ "151663": {
167
+ "content": "<|repo_name|>",
168
+ "lstrip": false,
169
+ "normalized": false,
170
+ "rstrip": false,
171
+ "single_word": false,
172
+ "special": false
173
+ },
174
+ "151664": {
175
+ "content": "<|file_sep|>",
176
+ "lstrip": false,
177
+ "normalized": false,
178
+ "rstrip": false,
179
+ "single_word": false,
180
+ "special": false
181
+ }
182
+ },
183
+ "bos_token": "<｜begin▁of▁sentence｜>",
184
+ "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='') %}{%- for message in messages %}{%- if message['role'] == 'system' %}{% set ns.system_prompt = message['content'] %}{%- endif %}{%- endfor %}{{bos_token}}{{ns.system_prompt}}{%- for message in messages %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{{'<｜User｜>' + message['content']}}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is none %}{%- set ns.is_tool = false -%}{%- for tool in message['tool_calls']%}{%- if not ns.is_first %}{{'<｜Assistant｜><｜tool▁calls▁begin｜><｜tool▁call▁begin��>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{%- set ns.is_first = true -%}{%- else %}{{'\\n' + '<｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{{'<｜tool▁calls▁end｜><｜end▁of▁sentence｜>'}}{%- endif %}{%- endfor %}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is not none %}{%- if ns.is_tool %}{{'<｜tool▁outputs▁end｜>' + message['content'] + '<｜end▁of▁sentence｜>'}}{%- set ns.is_tool = false -%}{%- else %}{% set content = message['content'] %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{% endif %}{{'<｜Assistant｜>' + content + '<｜end▁of▁sentence｜>'}}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<｜tool▁outputs▁begin｜><｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- set ns.is_output_first = false %}{%- else %}{{'\\n<｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<｜tool▁outputs▁end｜>'}}{% endif %}{% if add_generation_prompt and not ns.is_tool %}{{'<｜Assistant｜><think>\\n'}}{% endif %}",
185
+ "clean_up_tokenization_spaces": false,
186
+ "eos_token": "<｜end▁of▁sentence｜>",
187
+ "extra_special_tokens": {},
188
+ "legacy": true,
189
+ "model_max_length": 16384,
190
+ "pad_token": "<｜end▁of▁sentence｜>",
191
+ "sp_model_kwargs": {},
192
+ "tokenizer_class": "LlamaTokenizerFast",
193
+ "unk_token": null,
194
+ "use_default_system_prompt": false
195
+ }
DeepSeek-R1-Distill-Qwen-14B-3bit/config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151643,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 5120,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 13824,
12
+ "max_position_embeddings": 131072,
13
+ "max_window_layers": 48,
14
+ "model_type": "qwen2",
15
+ "num_attention_heads": 40,
16
+ "num_hidden_layers": 48,
17
+ "num_key_value_heads": 8,
18
+ "quantization": {
19
+ "group_size": 64,
20
+ "bits": 3
21
+ },
22
+ "quantization_config": {
23
+ "group_size": 64,
24
+ "bits": 3
25
+ },
26
+ "rms_norm_eps": 1e-05,
27
+ "rope_theta": 1000000.0,
28
+ "sliding_window": 131072,
29
+ "tie_word_embeddings": false,
30
+ "torch_dtype": "bfloat16",
31
+ "transformers_version": "4.43.1",
32
+ "use_cache": true,
33
+ "use_sliding_window": false,
34
+ "vocab_size": 152064
35
+ }
DeepSeek-R1-Distill-Qwen-14B-3bit/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9688fa20775d61c528a48a239a4a57e58547c5a4f577273a5bf4ec00ad971bf9
3
+ size 5366751564
DeepSeek-R1-Distill-Qwen-14B-3bit/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f7dc1b46731e6536b10a3b39388101040e4b0c76235aef48a9c6110c698fbcf
3
+ size 1096593521
DeepSeek-R1-Distill-Qwen-14B-3bit/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
DeepSeek-R1-Distill-Qwen-14B-3bit/special_tokens_map.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<｜begin▁of▁sentence｜>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<｜end▁of▁sentence｜>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<｜end▁of▁sentence｜>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ }
23
+ }
DeepSeek-R1-Distill-Qwen-14B-3bit/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e20ddafc659ba90242154b55275402edeca0715e5dbb30f56815a4ce081f4893
3
+ size 11422778
DeepSeek-R1-Distill-Qwen-14B-3bit/tokenizer_config.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "151643": {
7
+ "content": "<｜end▁of▁sentence｜>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "151644": {
15
+ "content": "<｜User｜>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": false
21
+ },
22
+ "151645": {
23
+ "content": "<｜Assistant｜>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": false
29
+ },
30
+ "151646": {
31
+ "content": "<｜begin▁of▁sentence｜>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false,
36
+ "special": true
37
+ },
38
+ "151647": {
39
+ "content": "<|EOT|>",
40
+ "lstrip": false,
41
+ "normalized": false,
42
+ "rstrip": false,
43
+ "single_word": false,
44
+ "special": false
45
+ },
46
+ "151648": {
47
+ "content": "<think>",
48
+ "lstrip": false,
49
+ "normalized": false,
50
+ "rstrip": false,
51
+ "single_word": false,
52
+ "special": false
53
+ },
54
+ "151649": {
55
+ "content": "</think>",
56
+ "lstrip": false,
57
+ "normalized": false,
58
+ "rstrip": false,
59
+ "single_word": false,
60
+ "special": false
61
+ },
62
+ "151650": {
63
+ "content": "<|quad_start|>",
64
+ "lstrip": false,
65
+ "normalized": false,
66
+ "rstrip": false,
67
+ "single_word": false,
68
+ "special": true
69
+ },
70
+ "151651": {
71
+ "content": "<|quad_end|>",
72
+ "lstrip": false,
73
+ "normalized": false,
74
+ "rstrip": false,
75
+ "single_word": false,
76
+ "special": true
77
+ },
78
+ "151652": {
79
+ "content": "<|vision_start|>",
80
+ "lstrip": false,
81
+ "normalized": false,
82
+ "rstrip": false,
83
+ "single_word": false,
84
+ "special": true
85
+ },
86
+ "151653": {
87
+ "content": "<|vision_end|>",
88
+ "lstrip": false,
89
+ "normalized": false,
90
+ "rstrip": false,
91
+ "single_word": false,
92
+ "special": true
93
+ },
94
+ "151654": {
95
+ "content": "<|vision_pad|>",
96
+ "lstrip": false,
97
+ "normalized": false,
98
+ "rstrip": false,
99
+ "single_word": false,
100
+ "special": true
101
+ },
102
+ "151655": {
103
+ "content": "<|image_pad|>",
104
+ "lstrip": false,
105
+ "normalized": false,
106
+ "rstrip": false,
107
+ "single_word": false,
108
+ "special": true
109
+ },
110
+ "151656": {
111
+ "content": "<|video_pad|>",
112
+ "lstrip": false,
113
+ "normalized": false,
114
+ "rstrip": false,
115
+ "single_word": false,
116
+ "special": true
117
+ },
118
+ "151657": {
119
+ "content": "<tool_call>",
120
+ "lstrip": false,
121
+ "normalized": false,
122
+ "rstrip": false,
123
+ "single_word": false,
124
+ "special": false
125
+ },
126
+ "151658": {
127
+ "content": "</tool_call>",
128
+ "lstrip": false,
129
+ "normalized": false,
130
+ "rstrip": false,
131
+ "single_word": false,
132
+ "special": false
133
+ },
134
+ "151659": {
135
+ "content": "<|fim_prefix|>",
136
+ "lstrip": false,
137
+ "normalized": false,
138
+ "rstrip": false,
139
+ "single_word": false,
140
+ "special": false
141
+ },
142
+ "151660": {
143
+ "content": "<|fim_middle|>",
144
+ "lstrip": false,
145
+ "normalized": false,
146
+ "rstrip": false,
147
+ "single_word": false,
148
+ "special": false
149
+ },
150
+ "151661": {
151
+ "content": "<|fim_suffix|>",
152
+ "lstrip": false,
153
+ "normalized": false,
154
+ "rstrip": false,
155
+ "single_word": false,
156
+ "special": false
157
+ },
158
+ "151662": {
159
+ "content": "<|fim_pad|>",
160
+ "lstrip": false,
161
+ "normalized": false,
162
+ "rstrip": false,
163
+ "single_word": false,
164
+ "special": false
165
+ },
166
+ "151663": {
167
+ "content": "<|repo_name|>",
168
+ "lstrip": false,
169
+ "normalized": false,
170
+ "rstrip": false,
171
+ "single_word": false,
172
+ "special": false
173
+ },
174
+ "151664": {
175
+ "content": "<|file_sep|>",
176
+ "lstrip": false,
177
+ "normalized": false,
178
+ "rstrip": false,
179
+ "single_word": false,
180
+ "special": false
181
+ }
182
+ },
183
+ "bos_token": "<｜begin▁of▁sentence｜>",
184
+ "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='') %}{%- for message in messages %}{%- if message['role'] == 'system' %}{% set ns.system_prompt = message['content'] %}{%- endif %}{%- endfor %}{{bos_token}}{{ns.system_prompt}}{%- for message in messages %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{{'<｜User｜>' + message['content']}}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is none %}{%- set ns.is_tool = false -%}{%- for tool in message['tool_calls']%}{%- if not ns.is_first %}{{'<｜Assistant｜><｜tool▁calls▁begin｜><｜tool▁call▁begin��>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{%- set ns.is_first = true -%}{%- else %}{{'\\n' + '<｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{{'<｜tool▁calls▁end｜><｜end▁of▁sentence｜>'}}{%- endif %}{%- endfor %}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is not none %}{%- if ns.is_tool %}{{'<｜tool▁outputs▁end｜>' + message['content'] + '<｜end▁of▁sentence｜>'}}{%- set ns.is_tool = false -%}{%- else %}{% set content = message['content'] %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{% endif %}{{'<｜Assistant｜>' + content + '<｜end▁of▁sentence｜>'}}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<｜tool▁outputs▁begin｜><｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- set ns.is_output_first = false %}{%- else %}{{'\\n<｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<｜tool▁outputs▁end｜>'}}{% endif %}{% if add_generation_prompt and not ns.is_tool %}{{'<｜Assistant｜><think>\\n'}}{% endif %}",
185
+ "clean_up_tokenization_spaces": false,
186
+ "eos_token": "<｜end▁of▁sentence｜>",
187
+ "extra_special_tokens": {},
188
+ "legacy": true,
189
+ "model_max_length": 16384,
190
+ "pad_token": "<｜end▁of▁sentence｜>",
191
+ "sp_model_kwargs": {},
192
+ "tokenizer_class": "LlamaTokenizerFast",
193
+ "unk_token": null,
194
+ "use_default_system_prompt": false
195
+ }
DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/config.json ADDED
@@ -0,0 +1,3029 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151643,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 5120,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 13824,
12
+ "max_position_embeddings": 131072,
13
+ "max_window_layers": 48,
14
+ "model_type": "qwen2",
15
+ "num_attention_heads": 40,
16
+ "num_hidden_layers": 48,
17
+ "num_key_value_heads": 8,
18
+ "quantization": {
19
+ "group_size": 64,
20
+ "bits": null,
21
+ "model.embed_tokens": {
22
+ "group_size": 64,
23
+ "bits": 4
24
+ },
25
+ "model.layers.0.self_attn.q_proj": {
26
+ "group_size": 64,
27
+ "bits": 4
28
+ },
29
+ "model.layers.0.self_attn.k_proj": {
30
+ "group_size": 64,
31
+ "bits": 4
32
+ },
33
+ "model.layers.0.self_attn.v_proj": {
34
+ "group_size": 64,
35
+ "bits": 6
36
+ },
37
+ "model.layers.0.self_attn.o_proj": {
38
+ "group_size": 64,
39
+ "bits": 4
40
+ },
41
+ "model.layers.0.self_attn.rope": false,
42
+ "model.layers.0.mlp.gate_proj": {
43
+ "group_size": 64,
44
+ "bits": 4
45
+ },
46
+ "model.layers.0.mlp.down_proj": {
47
+ "group_size": 64,
48
+ "bits": 6
49
+ },
50
+ "model.layers.0.mlp.up_proj": {
51
+ "group_size": 64,
52
+ "bits": 4
53
+ },
54
+ "model.layers.0.input_layernorm": false,
55
+ "model.layers.0.post_attention_layernorm": false,
56
+ "model.layers.1.self_attn.q_proj": {
57
+ "group_size": 64,
58
+ "bits": 4
59
+ },
60
+ "model.layers.1.self_attn.k_proj": {
61
+ "group_size": 64,
62
+ "bits": 4
63
+ },
64
+ "model.layers.1.self_attn.v_proj": {
65
+ "group_size": 64,
66
+ "bits": 6
67
+ },
68
+ "model.layers.1.self_attn.o_proj": {
69
+ "group_size": 64,
70
+ "bits": 4
71
+ },
72
+ "model.layers.1.self_attn.rope": false,
73
+ "model.layers.1.mlp.gate_proj": {
74
+ "group_size": 64,
75
+ "bits": 4
76
+ },
77
+ "model.layers.1.mlp.down_proj": {
78
+ "group_size": 64,
79
+ "bits": 6
80
+ },
81
+ "model.layers.1.mlp.up_proj": {
82
+ "group_size": 64,
83
+ "bits": 4
84
+ },
85
+ "model.layers.1.input_layernorm": false,
86
+ "model.layers.1.post_attention_layernorm": false,
87
+ "model.layers.2.self_attn.q_proj": {
88
+ "group_size": 64,
89
+ "bits": 4
90
+ },
91
+ "model.layers.2.self_attn.k_proj": {
92
+ "group_size": 64,
93
+ "bits": 4
94
+ },
95
+ "model.layers.2.self_attn.v_proj": {
96
+ "group_size": 64,
97
+ "bits": 6
98
+ },
99
+ "model.layers.2.self_attn.o_proj": {
100
+ "group_size": 64,
101
+ "bits": 4
102
+ },
103
+ "model.layers.2.self_attn.rope": false,
104
+ "model.layers.2.mlp.gate_proj": {
105
+ "group_size": 64,
106
+ "bits": 4
107
+ },
108
+ "model.layers.2.mlp.down_proj": {
109
+ "group_size": 64,
110
+ "bits": 6
111
+ },
112
+ "model.layers.2.mlp.up_proj": {
113
+ "group_size": 64,
114
+ "bits": 4
115
+ },
116
+ "model.layers.2.input_layernorm": false,
117
+ "model.layers.2.post_attention_layernorm": false,
118
+ "model.layers.3.self_attn.q_proj": {
119
+ "group_size": 64,
120
+ "bits": 4
121
+ },
122
+ "model.layers.3.self_attn.k_proj": {
123
+ "group_size": 64,
124
+ "bits": 4
125
+ },
126
+ "model.layers.3.self_attn.v_proj": {
127
+ "group_size": 64,
128
+ "bits": 6
129
+ },
130
+ "model.layers.3.self_attn.o_proj": {
131
+ "group_size": 64,
132
+ "bits": 4
133
+ },
134
+ "model.layers.3.self_attn.rope": false,
135
+ "model.layers.3.mlp.gate_proj": {
136
+ "group_size": 64,
137
+ "bits": 4
138
+ },
139
+ "model.layers.3.mlp.down_proj": {
140
+ "group_size": 64,
141
+ "bits": 6
142
+ },
143
+ "model.layers.3.mlp.up_proj": {
144
+ "group_size": 64,
145
+ "bits": 4
146
+ },
147
+ "model.layers.3.input_layernorm": false,
148
+ "model.layers.3.post_attention_layernorm": false,
149
+ "model.layers.4.self_attn.q_proj": {
150
+ "group_size": 64,
151
+ "bits": 4
152
+ },
153
+ "model.layers.4.self_attn.k_proj": {
154
+ "group_size": 64,
155
+ "bits": 4
156
+ },
157
+ "model.layers.4.self_attn.v_proj": {
158
+ "group_size": 64,
159
+ "bits": 6
160
+ },
161
+ "model.layers.4.self_attn.o_proj": {
162
+ "group_size": 64,
163
+ "bits": 4
164
+ },
165
+ "model.layers.4.self_attn.rope": false,
166
+ "model.layers.4.mlp.gate_proj": {
167
+ "group_size": 64,
168
+ "bits": 4
169
+ },
170
+ "model.layers.4.mlp.down_proj": {
171
+ "group_size": 64,
172
+ "bits": 6
173
+ },
174
+ "model.layers.4.mlp.up_proj": {
175
+ "group_size": 64,
176
+ "bits": 4
177
+ },
178
+ "model.layers.4.input_layernorm": false,
179
+ "model.layers.4.post_attention_layernorm": false,
180
+ "model.layers.5.self_attn.q_proj": {
181
+ "group_size": 64,
182
+ "bits": 4
183
+ },
184
+ "model.layers.5.self_attn.k_proj": {
185
+ "group_size": 64,
186
+ "bits": 4
187
+ },
188
+ "model.layers.5.self_attn.v_proj": {
189
+ "group_size": 64,
190
+ "bits": 6
191
+ },
192
+ "model.layers.5.self_attn.o_proj": {
193
+ "group_size": 64,
194
+ "bits": 4
195
+ },
196
+ "model.layers.5.self_attn.rope": false,
197
+ "model.layers.5.mlp.gate_proj": {
198
+ "group_size": 64,
199
+ "bits": 4
200
+ },
201
+ "model.layers.5.mlp.down_proj": {
202
+ "group_size": 64,
203
+ "bits": 6
204
+ },
205
+ "model.layers.5.mlp.up_proj": {
206
+ "group_size": 64,
207
+ "bits": 4
208
+ },
209
+ "model.layers.5.input_layernorm": false,
210
+ "model.layers.5.post_attention_layernorm": false,
211
+ "model.layers.6.self_attn.q_proj": {
212
+ "group_size": 64,
213
+ "bits": 4
214
+ },
215
+ "model.layers.6.self_attn.k_proj": {
216
+ "group_size": 64,
217
+ "bits": 4
218
+ },
219
+ "model.layers.6.self_attn.v_proj": {
220
+ "group_size": 64,
221
+ "bits": 4
222
+ },
223
+ "model.layers.6.self_attn.o_proj": {
224
+ "group_size": 64,
225
+ "bits": 4
226
+ },
227
+ "model.layers.6.self_attn.rope": false,
228
+ "model.layers.6.mlp.gate_proj": {
229
+ "group_size": 64,
230
+ "bits": 4
231
+ },
232
+ "model.layers.6.mlp.down_proj": {
233
+ "group_size": 64,
234
+ "bits": 4
235
+ },
236
+ "model.layers.6.mlp.up_proj": {
237
+ "group_size": 64,
238
+ "bits": 4
239
+ },
240
+ "model.layers.6.input_layernorm": false,
241
+ "model.layers.6.post_attention_layernorm": false,
242
+ "model.layers.7.self_attn.q_proj": {
243
+ "group_size": 64,
244
+ "bits": 4
245
+ },
246
+ "model.layers.7.self_attn.k_proj": {
247
+ "group_size": 64,
248
+ "bits": 4
249
+ },
250
+ "model.layers.7.self_attn.v_proj": {
251
+ "group_size": 64,
252
+ "bits": 4
253
+ },
254
+ "model.layers.7.self_attn.o_proj": {
255
+ "group_size": 64,
256
+ "bits": 4
257
+ },
258
+ "model.layers.7.self_attn.rope": false,
259
+ "model.layers.7.mlp.gate_proj": {
260
+ "group_size": 64,
261
+ "bits": 4
262
+ },
263
+ "model.layers.7.mlp.down_proj": {
264
+ "group_size": 64,
265
+ "bits": 4
266
+ },
267
+ "model.layers.7.mlp.up_proj": {
268
+ "group_size": 64,
269
+ "bits": 4
270
+ },
271
+ "model.layers.7.input_layernorm": false,
272
+ "model.layers.7.post_attention_layernorm": false,
273
+ "model.layers.8.self_attn.q_proj": {
274
+ "group_size": 64,
275
+ "bits": 4
276
+ },
277
+ "model.layers.8.self_attn.k_proj": {
278
+ "group_size": 64,
279
+ "bits": 4
280
+ },
281
+ "model.layers.8.self_attn.v_proj": {
282
+ "group_size": 64,
283
+ "bits": 6
284
+ },
285
+ "model.layers.8.self_attn.o_proj": {
286
+ "group_size": 64,
287
+ "bits": 4
288
+ },
289
+ "model.layers.8.self_attn.rope": false,
290
+ "model.layers.8.mlp.gate_proj": {
291
+ "group_size": 64,
292
+ "bits": 4
293
+ },
294
+ "model.layers.8.mlp.down_proj": {
295
+ "group_size": 64,
296
+ "bits": 6
297
+ },
298
+ "model.layers.8.mlp.up_proj": {
299
+ "group_size": 64,
300
+ "bits": 4
301
+ },
302
+ "model.layers.8.input_layernorm": false,
303
+ "model.layers.8.post_attention_layernorm": false,
304
+ "model.layers.9.self_attn.q_proj": {
305
+ "group_size": 64,
306
+ "bits": 4
307
+ },
308
+ "model.layers.9.self_attn.k_proj": {
309
+ "group_size": 64,
310
+ "bits": 4
311
+ },
312
+ "model.layers.9.self_attn.v_proj": {
313
+ "group_size": 64,
314
+ "bits": 4
315
+ },
316
+ "model.layers.9.self_attn.o_proj": {
317
+ "group_size": 64,
318
+ "bits": 4
319
+ },
320
+ "model.layers.9.self_attn.rope": false,
321
+ "model.layers.9.mlp.gate_proj": {
322
+ "group_size": 64,
323
+ "bits": 4
324
+ },
325
+ "model.layers.9.mlp.down_proj": {
326
+ "group_size": 64,
327
+ "bits": 4
328
+ },
329
+ "model.layers.9.mlp.up_proj": {
330
+ "group_size": 64,
331
+ "bits": 4
332
+ },
333
+ "model.layers.9.input_layernorm": false,
334
+ "model.layers.9.post_attention_layernorm": false,
335
+ "model.layers.10.self_attn.q_proj": {
336
+ "group_size": 64,
337
+ "bits": 4
338
+ },
339
+ "model.layers.10.self_attn.k_proj": {
340
+ "group_size": 64,
341
+ "bits": 4
342
+ },
343
+ "model.layers.10.self_attn.v_proj": {
344
+ "group_size": 64,
345
+ "bits": 4
346
+ },
347
+ "model.layers.10.self_attn.o_proj": {
348
+ "group_size": 64,
349
+ "bits": 4
350
+ },
351
+ "model.layers.10.self_attn.rope": false,
352
+ "model.layers.10.mlp.gate_proj": {
353
+ "group_size": 64,
354
+ "bits": 4
355
+ },
356
+ "model.layers.10.mlp.down_proj": {
357
+ "group_size": 64,
358
+ "bits": 4
359
+ },
360
+ "model.layers.10.mlp.up_proj": {
361
+ "group_size": 64,
362
+ "bits": 4
363
+ },
364
+ "model.layers.10.input_layernorm": false,
365
+ "model.layers.10.post_attention_layernorm": false,
366
+ "model.layers.11.self_attn.q_proj": {
367
+ "group_size": 64,
368
+ "bits": 4
369
+ },
370
+ "model.layers.11.self_attn.k_proj": {
371
+ "group_size": 64,
372
+ "bits": 4
373
+ },
374
+ "model.layers.11.self_attn.v_proj": {
375
+ "group_size": 64,
376
+ "bits": 6
377
+ },
378
+ "model.layers.11.self_attn.o_proj": {
379
+ "group_size": 64,
380
+ "bits": 4
381
+ },
382
+ "model.layers.11.self_attn.rope": false,
383
+ "model.layers.11.mlp.gate_proj": {
384
+ "group_size": 64,
385
+ "bits": 4
386
+ },
387
+ "model.layers.11.mlp.down_proj": {
388
+ "group_size": 64,
389
+ "bits": 6
390
+ },
391
+ "model.layers.11.mlp.up_proj": {
392
+ "group_size": 64,
393
+ "bits": 4
394
+ },
395
+ "model.layers.11.input_layernorm": false,
396
+ "model.layers.11.post_attention_layernorm": false,
397
+ "model.layers.12.self_attn.q_proj": {
398
+ "group_size": 64,
399
+ "bits": 4
400
+ },
401
+ "model.layers.12.self_attn.k_proj": {
402
+ "group_size": 64,
403
+ "bits": 4
404
+ },
405
+ "model.layers.12.self_attn.v_proj": {
406
+ "group_size": 64,
407
+ "bits": 4
408
+ },
409
+ "model.layers.12.self_attn.o_proj": {
410
+ "group_size": 64,
411
+ "bits": 4
412
+ },
413
+ "model.layers.12.self_attn.rope": false,
414
+ "model.layers.12.mlp.gate_proj": {
415
+ "group_size": 64,
416
+ "bits": 4
417
+ },
418
+ "model.layers.12.mlp.down_proj": {
419
+ "group_size": 64,
420
+ "bits": 4
421
+ },
422
+ "model.layers.12.mlp.up_proj": {
423
+ "group_size": 64,
424
+ "bits": 4
425
+ },
426
+ "model.layers.12.input_layernorm": false,
427
+ "model.layers.12.post_attention_layernorm": false,
428
+ "model.layers.13.self_attn.q_proj": {
429
+ "group_size": 64,
430
+ "bits": 4
431
+ },
432
+ "model.layers.13.self_attn.k_proj": {
433
+ "group_size": 64,
434
+ "bits": 4
435
+ },
436
+ "model.layers.13.self_attn.v_proj": {
437
+ "group_size": 64,
438
+ "bits": 4
439
+ },
440
+ "model.layers.13.self_attn.o_proj": {
441
+ "group_size": 64,
442
+ "bits": 4
443
+ },
444
+ "model.layers.13.self_attn.rope": false,
445
+ "model.layers.13.mlp.gate_proj": {
446
+ "group_size": 64,
447
+ "bits": 4
448
+ },
449
+ "model.layers.13.mlp.down_proj": {
450
+ "group_size": 64,
451
+ "bits": 4
452
+ },
453
+ "model.layers.13.mlp.up_proj": {
454
+ "group_size": 64,
455
+ "bits": 4
456
+ },
457
+ "model.layers.13.input_layernorm": false,
458
+ "model.layers.13.post_attention_layernorm": false,
459
+ "model.layers.14.self_attn.q_proj": {
460
+ "group_size": 64,
461
+ "bits": 4
462
+ },
463
+ "model.layers.14.self_attn.k_proj": {
464
+ "group_size": 64,
465
+ "bits": 4
466
+ },
467
+ "model.layers.14.self_attn.v_proj": {
468
+ "group_size": 64,
469
+ "bits": 6
470
+ },
471
+ "model.layers.14.self_attn.o_proj": {
472
+ "group_size": 64,
473
+ "bits": 4
474
+ },
475
+ "model.layers.14.self_attn.rope": false,
476
+ "model.layers.14.mlp.gate_proj": {
477
+ "group_size": 64,
478
+ "bits": 4
479
+ },
480
+ "model.layers.14.mlp.down_proj": {
481
+ "group_size": 64,
482
+ "bits": 6
483
+ },
484
+ "model.layers.14.mlp.up_proj": {
485
+ "group_size": 64,
486
+ "bits": 4
487
+ },
488
+ "model.layers.14.input_layernorm": false,
489
+ "model.layers.14.post_attention_layernorm": false,
490
+ "model.layers.15.self_attn.q_proj": {
491
+ "group_size": 64,
492
+ "bits": 4
493
+ },
494
+ "model.layers.15.self_attn.k_proj": {
495
+ "group_size": 64,
496
+ "bits": 4
497
+ },
498
+ "model.layers.15.self_attn.v_proj": {
499
+ "group_size": 64,
500
+ "bits": 4
501
+ },
502
+ "model.layers.15.self_attn.o_proj": {
503
+ "group_size": 64,
504
+ "bits": 4
505
+ },
506
+ "model.layers.15.self_attn.rope": false,
507
+ "model.layers.15.mlp.gate_proj": {
508
+ "group_size": 64,
509
+ "bits": 4
510
+ },
511
+ "model.layers.15.mlp.down_proj": {
512
+ "group_size": 64,
513
+ "bits": 4
514
+ },
515
+ "model.layers.15.mlp.up_proj": {
516
+ "group_size": 64,
517
+ "bits": 4
518
+ },
519
+ "model.layers.15.input_layernorm": false,
520
+ "model.layers.15.post_attention_layernorm": false,
521
+ "model.layers.16.self_attn.q_proj": {
522
+ "group_size": 64,
523
+ "bits": 4
524
+ },
525
+ "model.layers.16.self_attn.k_proj": {
526
+ "group_size": 64,
527
+ "bits": 4
528
+ },
529
+ "model.layers.16.self_attn.v_proj": {
530
+ "group_size": 64,
531
+ "bits": 4
532
+ },
533
+ "model.layers.16.self_attn.o_proj": {
534
+ "group_size": 64,
535
+ "bits": 4
536
+ },
537
+ "model.layers.16.self_attn.rope": false,
538
+ "model.layers.16.mlp.gate_proj": {
539
+ "group_size": 64,
540
+ "bits": 4
541
+ },
542
+ "model.layers.16.mlp.down_proj": {
543
+ "group_size": 64,
544
+ "bits": 4
545
+ },
546
+ "model.layers.16.mlp.up_proj": {
547
+ "group_size": 64,
548
+ "bits": 4
549
+ },
550
+ "model.layers.16.input_layernorm": false,
551
+ "model.layers.16.post_attention_layernorm": false,
552
+ "model.layers.17.self_attn.q_proj": {
553
+ "group_size": 64,
554
+ "bits": 4
555
+ },
556
+ "model.layers.17.self_attn.k_proj": {
557
+ "group_size": 64,
558
+ "bits": 4
559
+ },
560
+ "model.layers.17.self_attn.v_proj": {
561
+ "group_size": 64,
562
+ "bits": 6
563
+ },
564
+ "model.layers.17.self_attn.o_proj": {
565
+ "group_size": 64,
566
+ "bits": 4
567
+ },
568
+ "model.layers.17.self_attn.rope": false,
569
+ "model.layers.17.mlp.gate_proj": {
570
+ "group_size": 64,
571
+ "bits": 4
572
+ },
573
+ "model.layers.17.mlp.down_proj": {
574
+ "group_size": 64,
575
+ "bits": 6
576
+ },
577
+ "model.layers.17.mlp.up_proj": {
578
+ "group_size": 64,
579
+ "bits": 4
580
+ },
581
+ "model.layers.17.input_layernorm": false,
582
+ "model.layers.17.post_attention_layernorm": false,
583
+ "model.layers.18.self_attn.q_proj": {
584
+ "group_size": 64,
585
+ "bits": 4
586
+ },
587
+ "model.layers.18.self_attn.k_proj": {
588
+ "group_size": 64,
589
+ "bits": 4
590
+ },
591
+ "model.layers.18.self_attn.v_proj": {
592
+ "group_size": 64,
593
+ "bits": 4
594
+ },
595
+ "model.layers.18.self_attn.o_proj": {
596
+ "group_size": 64,
597
+ "bits": 4
598
+ },
599
+ "model.layers.18.self_attn.rope": false,
600
+ "model.layers.18.mlp.gate_proj": {
601
+ "group_size": 64,
602
+ "bits": 4
603
+ },
604
+ "model.layers.18.mlp.down_proj": {
605
+ "group_size": 64,
606
+ "bits": 4
607
+ },
608
+ "model.layers.18.mlp.up_proj": {
609
+ "group_size": 64,
610
+ "bits": 4
611
+ },
612
+ "model.layers.18.input_layernorm": false,
613
+ "model.layers.18.post_attention_layernorm": false,
614
+ "model.layers.19.self_attn.q_proj": {
615
+ "group_size": 64,
616
+ "bits": 4
617
+ },
618
+ "model.layers.19.self_attn.k_proj": {
619
+ "group_size": 64,
620
+ "bits": 4
621
+ },
622
+ "model.layers.19.self_attn.v_proj": {
623
+ "group_size": 64,
624
+ "bits": 4
625
+ },
626
+ "model.layers.19.self_attn.o_proj": {
627
+ "group_size": 64,
628
+ "bits": 4
629
+ },
630
+ "model.layers.19.self_attn.rope": false,
631
+ "model.layers.19.mlp.gate_proj": {
632
+ "group_size": 64,
633
+ "bits": 4
634
+ },
635
+ "model.layers.19.mlp.down_proj": {
636
+ "group_size": 64,
637
+ "bits": 4
638
+ },
639
+ "model.layers.19.mlp.up_proj": {
640
+ "group_size": 64,
641
+ "bits": 4
642
+ },
643
+ "model.layers.19.input_layernorm": false,
644
+ "model.layers.19.post_attention_layernorm": false,
645
+ "model.layers.20.self_attn.q_proj": {
646
+ "group_size": 64,
647
+ "bits": 4
648
+ },
649
+ "model.layers.20.self_attn.k_proj": {
650
+ "group_size": 64,
651
+ "bits": 4
652
+ },
653
+ "model.layers.20.self_attn.v_proj": {
654
+ "group_size": 64,
655
+ "bits": 6
656
+ },
657
+ "model.layers.20.self_attn.o_proj": {
658
+ "group_size": 64,
659
+ "bits": 4
660
+ },
661
+ "model.layers.20.self_attn.rope": false,
662
+ "model.layers.20.mlp.gate_proj": {
663
+ "group_size": 64,
664
+ "bits": 4
665
+ },
666
+ "model.layers.20.mlp.down_proj": {
667
+ "group_size": 64,
668
+ "bits": 6
669
+ },
670
+ "model.layers.20.mlp.up_proj": {
671
+ "group_size": 64,
672
+ "bits": 4
673
+ },
674
+ "model.layers.20.input_layernorm": false,
675
+ "model.layers.20.post_attention_layernorm": false,
676
+ "model.layers.21.self_attn.q_proj": {
677
+ "group_size": 64,
678
+ "bits": 4
679
+ },
680
+ "model.layers.21.self_attn.k_proj": {
681
+ "group_size": 64,
682
+ "bits": 4
683
+ },
684
+ "model.layers.21.self_attn.v_proj": {
685
+ "group_size": 64,
686
+ "bits": 4
687
+ },
688
+ "model.layers.21.self_attn.o_proj": {
689
+ "group_size": 64,
690
+ "bits": 4
691
+ },
692
+ "model.layers.21.self_attn.rope": false,
693
+ "model.layers.21.mlp.gate_proj": {
694
+ "group_size": 64,
695
+ "bits": 4
696
+ },
697
+ "model.layers.21.mlp.down_proj": {
698
+ "group_size": 64,
699
+ "bits": 4
700
+ },
701
+ "model.layers.21.mlp.up_proj": {
702
+ "group_size": 64,
703
+ "bits": 4
704
+ },
705
+ "model.layers.21.input_layernorm": false,
706
+ "model.layers.21.post_attention_layernorm": false,
707
+ "model.layers.22.self_attn.q_proj": {
708
+ "group_size": 64,
709
+ "bits": 4
710
+ },
711
+ "model.layers.22.self_attn.k_proj": {
712
+ "group_size": 64,
713
+ "bits": 4
714
+ },
715
+ "model.layers.22.self_attn.v_proj": {
716
+ "group_size": 64,
717
+ "bits": 4
718
+ },
719
+ "model.layers.22.self_attn.o_proj": {
720
+ "group_size": 64,
721
+ "bits": 4
722
+ },
723
+ "model.layers.22.self_attn.rope": false,
724
+ "model.layers.22.mlp.gate_proj": {
725
+ "group_size": 64,
726
+ "bits": 4
727
+ },
728
+ "model.layers.22.mlp.down_proj": {
729
+ "group_size": 64,
730
+ "bits": 4
731
+ },
732
+ "model.layers.22.mlp.up_proj": {
733
+ "group_size": 64,
734
+ "bits": 4
735
+ },
736
+ "model.layers.22.input_layernorm": false,
737
+ "model.layers.22.post_attention_layernorm": false,
738
+ "model.layers.23.self_attn.q_proj": {
739
+ "group_size": 64,
740
+ "bits": 4
741
+ },
742
+ "model.layers.23.self_attn.k_proj": {
743
+ "group_size": 64,
744
+ "bits": 4
745
+ },
746
+ "model.layers.23.self_attn.v_proj": {
747
+ "group_size": 64,
748
+ "bits": 6
749
+ },
750
+ "model.layers.23.self_attn.o_proj": {
751
+ "group_size": 64,
752
+ "bits": 4
753
+ },
754
+ "model.layers.23.self_attn.rope": false,
755
+ "model.layers.23.mlp.gate_proj": {
756
+ "group_size": 64,
757
+ "bits": 4
758
+ },
759
+ "model.layers.23.mlp.down_proj": {
760
+ "group_size": 64,
761
+ "bits": 6
762
+ },
763
+ "model.layers.23.mlp.up_proj": {
764
+ "group_size": 64,
765
+ "bits": 4
766
+ },
767
+ "model.layers.23.input_layernorm": false,
768
+ "model.layers.23.post_attention_layernorm": false,
769
+ "model.layers.24.self_attn.q_proj": {
770
+ "group_size": 64,
771
+ "bits": 4
772
+ },
773
+ "model.layers.24.self_attn.k_proj": {
774
+ "group_size": 64,
775
+ "bits": 4
776
+ },
777
+ "model.layers.24.self_attn.v_proj": {
778
+ "group_size": 64,
779
+ "bits": 4
780
+ },
781
+ "model.layers.24.self_attn.o_proj": {
782
+ "group_size": 64,
783
+ "bits": 4
784
+ },
785
+ "model.layers.24.self_attn.rope": false,
786
+ "model.layers.24.mlp.gate_proj": {
787
+ "group_size": 64,
788
+ "bits": 4
789
+ },
790
+ "model.layers.24.mlp.down_proj": {
791
+ "group_size": 64,
792
+ "bits": 4
793
+ },
794
+ "model.layers.24.mlp.up_proj": {
795
+ "group_size": 64,
796
+ "bits": 4
797
+ },
798
+ "model.layers.24.input_layernorm": false,
799
+ "model.layers.24.post_attention_layernorm": false,
800
+ "model.layers.25.self_attn.q_proj": {
801
+ "group_size": 64,
802
+ "bits": 4
803
+ },
804
+ "model.layers.25.self_attn.k_proj": {
805
+ "group_size": 64,
806
+ "bits": 4
807
+ },
808
+ "model.layers.25.self_attn.v_proj": {
809
+ "group_size": 64,
810
+ "bits": 4
811
+ },
812
+ "model.layers.25.self_attn.o_proj": {
813
+ "group_size": 64,
814
+ "bits": 4
815
+ },
816
+ "model.layers.25.self_attn.rope": false,
817
+ "model.layers.25.mlp.gate_proj": {
818
+ "group_size": 64,
819
+ "bits": 4
820
+ },
821
+ "model.layers.25.mlp.down_proj": {
822
+ "group_size": 64,
823
+ "bits": 4
824
+ },
825
+ "model.layers.25.mlp.up_proj": {
826
+ "group_size": 64,
827
+ "bits": 4
828
+ },
829
+ "model.layers.25.input_layernorm": false,
830
+ "model.layers.25.post_attention_layernorm": false,
831
+ "model.layers.26.self_attn.q_proj": {
832
+ "group_size": 64,
833
+ "bits": 4
834
+ },
835
+ "model.layers.26.self_attn.k_proj": {
836
+ "group_size": 64,
837
+ "bits": 4
838
+ },
839
+ "model.layers.26.self_attn.v_proj": {
840
+ "group_size": 64,
841
+ "bits": 6
842
+ },
843
+ "model.layers.26.self_attn.o_proj": {
844
+ "group_size": 64,
845
+ "bits": 4
846
+ },
847
+ "model.layers.26.self_attn.rope": false,
848
+ "model.layers.26.mlp.gate_proj": {
849
+ "group_size": 64,
850
+ "bits": 4
851
+ },
852
+ "model.layers.26.mlp.down_proj": {
853
+ "group_size": 64,
854
+ "bits": 6
855
+ },
856
+ "model.layers.26.mlp.up_proj": {
857
+ "group_size": 64,
858
+ "bits": 4
859
+ },
860
+ "model.layers.26.input_layernorm": false,
861
+ "model.layers.26.post_attention_layernorm": false,
862
+ "model.layers.27.self_attn.q_proj": {
863
+ "group_size": 64,
864
+ "bits": 4
865
+ },
866
+ "model.layers.27.self_attn.k_proj": {
867
+ "group_size": 64,
868
+ "bits": 4
869
+ },
870
+ "model.layers.27.self_attn.v_proj": {
871
+ "group_size": 64,
872
+ "bits": 4
873
+ },
874
+ "model.layers.27.self_attn.o_proj": {
875
+ "group_size": 64,
876
+ "bits": 4
877
+ },
878
+ "model.layers.27.self_attn.rope": false,
879
+ "model.layers.27.mlp.gate_proj": {
880
+ "group_size": 64,
881
+ "bits": 4
882
+ },
883
+ "model.layers.27.mlp.down_proj": {
884
+ "group_size": 64,
885
+ "bits": 4
886
+ },
887
+ "model.layers.27.mlp.up_proj": {
888
+ "group_size": 64,
889
+ "bits": 4
890
+ },
891
+ "model.layers.27.input_layernorm": false,
892
+ "model.layers.27.post_attention_layernorm": false,
893
+ "model.layers.28.self_attn.q_proj": {
894
+ "group_size": 64,
895
+ "bits": 4
896
+ },
897
+ "model.layers.28.self_attn.k_proj": {
898
+ "group_size": 64,
899
+ "bits": 4
900
+ },
901
+ "model.layers.28.self_attn.v_proj": {
902
+ "group_size": 64,
903
+ "bits": 4
904
+ },
905
+ "model.layers.28.self_attn.o_proj": {
906
+ "group_size": 64,
907
+ "bits": 4
908
+ },
909
+ "model.layers.28.self_attn.rope": false,
910
+ "model.layers.28.mlp.gate_proj": {
911
+ "group_size": 64,
912
+ "bits": 4
913
+ },
914
+ "model.layers.28.mlp.down_proj": {
915
+ "group_size": 64,
916
+ "bits": 4
917
+ },
918
+ "model.layers.28.mlp.up_proj": {
919
+ "group_size": 64,
920
+ "bits": 4
921
+ },
922
+ "model.layers.28.input_layernorm": false,
923
+ "model.layers.28.post_attention_layernorm": false,
924
+ "model.layers.29.self_attn.q_proj": {
925
+ "group_size": 64,
926
+ "bits": 4
927
+ },
928
+ "model.layers.29.self_attn.k_proj": {
929
+ "group_size": 64,
930
+ "bits": 4
931
+ },
932
+ "model.layers.29.self_attn.v_proj": {
933
+ "group_size": 64,
934
+ "bits": 6
935
+ },
936
+ "model.layers.29.self_attn.o_proj": {
937
+ "group_size": 64,
938
+ "bits": 4
939
+ },
940
+ "model.layers.29.self_attn.rope": false,
941
+ "model.layers.29.mlp.gate_proj": {
942
+ "group_size": 64,
943
+ "bits": 4
944
+ },
945
+ "model.layers.29.mlp.down_proj": {
946
+ "group_size": 64,
947
+ "bits": 6
948
+ },
949
+ "model.layers.29.mlp.up_proj": {
950
+ "group_size": 64,
951
+ "bits": 4
952
+ },
953
+ "model.layers.29.input_layernorm": false,
954
+ "model.layers.29.post_attention_layernorm": false,
955
+ "model.layers.30.self_attn.q_proj": {
956
+ "group_size": 64,
957
+ "bits": 4
958
+ },
959
+ "model.layers.30.self_attn.k_proj": {
960
+ "group_size": 64,
961
+ "bits": 4
962
+ },
963
+ "model.layers.30.self_attn.v_proj": {
964
+ "group_size": 64,
965
+ "bits": 4
966
+ },
967
+ "model.layers.30.self_attn.o_proj": {
968
+ "group_size": 64,
969
+ "bits": 4
970
+ },
971
+ "model.layers.30.self_attn.rope": false,
972
+ "model.layers.30.mlp.gate_proj": {
973
+ "group_size": 64,
974
+ "bits": 4
975
+ },
976
+ "model.layers.30.mlp.down_proj": {
977
+ "group_size": 64,
978
+ "bits": 4
979
+ },
980
+ "model.layers.30.mlp.up_proj": {
981
+ "group_size": 64,
982
+ "bits": 4
983
+ },
984
+ "model.layers.30.input_layernorm": false,
985
+ "model.layers.30.post_attention_layernorm": false,
986
+ "model.layers.31.self_attn.q_proj": {
987
+ "group_size": 64,
988
+ "bits": 4
989
+ },
990
+ "model.layers.31.self_attn.k_proj": {
991
+ "group_size": 64,
992
+ "bits": 4
993
+ },
994
+ "model.layers.31.self_attn.v_proj": {
995
+ "group_size": 64,
996
+ "bits": 4
997
+ },
998
+ "model.layers.31.self_attn.o_proj": {
999
+ "group_size": 64,
1000
+ "bits": 4
1001
+ },
1002
+ "model.layers.31.self_attn.rope": false,
1003
+ "model.layers.31.mlp.gate_proj": {
1004
+ "group_size": 64,
1005
+ "bits": 4
1006
+ },
1007
+ "model.layers.31.mlp.down_proj": {
1008
+ "group_size": 64,
1009
+ "bits": 4
1010
+ },
1011
+ "model.layers.31.mlp.up_proj": {
1012
+ "group_size": 64,
1013
+ "bits": 4
1014
+ },
1015
+ "model.layers.31.input_layernorm": false,
1016
+ "model.layers.31.post_attention_layernorm": false,
1017
+ "model.layers.32.self_attn.q_proj": {
1018
+ "group_size": 64,
1019
+ "bits": 4
1020
+ },
1021
+ "model.layers.32.self_attn.k_proj": {
1022
+ "group_size": 64,
1023
+ "bits": 4
1024
+ },
1025
+ "model.layers.32.self_attn.v_proj": {
1026
+ "group_size": 64,
1027
+ "bits": 6
1028
+ },
1029
+ "model.layers.32.self_attn.o_proj": {
1030
+ "group_size": 64,
1031
+ "bits": 4
1032
+ },
1033
+ "model.layers.32.self_attn.rope": false,
1034
+ "model.layers.32.mlp.gate_proj": {
1035
+ "group_size": 64,
1036
+ "bits": 4
1037
+ },
1038
+ "model.layers.32.mlp.down_proj": {
1039
+ "group_size": 64,
1040
+ "bits": 6
1041
+ },
1042
+ "model.layers.32.mlp.up_proj": {
1043
+ "group_size": 64,
1044
+ "bits": 4
1045
+ },
1046
+ "model.layers.32.input_layernorm": false,
1047
+ "model.layers.32.post_attention_layernorm": false,
1048
+ "model.layers.33.self_attn.q_proj": {
1049
+ "group_size": 64,
1050
+ "bits": 4
1051
+ },
1052
+ "model.layers.33.self_attn.k_proj": {
1053
+ "group_size": 64,
1054
+ "bits": 4
1055
+ },
1056
+ "model.layers.33.self_attn.v_proj": {
1057
+ "group_size": 64,
1058
+ "bits": 4
1059
+ },
1060
+ "model.layers.33.self_attn.o_proj": {
1061
+ "group_size": 64,
1062
+ "bits": 4
1063
+ },
1064
+ "model.layers.33.self_attn.rope": false,
1065
+ "model.layers.33.mlp.gate_proj": {
1066
+ "group_size": 64,
1067
+ "bits": 4
1068
+ },
1069
+ "model.layers.33.mlp.down_proj": {
1070
+ "group_size": 64,
1071
+ "bits": 4
1072
+ },
1073
+ "model.layers.33.mlp.up_proj": {
1074
+ "group_size": 64,
1075
+ "bits": 4
1076
+ },
1077
+ "model.layers.33.input_layernorm": false,
1078
+ "model.layers.33.post_attention_layernorm": false,
1079
+ "model.layers.34.self_attn.q_proj": {
1080
+ "group_size": 64,
1081
+ "bits": 4
1082
+ },
1083
+ "model.layers.34.self_attn.k_proj": {
1084
+ "group_size": 64,
1085
+ "bits": 4
1086
+ },
1087
+ "model.layers.34.self_attn.v_proj": {
1088
+ "group_size": 64,
1089
+ "bits": 4
1090
+ },
1091
+ "model.layers.34.self_attn.o_proj": {
1092
+ "group_size": 64,
1093
+ "bits": 4
1094
+ },
1095
+ "model.layers.34.self_attn.rope": false,
1096
+ "model.layers.34.mlp.gate_proj": {
1097
+ "group_size": 64,
1098
+ "bits": 4
1099
+ },
1100
+ "model.layers.34.mlp.down_proj": {
1101
+ "group_size": 64,
1102
+ "bits": 4
1103
+ },
1104
+ "model.layers.34.mlp.up_proj": {
1105
+ "group_size": 64,
1106
+ "bits": 4
1107
+ },
1108
+ "model.layers.34.input_layernorm": false,
1109
+ "model.layers.34.post_attention_layernorm": false,
1110
+ "model.layers.35.self_attn.q_proj": {
1111
+ "group_size": 64,
1112
+ "bits": 4
1113
+ },
1114
+ "model.layers.35.self_attn.k_proj": {
1115
+ "group_size": 64,
1116
+ "bits": 4
1117
+ },
1118
+ "model.layers.35.self_attn.v_proj": {
1119
+ "group_size": 64,
1120
+ "bits": 6
1121
+ },
1122
+ "model.layers.35.self_attn.o_proj": {
1123
+ "group_size": 64,
1124
+ "bits": 4
1125
+ },
1126
+ "model.layers.35.self_attn.rope": false,
1127
+ "model.layers.35.mlp.gate_proj": {
1128
+ "group_size": 64,
1129
+ "bits": 4
1130
+ },
1131
+ "model.layers.35.mlp.down_proj": {
1132
+ "group_size": 64,
1133
+ "bits": 6
1134
+ },
1135
+ "model.layers.35.mlp.up_proj": {
1136
+ "group_size": 64,
1137
+ "bits": 4
1138
+ },
1139
+ "model.layers.35.input_layernorm": false,
1140
+ "model.layers.35.post_attention_layernorm": false,
1141
+ "model.layers.36.self_attn.q_proj": {
1142
+ "group_size": 64,
1143
+ "bits": 4
1144
+ },
1145
+ "model.layers.36.self_attn.k_proj": {
1146
+ "group_size": 64,
1147
+ "bits": 4
1148
+ },
1149
+ "model.layers.36.self_attn.v_proj": {
1150
+ "group_size": 64,
1151
+ "bits": 4
1152
+ },
1153
+ "model.layers.36.self_attn.o_proj": {
1154
+ "group_size": 64,
1155
+ "bits": 4
1156
+ },
1157
+ "model.layers.36.self_attn.rope": false,
1158
+ "model.layers.36.mlp.gate_proj": {
1159
+ "group_size": 64,
1160
+ "bits": 4
1161
+ },
1162
+ "model.layers.36.mlp.down_proj": {
1163
+ "group_size": 64,
1164
+ "bits": 4
1165
+ },
1166
+ "model.layers.36.mlp.up_proj": {
1167
+ "group_size": 64,
1168
+ "bits": 4
1169
+ },
1170
+ "model.layers.36.input_layernorm": false,
1171
+ "model.layers.36.post_attention_layernorm": false,
1172
+ "model.layers.37.self_attn.q_proj": {
1173
+ "group_size": 64,
1174
+ "bits": 4
1175
+ },
1176
+ "model.layers.37.self_attn.k_proj": {
1177
+ "group_size": 64,
1178
+ "bits": 4
1179
+ },
1180
+ "model.layers.37.self_attn.v_proj": {
1181
+ "group_size": 64,
1182
+ "bits": 4
1183
+ },
1184
+ "model.layers.37.self_attn.o_proj": {
1185
+ "group_size": 64,
1186
+ "bits": 4
1187
+ },
1188
+ "model.layers.37.self_attn.rope": false,
1189
+ "model.layers.37.mlp.gate_proj": {
1190
+ "group_size": 64,
1191
+ "bits": 4
1192
+ },
1193
+ "model.layers.37.mlp.down_proj": {
1194
+ "group_size": 64,
1195
+ "bits": 4
1196
+ },
1197
+ "model.layers.37.mlp.up_proj": {
1198
+ "group_size": 64,
1199
+ "bits": 4
1200
+ },
1201
+ "model.layers.37.input_layernorm": false,
1202
+ "model.layers.37.post_attention_layernorm": false,
1203
+ "model.layers.38.self_attn.q_proj": {
1204
+ "group_size": 64,
1205
+ "bits": 4
1206
+ },
1207
+ "model.layers.38.self_attn.k_proj": {
1208
+ "group_size": 64,
1209
+ "bits": 4
1210
+ },
1211
+ "model.layers.38.self_attn.v_proj": {
1212
+ "group_size": 64,
1213
+ "bits": 6
1214
+ },
1215
+ "model.layers.38.self_attn.o_proj": {
1216
+ "group_size": 64,
1217
+ "bits": 4
1218
+ },
1219
+ "model.layers.38.self_attn.rope": false,
1220
+ "model.layers.38.mlp.gate_proj": {
1221
+ "group_size": 64,
1222
+ "bits": 4
1223
+ },
1224
+ "model.layers.38.mlp.down_proj": {
1225
+ "group_size": 64,
1226
+ "bits": 6
1227
+ },
1228
+ "model.layers.38.mlp.up_proj": {
1229
+ "group_size": 64,
1230
+ "bits": 4
1231
+ },
1232
+ "model.layers.38.input_layernorm": false,
1233
+ "model.layers.38.post_attention_layernorm": false,
1234
+ "model.layers.39.self_attn.q_proj": {
1235
+ "group_size": 64,
1236
+ "bits": 4
1237
+ },
1238
+ "model.layers.39.self_attn.k_proj": {
1239
+ "group_size": 64,
1240
+ "bits": 4
1241
+ },
1242
+ "model.layers.39.self_attn.v_proj": {
1243
+ "group_size": 64,
1244
+ "bits": 4
1245
+ },
1246
+ "model.layers.39.self_attn.o_proj": {
1247
+ "group_size": 64,
1248
+ "bits": 4
1249
+ },
1250
+ "model.layers.39.self_attn.rope": false,
1251
+ "model.layers.39.mlp.gate_proj": {
1252
+ "group_size": 64,
1253
+ "bits": 4
1254
+ },
1255
+ "model.layers.39.mlp.down_proj": {
1256
+ "group_size": 64,
1257
+ "bits": 4
1258
+ },
1259
+ "model.layers.39.mlp.up_proj": {
1260
+ "group_size": 64,
1261
+ "bits": 4
1262
+ },
1263
+ "model.layers.39.input_layernorm": false,
1264
+ "model.layers.39.post_attention_layernorm": false,
1265
+ "model.layers.40.self_attn.q_proj": {
1266
+ "group_size": 64,
1267
+ "bits": 4
1268
+ },
1269
+ "model.layers.40.self_attn.k_proj": {
1270
+ "group_size": 64,
1271
+ "bits": 4
1272
+ },
1273
+ "model.layers.40.self_attn.v_proj": {
1274
+ "group_size": 64,
1275
+ "bits": 4
1276
+ },
1277
+ "model.layers.40.self_attn.o_proj": {
1278
+ "group_size": 64,
1279
+ "bits": 4
1280
+ },
1281
+ "model.layers.40.self_attn.rope": false,
1282
+ "model.layers.40.mlp.gate_proj": {
1283
+ "group_size": 64,
1284
+ "bits": 4
1285
+ },
1286
+ "model.layers.40.mlp.down_proj": {
1287
+ "group_size": 64,
1288
+ "bits": 4
1289
+ },
1290
+ "model.layers.40.mlp.up_proj": {
1291
+ "group_size": 64,
1292
+ "bits": 4
1293
+ },
1294
+ "model.layers.40.input_layernorm": false,
1295
+ "model.layers.40.post_attention_layernorm": false,
1296
+ "model.layers.41.self_attn.q_proj": {
1297
+ "group_size": 64,
1298
+ "bits": 4
1299
+ },
1300
+ "model.layers.41.self_attn.k_proj": {
1301
+ "group_size": 64,
1302
+ "bits": 4
1303
+ },
1304
+ "model.layers.41.self_attn.v_proj": {
1305
+ "group_size": 64,
1306
+ "bits": 6
1307
+ },
1308
+ "model.layers.41.self_attn.o_proj": {
1309
+ "group_size": 64,
1310
+ "bits": 4
1311
+ },
1312
+ "model.layers.41.self_attn.rope": false,
1313
+ "model.layers.41.mlp.gate_proj": {
1314
+ "group_size": 64,
1315
+ "bits": 4
1316
+ },
1317
+ "model.layers.41.mlp.down_proj": {
1318
+ "group_size": 64,
1319
+ "bits": 6
1320
+ },
1321
+ "model.layers.41.mlp.up_proj": {
1322
+ "group_size": 64,
1323
+ "bits": 4
1324
+ },
1325
+ "model.layers.41.input_layernorm": false,
1326
+ "model.layers.41.post_attention_layernorm": false,
1327
+ "model.layers.42.self_attn.q_proj": {
1328
+ "group_size": 64,
1329
+ "bits": 4
1330
+ },
1331
+ "model.layers.42.self_attn.k_proj": {
1332
+ "group_size": 64,
1333
+ "bits": 4
1334
+ },
1335
+ "model.layers.42.self_attn.v_proj": {
1336
+ "group_size": 64,
1337
+ "bits": 6
1338
+ },
1339
+ "model.layers.42.self_attn.o_proj": {
1340
+ "group_size": 64,
1341
+ "bits": 4
1342
+ },
1343
+ "model.layers.42.self_attn.rope": false,
1344
+ "model.layers.42.mlp.gate_proj": {
1345
+ "group_size": 64,
1346
+ "bits": 4
1347
+ },
1348
+ "model.layers.42.mlp.down_proj": {
1349
+ "group_size": 64,
1350
+ "bits": 6
1351
+ },
1352
+ "model.layers.42.mlp.up_proj": {
1353
+ "group_size": 64,
1354
+ "bits": 4
1355
+ },
1356
+ "model.layers.42.input_layernorm": false,
1357
+ "model.layers.42.post_attention_layernorm": false,
1358
+ "model.layers.43.self_attn.q_proj": {
1359
+ "group_size": 64,
1360
+ "bits": 4
1361
+ },
1362
+ "model.layers.43.self_attn.k_proj": {
1363
+ "group_size": 64,
1364
+ "bits": 4
1365
+ },
1366
+ "model.layers.43.self_attn.v_proj": {
1367
+ "group_size": 64,
1368
+ "bits": 6
1369
+ },
1370
+ "model.layers.43.self_attn.o_proj": {
1371
+ "group_size": 64,
1372
+ "bits": 4
1373
+ },
1374
+ "model.layers.43.self_attn.rope": false,
1375
+ "model.layers.43.mlp.gate_proj": {
1376
+ "group_size": 64,
1377
+ "bits": 4
1378
+ },
1379
+ "model.layers.43.mlp.down_proj": {
1380
+ "group_size": 64,
1381
+ "bits": 6
1382
+ },
1383
+ "model.layers.43.mlp.up_proj": {
1384
+ "group_size": 64,
1385
+ "bits": 4
1386
+ },
1387
+ "model.layers.43.input_layernorm": false,
1388
+ "model.layers.43.post_attention_layernorm": false,
1389
+ "model.layers.44.self_attn.q_proj": {
1390
+ "group_size": 64,
1391
+ "bits": 4
1392
+ },
1393
+ "model.layers.44.self_attn.k_proj": {
1394
+ "group_size": 64,
1395
+ "bits": 4
1396
+ },
1397
+ "model.layers.44.self_attn.v_proj": {
1398
+ "group_size": 64,
1399
+ "bits": 6
1400
+ },
1401
+ "model.layers.44.self_attn.o_proj": {
1402
+ "group_size": 64,
1403
+ "bits": 4
1404
+ },
1405
+ "model.layers.44.self_attn.rope": false,
1406
+ "model.layers.44.mlp.gate_proj": {
1407
+ "group_size": 64,
1408
+ "bits": 4
1409
+ },
1410
+ "model.layers.44.mlp.down_proj": {
1411
+ "group_size": 64,
1412
+ "bits": 6
1413
+ },
1414
+ "model.layers.44.mlp.up_proj": {
1415
+ "group_size": 64,
1416
+ "bits": 4
1417
+ },
1418
+ "model.layers.44.input_layernorm": false,
1419
+ "model.layers.44.post_attention_layernorm": false,
1420
+ "model.layers.45.self_attn.q_proj": {
1421
+ "group_size": 64,
1422
+ "bits": 4
1423
+ },
1424
+ "model.layers.45.self_attn.k_proj": {
1425
+ "group_size": 64,
1426
+ "bits": 4
1427
+ },
1428
+ "model.layers.45.self_attn.v_proj": {
1429
+ "group_size": 64,
1430
+ "bits": 6
1431
+ },
1432
+ "model.layers.45.self_attn.o_proj": {
1433
+ "group_size": 64,
1434
+ "bits": 4
1435
+ },
1436
+ "model.layers.45.self_attn.rope": false,
1437
+ "model.layers.45.mlp.gate_proj": {
1438
+ "group_size": 64,
1439
+ "bits": 4
1440
+ },
1441
+ "model.layers.45.mlp.down_proj": {
1442
+ "group_size": 64,
1443
+ "bits": 6
1444
+ },
1445
+ "model.layers.45.mlp.up_proj": {
1446
+ "group_size": 64,
1447
+ "bits": 4
1448
+ },
1449
+ "model.layers.45.input_layernorm": false,
1450
+ "model.layers.45.post_attention_layernorm": false,
1451
+ "model.layers.46.self_attn.q_proj": {
1452
+ "group_size": 64,
1453
+ "bits": 4
1454
+ },
1455
+ "model.layers.46.self_attn.k_proj": {
1456
+ "group_size": 64,
1457
+ "bits": 4
1458
+ },
1459
+ "model.layers.46.self_attn.v_proj": {
1460
+ "group_size": 64,
1461
+ "bits": 6
1462
+ },
1463
+ "model.layers.46.self_attn.o_proj": {
1464
+ "group_size": 64,
1465
+ "bits": 4
1466
+ },
1467
+ "model.layers.46.self_attn.rope": false,
1468
+ "model.layers.46.mlp.gate_proj": {
1469
+ "group_size": 64,
1470
+ "bits": 4
1471
+ },
1472
+ "model.layers.46.mlp.down_proj": {
1473
+ "group_size": 64,
1474
+ "bits": 6
1475
+ },
1476
+ "model.layers.46.mlp.up_proj": {
1477
+ "group_size": 64,
1478
+ "bits": 4
1479
+ },
1480
+ "model.layers.46.input_layernorm": false,
1481
+ "model.layers.46.post_attention_layernorm": false,
1482
+ "model.layers.47.self_attn.q_proj": {
1483
+ "group_size": 64,
1484
+ "bits": 4
1485
+ },
1486
+ "model.layers.47.self_attn.k_proj": {
1487
+ "group_size": 64,
1488
+ "bits": 4
1489
+ },
1490
+ "model.layers.47.self_attn.v_proj": {
1491
+ "group_size": 64,
1492
+ "bits": 6
1493
+ },
1494
+ "model.layers.47.self_attn.o_proj": {
1495
+ "group_size": 64,
1496
+ "bits": 4
1497
+ },
1498
+ "model.layers.47.self_attn.rope": false,
1499
+ "model.layers.47.mlp.gate_proj": {
1500
+ "group_size": 64,
1501
+ "bits": 4
1502
+ },
1503
+ "model.layers.47.mlp.down_proj": {
1504
+ "group_size": 64,
1505
+ "bits": 6
1506
+ },
1507
+ "model.layers.47.mlp.up_proj": {
1508
+ "group_size": 64,
1509
+ "bits": 4
1510
+ },
1511
+ "model.layers.47.input_layernorm": false,
1512
+ "model.layers.47.post_attention_layernorm": false,
1513
+ "model.norm": false,
1514
+ "lm_head": {
1515
+ "group_size": 64,
1516
+ "bits": 6
1517
+ }
1518
+ },
1519
+ "quantization_config": {
1520
+ "group_size": 64,
1521
+ "bits": null,
1522
+ "model.embed_tokens": {
1523
+ "group_size": 64,
1524
+ "bits": 4
1525
+ },
1526
+ "model.layers.0.self_attn.q_proj": {
1527
+ "group_size": 64,
1528
+ "bits": 4
1529
+ },
1530
+ "model.layers.0.self_attn.k_proj": {
1531
+ "group_size": 64,
1532
+ "bits": 4
1533
+ },
1534
+ "model.layers.0.self_attn.v_proj": {
1535
+ "group_size": 64,
1536
+ "bits": 6
1537
+ },
1538
+ "model.layers.0.self_attn.o_proj": {
1539
+ "group_size": 64,
1540
+ "bits": 4
1541
+ },
1542
+ "model.layers.0.self_attn.rope": false,
1543
+ "model.layers.0.mlp.gate_proj": {
1544
+ "group_size": 64,
1545
+ "bits": 4
1546
+ },
1547
+ "model.layers.0.mlp.down_proj": {
1548
+ "group_size": 64,
1549
+ "bits": 6
1550
+ },
1551
+ "model.layers.0.mlp.up_proj": {
1552
+ "group_size": 64,
1553
+ "bits": 4
1554
+ },
1555
+ "model.layers.0.input_layernorm": false,
1556
+ "model.layers.0.post_attention_layernorm": false,
1557
+ "model.layers.1.self_attn.q_proj": {
1558
+ "group_size": 64,
1559
+ "bits": 4
1560
+ },
1561
+ "model.layers.1.self_attn.k_proj": {
1562
+ "group_size": 64,
1563
+ "bits": 4
1564
+ },
1565
+ "model.layers.1.self_attn.v_proj": {
1566
+ "group_size": 64,
1567
+ "bits": 6
1568
+ },
1569
+ "model.layers.1.self_attn.o_proj": {
1570
+ "group_size": 64,
1571
+ "bits": 4
1572
+ },
1573
+ "model.layers.1.self_attn.rope": false,
1574
+ "model.layers.1.mlp.gate_proj": {
1575
+ "group_size": 64,
1576
+ "bits": 4
1577
+ },
1578
+ "model.layers.1.mlp.down_proj": {
1579
+ "group_size": 64,
1580
+ "bits": 6
1581
+ },
1582
+ "model.layers.1.mlp.up_proj": {
1583
+ "group_size": 64,
1584
+ "bits": 4
1585
+ },
1586
+ "model.layers.1.input_layernorm": false,
1587
+ "model.layers.1.post_attention_layernorm": false,
1588
+ "model.layers.2.self_attn.q_proj": {
1589
+ "group_size": 64,
1590
+ "bits": 4
1591
+ },
1592
+ "model.layers.2.self_attn.k_proj": {
1593
+ "group_size": 64,
1594
+ "bits": 4
1595
+ },
1596
+ "model.layers.2.self_attn.v_proj": {
1597
+ "group_size": 64,
1598
+ "bits": 6
1599
+ },
1600
+ "model.layers.2.self_attn.o_proj": {
1601
+ "group_size": 64,
1602
+ "bits": 4
1603
+ },
1604
+ "model.layers.2.self_attn.rope": false,
1605
+ "model.layers.2.mlp.gate_proj": {
1606
+ "group_size": 64,
1607
+ "bits": 4
1608
+ },
1609
+ "model.layers.2.mlp.down_proj": {
1610
+ "group_size": 64,
1611
+ "bits": 6
1612
+ },
1613
+ "model.layers.2.mlp.up_proj": {
1614
+ "group_size": 64,
1615
+ "bits": 4
1616
+ },
1617
+ "model.layers.2.input_layernorm": false,
1618
+ "model.layers.2.post_attention_layernorm": false,
1619
+ "model.layers.3.self_attn.q_proj": {
1620
+ "group_size": 64,
1621
+ "bits": 4
1622
+ },
1623
+ "model.layers.3.self_attn.k_proj": {
1624
+ "group_size": 64,
1625
+ "bits": 4
1626
+ },
1627
+ "model.layers.3.self_attn.v_proj": {
1628
+ "group_size": 64,
1629
+ "bits": 6
1630
+ },
1631
+ "model.layers.3.self_attn.o_proj": {
1632
+ "group_size": 64,
1633
+ "bits": 4
1634
+ },
1635
+ "model.layers.3.self_attn.rope": false,
1636
+ "model.layers.3.mlp.gate_proj": {
1637
+ "group_size": 64,
1638
+ "bits": 4
1639
+ },
1640
+ "model.layers.3.mlp.down_proj": {
1641
+ "group_size": 64,
1642
+ "bits": 6
1643
+ },
1644
+ "model.layers.3.mlp.up_proj": {
1645
+ "group_size": 64,
1646
+ "bits": 4
1647
+ },
1648
+ "model.layers.3.input_layernorm": false,
1649
+ "model.layers.3.post_attention_layernorm": false,
1650
+ "model.layers.4.self_attn.q_proj": {
1651
+ "group_size": 64,
1652
+ "bits": 4
1653
+ },
1654
+ "model.layers.4.self_attn.k_proj": {
1655
+ "group_size": 64,
1656
+ "bits": 4
1657
+ },
1658
+ "model.layers.4.self_attn.v_proj": {
1659
+ "group_size": 64,
1660
+ "bits": 6
1661
+ },
1662
+ "model.layers.4.self_attn.o_proj": {
1663
+ "group_size": 64,
1664
+ "bits": 4
1665
+ },
1666
+ "model.layers.4.self_attn.rope": false,
1667
+ "model.layers.4.mlp.gate_proj": {
1668
+ "group_size": 64,
1669
+ "bits": 4
1670
+ },
1671
+ "model.layers.4.mlp.down_proj": {
1672
+ "group_size": 64,
1673
+ "bits": 6
1674
+ },
1675
+ "model.layers.4.mlp.up_proj": {
1676
+ "group_size": 64,
1677
+ "bits": 4
1678
+ },
1679
+ "model.layers.4.input_layernorm": false,
1680
+ "model.layers.4.post_attention_layernorm": false,
1681
+ "model.layers.5.self_attn.q_proj": {
1682
+ "group_size": 64,
1683
+ "bits": 4
1684
+ },
1685
+ "model.layers.5.self_attn.k_proj": {
1686
+ "group_size": 64,
1687
+ "bits": 4
1688
+ },
1689
+ "model.layers.5.self_attn.v_proj": {
1690
+ "group_size": 64,
1691
+ "bits": 6
1692
+ },
1693
+ "model.layers.5.self_attn.o_proj": {
1694
+ "group_size": 64,
1695
+ "bits": 4
1696
+ },
1697
+ "model.layers.5.self_attn.rope": false,
1698
+ "model.layers.5.mlp.gate_proj": {
1699
+ "group_size": 64,
1700
+ "bits": 4
1701
+ },
1702
+ "model.layers.5.mlp.down_proj": {
1703
+ "group_size": 64,
1704
+ "bits": 6
1705
+ },
1706
+ "model.layers.5.mlp.up_proj": {
1707
+ "group_size": 64,
1708
+ "bits": 4
1709
+ },
1710
+ "model.layers.5.input_layernorm": false,
1711
+ "model.layers.5.post_attention_layernorm": false,
1712
+ "model.layers.6.self_attn.q_proj": {
1713
+ "group_size": 64,
1714
+ "bits": 4
1715
+ },
1716
+ "model.layers.6.self_attn.k_proj": {
1717
+ "group_size": 64,
1718
+ "bits": 4
1719
+ },
1720
+ "model.layers.6.self_attn.v_proj": {
1721
+ "group_size": 64,
1722
+ "bits": 4
1723
+ },
1724
+ "model.layers.6.self_attn.o_proj": {
1725
+ "group_size": 64,
1726
+ "bits": 4
1727
+ },
1728
+ "model.layers.6.self_attn.rope": false,
1729
+ "model.layers.6.mlp.gate_proj": {
1730
+ "group_size": 64,
1731
+ "bits": 4
1732
+ },
1733
+ "model.layers.6.mlp.down_proj": {
1734
+ "group_size": 64,
1735
+ "bits": 4
1736
+ },
1737
+ "model.layers.6.mlp.up_proj": {
1738
+ "group_size": 64,
1739
+ "bits": 4
1740
+ },
1741
+ "model.layers.6.input_layernorm": false,
1742
+ "model.layers.6.post_attention_layernorm": false,
1743
+ "model.layers.7.self_attn.q_proj": {
1744
+ "group_size": 64,
1745
+ "bits": 4
1746
+ },
1747
+ "model.layers.7.self_attn.k_proj": {
1748
+ "group_size": 64,
1749
+ "bits": 4
1750
+ },
1751
+ "model.layers.7.self_attn.v_proj": {
1752
+ "group_size": 64,
1753
+ "bits": 4
1754
+ },
1755
+ "model.layers.7.self_attn.o_proj": {
1756
+ "group_size": 64,
1757
+ "bits": 4
1758
+ },
1759
+ "model.layers.7.self_attn.rope": false,
1760
+ "model.layers.7.mlp.gate_proj": {
1761
+ "group_size": 64,
1762
+ "bits": 4
1763
+ },
1764
+ "model.layers.7.mlp.down_proj": {
1765
+ "group_size": 64,
1766
+ "bits": 4
1767
+ },
1768
+ "model.layers.7.mlp.up_proj": {
1769
+ "group_size": 64,
1770
+ "bits": 4
1771
+ },
1772
+ "model.layers.7.input_layernorm": false,
1773
+ "model.layers.7.post_attention_layernorm": false,
1774
+ "model.layers.8.self_attn.q_proj": {
1775
+ "group_size": 64,
1776
+ "bits": 4
1777
+ },
1778
+ "model.layers.8.self_attn.k_proj": {
1779
+ "group_size": 64,
1780
+ "bits": 4
1781
+ },
1782
+ "model.layers.8.self_attn.v_proj": {
1783
+ "group_size": 64,
1784
+ "bits": 6
1785
+ },
1786
+ "model.layers.8.self_attn.o_proj": {
1787
+ "group_size": 64,
1788
+ "bits": 4
1789
+ },
1790
+ "model.layers.8.self_attn.rope": false,
1791
+ "model.layers.8.mlp.gate_proj": {
1792
+ "group_size": 64,
1793
+ "bits": 4
1794
+ },
1795
+ "model.layers.8.mlp.down_proj": {
1796
+ "group_size": 64,
1797
+ "bits": 6
1798
+ },
1799
+ "model.layers.8.mlp.up_proj": {
1800
+ "group_size": 64,
1801
+ "bits": 4
1802
+ },
1803
+ "model.layers.8.input_layernorm": false,
1804
+ "model.layers.8.post_attention_layernorm": false,
1805
+ "model.layers.9.self_attn.q_proj": {
1806
+ "group_size": 64,
1807
+ "bits": 4
1808
+ },
1809
+ "model.layers.9.self_attn.k_proj": {
1810
+ "group_size": 64,
1811
+ "bits": 4
1812
+ },
1813
+ "model.layers.9.self_attn.v_proj": {
1814
+ "group_size": 64,
1815
+ "bits": 4
1816
+ },
1817
+ "model.layers.9.self_attn.o_proj": {
1818
+ "group_size": 64,
1819
+ "bits": 4
1820
+ },
1821
+ "model.layers.9.self_attn.rope": false,
1822
+ "model.layers.9.mlp.gate_proj": {
1823
+ "group_size": 64,
1824
+ "bits": 4
1825
+ },
1826
+ "model.layers.9.mlp.down_proj": {
1827
+ "group_size": 64,
1828
+ "bits": 4
1829
+ },
1830
+ "model.layers.9.mlp.up_proj": {
1831
+ "group_size": 64,
1832
+ "bits": 4
1833
+ },
1834
+ "model.layers.9.input_layernorm": false,
1835
+ "model.layers.9.post_attention_layernorm": false,
1836
+ "model.layers.10.self_attn.q_proj": {
1837
+ "group_size": 64,
1838
+ "bits": 4
1839
+ },
1840
+ "model.layers.10.self_attn.k_proj": {
1841
+ "group_size": 64,
1842
+ "bits": 4
1843
+ },
1844
+ "model.layers.10.self_attn.v_proj": {
1845
+ "group_size": 64,
1846
+ "bits": 4
1847
+ },
1848
+ "model.layers.10.self_attn.o_proj": {
1849
+ "group_size": 64,
1850
+ "bits": 4
1851
+ },
1852
+ "model.layers.10.self_attn.rope": false,
1853
+ "model.layers.10.mlp.gate_proj": {
1854
+ "group_size": 64,
1855
+ "bits": 4
1856
+ },
1857
+ "model.layers.10.mlp.down_proj": {
1858
+ "group_size": 64,
1859
+ "bits": 4
1860
+ },
1861
+ "model.layers.10.mlp.up_proj": {
1862
+ "group_size": 64,
1863
+ "bits": 4
1864
+ },
1865
+ "model.layers.10.input_layernorm": false,
1866
+ "model.layers.10.post_attention_layernorm": false,
1867
+ "model.layers.11.self_attn.q_proj": {
1868
+ "group_size": 64,
1869
+ "bits": 4
1870
+ },
1871
+ "model.layers.11.self_attn.k_proj": {
1872
+ "group_size": 64,
1873
+ "bits": 4
1874
+ },
1875
+ "model.layers.11.self_attn.v_proj": {
1876
+ "group_size": 64,
1877
+ "bits": 6
1878
+ },
1879
+ "model.layers.11.self_attn.o_proj": {
1880
+ "group_size": 64,
1881
+ "bits": 4
1882
+ },
1883
+ "model.layers.11.self_attn.rope": false,
1884
+ "model.layers.11.mlp.gate_proj": {
1885
+ "group_size": 64,
1886
+ "bits": 4
1887
+ },
1888
+ "model.layers.11.mlp.down_proj": {
1889
+ "group_size": 64,
1890
+ "bits": 6
1891
+ },
1892
+ "model.layers.11.mlp.up_proj": {
1893
+ "group_size": 64,
1894
+ "bits": 4
1895
+ },
1896
+ "model.layers.11.input_layernorm": false,
1897
+ "model.layers.11.post_attention_layernorm": false,
1898
+ "model.layers.12.self_attn.q_proj": {
1899
+ "group_size": 64,
1900
+ "bits": 4
1901
+ },
1902
+ "model.layers.12.self_attn.k_proj": {
1903
+ "group_size": 64,
1904
+ "bits": 4
1905
+ },
1906
+ "model.layers.12.self_attn.v_proj": {
1907
+ "group_size": 64,
1908
+ "bits": 4
1909
+ },
1910
+ "model.layers.12.self_attn.o_proj": {
1911
+ "group_size": 64,
1912
+ "bits": 4
1913
+ },
1914
+ "model.layers.12.self_attn.rope": false,
1915
+ "model.layers.12.mlp.gate_proj": {
1916
+ "group_size": 64,
1917
+ "bits": 4
1918
+ },
1919
+ "model.layers.12.mlp.down_proj": {
1920
+ "group_size": 64,
1921
+ "bits": 4
1922
+ },
1923
+ "model.layers.12.mlp.up_proj": {
1924
+ "group_size": 64,
1925
+ "bits": 4
1926
+ },
1927
+ "model.layers.12.input_layernorm": false,
1928
+ "model.layers.12.post_attention_layernorm": false,
1929
+ "model.layers.13.self_attn.q_proj": {
1930
+ "group_size": 64,
1931
+ "bits": 4
1932
+ },
1933
+ "model.layers.13.self_attn.k_proj": {
1934
+ "group_size": 64,
1935
+ "bits": 4
1936
+ },
1937
+ "model.layers.13.self_attn.v_proj": {
1938
+ "group_size": 64,
1939
+ "bits": 4
1940
+ },
1941
+ "model.layers.13.self_attn.o_proj": {
1942
+ "group_size": 64,
1943
+ "bits": 4
1944
+ },
1945
+ "model.layers.13.self_attn.rope": false,
1946
+ "model.layers.13.mlp.gate_proj": {
1947
+ "group_size": 64,
1948
+ "bits": 4
1949
+ },
1950
+ "model.layers.13.mlp.down_proj": {
1951
+ "group_size": 64,
1952
+ "bits": 4
1953
+ },
1954
+ "model.layers.13.mlp.up_proj": {
1955
+ "group_size": 64,
1956
+ "bits": 4
1957
+ },
1958
+ "model.layers.13.input_layernorm": false,
1959
+ "model.layers.13.post_attention_layernorm": false,
1960
+ "model.layers.14.self_attn.q_proj": {
1961
+ "group_size": 64,
1962
+ "bits": 4
1963
+ },
1964
+ "model.layers.14.self_attn.k_proj": {
1965
+ "group_size": 64,
1966
+ "bits": 4
1967
+ },
1968
+ "model.layers.14.self_attn.v_proj": {
1969
+ "group_size": 64,
1970
+ "bits": 6
1971
+ },
1972
+ "model.layers.14.self_attn.o_proj": {
1973
+ "group_size": 64,
1974
+ "bits": 4
1975
+ },
1976
+ "model.layers.14.self_attn.rope": false,
1977
+ "model.layers.14.mlp.gate_proj": {
1978
+ "group_size": 64,
1979
+ "bits": 4
1980
+ },
1981
+ "model.layers.14.mlp.down_proj": {
1982
+ "group_size": 64,
1983
+ "bits": 6
1984
+ },
1985
+ "model.layers.14.mlp.up_proj": {
1986
+ "group_size": 64,
1987
+ "bits": 4
1988
+ },
1989
+ "model.layers.14.input_layernorm": false,
1990
+ "model.layers.14.post_attention_layernorm": false,
1991
+ "model.layers.15.self_attn.q_proj": {
1992
+ "group_size": 64,
1993
+ "bits": 4
1994
+ },
1995
+ "model.layers.15.self_attn.k_proj": {
1996
+ "group_size": 64,
1997
+ "bits": 4
1998
+ },
1999
+ "model.layers.15.self_attn.v_proj": {
2000
+ "group_size": 64,
2001
+ "bits": 4
2002
+ },
2003
+ "model.layers.15.self_attn.o_proj": {
2004
+ "group_size": 64,
2005
+ "bits": 4
2006
+ },
2007
+ "model.layers.15.self_attn.rope": false,
2008
+ "model.layers.15.mlp.gate_proj": {
2009
+ "group_size": 64,
2010
+ "bits": 4
2011
+ },
2012
+ "model.layers.15.mlp.down_proj": {
2013
+ "group_size": 64,
2014
+ "bits": 4
2015
+ },
2016
+ "model.layers.15.mlp.up_proj": {
2017
+ "group_size": 64,
2018
+ "bits": 4
2019
+ },
2020
+ "model.layers.15.input_layernorm": false,
2021
+ "model.layers.15.post_attention_layernorm": false,
2022
+ "model.layers.16.self_attn.q_proj": {
2023
+ "group_size": 64,
2024
+ "bits": 4
2025
+ },
2026
+ "model.layers.16.self_attn.k_proj": {
2027
+ "group_size": 64,
2028
+ "bits": 4
2029
+ },
2030
+ "model.layers.16.self_attn.v_proj": {
2031
+ "group_size": 64,
2032
+ "bits": 4
2033
+ },
2034
+ "model.layers.16.self_attn.o_proj": {
2035
+ "group_size": 64,
2036
+ "bits": 4
2037
+ },
2038
+ "model.layers.16.self_attn.rope": false,
2039
+ "model.layers.16.mlp.gate_proj": {
2040
+ "group_size": 64,
2041
+ "bits": 4
2042
+ },
2043
+ "model.layers.16.mlp.down_proj": {
2044
+ "group_size": 64,
2045
+ "bits": 4
2046
+ },
2047
+ "model.layers.16.mlp.up_proj": {
2048
+ "group_size": 64,
2049
+ "bits": 4
2050
+ },
2051
+ "model.layers.16.input_layernorm": false,
2052
+ "model.layers.16.post_attention_layernorm": false,
2053
+ "model.layers.17.self_attn.q_proj": {
2054
+ "group_size": 64,
2055
+ "bits": 4
2056
+ },
2057
+ "model.layers.17.self_attn.k_proj": {
2058
+ "group_size": 64,
2059
+ "bits": 4
2060
+ },
2061
+ "model.layers.17.self_attn.v_proj": {
2062
+ "group_size": 64,
2063
+ "bits": 6
2064
+ },
2065
+ "model.layers.17.self_attn.o_proj": {
2066
+ "group_size": 64,
2067
+ "bits": 4
2068
+ },
2069
+ "model.layers.17.self_attn.rope": false,
2070
+ "model.layers.17.mlp.gate_proj": {
2071
+ "group_size": 64,
2072
+ "bits": 4
2073
+ },
2074
+ "model.layers.17.mlp.down_proj": {
2075
+ "group_size": 64,
2076
+ "bits": 6
2077
+ },
2078
+ "model.layers.17.mlp.up_proj": {
2079
+ "group_size": 64,
2080
+ "bits": 4
2081
+ },
2082
+ "model.layers.17.input_layernorm": false,
2083
+ "model.layers.17.post_attention_layernorm": false,
2084
+ "model.layers.18.self_attn.q_proj": {
2085
+ "group_size": 64,
2086
+ "bits": 4
2087
+ },
2088
+ "model.layers.18.self_attn.k_proj": {
2089
+ "group_size": 64,
2090
+ "bits": 4
2091
+ },
2092
+ "model.layers.18.self_attn.v_proj": {
2093
+ "group_size": 64,
2094
+ "bits": 4
2095
+ },
2096
+ "model.layers.18.self_attn.o_proj": {
2097
+ "group_size": 64,
2098
+ "bits": 4
2099
+ },
2100
+ "model.layers.18.self_attn.rope": false,
2101
+ "model.layers.18.mlp.gate_proj": {
2102
+ "group_size": 64,
2103
+ "bits": 4
2104
+ },
2105
+ "model.layers.18.mlp.down_proj": {
2106
+ "group_size": 64,
2107
+ "bits": 4
2108
+ },
2109
+ "model.layers.18.mlp.up_proj": {
2110
+ "group_size": 64,
2111
+ "bits": 4
2112
+ },
2113
+ "model.layers.18.input_layernorm": false,
2114
+ "model.layers.18.post_attention_layernorm": false,
2115
+ "model.layers.19.self_attn.q_proj": {
2116
+ "group_size": 64,
2117
+ "bits": 4
2118
+ },
2119
+ "model.layers.19.self_attn.k_proj": {
2120
+ "group_size": 64,
2121
+ "bits": 4
2122
+ },
2123
+ "model.layers.19.self_attn.v_proj": {
2124
+ "group_size": 64,
2125
+ "bits": 4
2126
+ },
2127
+ "model.layers.19.self_attn.o_proj": {
2128
+ "group_size": 64,
2129
+ "bits": 4
2130
+ },
2131
+ "model.layers.19.self_attn.rope": false,
2132
+ "model.layers.19.mlp.gate_proj": {
2133
+ "group_size": 64,
2134
+ "bits": 4
2135
+ },
2136
+ "model.layers.19.mlp.down_proj": {
2137
+ "group_size": 64,
2138
+ "bits": 4
2139
+ },
2140
+ "model.layers.19.mlp.up_proj": {
2141
+ "group_size": 64,
2142
+ "bits": 4
2143
+ },
2144
+ "model.layers.19.input_layernorm": false,
2145
+ "model.layers.19.post_attention_layernorm": false,
2146
+ "model.layers.20.self_attn.q_proj": {
2147
+ "group_size": 64,
2148
+ "bits": 4
2149
+ },
2150
+ "model.layers.20.self_attn.k_proj": {
2151
+ "group_size": 64,
2152
+ "bits": 4
2153
+ },
2154
+ "model.layers.20.self_attn.v_proj": {
2155
+ "group_size": 64,
2156
+ "bits": 6
2157
+ },
2158
+ "model.layers.20.self_attn.o_proj": {
2159
+ "group_size": 64,
2160
+ "bits": 4
2161
+ },
2162
+ "model.layers.20.self_attn.rope": false,
2163
+ "model.layers.20.mlp.gate_proj": {
2164
+ "group_size": 64,
2165
+ "bits": 4
2166
+ },
2167
+ "model.layers.20.mlp.down_proj": {
2168
+ "group_size": 64,
2169
+ "bits": 6
2170
+ },
2171
+ "model.layers.20.mlp.up_proj": {
2172
+ "group_size": 64,
2173
+ "bits": 4
2174
+ },
2175
+ "model.layers.20.input_layernorm": false,
2176
+ "model.layers.20.post_attention_layernorm": false,
2177
+ "model.layers.21.self_attn.q_proj": {
2178
+ "group_size": 64,
2179
+ "bits": 4
2180
+ },
2181
+ "model.layers.21.self_attn.k_proj": {
2182
+ "group_size": 64,
2183
+ "bits": 4
2184
+ },
2185
+ "model.layers.21.self_attn.v_proj": {
2186
+ "group_size": 64,
2187
+ "bits": 4
2188
+ },
2189
+ "model.layers.21.self_attn.o_proj": {
2190
+ "group_size": 64,
2191
+ "bits": 4
2192
+ },
2193
+ "model.layers.21.self_attn.rope": false,
2194
+ "model.layers.21.mlp.gate_proj": {
2195
+ "group_size": 64,
2196
+ "bits": 4
2197
+ },
2198
+ "model.layers.21.mlp.down_proj": {
2199
+ "group_size": 64,
2200
+ "bits": 4
2201
+ },
2202
+ "model.layers.21.mlp.up_proj": {
2203
+ "group_size": 64,
2204
+ "bits": 4
2205
+ },
2206
+ "model.layers.21.input_layernorm": false,
2207
+ "model.layers.21.post_attention_layernorm": false,
2208
+ "model.layers.22.self_attn.q_proj": {
2209
+ "group_size": 64,
2210
+ "bits": 4
2211
+ },
2212
+ "model.layers.22.self_attn.k_proj": {
2213
+ "group_size": 64,
2214
+ "bits": 4
2215
+ },
2216
+ "model.layers.22.self_attn.v_proj": {
2217
+ "group_size": 64,
2218
+ "bits": 4
2219
+ },
2220
+ "model.layers.22.self_attn.o_proj": {
2221
+ "group_size": 64,
2222
+ "bits": 4
2223
+ },
2224
+ "model.layers.22.self_attn.rope": false,
2225
+ "model.layers.22.mlp.gate_proj": {
2226
+ "group_size": 64,
2227
+ "bits": 4
2228
+ },
2229
+ "model.layers.22.mlp.down_proj": {
2230
+ "group_size": 64,
2231
+ "bits": 4
2232
+ },
2233
+ "model.layers.22.mlp.up_proj": {
2234
+ "group_size": 64,
2235
+ "bits": 4
2236
+ },
2237
+ "model.layers.22.input_layernorm": false,
2238
+ "model.layers.22.post_attention_layernorm": false,
2239
+ "model.layers.23.self_attn.q_proj": {
2240
+ "group_size": 64,
2241
+ "bits": 4
2242
+ },
2243
+ "model.layers.23.self_attn.k_proj": {
2244
+ "group_size": 64,
2245
+ "bits": 4
2246
+ },
2247
+ "model.layers.23.self_attn.v_proj": {
2248
+ "group_size": 64,
2249
+ "bits": 6
2250
+ },
2251
+ "model.layers.23.self_attn.o_proj": {
2252
+ "group_size": 64,
2253
+ "bits": 4
2254
+ },
2255
+ "model.layers.23.self_attn.rope": false,
2256
+ "model.layers.23.mlp.gate_proj": {
2257
+ "group_size": 64,
2258
+ "bits": 4
2259
+ },
2260
+ "model.layers.23.mlp.down_proj": {
2261
+ "group_size": 64,
2262
+ "bits": 6
2263
+ },
2264
+ "model.layers.23.mlp.up_proj": {
2265
+ "group_size": 64,
2266
+ "bits": 4
2267
+ },
2268
+ "model.layers.23.input_layernorm": false,
2269
+ "model.layers.23.post_attention_layernorm": false,
2270
+ "model.layers.24.self_attn.q_proj": {
2271
+ "group_size": 64,
2272
+ "bits": 4
2273
+ },
2274
+ "model.layers.24.self_attn.k_proj": {
2275
+ "group_size": 64,
2276
+ "bits": 4
2277
+ },
2278
+ "model.layers.24.self_attn.v_proj": {
2279
+ "group_size": 64,
2280
+ "bits": 4
2281
+ },
2282
+ "model.layers.24.self_attn.o_proj": {
2283
+ "group_size": 64,
2284
+ "bits": 4
2285
+ },
2286
+ "model.layers.24.self_attn.rope": false,
2287
+ "model.layers.24.mlp.gate_proj": {
2288
+ "group_size": 64,
2289
+ "bits": 4
2290
+ },
2291
+ "model.layers.24.mlp.down_proj": {
2292
+ "group_size": 64,
2293
+ "bits": 4
2294
+ },
2295
+ "model.layers.24.mlp.up_proj": {
2296
+ "group_size": 64,
2297
+ "bits": 4
2298
+ },
2299
+ "model.layers.24.input_layernorm": false,
2300
+ "model.layers.24.post_attention_layernorm": false,
2301
+ "model.layers.25.self_attn.q_proj": {
2302
+ "group_size": 64,
2303
+ "bits": 4
2304
+ },
2305
+ "model.layers.25.self_attn.k_proj": {
2306
+ "group_size": 64,
2307
+ "bits": 4
2308
+ },
2309
+ "model.layers.25.self_attn.v_proj": {
2310
+ "group_size": 64,
2311
+ "bits": 4
2312
+ },
2313
+ "model.layers.25.self_attn.o_proj": {
2314
+ "group_size": 64,
2315
+ "bits": 4
2316
+ },
2317
+ "model.layers.25.self_attn.rope": false,
2318
+ "model.layers.25.mlp.gate_proj": {
2319
+ "group_size": 64,
2320
+ "bits": 4
2321
+ },
2322
+ "model.layers.25.mlp.down_proj": {
2323
+ "group_size": 64,
2324
+ "bits": 4
2325
+ },
2326
+ "model.layers.25.mlp.up_proj": {
2327
+ "group_size": 64,
2328
+ "bits": 4
2329
+ },
2330
+ "model.layers.25.input_layernorm": false,
2331
+ "model.layers.25.post_attention_layernorm": false,
2332
+ "model.layers.26.self_attn.q_proj": {
2333
+ "group_size": 64,
2334
+ "bits": 4
2335
+ },
2336
+ "model.layers.26.self_attn.k_proj": {
2337
+ "group_size": 64,
2338
+ "bits": 4
2339
+ },
2340
+ "model.layers.26.self_attn.v_proj": {
2341
+ "group_size": 64,
2342
+ "bits": 6
2343
+ },
2344
+ "model.layers.26.self_attn.o_proj": {
2345
+ "group_size": 64,
2346
+ "bits": 4
2347
+ },
2348
+ "model.layers.26.self_attn.rope": false,
2349
+ "model.layers.26.mlp.gate_proj": {
2350
+ "group_size": 64,
2351
+ "bits": 4
2352
+ },
2353
+ "model.layers.26.mlp.down_proj": {
2354
+ "group_size": 64,
2355
+ "bits": 6
2356
+ },
2357
+ "model.layers.26.mlp.up_proj": {
2358
+ "group_size": 64,
2359
+ "bits": 4
2360
+ },
2361
+ "model.layers.26.input_layernorm": false,
2362
+ "model.layers.26.post_attention_layernorm": false,
2363
+ "model.layers.27.self_attn.q_proj": {
2364
+ "group_size": 64,
2365
+ "bits": 4
2366
+ },
2367
+ "model.layers.27.self_attn.k_proj": {
2368
+ "group_size": 64,
2369
+ "bits": 4
2370
+ },
2371
+ "model.layers.27.self_attn.v_proj": {
2372
+ "group_size": 64,
2373
+ "bits": 4
2374
+ },
2375
+ "model.layers.27.self_attn.o_proj": {
2376
+ "group_size": 64,
2377
+ "bits": 4
2378
+ },
2379
+ "model.layers.27.self_attn.rope": false,
2380
+ "model.layers.27.mlp.gate_proj": {
2381
+ "group_size": 64,
2382
+ "bits": 4
2383
+ },
2384
+ "model.layers.27.mlp.down_proj": {
2385
+ "group_size": 64,
2386
+ "bits": 4
2387
+ },
2388
+ "model.layers.27.mlp.up_proj": {
2389
+ "group_size": 64,
2390
+ "bits": 4
2391
+ },
2392
+ "model.layers.27.input_layernorm": false,
2393
+ "model.layers.27.post_attention_layernorm": false,
2394
+ "model.layers.28.self_attn.q_proj": {
2395
+ "group_size": 64,
2396
+ "bits": 4
2397
+ },
2398
+ "model.layers.28.self_attn.k_proj": {
2399
+ "group_size": 64,
2400
+ "bits": 4
2401
+ },
2402
+ "model.layers.28.self_attn.v_proj": {
2403
+ "group_size": 64,
2404
+ "bits": 4
2405
+ },
2406
+ "model.layers.28.self_attn.o_proj": {
2407
+ "group_size": 64,
2408
+ "bits": 4
2409
+ },
2410
+ "model.layers.28.self_attn.rope": false,
2411
+ "model.layers.28.mlp.gate_proj": {
2412
+ "group_size": 64,
2413
+ "bits": 4
2414
+ },
2415
+ "model.layers.28.mlp.down_proj": {
2416
+ "group_size": 64,
2417
+ "bits": 4
2418
+ },
2419
+ "model.layers.28.mlp.up_proj": {
2420
+ "group_size": 64,
2421
+ "bits": 4
2422
+ },
2423
+ "model.layers.28.input_layernorm": false,
2424
+ "model.layers.28.post_attention_layernorm": false,
2425
+ "model.layers.29.self_attn.q_proj": {
2426
+ "group_size": 64,
2427
+ "bits": 4
2428
+ },
2429
+ "model.layers.29.self_attn.k_proj": {
2430
+ "group_size": 64,
2431
+ "bits": 4
2432
+ },
2433
+ "model.layers.29.self_attn.v_proj": {
2434
+ "group_size": 64,
2435
+ "bits": 6
2436
+ },
2437
+ "model.layers.29.self_attn.o_proj": {
2438
+ "group_size": 64,
2439
+ "bits": 4
2440
+ },
2441
+ "model.layers.29.self_attn.rope": false,
2442
+ "model.layers.29.mlp.gate_proj": {
2443
+ "group_size": 64,
2444
+ "bits": 4
2445
+ },
2446
+ "model.layers.29.mlp.down_proj": {
2447
+ "group_size": 64,
2448
+ "bits": 6
2449
+ },
2450
+ "model.layers.29.mlp.up_proj": {
2451
+ "group_size": 64,
2452
+ "bits": 4
2453
+ },
2454
+ "model.layers.29.input_layernorm": false,
2455
+ "model.layers.29.post_attention_layernorm": false,
2456
+ "model.layers.30.self_attn.q_proj": {
2457
+ "group_size": 64,
2458
+ "bits": 4
2459
+ },
2460
+ "model.layers.30.self_attn.k_proj": {
2461
+ "group_size": 64,
2462
+ "bits": 4
2463
+ },
2464
+ "model.layers.30.self_attn.v_proj": {
2465
+ "group_size": 64,
2466
+ "bits": 4
2467
+ },
2468
+ "model.layers.30.self_attn.o_proj": {
2469
+ "group_size": 64,
2470
+ "bits": 4
2471
+ },
2472
+ "model.layers.30.self_attn.rope": false,
2473
+ "model.layers.30.mlp.gate_proj": {
2474
+ "group_size": 64,
2475
+ "bits": 4
2476
+ },
2477
+ "model.layers.30.mlp.down_proj": {
2478
+ "group_size": 64,
2479
+ "bits": 4
2480
+ },
2481
+ "model.layers.30.mlp.up_proj": {
2482
+ "group_size": 64,
2483
+ "bits": 4
2484
+ },
2485
+ "model.layers.30.input_layernorm": false,
2486
+ "model.layers.30.post_attention_layernorm": false,
2487
+ "model.layers.31.self_attn.q_proj": {
2488
+ "group_size": 64,
2489
+ "bits": 4
2490
+ },
2491
+ "model.layers.31.self_attn.k_proj": {
2492
+ "group_size": 64,
2493
+ "bits": 4
2494
+ },
2495
+ "model.layers.31.self_attn.v_proj": {
2496
+ "group_size": 64,
2497
+ "bits": 4
2498
+ },
2499
+ "model.layers.31.self_attn.o_proj": {
2500
+ "group_size": 64,
2501
+ "bits": 4
2502
+ },
2503
+ "model.layers.31.self_attn.rope": false,
2504
+ "model.layers.31.mlp.gate_proj": {
2505
+ "group_size": 64,
2506
+ "bits": 4
2507
+ },
2508
+ "model.layers.31.mlp.down_proj": {
2509
+ "group_size": 64,
2510
+ "bits": 4
2511
+ },
2512
+ "model.layers.31.mlp.up_proj": {
2513
+ "group_size": 64,
2514
+ "bits": 4
2515
+ },
2516
+ "model.layers.31.input_layernorm": false,
2517
+ "model.layers.31.post_attention_layernorm": false,
2518
+ "model.layers.32.self_attn.q_proj": {
2519
+ "group_size": 64,
2520
+ "bits": 4
2521
+ },
2522
+ "model.layers.32.self_attn.k_proj": {
2523
+ "group_size": 64,
2524
+ "bits": 4
2525
+ },
2526
+ "model.layers.32.self_attn.v_proj": {
2527
+ "group_size": 64,
2528
+ "bits": 6
2529
+ },
2530
+ "model.layers.32.self_attn.o_proj": {
2531
+ "group_size": 64,
2532
+ "bits": 4
2533
+ },
2534
+ "model.layers.32.self_attn.rope": false,
2535
+ "model.layers.32.mlp.gate_proj": {
2536
+ "group_size": 64,
2537
+ "bits": 4
2538
+ },
2539
+ "model.layers.32.mlp.down_proj": {
2540
+ "group_size": 64,
2541
+ "bits": 6
2542
+ },
2543
+ "model.layers.32.mlp.up_proj": {
2544
+ "group_size": 64,
2545
+ "bits": 4
2546
+ },
2547
+ "model.layers.32.input_layernorm": false,
2548
+ "model.layers.32.post_attention_layernorm": false,
2549
+ "model.layers.33.self_attn.q_proj": {
2550
+ "group_size": 64,
2551
+ "bits": 4
2552
+ },
2553
+ "model.layers.33.self_attn.k_proj": {
2554
+ "group_size": 64,
2555
+ "bits": 4
2556
+ },
2557
+ "model.layers.33.self_attn.v_proj": {
2558
+ "group_size": 64,
2559
+ "bits": 4
2560
+ },
2561
+ "model.layers.33.self_attn.o_proj": {
2562
+ "group_size": 64,
2563
+ "bits": 4
2564
+ },
2565
+ "model.layers.33.self_attn.rope": false,
2566
+ "model.layers.33.mlp.gate_proj": {
2567
+ "group_size": 64,
2568
+ "bits": 4
2569
+ },
2570
+ "model.layers.33.mlp.down_proj": {
2571
+ "group_size": 64,
2572
+ "bits": 4
2573
+ },
2574
+ "model.layers.33.mlp.up_proj": {
2575
+ "group_size": 64,
2576
+ "bits": 4
2577
+ },
2578
+ "model.layers.33.input_layernorm": false,
2579
+ "model.layers.33.post_attention_layernorm": false,
2580
+ "model.layers.34.self_attn.q_proj": {
2581
+ "group_size": 64,
2582
+ "bits": 4
2583
+ },
2584
+ "model.layers.34.self_attn.k_proj": {
2585
+ "group_size": 64,
2586
+ "bits": 4
2587
+ },
2588
+ "model.layers.34.self_attn.v_proj": {
2589
+ "group_size": 64,
2590
+ "bits": 4
2591
+ },
2592
+ "model.layers.34.self_attn.o_proj": {
2593
+ "group_size": 64,
2594
+ "bits": 4
2595
+ },
2596
+ "model.layers.34.self_attn.rope": false,
2597
+ "model.layers.34.mlp.gate_proj": {
2598
+ "group_size": 64,
2599
+ "bits": 4
2600
+ },
2601
+ "model.layers.34.mlp.down_proj": {
2602
+ "group_size": 64,
2603
+ "bits": 4
2604
+ },
2605
+ "model.layers.34.mlp.up_proj": {
2606
+ "group_size": 64,
2607
+ "bits": 4
2608
+ },
2609
+ "model.layers.34.input_layernorm": false,
2610
+ "model.layers.34.post_attention_layernorm": false,
2611
+ "model.layers.35.self_attn.q_proj": {
2612
+ "group_size": 64,
2613
+ "bits": 4
2614
+ },
2615
+ "model.layers.35.self_attn.k_proj": {
2616
+ "group_size": 64,
2617
+ "bits": 4
2618
+ },
2619
+ "model.layers.35.self_attn.v_proj": {
2620
+ "group_size": 64,
2621
+ "bits": 6
2622
+ },
2623
+ "model.layers.35.self_attn.o_proj": {
2624
+ "group_size": 64,
2625
+ "bits": 4
2626
+ },
2627
+ "model.layers.35.self_attn.rope": false,
2628
+ "model.layers.35.mlp.gate_proj": {
2629
+ "group_size": 64,
2630
+ "bits": 4
2631
+ },
2632
+ "model.layers.35.mlp.down_proj": {
2633
+ "group_size": 64,
2634
+ "bits": 6
2635
+ },
2636
+ "model.layers.35.mlp.up_proj": {
2637
+ "group_size": 64,
2638
+ "bits": 4
2639
+ },
2640
+ "model.layers.35.input_layernorm": false,
2641
+ "model.layers.35.post_attention_layernorm": false,
2642
+ "model.layers.36.self_attn.q_proj": {
2643
+ "group_size": 64,
2644
+ "bits": 4
2645
+ },
2646
+ "model.layers.36.self_attn.k_proj": {
2647
+ "group_size": 64,
2648
+ "bits": 4
2649
+ },
2650
+ "model.layers.36.self_attn.v_proj": {
2651
+ "group_size": 64,
2652
+ "bits": 4
2653
+ },
2654
+ "model.layers.36.self_attn.o_proj": {
2655
+ "group_size": 64,
2656
+ "bits": 4
2657
+ },
2658
+ "model.layers.36.self_attn.rope": false,
2659
+ "model.layers.36.mlp.gate_proj": {
2660
+ "group_size": 64,
2661
+ "bits": 4
2662
+ },
2663
+ "model.layers.36.mlp.down_proj": {
2664
+ "group_size": 64,
2665
+ "bits": 4
2666
+ },
2667
+ "model.layers.36.mlp.up_proj": {
2668
+ "group_size": 64,
2669
+ "bits": 4
2670
+ },
2671
+ "model.layers.36.input_layernorm": false,
2672
+ "model.layers.36.post_attention_layernorm": false,
2673
+ "model.layers.37.self_attn.q_proj": {
2674
+ "group_size": 64,
2675
+ "bits": 4
2676
+ },
2677
+ "model.layers.37.self_attn.k_proj": {
2678
+ "group_size": 64,
2679
+ "bits": 4
2680
+ },
2681
+ "model.layers.37.self_attn.v_proj": {
2682
+ "group_size": 64,
2683
+ "bits": 4
2684
+ },
2685
+ "model.layers.37.self_attn.o_proj": {
2686
+ "group_size": 64,
2687
+ "bits": 4
2688
+ },
2689
+ "model.layers.37.self_attn.rope": false,
2690
+ "model.layers.37.mlp.gate_proj": {
2691
+ "group_size": 64,
2692
+ "bits": 4
2693
+ },
2694
+ "model.layers.37.mlp.down_proj": {
2695
+ "group_size": 64,
2696
+ "bits": 4
2697
+ },
2698
+ "model.layers.37.mlp.up_proj": {
2699
+ "group_size": 64,
2700
+ "bits": 4
2701
+ },
2702
+ "model.layers.37.input_layernorm": false,
2703
+ "model.layers.37.post_attention_layernorm": false,
2704
+ "model.layers.38.self_attn.q_proj": {
2705
+ "group_size": 64,
2706
+ "bits": 4
2707
+ },
2708
+ "model.layers.38.self_attn.k_proj": {
2709
+ "group_size": 64,
2710
+ "bits": 4
2711
+ },
2712
+ "model.layers.38.self_attn.v_proj": {
2713
+ "group_size": 64,
2714
+ "bits": 6
2715
+ },
2716
+ "model.layers.38.self_attn.o_proj": {
2717
+ "group_size": 64,
2718
+ "bits": 4
2719
+ },
2720
+ "model.layers.38.self_attn.rope": false,
2721
+ "model.layers.38.mlp.gate_proj": {
2722
+ "group_size": 64,
2723
+ "bits": 4
2724
+ },
2725
+ "model.layers.38.mlp.down_proj": {
2726
+ "group_size": 64,
2727
+ "bits": 6
2728
+ },
2729
+ "model.layers.38.mlp.up_proj": {
2730
+ "group_size": 64,
2731
+ "bits": 4
2732
+ },
2733
+ "model.layers.38.input_layernorm": false,
2734
+ "model.layers.38.post_attention_layernorm": false,
2735
+ "model.layers.39.self_attn.q_proj": {
2736
+ "group_size": 64,
2737
+ "bits": 4
2738
+ },
2739
+ "model.layers.39.self_attn.k_proj": {
2740
+ "group_size": 64,
2741
+ "bits": 4
2742
+ },
2743
+ "model.layers.39.self_attn.v_proj": {
2744
+ "group_size": 64,
2745
+ "bits": 4
2746
+ },
2747
+ "model.layers.39.self_attn.o_proj": {
2748
+ "group_size": 64,
2749
+ "bits": 4
2750
+ },
2751
+ "model.layers.39.self_attn.rope": false,
2752
+ "model.layers.39.mlp.gate_proj": {
2753
+ "group_size": 64,
2754
+ "bits": 4
2755
+ },
2756
+ "model.layers.39.mlp.down_proj": {
2757
+ "group_size": 64,
2758
+ "bits": 4
2759
+ },
2760
+ "model.layers.39.mlp.up_proj": {
2761
+ "group_size": 64,
2762
+ "bits": 4
2763
+ },
2764
+ "model.layers.39.input_layernorm": false,
2765
+ "model.layers.39.post_attention_layernorm": false,
2766
+ "model.layers.40.self_attn.q_proj": {
2767
+ "group_size": 64,
2768
+ "bits": 4
2769
+ },
2770
+ "model.layers.40.self_attn.k_proj": {
2771
+ "group_size": 64,
2772
+ "bits": 4
2773
+ },
2774
+ "model.layers.40.self_attn.v_proj": {
2775
+ "group_size": 64,
2776
+ "bits": 4
2777
+ },
2778
+ "model.layers.40.self_attn.o_proj": {
2779
+ "group_size": 64,
2780
+ "bits": 4
2781
+ },
2782
+ "model.layers.40.self_attn.rope": false,
2783
+ "model.layers.40.mlp.gate_proj": {
2784
+ "group_size": 64,
2785
+ "bits": 4
2786
+ },
2787
+ "model.layers.40.mlp.down_proj": {
2788
+ "group_size": 64,
2789
+ "bits": 4
2790
+ },
2791
+ "model.layers.40.mlp.up_proj": {
2792
+ "group_size": 64,
2793
+ "bits": 4
2794
+ },
2795
+ "model.layers.40.input_layernorm": false,
2796
+ "model.layers.40.post_attention_layernorm": false,
2797
+ "model.layers.41.self_attn.q_proj": {
2798
+ "group_size": 64,
2799
+ "bits": 4
2800
+ },
2801
+ "model.layers.41.self_attn.k_proj": {
2802
+ "group_size": 64,
2803
+ "bits": 4
2804
+ },
2805
+ "model.layers.41.self_attn.v_proj": {
2806
+ "group_size": 64,
2807
+ "bits": 6
2808
+ },
2809
+ "model.layers.41.self_attn.o_proj": {
2810
+ "group_size": 64,
2811
+ "bits": 4
2812
+ },
2813
+ "model.layers.41.self_attn.rope": false,
2814
+ "model.layers.41.mlp.gate_proj": {
2815
+ "group_size": 64,
2816
+ "bits": 4
2817
+ },
2818
+ "model.layers.41.mlp.down_proj": {
2819
+ "group_size": 64,
2820
+ "bits": 6
2821
+ },
2822
+ "model.layers.41.mlp.up_proj": {
2823
+ "group_size": 64,
2824
+ "bits": 4
2825
+ },
2826
+ "model.layers.41.input_layernorm": false,
2827
+ "model.layers.41.post_attention_layernorm": false,
2828
+ "model.layers.42.self_attn.q_proj": {
2829
+ "group_size": 64,
2830
+ "bits": 4
2831
+ },
2832
+ "model.layers.42.self_attn.k_proj": {
2833
+ "group_size": 64,
2834
+ "bits": 4
2835
+ },
2836
+ "model.layers.42.self_attn.v_proj": {
2837
+ "group_size": 64,
2838
+ "bits": 6
2839
+ },
2840
+ "model.layers.42.self_attn.o_proj": {
2841
+ "group_size": 64,
2842
+ "bits": 4
2843
+ },
2844
+ "model.layers.42.self_attn.rope": false,
2845
+ "model.layers.42.mlp.gate_proj": {
2846
+ "group_size": 64,
2847
+ "bits": 4
2848
+ },
2849
+ "model.layers.42.mlp.down_proj": {
2850
+ "group_size": 64,
2851
+ "bits": 6
2852
+ },
2853
+ "model.layers.42.mlp.up_proj": {
2854
+ "group_size": 64,
2855
+ "bits": 4
2856
+ },
2857
+ "model.layers.42.input_layernorm": false,
2858
+ "model.layers.42.post_attention_layernorm": false,
2859
+ "model.layers.43.self_attn.q_proj": {
2860
+ "group_size": 64,
2861
+ "bits": 4
2862
+ },
2863
+ "model.layers.43.self_attn.k_proj": {
2864
+ "group_size": 64,
2865
+ "bits": 4
2866
+ },
2867
+ "model.layers.43.self_attn.v_proj": {
2868
+ "group_size": 64,
2869
+ "bits": 6
2870
+ },
2871
+ "model.layers.43.self_attn.o_proj": {
2872
+ "group_size": 64,
2873
+ "bits": 4
2874
+ },
2875
+ "model.layers.43.self_attn.rope": false,
2876
+ "model.layers.43.mlp.gate_proj": {
2877
+ "group_size": 64,
2878
+ "bits": 4
2879
+ },
2880
+ "model.layers.43.mlp.down_proj": {
2881
+ "group_size": 64,
2882
+ "bits": 6
2883
+ },
2884
+ "model.layers.43.mlp.up_proj": {
2885
+ "group_size": 64,
2886
+ "bits": 4
2887
+ },
2888
+ "model.layers.43.input_layernorm": false,
2889
+ "model.layers.43.post_attention_layernorm": false,
2890
+ "model.layers.44.self_attn.q_proj": {
2891
+ "group_size": 64,
2892
+ "bits": 4
2893
+ },
2894
+ "model.layers.44.self_attn.k_proj": {
2895
+ "group_size": 64,
2896
+ "bits": 4
2897
+ },
2898
+ "model.layers.44.self_attn.v_proj": {
2899
+ "group_size": 64,
2900
+ "bits": 6
2901
+ },
2902
+ "model.layers.44.self_attn.o_proj": {
2903
+ "group_size": 64,
2904
+ "bits": 4
2905
+ },
2906
+ "model.layers.44.self_attn.rope": false,
2907
+ "model.layers.44.mlp.gate_proj": {
2908
+ "group_size": 64,
2909
+ "bits": 4
2910
+ },
2911
+ "model.layers.44.mlp.down_proj": {
2912
+ "group_size": 64,
2913
+ "bits": 6
2914
+ },
2915
+ "model.layers.44.mlp.up_proj": {
2916
+ "group_size": 64,
2917
+ "bits": 4
2918
+ },
2919
+ "model.layers.44.input_layernorm": false,
2920
+ "model.layers.44.post_attention_layernorm": false,
2921
+ "model.layers.45.self_attn.q_proj": {
2922
+ "group_size": 64,
2923
+ "bits": 4
2924
+ },
2925
+ "model.layers.45.self_attn.k_proj": {
2926
+ "group_size": 64,
2927
+ "bits": 4
2928
+ },
2929
+ "model.layers.45.self_attn.v_proj": {
2930
+ "group_size": 64,
2931
+ "bits": 6
2932
+ },
2933
+ "model.layers.45.self_attn.o_proj": {
2934
+ "group_size": 64,
2935
+ "bits": 4
2936
+ },
2937
+ "model.layers.45.self_attn.rope": false,
2938
+ "model.layers.45.mlp.gate_proj": {
2939
+ "group_size": 64,
2940
+ "bits": 4
2941
+ },
2942
+ "model.layers.45.mlp.down_proj": {
2943
+ "group_size": 64,
2944
+ "bits": 6
2945
+ },
2946
+ "model.layers.45.mlp.up_proj": {
2947
+ "group_size": 64,
2948
+ "bits": 4
2949
+ },
2950
+ "model.layers.45.input_layernorm": false,
2951
+ "model.layers.45.post_attention_layernorm": false,
2952
+ "model.layers.46.self_attn.q_proj": {
2953
+ "group_size": 64,
2954
+ "bits": 4
2955
+ },
2956
+ "model.layers.46.self_attn.k_proj": {
2957
+ "group_size": 64,
2958
+ "bits": 4
2959
+ },
2960
+ "model.layers.46.self_attn.v_proj": {
2961
+ "group_size": 64,
2962
+ "bits": 6
2963
+ },
2964
+ "model.layers.46.self_attn.o_proj": {
2965
+ "group_size": 64,
2966
+ "bits": 4
2967
+ },
2968
+ "model.layers.46.self_attn.rope": false,
2969
+ "model.layers.46.mlp.gate_proj": {
2970
+ "group_size": 64,
2971
+ "bits": 4
2972
+ },
2973
+ "model.layers.46.mlp.down_proj": {
2974
+ "group_size": 64,
2975
+ "bits": 6
2976
+ },
2977
+ "model.layers.46.mlp.up_proj": {
2978
+ "group_size": 64,
2979
+ "bits": 4
2980
+ },
2981
+ "model.layers.46.input_layernorm": false,
2982
+ "model.layers.46.post_attention_layernorm": false,
2983
+ "model.layers.47.self_attn.q_proj": {
2984
+ "group_size": 64,
2985
+ "bits": 4
2986
+ },
2987
+ "model.layers.47.self_attn.k_proj": {
2988
+ "group_size": 64,
2989
+ "bits": 4
2990
+ },
2991
+ "model.layers.47.self_attn.v_proj": {
2992
+ "group_size": 64,
2993
+ "bits": 6
2994
+ },
2995
+ "model.layers.47.self_attn.o_proj": {
2996
+ "group_size": 64,
2997
+ "bits": 4
2998
+ },
2999
+ "model.layers.47.self_attn.rope": false,
3000
+ "model.layers.47.mlp.gate_proj": {
3001
+ "group_size": 64,
3002
+ "bits": 4
3003
+ },
3004
+ "model.layers.47.mlp.down_proj": {
3005
+ "group_size": 64,
3006
+ "bits": 6
3007
+ },
3008
+ "model.layers.47.mlp.up_proj": {
3009
+ "group_size": 64,
3010
+ "bits": 4
3011
+ },
3012
+ "model.layers.47.input_layernorm": false,
3013
+ "model.layers.47.post_attention_layernorm": false,
3014
+ "model.norm": false,
3015
+ "lm_head": {
3016
+ "group_size": 64,
3017
+ "bits": 6
3018
+ }
3019
+ },
3020
+ "rms_norm_eps": 1e-05,
3021
+ "rope_theta": 1000000.0,
3022
+ "sliding_window": 131072,
3023
+ "tie_word_embeddings": false,
3024
+ "torch_dtype": "bfloat16",
3025
+ "transformers_version": "4.43.1",
3026
+ "use_cache": true,
3027
+ "use_sliding_window": false,
3028
+ "vocab_size": 152064
3029
+ }
DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c71c01a2c8003f356e45fd63b08b78c2d28102aa026bcdea361ee6109338c7c
3
+ size 5367725398
DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d24884c671d51068e3d9c245cabadf0bf8970d99d3a11d1da5227a5b87da7ef
3
+ size 3592541399
DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/special_tokens_map.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<｜begin▁of▁sentence｜>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<｜end▁of▁sentence｜>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<｜end▁of▁sentence｜>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ }
23
+ }
DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e20ddafc659ba90242154b55275402edeca0715e5dbb30f56815a4ce081f4893
3
+ size 11422778
DeepSeek-R1-Distill-Qwen-14B-4,6_mixed/tokenizer_config.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "151643": {
7
+ "content": "<｜end▁of▁sentence｜>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "151644": {
15
+ "content": "<｜User｜>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": false
21
+ },
22
+ "151645": {
23
+ "content": "<｜Assistant｜>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": false
29
+ },
30
+ "151646": {
31
+ "content": "<｜begin▁of▁sentence｜>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false,
36
+ "special": true
37
+ },
38
+ "151647": {
39
+ "content": "<|EOT|>",
40
+ "lstrip": false,
41
+ "normalized": false,
42
+ "rstrip": false,
43
+ "single_word": false,
44
+ "special": false
45
+ },
46
+ "151648": {
47
+ "content": "<think>",
48
+ "lstrip": false,
49
+ "normalized": false,
50
+ "rstrip": false,
51
+ "single_word": false,
52
+ "special": false
53
+ },
54
+ "151649": {
55
+ "content": "</think>",
56
+ "lstrip": false,
57
+ "normalized": false,
58
+ "rstrip": false,
59
+ "single_word": false,
60
+ "special": false
61
+ },
62
+ "151650": {
63
+ "content": "<|quad_start|>",
64
+ "lstrip": false,
65
+ "normalized": false,
66
+ "rstrip": false,
67
+ "single_word": false,
68
+ "special": true
69
+ },
70
+ "151651": {
71
+ "content": "<|quad_end|>",
72
+ "lstrip": false,
73
+ "normalized": false,
74
+ "rstrip": false,
75
+ "single_word": false,
76
+ "special": true
77
+ },
78
+ "151652": {
79
+ "content": "<|vision_start|>",
80
+ "lstrip": false,
81
+ "normalized": false,
82
+ "rstrip": false,
83
+ "single_word": false,
84
+ "special": true
85
+ },
86
+ "151653": {
87
+ "content": "<|vision_end|>",
88
+ "lstrip": false,
89
+ "normalized": false,
90
+ "rstrip": false,
91
+ "single_word": false,
92
+ "special": true
93
+ },
94
+ "151654": {
95
+ "content": "<|vision_pad|>",
96
+ "lstrip": false,
97
+ "normalized": false,
98
+ "rstrip": false,
99
+ "single_word": false,
100
+ "special": true
101
+ },
102
+ "151655": {
103
+ "content": "<|image_pad|>",
104
+ "lstrip": false,
105
+ "normalized": false,
106
+ "rstrip": false,
107
+ "single_word": false,
108
+ "special": true
109
+ },
110
+ "151656": {
111
+ "content": "<|video_pad|>",
112
+ "lstrip": false,
113
+ "normalized": false,
114
+ "rstrip": false,
115
+ "single_word": false,
116
+ "special": true
117
+ },
118
+ "151657": {
119
+ "content": "<tool_call>",
120
+ "lstrip": false,
121
+ "normalized": false,
122
+ "rstrip": false,
123
+ "single_word": false,
124
+ "special": false
125
+ },
126
+ "151658": {
127
+ "content": "</tool_call>",
128
+ "lstrip": false,
129
+ "normalized": false,
130
+ "rstrip": false,
131
+ "single_word": false,
132
+ "special": false
133
+ },
134
+ "151659": {
135
+ "content": "<|fim_prefix|>",
136
+ "lstrip": false,
137
+ "normalized": false,
138
+ "rstrip": false,
139
+ "single_word": false,
140
+ "special": false
141
+ },
142
+ "151660": {
143
+ "content": "<|fim_middle|>",
144
+ "lstrip": false,
145
+ "normalized": false,
146
+ "rstrip": false,
147
+ "single_word": false,
148
+ "special": false
149
+ },
150
+ "151661": {
151
+ "content": "<|fim_suffix|>",
152
+ "lstrip": false,
153
+ "normalized": false,
154
+ "rstrip": false,
155
+ "single_word": false,
156
+ "special": false
157
+ },
158
+ "151662": {
159
+ "content": "<|fim_pad|>",
160
+ "lstrip": false,
161
+ "normalized": false,
162
+ "rstrip": false,
163
+ "single_word": false,
164
+ "special": false
165
+ },
166
+ "151663": {
167
+ "content": "<|repo_name|>",
168
+ "lstrip": false,
169
+ "normalized": false,
170
+ "rstrip": false,
171
+ "single_word": false,
172
+ "special": false
173
+ },
174
+ "151664": {
175
+ "content": "<|file_sep|>",
176
+ "lstrip": false,
177
+ "normalized": false,
178
+ "rstrip": false,
179
+ "single_word": false,
180
+ "special": false
181
+ }
182
+ },
183
+ "bos_token": "<｜begin▁of▁sentence｜>",
184
+ "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='') %}{%- for message in messages %}{%- if message['role'] == 'system' %}{% set ns.system_prompt = message['content'] %}{%- endif %}{%- endfor %}{{bos_token}}{{ns.system_prompt}}{%- for message in messages %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{{'<｜User｜>' + message['content']}}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is none %}{%- set ns.is_tool = false -%}{%- for tool in message['tool_calls']%}{%- if not ns.is_first %}{{'<｜Assistant｜><｜tool▁calls▁begin｜><｜tool▁call▁begin��>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{%- set ns.is_first = true -%}{%- else %}{{'\\n' + '<｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{{'<｜tool▁calls▁end｜><｜end▁of▁sentence｜>'}}{%- endif %}{%- endfor %}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is not none %}{%- if ns.is_tool %}{{'<｜tool▁outputs▁end｜>' + message['content'] + '<｜end▁of▁sentence｜>'}}{%- set ns.is_tool = false -%}{%- else %}{% set content = message['content'] %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{% endif %}{{'<｜Assistant｜>' + content + '<｜end▁of▁sentence｜>'}}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<｜tool▁outputs▁begin｜><｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- set ns.is_output_first = false %}{%- else %}{{'\\n<｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<｜tool▁outputs▁end｜>'}}{% endif %}{% if add_generation_prompt and not ns.is_tool %}{{'<｜Assistant｜><think>\\n'}}{% endif %}",
185
+ "clean_up_tokenization_spaces": false,
186
+ "eos_token": "<｜end▁of▁sentence｜>",
187
+ "extra_special_tokens": {},
188
+ "legacy": true,
189
+ "model_max_length": 16384,
190
+ "pad_token": "<｜end▁of▁sentence｜>",
191
+ "sp_model_kwargs": {},
192
+ "tokenizer_class": "LlamaTokenizerFast",
193
+ "unk_token": null,
194
+ "use_default_system_prompt": false
195
+ }
DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/config.json ADDED
@@ -0,0 +1,3029 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151643,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 5120,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 13824,
12
+ "max_position_embeddings": 131072,
13
+ "max_window_layers": 48,
14
+ "model_type": "qwen2",
15
+ "num_attention_heads": 40,
16
+ "num_hidden_layers": 48,
17
+ "num_key_value_heads": 8,
18
+ "quantization": {
19
+ "group_size": 64,
20
+ "bits": null,
21
+ "model.embed_tokens": {
22
+ "group_size": 64,
23
+ "bits": 4
24
+ },
25
+ "model.layers.0.self_attn.q_proj": {
26
+ "group_size": 64,
27
+ "bits": 4
28
+ },
29
+ "model.layers.0.self_attn.k_proj": {
30
+ "group_size": 64,
31
+ "bits": 4
32
+ },
33
+ "model.layers.0.self_attn.v_proj": {
34
+ "group_size": 64,
35
+ "bits": 8
36
+ },
37
+ "model.layers.0.self_attn.o_proj": {
38
+ "group_size": 64,
39
+ "bits": 4
40
+ },
41
+ "model.layers.0.self_attn.rope": false,
42
+ "model.layers.0.mlp.gate_proj": {
43
+ "group_size": 64,
44
+ "bits": 4
45
+ },
46
+ "model.layers.0.mlp.down_proj": {
47
+ "group_size": 64,
48
+ "bits": 8
49
+ },
50
+ "model.layers.0.mlp.up_proj": {
51
+ "group_size": 64,
52
+ "bits": 4
53
+ },
54
+ "model.layers.0.input_layernorm": false,
55
+ "model.layers.0.post_attention_layernorm": false,
56
+ "model.layers.1.self_attn.q_proj": {
57
+ "group_size": 64,
58
+ "bits": 4
59
+ },
60
+ "model.layers.1.self_attn.k_proj": {
61
+ "group_size": 64,
62
+ "bits": 4
63
+ },
64
+ "model.layers.1.self_attn.v_proj": {
65
+ "group_size": 64,
66
+ "bits": 8
67
+ },
68
+ "model.layers.1.self_attn.o_proj": {
69
+ "group_size": 64,
70
+ "bits": 4
71
+ },
72
+ "model.layers.1.self_attn.rope": false,
73
+ "model.layers.1.mlp.gate_proj": {
74
+ "group_size": 64,
75
+ "bits": 4
76
+ },
77
+ "model.layers.1.mlp.down_proj": {
78
+ "group_size": 64,
79
+ "bits": 8
80
+ },
81
+ "model.layers.1.mlp.up_proj": {
82
+ "group_size": 64,
83
+ "bits": 4
84
+ },
85
+ "model.layers.1.input_layernorm": false,
86
+ "model.layers.1.post_attention_layernorm": false,
87
+ "model.layers.2.self_attn.q_proj": {
88
+ "group_size": 64,
89
+ "bits": 4
90
+ },
91
+ "model.layers.2.self_attn.k_proj": {
92
+ "group_size": 64,
93
+ "bits": 4
94
+ },
95
+ "model.layers.2.self_attn.v_proj": {
96
+ "group_size": 64,
97
+ "bits": 8
98
+ },
99
+ "model.layers.2.self_attn.o_proj": {
100
+ "group_size": 64,
101
+ "bits": 4
102
+ },
103
+ "model.layers.2.self_attn.rope": false,
104
+ "model.layers.2.mlp.gate_proj": {
105
+ "group_size": 64,
106
+ "bits": 4
107
+ },
108
+ "model.layers.2.mlp.down_proj": {
109
+ "group_size": 64,
110
+ "bits": 8
111
+ },
112
+ "model.layers.2.mlp.up_proj": {
113
+ "group_size": 64,
114
+ "bits": 4
115
+ },
116
+ "model.layers.2.input_layernorm": false,
117
+ "model.layers.2.post_attention_layernorm": false,
118
+ "model.layers.3.self_attn.q_proj": {
119
+ "group_size": 64,
120
+ "bits": 4
121
+ },
122
+ "model.layers.3.self_attn.k_proj": {
123
+ "group_size": 64,
124
+ "bits": 4
125
+ },
126
+ "model.layers.3.self_attn.v_proj": {
127
+ "group_size": 64,
128
+ "bits": 8
129
+ },
130
+ "model.layers.3.self_attn.o_proj": {
131
+ "group_size": 64,
132
+ "bits": 4
133
+ },
134
+ "model.layers.3.self_attn.rope": false,
135
+ "model.layers.3.mlp.gate_proj": {
136
+ "group_size": 64,
137
+ "bits": 4
138
+ },
139
+ "model.layers.3.mlp.down_proj": {
140
+ "group_size": 64,
141
+ "bits": 8
142
+ },
143
+ "model.layers.3.mlp.up_proj": {
144
+ "group_size": 64,
145
+ "bits": 4
146
+ },
147
+ "model.layers.3.input_layernorm": false,
148
+ "model.layers.3.post_attention_layernorm": false,
149
+ "model.layers.4.self_attn.q_proj": {
150
+ "group_size": 64,
151
+ "bits": 4
152
+ },
153
+ "model.layers.4.self_attn.k_proj": {
154
+ "group_size": 64,
155
+ "bits": 4
156
+ },
157
+ "model.layers.4.self_attn.v_proj": {
158
+ "group_size": 64,
159
+ "bits": 8
160
+ },
161
+ "model.layers.4.self_attn.o_proj": {
162
+ "group_size": 64,
163
+ "bits": 4
164
+ },
165
+ "model.layers.4.self_attn.rope": false,
166
+ "model.layers.4.mlp.gate_proj": {
167
+ "group_size": 64,
168
+ "bits": 4
169
+ },
170
+ "model.layers.4.mlp.down_proj": {
171
+ "group_size": 64,
172
+ "bits": 8
173
+ },
174
+ "model.layers.4.mlp.up_proj": {
175
+ "group_size": 64,
176
+ "bits": 4
177
+ },
178
+ "model.layers.4.input_layernorm": false,
179
+ "model.layers.4.post_attention_layernorm": false,
180
+ "model.layers.5.self_attn.q_proj": {
181
+ "group_size": 64,
182
+ "bits": 4
183
+ },
184
+ "model.layers.5.self_attn.k_proj": {
185
+ "group_size": 64,
186
+ "bits": 4
187
+ },
188
+ "model.layers.5.self_attn.v_proj": {
189
+ "group_size": 64,
190
+ "bits": 8
191
+ },
192
+ "model.layers.5.self_attn.o_proj": {
193
+ "group_size": 64,
194
+ "bits": 4
195
+ },
196
+ "model.layers.5.self_attn.rope": false,
197
+ "model.layers.5.mlp.gate_proj": {
198
+ "group_size": 64,
199
+ "bits": 4
200
+ },
201
+ "model.layers.5.mlp.down_proj": {
202
+ "group_size": 64,
203
+ "bits": 8
204
+ },
205
+ "model.layers.5.mlp.up_proj": {
206
+ "group_size": 64,
207
+ "bits": 4
208
+ },
209
+ "model.layers.5.input_layernorm": false,
210
+ "model.layers.5.post_attention_layernorm": false,
211
+ "model.layers.6.self_attn.q_proj": {
212
+ "group_size": 64,
213
+ "bits": 4
214
+ },
215
+ "model.layers.6.self_attn.k_proj": {
216
+ "group_size": 64,
217
+ "bits": 4
218
+ },
219
+ "model.layers.6.self_attn.v_proj": {
220
+ "group_size": 64,
221
+ "bits": 4
222
+ },
223
+ "model.layers.6.self_attn.o_proj": {
224
+ "group_size": 64,
225
+ "bits": 4
226
+ },
227
+ "model.layers.6.self_attn.rope": false,
228
+ "model.layers.6.mlp.gate_proj": {
229
+ "group_size": 64,
230
+ "bits": 4
231
+ },
232
+ "model.layers.6.mlp.down_proj": {
233
+ "group_size": 64,
234
+ "bits": 4
235
+ },
236
+ "model.layers.6.mlp.up_proj": {
237
+ "group_size": 64,
238
+ "bits": 4
239
+ },
240
+ "model.layers.6.input_layernorm": false,
241
+ "model.layers.6.post_attention_layernorm": false,
242
+ "model.layers.7.self_attn.q_proj": {
243
+ "group_size": 64,
244
+ "bits": 4
245
+ },
246
+ "model.layers.7.self_attn.k_proj": {
247
+ "group_size": 64,
248
+ "bits": 4
249
+ },
250
+ "model.layers.7.self_attn.v_proj": {
251
+ "group_size": 64,
252
+ "bits": 4
253
+ },
254
+ "model.layers.7.self_attn.o_proj": {
255
+ "group_size": 64,
256
+ "bits": 4
257
+ },
258
+ "model.layers.7.self_attn.rope": false,
259
+ "model.layers.7.mlp.gate_proj": {
260
+ "group_size": 64,
261
+ "bits": 4
262
+ },
263
+ "model.layers.7.mlp.down_proj": {
264
+ "group_size": 64,
265
+ "bits": 4
266
+ },
267
+ "model.layers.7.mlp.up_proj": {
268
+ "group_size": 64,
269
+ "bits": 4
270
+ },
271
+ "model.layers.7.input_layernorm": false,
272
+ "model.layers.7.post_attention_layernorm": false,
273
+ "model.layers.8.self_attn.q_proj": {
274
+ "group_size": 64,
275
+ "bits": 4
276
+ },
277
+ "model.layers.8.self_attn.k_proj": {
278
+ "group_size": 64,
279
+ "bits": 4
280
+ },
281
+ "model.layers.8.self_attn.v_proj": {
282
+ "group_size": 64,
283
+ "bits": 8
284
+ },
285
+ "model.layers.8.self_attn.o_proj": {
286
+ "group_size": 64,
287
+ "bits": 4
288
+ },
289
+ "model.layers.8.self_attn.rope": false,
290
+ "model.layers.8.mlp.gate_proj": {
291
+ "group_size": 64,
292
+ "bits": 4
293
+ },
294
+ "model.layers.8.mlp.down_proj": {
295
+ "group_size": 64,
296
+ "bits": 8
297
+ },
298
+ "model.layers.8.mlp.up_proj": {
299
+ "group_size": 64,
300
+ "bits": 4
301
+ },
302
+ "model.layers.8.input_layernorm": false,
303
+ "model.layers.8.post_attention_layernorm": false,
304
+ "model.layers.9.self_attn.q_proj": {
305
+ "group_size": 64,
306
+ "bits": 4
307
+ },
308
+ "model.layers.9.self_attn.k_proj": {
309
+ "group_size": 64,
310
+ "bits": 4
311
+ },
312
+ "model.layers.9.self_attn.v_proj": {
313
+ "group_size": 64,
314
+ "bits": 4
315
+ },
316
+ "model.layers.9.self_attn.o_proj": {
317
+ "group_size": 64,
318
+ "bits": 4
319
+ },
320
+ "model.layers.9.self_attn.rope": false,
321
+ "model.layers.9.mlp.gate_proj": {
322
+ "group_size": 64,
323
+ "bits": 4
324
+ },
325
+ "model.layers.9.mlp.down_proj": {
326
+ "group_size": 64,
327
+ "bits": 4
328
+ },
329
+ "model.layers.9.mlp.up_proj": {
330
+ "group_size": 64,
331
+ "bits": 4
332
+ },
333
+ "model.layers.9.input_layernorm": false,
334
+ "model.layers.9.post_attention_layernorm": false,
335
+ "model.layers.10.self_attn.q_proj": {
336
+ "group_size": 64,
337
+ "bits": 4
338
+ },
339
+ "model.layers.10.self_attn.k_proj": {
340
+ "group_size": 64,
341
+ "bits": 4
342
+ },
343
+ "model.layers.10.self_attn.v_proj": {
344
+ "group_size": 64,
345
+ "bits": 4
346
+ },
347
+ "model.layers.10.self_attn.o_proj": {
348
+ "group_size": 64,
349
+ "bits": 4
350
+ },
351
+ "model.layers.10.self_attn.rope": false,
352
+ "model.layers.10.mlp.gate_proj": {
353
+ "group_size": 64,
354
+ "bits": 4
355
+ },
356
+ "model.layers.10.mlp.down_proj": {
357
+ "group_size": 64,
358
+ "bits": 4
359
+ },
360
+ "model.layers.10.mlp.up_proj": {
361
+ "group_size": 64,
362
+ "bits": 4
363
+ },
364
+ "model.layers.10.input_layernorm": false,
365
+ "model.layers.10.post_attention_layernorm": false,
366
+ "model.layers.11.self_attn.q_proj": {
367
+ "group_size": 64,
368
+ "bits": 4
369
+ },
370
+ "model.layers.11.self_attn.k_proj": {
371
+ "group_size": 64,
372
+ "bits": 4
373
+ },
374
+ "model.layers.11.self_attn.v_proj": {
375
+ "group_size": 64,
376
+ "bits": 8
377
+ },
378
+ "model.layers.11.self_attn.o_proj": {
379
+ "group_size": 64,
380
+ "bits": 4
381
+ },
382
+ "model.layers.11.self_attn.rope": false,
383
+ "model.layers.11.mlp.gate_proj": {
384
+ "group_size": 64,
385
+ "bits": 4
386
+ },
387
+ "model.layers.11.mlp.down_proj": {
388
+ "group_size": 64,
389
+ "bits": 8
390
+ },
391
+ "model.layers.11.mlp.up_proj": {
392
+ "group_size": 64,
393
+ "bits": 4
394
+ },
395
+ "model.layers.11.input_layernorm": false,
396
+ "model.layers.11.post_attention_layernorm": false,
397
+ "model.layers.12.self_attn.q_proj": {
398
+ "group_size": 64,
399
+ "bits": 4
400
+ },
401
+ "model.layers.12.self_attn.k_proj": {
402
+ "group_size": 64,
403
+ "bits": 4
404
+ },
405
+ "model.layers.12.self_attn.v_proj": {
406
+ "group_size": 64,
407
+ "bits": 4
408
+ },
409
+ "model.layers.12.self_attn.o_proj": {
410
+ "group_size": 64,
411
+ "bits": 4
412
+ },
413
+ "model.layers.12.self_attn.rope": false,
414
+ "model.layers.12.mlp.gate_proj": {
415
+ "group_size": 64,
416
+ "bits": 4
417
+ },
418
+ "model.layers.12.mlp.down_proj": {
419
+ "group_size": 64,
420
+ "bits": 4
421
+ },
422
+ "model.layers.12.mlp.up_proj": {
423
+ "group_size": 64,
424
+ "bits": 4
425
+ },
426
+ "model.layers.12.input_layernorm": false,
427
+ "model.layers.12.post_attention_layernorm": false,
428
+ "model.layers.13.self_attn.q_proj": {
429
+ "group_size": 64,
430
+ "bits": 4
431
+ },
432
+ "model.layers.13.self_attn.k_proj": {
433
+ "group_size": 64,
434
+ "bits": 4
435
+ },
436
+ "model.layers.13.self_attn.v_proj": {
437
+ "group_size": 64,
438
+ "bits": 4
439
+ },
440
+ "model.layers.13.self_attn.o_proj": {
441
+ "group_size": 64,
442
+ "bits": 4
443
+ },
444
+ "model.layers.13.self_attn.rope": false,
445
+ "model.layers.13.mlp.gate_proj": {
446
+ "group_size": 64,
447
+ "bits": 4
448
+ },
449
+ "model.layers.13.mlp.down_proj": {
450
+ "group_size": 64,
451
+ "bits": 4
452
+ },
453
+ "model.layers.13.mlp.up_proj": {
454
+ "group_size": 64,
455
+ "bits": 4
456
+ },
457
+ "model.layers.13.input_layernorm": false,
458
+ "model.layers.13.post_attention_layernorm": false,
459
+ "model.layers.14.self_attn.q_proj": {
460
+ "group_size": 64,
461
+ "bits": 4
462
+ },
463
+ "model.layers.14.self_attn.k_proj": {
464
+ "group_size": 64,
465
+ "bits": 4
466
+ },
467
+ "model.layers.14.self_attn.v_proj": {
468
+ "group_size": 64,
469
+ "bits": 8
470
+ },
471
+ "model.layers.14.self_attn.o_proj": {
472
+ "group_size": 64,
473
+ "bits": 4
474
+ },
475
+ "model.layers.14.self_attn.rope": false,
476
+ "model.layers.14.mlp.gate_proj": {
477
+ "group_size": 64,
478
+ "bits": 4
479
+ },
480
+ "model.layers.14.mlp.down_proj": {
481
+ "group_size": 64,
482
+ "bits": 8
483
+ },
484
+ "model.layers.14.mlp.up_proj": {
485
+ "group_size": 64,
486
+ "bits": 4
487
+ },
488
+ "model.layers.14.input_layernorm": false,
489
+ "model.layers.14.post_attention_layernorm": false,
490
+ "model.layers.15.self_attn.q_proj": {
491
+ "group_size": 64,
492
+ "bits": 4
493
+ },
494
+ "model.layers.15.self_attn.k_proj": {
495
+ "group_size": 64,
496
+ "bits": 4
497
+ },
498
+ "model.layers.15.self_attn.v_proj": {
499
+ "group_size": 64,
500
+ "bits": 4
501
+ },
502
+ "model.layers.15.self_attn.o_proj": {
503
+ "group_size": 64,
504
+ "bits": 4
505
+ },
506
+ "model.layers.15.self_attn.rope": false,
507
+ "model.layers.15.mlp.gate_proj": {
508
+ "group_size": 64,
509
+ "bits": 4
510
+ },
511
+ "model.layers.15.mlp.down_proj": {
512
+ "group_size": 64,
513
+ "bits": 4
514
+ },
515
+ "model.layers.15.mlp.up_proj": {
516
+ "group_size": 64,
517
+ "bits": 4
518
+ },
519
+ "model.layers.15.input_layernorm": false,
520
+ "model.layers.15.post_attention_layernorm": false,
521
+ "model.layers.16.self_attn.q_proj": {
522
+ "group_size": 64,
523
+ "bits": 4
524
+ },
525
+ "model.layers.16.self_attn.k_proj": {
526
+ "group_size": 64,
527
+ "bits": 4
528
+ },
529
+ "model.layers.16.self_attn.v_proj": {
530
+ "group_size": 64,
531
+ "bits": 4
532
+ },
533
+ "model.layers.16.self_attn.o_proj": {
534
+ "group_size": 64,
535
+ "bits": 4
536
+ },
537
+ "model.layers.16.self_attn.rope": false,
538
+ "model.layers.16.mlp.gate_proj": {
539
+ "group_size": 64,
540
+ "bits": 4
541
+ },
542
+ "model.layers.16.mlp.down_proj": {
543
+ "group_size": 64,
544
+ "bits": 4
545
+ },
546
+ "model.layers.16.mlp.up_proj": {
547
+ "group_size": 64,
548
+ "bits": 4
549
+ },
550
+ "model.layers.16.input_layernorm": false,
551
+ "model.layers.16.post_attention_layernorm": false,
552
+ "model.layers.17.self_attn.q_proj": {
553
+ "group_size": 64,
554
+ "bits": 4
555
+ },
556
+ "model.layers.17.self_attn.k_proj": {
557
+ "group_size": 64,
558
+ "bits": 4
559
+ },
560
+ "model.layers.17.self_attn.v_proj": {
561
+ "group_size": 64,
562
+ "bits": 8
563
+ },
564
+ "model.layers.17.self_attn.o_proj": {
565
+ "group_size": 64,
566
+ "bits": 4
567
+ },
568
+ "model.layers.17.self_attn.rope": false,
569
+ "model.layers.17.mlp.gate_proj": {
570
+ "group_size": 64,
571
+ "bits": 4
572
+ },
573
+ "model.layers.17.mlp.down_proj": {
574
+ "group_size": 64,
575
+ "bits": 8
576
+ },
577
+ "model.layers.17.mlp.up_proj": {
578
+ "group_size": 64,
579
+ "bits": 4
580
+ },
581
+ "model.layers.17.input_layernorm": false,
582
+ "model.layers.17.post_attention_layernorm": false,
583
+ "model.layers.18.self_attn.q_proj": {
584
+ "group_size": 64,
585
+ "bits": 4
586
+ },
587
+ "model.layers.18.self_attn.k_proj": {
588
+ "group_size": 64,
589
+ "bits": 4
590
+ },
591
+ "model.layers.18.self_attn.v_proj": {
592
+ "group_size": 64,
593
+ "bits": 4
594
+ },
595
+ "model.layers.18.self_attn.o_proj": {
596
+ "group_size": 64,
597
+ "bits": 4
598
+ },
599
+ "model.layers.18.self_attn.rope": false,
600
+ "model.layers.18.mlp.gate_proj": {
601
+ "group_size": 64,
602
+ "bits": 4
603
+ },
604
+ "model.layers.18.mlp.down_proj": {
605
+ "group_size": 64,
606
+ "bits": 4
607
+ },
608
+ "model.layers.18.mlp.up_proj": {
609
+ "group_size": 64,
610
+ "bits": 4
611
+ },
612
+ "model.layers.18.input_layernorm": false,
613
+ "model.layers.18.post_attention_layernorm": false,
614
+ "model.layers.19.self_attn.q_proj": {
615
+ "group_size": 64,
616
+ "bits": 4
617
+ },
618
+ "model.layers.19.self_attn.k_proj": {
619
+ "group_size": 64,
620
+ "bits": 4
621
+ },
622
+ "model.layers.19.self_attn.v_proj": {
623
+ "group_size": 64,
624
+ "bits": 4
625
+ },
626
+ "model.layers.19.self_attn.o_proj": {
627
+ "group_size": 64,
628
+ "bits": 4
629
+ },
630
+ "model.layers.19.self_attn.rope": false,
631
+ "model.layers.19.mlp.gate_proj": {
632
+ "group_size": 64,
633
+ "bits": 4
634
+ },
635
+ "model.layers.19.mlp.down_proj": {
636
+ "group_size": 64,
637
+ "bits": 4
638
+ },
639
+ "model.layers.19.mlp.up_proj": {
640
+ "group_size": 64,
641
+ "bits": 4
642
+ },
643
+ "model.layers.19.input_layernorm": false,
644
+ "model.layers.19.post_attention_layernorm": false,
645
+ "model.layers.20.self_attn.q_proj": {
646
+ "group_size": 64,
647
+ "bits": 4
648
+ },
649
+ "model.layers.20.self_attn.k_proj": {
650
+ "group_size": 64,
651
+ "bits": 4
652
+ },
653
+ "model.layers.20.self_attn.v_proj": {
654
+ "group_size": 64,
655
+ "bits": 8
656
+ },
657
+ "model.layers.20.self_attn.o_proj": {
658
+ "group_size": 64,
659
+ "bits": 4
660
+ },
661
+ "model.layers.20.self_attn.rope": false,
662
+ "model.layers.20.mlp.gate_proj": {
663
+ "group_size": 64,
664
+ "bits": 4
665
+ },
666
+ "model.layers.20.mlp.down_proj": {
667
+ "group_size": 64,
668
+ "bits": 8
669
+ },
670
+ "model.layers.20.mlp.up_proj": {
671
+ "group_size": 64,
672
+ "bits": 4
673
+ },
674
+ "model.layers.20.input_layernorm": false,
675
+ "model.layers.20.post_attention_layernorm": false,
676
+ "model.layers.21.self_attn.q_proj": {
677
+ "group_size": 64,
678
+ "bits": 4
679
+ },
680
+ "model.layers.21.self_attn.k_proj": {
681
+ "group_size": 64,
682
+ "bits": 4
683
+ },
684
+ "model.layers.21.self_attn.v_proj": {
685
+ "group_size": 64,
686
+ "bits": 4
687
+ },
688
+ "model.layers.21.self_attn.o_proj": {
689
+ "group_size": 64,
690
+ "bits": 4
691
+ },
692
+ "model.layers.21.self_attn.rope": false,
693
+ "model.layers.21.mlp.gate_proj": {
694
+ "group_size": 64,
695
+ "bits": 4
696
+ },
697
+ "model.layers.21.mlp.down_proj": {
698
+ "group_size": 64,
699
+ "bits": 4
700
+ },
701
+ "model.layers.21.mlp.up_proj": {
702
+ "group_size": 64,
703
+ "bits": 4
704
+ },
705
+ "model.layers.21.input_layernorm": false,
706
+ "model.layers.21.post_attention_layernorm": false,
707
+ "model.layers.22.self_attn.q_proj": {
708
+ "group_size": 64,
709
+ "bits": 4
710
+ },
711
+ "model.layers.22.self_attn.k_proj": {
712
+ "group_size": 64,
713
+ "bits": 4
714
+ },
715
+ "model.layers.22.self_attn.v_proj": {
716
+ "group_size": 64,
717
+ "bits": 4
718
+ },
719
+ "model.layers.22.self_attn.o_proj": {
720
+ "group_size": 64,
721
+ "bits": 4
722
+ },
723
+ "model.layers.22.self_attn.rope": false,
724
+ "model.layers.22.mlp.gate_proj": {
725
+ "group_size": 64,
726
+ "bits": 4
727
+ },
728
+ "model.layers.22.mlp.down_proj": {
729
+ "group_size": 64,
730
+ "bits": 4
731
+ },
732
+ "model.layers.22.mlp.up_proj": {
733
+ "group_size": 64,
734
+ "bits": 4
735
+ },
736
+ "model.layers.22.input_layernorm": false,
737
+ "model.layers.22.post_attention_layernorm": false,
738
+ "model.layers.23.self_attn.q_proj": {
739
+ "group_size": 64,
740
+ "bits": 4
741
+ },
742
+ "model.layers.23.self_attn.k_proj": {
743
+ "group_size": 64,
744
+ "bits": 4
745
+ },
746
+ "model.layers.23.self_attn.v_proj": {
747
+ "group_size": 64,
748
+ "bits": 8
749
+ },
750
+ "model.layers.23.self_attn.o_proj": {
751
+ "group_size": 64,
752
+ "bits": 4
753
+ },
754
+ "model.layers.23.self_attn.rope": false,
755
+ "model.layers.23.mlp.gate_proj": {
756
+ "group_size": 64,
757
+ "bits": 4
758
+ },
759
+ "model.layers.23.mlp.down_proj": {
760
+ "group_size": 64,
761
+ "bits": 8
762
+ },
763
+ "model.layers.23.mlp.up_proj": {
764
+ "group_size": 64,
765
+ "bits": 4
766
+ },
767
+ "model.layers.23.input_layernorm": false,
768
+ "model.layers.23.post_attention_layernorm": false,
769
+ "model.layers.24.self_attn.q_proj": {
770
+ "group_size": 64,
771
+ "bits": 4
772
+ },
773
+ "model.layers.24.self_attn.k_proj": {
774
+ "group_size": 64,
775
+ "bits": 4
776
+ },
777
+ "model.layers.24.self_attn.v_proj": {
778
+ "group_size": 64,
779
+ "bits": 4
780
+ },
781
+ "model.layers.24.self_attn.o_proj": {
782
+ "group_size": 64,
783
+ "bits": 4
784
+ },
785
+ "model.layers.24.self_attn.rope": false,
786
+ "model.layers.24.mlp.gate_proj": {
787
+ "group_size": 64,
788
+ "bits": 4
789
+ },
790
+ "model.layers.24.mlp.down_proj": {
791
+ "group_size": 64,
792
+ "bits": 4
793
+ },
794
+ "model.layers.24.mlp.up_proj": {
795
+ "group_size": 64,
796
+ "bits": 4
797
+ },
798
+ "model.layers.24.input_layernorm": false,
799
+ "model.layers.24.post_attention_layernorm": false,
800
+ "model.layers.25.self_attn.q_proj": {
801
+ "group_size": 64,
802
+ "bits": 4
803
+ },
804
+ "model.layers.25.self_attn.k_proj": {
805
+ "group_size": 64,
806
+ "bits": 4
807
+ },
808
+ "model.layers.25.self_attn.v_proj": {
809
+ "group_size": 64,
810
+ "bits": 4
811
+ },
812
+ "model.layers.25.self_attn.o_proj": {
813
+ "group_size": 64,
814
+ "bits": 4
815
+ },
816
+ "model.layers.25.self_attn.rope": false,
817
+ "model.layers.25.mlp.gate_proj": {
818
+ "group_size": 64,
819
+ "bits": 4
820
+ },
821
+ "model.layers.25.mlp.down_proj": {
822
+ "group_size": 64,
823
+ "bits": 4
824
+ },
825
+ "model.layers.25.mlp.up_proj": {
826
+ "group_size": 64,
827
+ "bits": 4
828
+ },
829
+ "model.layers.25.input_layernorm": false,
830
+ "model.layers.25.post_attention_layernorm": false,
831
+ "model.layers.26.self_attn.q_proj": {
832
+ "group_size": 64,
833
+ "bits": 4
834
+ },
835
+ "model.layers.26.self_attn.k_proj": {
836
+ "group_size": 64,
837
+ "bits": 4
838
+ },
839
+ "model.layers.26.self_attn.v_proj": {
840
+ "group_size": 64,
841
+ "bits": 8
842
+ },
843
+ "model.layers.26.self_attn.o_proj": {
844
+ "group_size": 64,
845
+ "bits": 4
846
+ },
847
+ "model.layers.26.self_attn.rope": false,
848
+ "model.layers.26.mlp.gate_proj": {
849
+ "group_size": 64,
850
+ "bits": 4
851
+ },
852
+ "model.layers.26.mlp.down_proj": {
853
+ "group_size": 64,
854
+ "bits": 8
855
+ },
856
+ "model.layers.26.mlp.up_proj": {
857
+ "group_size": 64,
858
+ "bits": 4
859
+ },
860
+ "model.layers.26.input_layernorm": false,
861
+ "model.layers.26.post_attention_layernorm": false,
862
+ "model.layers.27.self_attn.q_proj": {
863
+ "group_size": 64,
864
+ "bits": 4
865
+ },
866
+ "model.layers.27.self_attn.k_proj": {
867
+ "group_size": 64,
868
+ "bits": 4
869
+ },
870
+ "model.layers.27.self_attn.v_proj": {
871
+ "group_size": 64,
872
+ "bits": 4
873
+ },
874
+ "model.layers.27.self_attn.o_proj": {
875
+ "group_size": 64,
876
+ "bits": 4
877
+ },
878
+ "model.layers.27.self_attn.rope": false,
879
+ "model.layers.27.mlp.gate_proj": {
880
+ "group_size": 64,
881
+ "bits": 4
882
+ },
883
+ "model.layers.27.mlp.down_proj": {
884
+ "group_size": 64,
885
+ "bits": 4
886
+ },
887
+ "model.layers.27.mlp.up_proj": {
888
+ "group_size": 64,
889
+ "bits": 4
890
+ },
891
+ "model.layers.27.input_layernorm": false,
892
+ "model.layers.27.post_attention_layernorm": false,
893
+ "model.layers.28.self_attn.q_proj": {
894
+ "group_size": 64,
895
+ "bits": 4
896
+ },
897
+ "model.layers.28.self_attn.k_proj": {
898
+ "group_size": 64,
899
+ "bits": 4
900
+ },
901
+ "model.layers.28.self_attn.v_proj": {
902
+ "group_size": 64,
903
+ "bits": 4
904
+ },
905
+ "model.layers.28.self_attn.o_proj": {
906
+ "group_size": 64,
907
+ "bits": 4
908
+ },
909
+ "model.layers.28.self_attn.rope": false,
910
+ "model.layers.28.mlp.gate_proj": {
911
+ "group_size": 64,
912
+ "bits": 4
913
+ },
914
+ "model.layers.28.mlp.down_proj": {
915
+ "group_size": 64,
916
+ "bits": 4
917
+ },
918
+ "model.layers.28.mlp.up_proj": {
919
+ "group_size": 64,
920
+ "bits": 4
921
+ },
922
+ "model.layers.28.input_layernorm": false,
923
+ "model.layers.28.post_attention_layernorm": false,
924
+ "model.layers.29.self_attn.q_proj": {
925
+ "group_size": 64,
926
+ "bits": 4
927
+ },
928
+ "model.layers.29.self_attn.k_proj": {
929
+ "group_size": 64,
930
+ "bits": 4
931
+ },
932
+ "model.layers.29.self_attn.v_proj": {
933
+ "group_size": 64,
934
+ "bits": 8
935
+ },
936
+ "model.layers.29.self_attn.o_proj": {
937
+ "group_size": 64,
938
+ "bits": 4
939
+ },
940
+ "model.layers.29.self_attn.rope": false,
941
+ "model.layers.29.mlp.gate_proj": {
942
+ "group_size": 64,
943
+ "bits": 4
944
+ },
945
+ "model.layers.29.mlp.down_proj": {
946
+ "group_size": 64,
947
+ "bits": 8
948
+ },
949
+ "model.layers.29.mlp.up_proj": {
950
+ "group_size": 64,
951
+ "bits": 4
952
+ },
953
+ "model.layers.29.input_layernorm": false,
954
+ "model.layers.29.post_attention_layernorm": false,
955
+ "model.layers.30.self_attn.q_proj": {
956
+ "group_size": 64,
957
+ "bits": 4
958
+ },
959
+ "model.layers.30.self_attn.k_proj": {
960
+ "group_size": 64,
961
+ "bits": 4
962
+ },
963
+ "model.layers.30.self_attn.v_proj": {
964
+ "group_size": 64,
965
+ "bits": 4
966
+ },
967
+ "model.layers.30.self_attn.o_proj": {
968
+ "group_size": 64,
969
+ "bits": 4
970
+ },
971
+ "model.layers.30.self_attn.rope": false,
972
+ "model.layers.30.mlp.gate_proj": {
973
+ "group_size": 64,
974
+ "bits": 4
975
+ },
976
+ "model.layers.30.mlp.down_proj": {
977
+ "group_size": 64,
978
+ "bits": 4
979
+ },
980
+ "model.layers.30.mlp.up_proj": {
981
+ "group_size": 64,
982
+ "bits": 4
983
+ },
984
+ "model.layers.30.input_layernorm": false,
985
+ "model.layers.30.post_attention_layernorm": false,
986
+ "model.layers.31.self_attn.q_proj": {
987
+ "group_size": 64,
988
+ "bits": 4
989
+ },
990
+ "model.layers.31.self_attn.k_proj": {
991
+ "group_size": 64,
992
+ "bits": 4
993
+ },
994
+ "model.layers.31.self_attn.v_proj": {
995
+ "group_size": 64,
996
+ "bits": 4
997
+ },
998
+ "model.layers.31.self_attn.o_proj": {
999
+ "group_size": 64,
1000
+ "bits": 4
1001
+ },
1002
+ "model.layers.31.self_attn.rope": false,
1003
+ "model.layers.31.mlp.gate_proj": {
1004
+ "group_size": 64,
1005
+ "bits": 4
1006
+ },
1007
+ "model.layers.31.mlp.down_proj": {
1008
+ "group_size": 64,
1009
+ "bits": 4
1010
+ },
1011
+ "model.layers.31.mlp.up_proj": {
1012
+ "group_size": 64,
1013
+ "bits": 4
1014
+ },
1015
+ "model.layers.31.input_layernorm": false,
1016
+ "model.layers.31.post_attention_layernorm": false,
1017
+ "model.layers.32.self_attn.q_proj": {
1018
+ "group_size": 64,
1019
+ "bits": 4
1020
+ },
1021
+ "model.layers.32.self_attn.k_proj": {
1022
+ "group_size": 64,
1023
+ "bits": 4
1024
+ },
1025
+ "model.layers.32.self_attn.v_proj": {
1026
+ "group_size": 64,
1027
+ "bits": 8
1028
+ },
1029
+ "model.layers.32.self_attn.o_proj": {
1030
+ "group_size": 64,
1031
+ "bits": 4
1032
+ },
1033
+ "model.layers.32.self_attn.rope": false,
1034
+ "model.layers.32.mlp.gate_proj": {
1035
+ "group_size": 64,
1036
+ "bits": 4
1037
+ },
1038
+ "model.layers.32.mlp.down_proj": {
1039
+ "group_size": 64,
1040
+ "bits": 8
1041
+ },
1042
+ "model.layers.32.mlp.up_proj": {
1043
+ "group_size": 64,
1044
+ "bits": 4
1045
+ },
1046
+ "model.layers.32.input_layernorm": false,
1047
+ "model.layers.32.post_attention_layernorm": false,
1048
+ "model.layers.33.self_attn.q_proj": {
1049
+ "group_size": 64,
1050
+ "bits": 4
1051
+ },
1052
+ "model.layers.33.self_attn.k_proj": {
1053
+ "group_size": 64,
1054
+ "bits": 4
1055
+ },
1056
+ "model.layers.33.self_attn.v_proj": {
1057
+ "group_size": 64,
1058
+ "bits": 4
1059
+ },
1060
+ "model.layers.33.self_attn.o_proj": {
1061
+ "group_size": 64,
1062
+ "bits": 4
1063
+ },
1064
+ "model.layers.33.self_attn.rope": false,
1065
+ "model.layers.33.mlp.gate_proj": {
1066
+ "group_size": 64,
1067
+ "bits": 4
1068
+ },
1069
+ "model.layers.33.mlp.down_proj": {
1070
+ "group_size": 64,
1071
+ "bits": 4
1072
+ },
1073
+ "model.layers.33.mlp.up_proj": {
1074
+ "group_size": 64,
1075
+ "bits": 4
1076
+ },
1077
+ "model.layers.33.input_layernorm": false,
1078
+ "model.layers.33.post_attention_layernorm": false,
1079
+ "model.layers.34.self_attn.q_proj": {
1080
+ "group_size": 64,
1081
+ "bits": 4
1082
+ },
1083
+ "model.layers.34.self_attn.k_proj": {
1084
+ "group_size": 64,
1085
+ "bits": 4
1086
+ },
1087
+ "model.layers.34.self_attn.v_proj": {
1088
+ "group_size": 64,
1089
+ "bits": 4
1090
+ },
1091
+ "model.layers.34.self_attn.o_proj": {
1092
+ "group_size": 64,
1093
+ "bits": 4
1094
+ },
1095
+ "model.layers.34.self_attn.rope": false,
1096
+ "model.layers.34.mlp.gate_proj": {
1097
+ "group_size": 64,
1098
+ "bits": 4
1099
+ },
1100
+ "model.layers.34.mlp.down_proj": {
1101
+ "group_size": 64,
1102
+ "bits": 4
1103
+ },
1104
+ "model.layers.34.mlp.up_proj": {
1105
+ "group_size": 64,
1106
+ "bits": 4
1107
+ },
1108
+ "model.layers.34.input_layernorm": false,
1109
+ "model.layers.34.post_attention_layernorm": false,
1110
+ "model.layers.35.self_attn.q_proj": {
1111
+ "group_size": 64,
1112
+ "bits": 4
1113
+ },
1114
+ "model.layers.35.self_attn.k_proj": {
1115
+ "group_size": 64,
1116
+ "bits": 4
1117
+ },
1118
+ "model.layers.35.self_attn.v_proj": {
1119
+ "group_size": 64,
1120
+ "bits": 8
1121
+ },
1122
+ "model.layers.35.self_attn.o_proj": {
1123
+ "group_size": 64,
1124
+ "bits": 4
1125
+ },
1126
+ "model.layers.35.self_attn.rope": false,
1127
+ "model.layers.35.mlp.gate_proj": {
1128
+ "group_size": 64,
1129
+ "bits": 4
1130
+ },
1131
+ "model.layers.35.mlp.down_proj": {
1132
+ "group_size": 64,
1133
+ "bits": 8
1134
+ },
1135
+ "model.layers.35.mlp.up_proj": {
1136
+ "group_size": 64,
1137
+ "bits": 4
1138
+ },
1139
+ "model.layers.35.input_layernorm": false,
1140
+ "model.layers.35.post_attention_layernorm": false,
1141
+ "model.layers.36.self_attn.q_proj": {
1142
+ "group_size": 64,
1143
+ "bits": 4
1144
+ },
1145
+ "model.layers.36.self_attn.k_proj": {
1146
+ "group_size": 64,
1147
+ "bits": 4
1148
+ },
1149
+ "model.layers.36.self_attn.v_proj": {
1150
+ "group_size": 64,
1151
+ "bits": 4
1152
+ },
1153
+ "model.layers.36.self_attn.o_proj": {
1154
+ "group_size": 64,
1155
+ "bits": 4
1156
+ },
1157
+ "model.layers.36.self_attn.rope": false,
1158
+ "model.layers.36.mlp.gate_proj": {
1159
+ "group_size": 64,
1160
+ "bits": 4
1161
+ },
1162
+ "model.layers.36.mlp.down_proj": {
1163
+ "group_size": 64,
1164
+ "bits": 4
1165
+ },
1166
+ "model.layers.36.mlp.up_proj": {
1167
+ "group_size": 64,
1168
+ "bits": 4
1169
+ },
1170
+ "model.layers.36.input_layernorm": false,
1171
+ "model.layers.36.post_attention_layernorm": false,
1172
+ "model.layers.37.self_attn.q_proj": {
1173
+ "group_size": 64,
1174
+ "bits": 4
1175
+ },
1176
+ "model.layers.37.self_attn.k_proj": {
1177
+ "group_size": 64,
1178
+ "bits": 4
1179
+ },
1180
+ "model.layers.37.self_attn.v_proj": {
1181
+ "group_size": 64,
1182
+ "bits": 4
1183
+ },
1184
+ "model.layers.37.self_attn.o_proj": {
1185
+ "group_size": 64,
1186
+ "bits": 4
1187
+ },
1188
+ "model.layers.37.self_attn.rope": false,
1189
+ "model.layers.37.mlp.gate_proj": {
1190
+ "group_size": 64,
1191
+ "bits": 4
1192
+ },
1193
+ "model.layers.37.mlp.down_proj": {
1194
+ "group_size": 64,
1195
+ "bits": 4
1196
+ },
1197
+ "model.layers.37.mlp.up_proj": {
1198
+ "group_size": 64,
1199
+ "bits": 4
1200
+ },
1201
+ "model.layers.37.input_layernorm": false,
1202
+ "model.layers.37.post_attention_layernorm": false,
1203
+ "model.layers.38.self_attn.q_proj": {
1204
+ "group_size": 64,
1205
+ "bits": 4
1206
+ },
1207
+ "model.layers.38.self_attn.k_proj": {
1208
+ "group_size": 64,
1209
+ "bits": 4
1210
+ },
1211
+ "model.layers.38.self_attn.v_proj": {
1212
+ "group_size": 64,
1213
+ "bits": 8
1214
+ },
1215
+ "model.layers.38.self_attn.o_proj": {
1216
+ "group_size": 64,
1217
+ "bits": 4
1218
+ },
1219
+ "model.layers.38.self_attn.rope": false,
1220
+ "model.layers.38.mlp.gate_proj": {
1221
+ "group_size": 64,
1222
+ "bits": 4
1223
+ },
1224
+ "model.layers.38.mlp.down_proj": {
1225
+ "group_size": 64,
1226
+ "bits": 8
1227
+ },
1228
+ "model.layers.38.mlp.up_proj": {
1229
+ "group_size": 64,
1230
+ "bits": 4
1231
+ },
1232
+ "model.layers.38.input_layernorm": false,
1233
+ "model.layers.38.post_attention_layernorm": false,
1234
+ "model.layers.39.self_attn.q_proj": {
1235
+ "group_size": 64,
1236
+ "bits": 4
1237
+ },
1238
+ "model.layers.39.self_attn.k_proj": {
1239
+ "group_size": 64,
1240
+ "bits": 4
1241
+ },
1242
+ "model.layers.39.self_attn.v_proj": {
1243
+ "group_size": 64,
1244
+ "bits": 4
1245
+ },
1246
+ "model.layers.39.self_attn.o_proj": {
1247
+ "group_size": 64,
1248
+ "bits": 4
1249
+ },
1250
+ "model.layers.39.self_attn.rope": false,
1251
+ "model.layers.39.mlp.gate_proj": {
1252
+ "group_size": 64,
1253
+ "bits": 4
1254
+ },
1255
+ "model.layers.39.mlp.down_proj": {
1256
+ "group_size": 64,
1257
+ "bits": 4
1258
+ },
1259
+ "model.layers.39.mlp.up_proj": {
1260
+ "group_size": 64,
1261
+ "bits": 4
1262
+ },
1263
+ "model.layers.39.input_layernorm": false,
1264
+ "model.layers.39.post_attention_layernorm": false,
1265
+ "model.layers.40.self_attn.q_proj": {
1266
+ "group_size": 64,
1267
+ "bits": 4
1268
+ },
1269
+ "model.layers.40.self_attn.k_proj": {
1270
+ "group_size": 64,
1271
+ "bits": 4
1272
+ },
1273
+ "model.layers.40.self_attn.v_proj": {
1274
+ "group_size": 64,
1275
+ "bits": 4
1276
+ },
1277
+ "model.layers.40.self_attn.o_proj": {
1278
+ "group_size": 64,
1279
+ "bits": 4
1280
+ },
1281
+ "model.layers.40.self_attn.rope": false,
1282
+ "model.layers.40.mlp.gate_proj": {
1283
+ "group_size": 64,
1284
+ "bits": 4
1285
+ },
1286
+ "model.layers.40.mlp.down_proj": {
1287
+ "group_size": 64,
1288
+ "bits": 4
1289
+ },
1290
+ "model.layers.40.mlp.up_proj": {
1291
+ "group_size": 64,
1292
+ "bits": 4
1293
+ },
1294
+ "model.layers.40.input_layernorm": false,
1295
+ "model.layers.40.post_attention_layernorm": false,
1296
+ "model.layers.41.self_attn.q_proj": {
1297
+ "group_size": 64,
1298
+ "bits": 4
1299
+ },
1300
+ "model.layers.41.self_attn.k_proj": {
1301
+ "group_size": 64,
1302
+ "bits": 4
1303
+ },
1304
+ "model.layers.41.self_attn.v_proj": {
1305
+ "group_size": 64,
1306
+ "bits": 8
1307
+ },
1308
+ "model.layers.41.self_attn.o_proj": {
1309
+ "group_size": 64,
1310
+ "bits": 4
1311
+ },
1312
+ "model.layers.41.self_attn.rope": false,
1313
+ "model.layers.41.mlp.gate_proj": {
1314
+ "group_size": 64,
1315
+ "bits": 4
1316
+ },
1317
+ "model.layers.41.mlp.down_proj": {
1318
+ "group_size": 64,
1319
+ "bits": 8
1320
+ },
1321
+ "model.layers.41.mlp.up_proj": {
1322
+ "group_size": 64,
1323
+ "bits": 4
1324
+ },
1325
+ "model.layers.41.input_layernorm": false,
1326
+ "model.layers.41.post_attention_layernorm": false,
1327
+ "model.layers.42.self_attn.q_proj": {
1328
+ "group_size": 64,
1329
+ "bits": 4
1330
+ },
1331
+ "model.layers.42.self_attn.k_proj": {
1332
+ "group_size": 64,
1333
+ "bits": 4
1334
+ },
1335
+ "model.layers.42.self_attn.v_proj": {
1336
+ "group_size": 64,
1337
+ "bits": 8
1338
+ },
1339
+ "model.layers.42.self_attn.o_proj": {
1340
+ "group_size": 64,
1341
+ "bits": 4
1342
+ },
1343
+ "model.layers.42.self_attn.rope": false,
1344
+ "model.layers.42.mlp.gate_proj": {
1345
+ "group_size": 64,
1346
+ "bits": 4
1347
+ },
1348
+ "model.layers.42.mlp.down_proj": {
1349
+ "group_size": 64,
1350
+ "bits": 8
1351
+ },
1352
+ "model.layers.42.mlp.up_proj": {
1353
+ "group_size": 64,
1354
+ "bits": 4
1355
+ },
1356
+ "model.layers.42.input_layernorm": false,
1357
+ "model.layers.42.post_attention_layernorm": false,
1358
+ "model.layers.43.self_attn.q_proj": {
1359
+ "group_size": 64,
1360
+ "bits": 4
1361
+ },
1362
+ "model.layers.43.self_attn.k_proj": {
1363
+ "group_size": 64,
1364
+ "bits": 4
1365
+ },
1366
+ "model.layers.43.self_attn.v_proj": {
1367
+ "group_size": 64,
1368
+ "bits": 8
1369
+ },
1370
+ "model.layers.43.self_attn.o_proj": {
1371
+ "group_size": 64,
1372
+ "bits": 4
1373
+ },
1374
+ "model.layers.43.self_attn.rope": false,
1375
+ "model.layers.43.mlp.gate_proj": {
1376
+ "group_size": 64,
1377
+ "bits": 4
1378
+ },
1379
+ "model.layers.43.mlp.down_proj": {
1380
+ "group_size": 64,
1381
+ "bits": 8
1382
+ },
1383
+ "model.layers.43.mlp.up_proj": {
1384
+ "group_size": 64,
1385
+ "bits": 4
1386
+ },
1387
+ "model.layers.43.input_layernorm": false,
1388
+ "model.layers.43.post_attention_layernorm": false,
1389
+ "model.layers.44.self_attn.q_proj": {
1390
+ "group_size": 64,
1391
+ "bits": 4
1392
+ },
1393
+ "model.layers.44.self_attn.k_proj": {
1394
+ "group_size": 64,
1395
+ "bits": 4
1396
+ },
1397
+ "model.layers.44.self_attn.v_proj": {
1398
+ "group_size": 64,
1399
+ "bits": 8
1400
+ },
1401
+ "model.layers.44.self_attn.o_proj": {
1402
+ "group_size": 64,
1403
+ "bits": 4
1404
+ },
1405
+ "model.layers.44.self_attn.rope": false,
1406
+ "model.layers.44.mlp.gate_proj": {
1407
+ "group_size": 64,
1408
+ "bits": 4
1409
+ },
1410
+ "model.layers.44.mlp.down_proj": {
1411
+ "group_size": 64,
1412
+ "bits": 8
1413
+ },
1414
+ "model.layers.44.mlp.up_proj": {
1415
+ "group_size": 64,
1416
+ "bits": 4
1417
+ },
1418
+ "model.layers.44.input_layernorm": false,
1419
+ "model.layers.44.post_attention_layernorm": false,
1420
+ "model.layers.45.self_attn.q_proj": {
1421
+ "group_size": 64,
1422
+ "bits": 4
1423
+ },
1424
+ "model.layers.45.self_attn.k_proj": {
1425
+ "group_size": 64,
1426
+ "bits": 4
1427
+ },
1428
+ "model.layers.45.self_attn.v_proj": {
1429
+ "group_size": 64,
1430
+ "bits": 8
1431
+ },
1432
+ "model.layers.45.self_attn.o_proj": {
1433
+ "group_size": 64,
1434
+ "bits": 4
1435
+ },
1436
+ "model.layers.45.self_attn.rope": false,
1437
+ "model.layers.45.mlp.gate_proj": {
1438
+ "group_size": 64,
1439
+ "bits": 4
1440
+ },
1441
+ "model.layers.45.mlp.down_proj": {
1442
+ "group_size": 64,
1443
+ "bits": 8
1444
+ },
1445
+ "model.layers.45.mlp.up_proj": {
1446
+ "group_size": 64,
1447
+ "bits": 4
1448
+ },
1449
+ "model.layers.45.input_layernorm": false,
1450
+ "model.layers.45.post_attention_layernorm": false,
1451
+ "model.layers.46.self_attn.q_proj": {
1452
+ "group_size": 64,
1453
+ "bits": 4
1454
+ },
1455
+ "model.layers.46.self_attn.k_proj": {
1456
+ "group_size": 64,
1457
+ "bits": 4
1458
+ },
1459
+ "model.layers.46.self_attn.v_proj": {
1460
+ "group_size": 64,
1461
+ "bits": 8
1462
+ },
1463
+ "model.layers.46.self_attn.o_proj": {
1464
+ "group_size": 64,
1465
+ "bits": 4
1466
+ },
1467
+ "model.layers.46.self_attn.rope": false,
1468
+ "model.layers.46.mlp.gate_proj": {
1469
+ "group_size": 64,
1470
+ "bits": 4
1471
+ },
1472
+ "model.layers.46.mlp.down_proj": {
1473
+ "group_size": 64,
1474
+ "bits": 8
1475
+ },
1476
+ "model.layers.46.mlp.up_proj": {
1477
+ "group_size": 64,
1478
+ "bits": 4
1479
+ },
1480
+ "model.layers.46.input_layernorm": false,
1481
+ "model.layers.46.post_attention_layernorm": false,
1482
+ "model.layers.47.self_attn.q_proj": {
1483
+ "group_size": 64,
1484
+ "bits": 4
1485
+ },
1486
+ "model.layers.47.self_attn.k_proj": {
1487
+ "group_size": 64,
1488
+ "bits": 4
1489
+ },
1490
+ "model.layers.47.self_attn.v_proj": {
1491
+ "group_size": 64,
1492
+ "bits": 8
1493
+ },
1494
+ "model.layers.47.self_attn.o_proj": {
1495
+ "group_size": 64,
1496
+ "bits": 4
1497
+ },
1498
+ "model.layers.47.self_attn.rope": false,
1499
+ "model.layers.47.mlp.gate_proj": {
1500
+ "group_size": 64,
1501
+ "bits": 4
1502
+ },
1503
+ "model.layers.47.mlp.down_proj": {
1504
+ "group_size": 64,
1505
+ "bits": 8
1506
+ },
1507
+ "model.layers.47.mlp.up_proj": {
1508
+ "group_size": 64,
1509
+ "bits": 4
1510
+ },
1511
+ "model.layers.47.input_layernorm": false,
1512
+ "model.layers.47.post_attention_layernorm": false,
1513
+ "model.norm": false,
1514
+ "lm_head": {
1515
+ "group_size": 64,
1516
+ "bits": 8
1517
+ }
1518
+ },
1519
+ "quantization_config": {
1520
+ "group_size": 64,
1521
+ "bits": null,
1522
+ "model.embed_tokens": {
1523
+ "group_size": 64,
1524
+ "bits": 4
1525
+ },
1526
+ "model.layers.0.self_attn.q_proj": {
1527
+ "group_size": 64,
1528
+ "bits": 4
1529
+ },
1530
+ "model.layers.0.self_attn.k_proj": {
1531
+ "group_size": 64,
1532
+ "bits": 4
1533
+ },
1534
+ "model.layers.0.self_attn.v_proj": {
1535
+ "group_size": 64,
1536
+ "bits": 8
1537
+ },
1538
+ "model.layers.0.self_attn.o_proj": {
1539
+ "group_size": 64,
1540
+ "bits": 4
1541
+ },
1542
+ "model.layers.0.self_attn.rope": false,
1543
+ "model.layers.0.mlp.gate_proj": {
1544
+ "group_size": 64,
1545
+ "bits": 4
1546
+ },
1547
+ "model.layers.0.mlp.down_proj": {
1548
+ "group_size": 64,
1549
+ "bits": 8
1550
+ },
1551
+ "model.layers.0.mlp.up_proj": {
1552
+ "group_size": 64,
1553
+ "bits": 4
1554
+ },
1555
+ "model.layers.0.input_layernorm": false,
1556
+ "model.layers.0.post_attention_layernorm": false,
1557
+ "model.layers.1.self_attn.q_proj": {
1558
+ "group_size": 64,
1559
+ "bits": 4
1560
+ },
1561
+ "model.layers.1.self_attn.k_proj": {
1562
+ "group_size": 64,
1563
+ "bits": 4
1564
+ },
1565
+ "model.layers.1.self_attn.v_proj": {
1566
+ "group_size": 64,
1567
+ "bits": 8
1568
+ },
1569
+ "model.layers.1.self_attn.o_proj": {
1570
+ "group_size": 64,
1571
+ "bits": 4
1572
+ },
1573
+ "model.layers.1.self_attn.rope": false,
1574
+ "model.layers.1.mlp.gate_proj": {
1575
+ "group_size": 64,
1576
+ "bits": 4
1577
+ },
1578
+ "model.layers.1.mlp.down_proj": {
1579
+ "group_size": 64,
1580
+ "bits": 8
1581
+ },
1582
+ "model.layers.1.mlp.up_proj": {
1583
+ "group_size": 64,
1584
+ "bits": 4
1585
+ },
1586
+ "model.layers.1.input_layernorm": false,
1587
+ "model.layers.1.post_attention_layernorm": false,
1588
+ "model.layers.2.self_attn.q_proj": {
1589
+ "group_size": 64,
1590
+ "bits": 4
1591
+ },
1592
+ "model.layers.2.self_attn.k_proj": {
1593
+ "group_size": 64,
1594
+ "bits": 4
1595
+ },
1596
+ "model.layers.2.self_attn.v_proj": {
1597
+ "group_size": 64,
1598
+ "bits": 8
1599
+ },
1600
+ "model.layers.2.self_attn.o_proj": {
1601
+ "group_size": 64,
1602
+ "bits": 4
1603
+ },
1604
+ "model.layers.2.self_attn.rope": false,
1605
+ "model.layers.2.mlp.gate_proj": {
1606
+ "group_size": 64,
1607
+ "bits": 4
1608
+ },
1609
+ "model.layers.2.mlp.down_proj": {
1610
+ "group_size": 64,
1611
+ "bits": 8
1612
+ },
1613
+ "model.layers.2.mlp.up_proj": {
1614
+ "group_size": 64,
1615
+ "bits": 4
1616
+ },
1617
+ "model.layers.2.input_layernorm": false,
1618
+ "model.layers.2.post_attention_layernorm": false,
1619
+ "model.layers.3.self_attn.q_proj": {
1620
+ "group_size": 64,
1621
+ "bits": 4
1622
+ },
1623
+ "model.layers.3.self_attn.k_proj": {
1624
+ "group_size": 64,
1625
+ "bits": 4
1626
+ },
1627
+ "model.layers.3.self_attn.v_proj": {
1628
+ "group_size": 64,
1629
+ "bits": 8
1630
+ },
1631
+ "model.layers.3.self_attn.o_proj": {
1632
+ "group_size": 64,
1633
+ "bits": 4
1634
+ },
1635
+ "model.layers.3.self_attn.rope": false,
1636
+ "model.layers.3.mlp.gate_proj": {
1637
+ "group_size": 64,
1638
+ "bits": 4
1639
+ },
1640
+ "model.layers.3.mlp.down_proj": {
1641
+ "group_size": 64,
1642
+ "bits": 8
1643
+ },
1644
+ "model.layers.3.mlp.up_proj": {
1645
+ "group_size": 64,
1646
+ "bits": 4
1647
+ },
1648
+ "model.layers.3.input_layernorm": false,
1649
+ "model.layers.3.post_attention_layernorm": false,
1650
+ "model.layers.4.self_attn.q_proj": {
1651
+ "group_size": 64,
1652
+ "bits": 4
1653
+ },
1654
+ "model.layers.4.self_attn.k_proj": {
1655
+ "group_size": 64,
1656
+ "bits": 4
1657
+ },
1658
+ "model.layers.4.self_attn.v_proj": {
1659
+ "group_size": 64,
1660
+ "bits": 8
1661
+ },
1662
+ "model.layers.4.self_attn.o_proj": {
1663
+ "group_size": 64,
1664
+ "bits": 4
1665
+ },
1666
+ "model.layers.4.self_attn.rope": false,
1667
+ "model.layers.4.mlp.gate_proj": {
1668
+ "group_size": 64,
1669
+ "bits": 4
1670
+ },
1671
+ "model.layers.4.mlp.down_proj": {
1672
+ "group_size": 64,
1673
+ "bits": 8
1674
+ },
1675
+ "model.layers.4.mlp.up_proj": {
1676
+ "group_size": 64,
1677
+ "bits": 4
1678
+ },
1679
+ "model.layers.4.input_layernorm": false,
1680
+ "model.layers.4.post_attention_layernorm": false,
1681
+ "model.layers.5.self_attn.q_proj": {
1682
+ "group_size": 64,
1683
+ "bits": 4
1684
+ },
1685
+ "model.layers.5.self_attn.k_proj": {
1686
+ "group_size": 64,
1687
+ "bits": 4
1688
+ },
1689
+ "model.layers.5.self_attn.v_proj": {
1690
+ "group_size": 64,
1691
+ "bits": 8
1692
+ },
1693
+ "model.layers.5.self_attn.o_proj": {
1694
+ "group_size": 64,
1695
+ "bits": 4
1696
+ },
1697
+ "model.layers.5.self_attn.rope": false,
1698
+ "model.layers.5.mlp.gate_proj": {
1699
+ "group_size": 64,
1700
+ "bits": 4
1701
+ },
1702
+ "model.layers.5.mlp.down_proj": {
1703
+ "group_size": 64,
1704
+ "bits": 8
1705
+ },
1706
+ "model.layers.5.mlp.up_proj": {
1707
+ "group_size": 64,
1708
+ "bits": 4
1709
+ },
1710
+ "model.layers.5.input_layernorm": false,
1711
+ "model.layers.5.post_attention_layernorm": false,
1712
+ "model.layers.6.self_attn.q_proj": {
1713
+ "group_size": 64,
1714
+ "bits": 4
1715
+ },
1716
+ "model.layers.6.self_attn.k_proj": {
1717
+ "group_size": 64,
1718
+ "bits": 4
1719
+ },
1720
+ "model.layers.6.self_attn.v_proj": {
1721
+ "group_size": 64,
1722
+ "bits": 4
1723
+ },
1724
+ "model.layers.6.self_attn.o_proj": {
1725
+ "group_size": 64,
1726
+ "bits": 4
1727
+ },
1728
+ "model.layers.6.self_attn.rope": false,
1729
+ "model.layers.6.mlp.gate_proj": {
1730
+ "group_size": 64,
1731
+ "bits": 4
1732
+ },
1733
+ "model.layers.6.mlp.down_proj": {
1734
+ "group_size": 64,
1735
+ "bits": 4
1736
+ },
1737
+ "model.layers.6.mlp.up_proj": {
1738
+ "group_size": 64,
1739
+ "bits": 4
1740
+ },
1741
+ "model.layers.6.input_layernorm": false,
1742
+ "model.layers.6.post_attention_layernorm": false,
1743
+ "model.layers.7.self_attn.q_proj": {
1744
+ "group_size": 64,
1745
+ "bits": 4
1746
+ },
1747
+ "model.layers.7.self_attn.k_proj": {
1748
+ "group_size": 64,
1749
+ "bits": 4
1750
+ },
1751
+ "model.layers.7.self_attn.v_proj": {
1752
+ "group_size": 64,
1753
+ "bits": 4
1754
+ },
1755
+ "model.layers.7.self_attn.o_proj": {
1756
+ "group_size": 64,
1757
+ "bits": 4
1758
+ },
1759
+ "model.layers.7.self_attn.rope": false,
1760
+ "model.layers.7.mlp.gate_proj": {
1761
+ "group_size": 64,
1762
+ "bits": 4
1763
+ },
1764
+ "model.layers.7.mlp.down_proj": {
1765
+ "group_size": 64,
1766
+ "bits": 4
1767
+ },
1768
+ "model.layers.7.mlp.up_proj": {
1769
+ "group_size": 64,
1770
+ "bits": 4
1771
+ },
1772
+ "model.layers.7.input_layernorm": false,
1773
+ "model.layers.7.post_attention_layernorm": false,
1774
+ "model.layers.8.self_attn.q_proj": {
1775
+ "group_size": 64,
1776
+ "bits": 4
1777
+ },
1778
+ "model.layers.8.self_attn.k_proj": {
1779
+ "group_size": 64,
1780
+ "bits": 4
1781
+ },
1782
+ "model.layers.8.self_attn.v_proj": {
1783
+ "group_size": 64,
1784
+ "bits": 8
1785
+ },
1786
+ "model.layers.8.self_attn.o_proj": {
1787
+ "group_size": 64,
1788
+ "bits": 4
1789
+ },
1790
+ "model.layers.8.self_attn.rope": false,
1791
+ "model.layers.8.mlp.gate_proj": {
1792
+ "group_size": 64,
1793
+ "bits": 4
1794
+ },
1795
+ "model.layers.8.mlp.down_proj": {
1796
+ "group_size": 64,
1797
+ "bits": 8
1798
+ },
1799
+ "model.layers.8.mlp.up_proj": {
1800
+ "group_size": 64,
1801
+ "bits": 4
1802
+ },
1803
+ "model.layers.8.input_layernorm": false,
1804
+ "model.layers.8.post_attention_layernorm": false,
1805
+ "model.layers.9.self_attn.q_proj": {
1806
+ "group_size": 64,
1807
+ "bits": 4
1808
+ },
1809
+ "model.layers.9.self_attn.k_proj": {
1810
+ "group_size": 64,
1811
+ "bits": 4
1812
+ },
1813
+ "model.layers.9.self_attn.v_proj": {
1814
+ "group_size": 64,
1815
+ "bits": 4
1816
+ },
1817
+ "model.layers.9.self_attn.o_proj": {
1818
+ "group_size": 64,
1819
+ "bits": 4
1820
+ },
1821
+ "model.layers.9.self_attn.rope": false,
1822
+ "model.layers.9.mlp.gate_proj": {
1823
+ "group_size": 64,
1824
+ "bits": 4
1825
+ },
1826
+ "model.layers.9.mlp.down_proj": {
1827
+ "group_size": 64,
1828
+ "bits": 4
1829
+ },
1830
+ "model.layers.9.mlp.up_proj": {
1831
+ "group_size": 64,
1832
+ "bits": 4
1833
+ },
1834
+ "model.layers.9.input_layernorm": false,
1835
+ "model.layers.9.post_attention_layernorm": false,
1836
+ "model.layers.10.self_attn.q_proj": {
1837
+ "group_size": 64,
1838
+ "bits": 4
1839
+ },
1840
+ "model.layers.10.self_attn.k_proj": {
1841
+ "group_size": 64,
1842
+ "bits": 4
1843
+ },
1844
+ "model.layers.10.self_attn.v_proj": {
1845
+ "group_size": 64,
1846
+ "bits": 4
1847
+ },
1848
+ "model.layers.10.self_attn.o_proj": {
1849
+ "group_size": 64,
1850
+ "bits": 4
1851
+ },
1852
+ "model.layers.10.self_attn.rope": false,
1853
+ "model.layers.10.mlp.gate_proj": {
1854
+ "group_size": 64,
1855
+ "bits": 4
1856
+ },
1857
+ "model.layers.10.mlp.down_proj": {
1858
+ "group_size": 64,
1859
+ "bits": 4
1860
+ },
1861
+ "model.layers.10.mlp.up_proj": {
1862
+ "group_size": 64,
1863
+ "bits": 4
1864
+ },
1865
+ "model.layers.10.input_layernorm": false,
1866
+ "model.layers.10.post_attention_layernorm": false,
1867
+ "model.layers.11.self_attn.q_proj": {
1868
+ "group_size": 64,
1869
+ "bits": 4
1870
+ },
1871
+ "model.layers.11.self_attn.k_proj": {
1872
+ "group_size": 64,
1873
+ "bits": 4
1874
+ },
1875
+ "model.layers.11.self_attn.v_proj": {
1876
+ "group_size": 64,
1877
+ "bits": 8
1878
+ },
1879
+ "model.layers.11.self_attn.o_proj": {
1880
+ "group_size": 64,
1881
+ "bits": 4
1882
+ },
1883
+ "model.layers.11.self_attn.rope": false,
1884
+ "model.layers.11.mlp.gate_proj": {
1885
+ "group_size": 64,
1886
+ "bits": 4
1887
+ },
1888
+ "model.layers.11.mlp.down_proj": {
1889
+ "group_size": 64,
1890
+ "bits": 8
1891
+ },
1892
+ "model.layers.11.mlp.up_proj": {
1893
+ "group_size": 64,
1894
+ "bits": 4
1895
+ },
1896
+ "model.layers.11.input_layernorm": false,
1897
+ "model.layers.11.post_attention_layernorm": false,
1898
+ "model.layers.12.self_attn.q_proj": {
1899
+ "group_size": 64,
1900
+ "bits": 4
1901
+ },
1902
+ "model.layers.12.self_attn.k_proj": {
1903
+ "group_size": 64,
1904
+ "bits": 4
1905
+ },
1906
+ "model.layers.12.self_attn.v_proj": {
1907
+ "group_size": 64,
1908
+ "bits": 4
1909
+ },
1910
+ "model.layers.12.self_attn.o_proj": {
1911
+ "group_size": 64,
1912
+ "bits": 4
1913
+ },
1914
+ "model.layers.12.self_attn.rope": false,
1915
+ "model.layers.12.mlp.gate_proj": {
1916
+ "group_size": 64,
1917
+ "bits": 4
1918
+ },
1919
+ "model.layers.12.mlp.down_proj": {
1920
+ "group_size": 64,
1921
+ "bits": 4
1922
+ },
1923
+ "model.layers.12.mlp.up_proj": {
1924
+ "group_size": 64,
1925
+ "bits": 4
1926
+ },
1927
+ "model.layers.12.input_layernorm": false,
1928
+ "model.layers.12.post_attention_layernorm": false,
1929
+ "model.layers.13.self_attn.q_proj": {
1930
+ "group_size": 64,
1931
+ "bits": 4
1932
+ },
1933
+ "model.layers.13.self_attn.k_proj": {
1934
+ "group_size": 64,
1935
+ "bits": 4
1936
+ },
1937
+ "model.layers.13.self_attn.v_proj": {
1938
+ "group_size": 64,
1939
+ "bits": 4
1940
+ },
1941
+ "model.layers.13.self_attn.o_proj": {
1942
+ "group_size": 64,
1943
+ "bits": 4
1944
+ },
1945
+ "model.layers.13.self_attn.rope": false,
1946
+ "model.layers.13.mlp.gate_proj": {
1947
+ "group_size": 64,
1948
+ "bits": 4
1949
+ },
1950
+ "model.layers.13.mlp.down_proj": {
1951
+ "group_size": 64,
1952
+ "bits": 4
1953
+ },
1954
+ "model.layers.13.mlp.up_proj": {
1955
+ "group_size": 64,
1956
+ "bits": 4
1957
+ },
1958
+ "model.layers.13.input_layernorm": false,
1959
+ "model.layers.13.post_attention_layernorm": false,
1960
+ "model.layers.14.self_attn.q_proj": {
1961
+ "group_size": 64,
1962
+ "bits": 4
1963
+ },
1964
+ "model.layers.14.self_attn.k_proj": {
1965
+ "group_size": 64,
1966
+ "bits": 4
1967
+ },
1968
+ "model.layers.14.self_attn.v_proj": {
1969
+ "group_size": 64,
1970
+ "bits": 8
1971
+ },
1972
+ "model.layers.14.self_attn.o_proj": {
1973
+ "group_size": 64,
1974
+ "bits": 4
1975
+ },
1976
+ "model.layers.14.self_attn.rope": false,
1977
+ "model.layers.14.mlp.gate_proj": {
1978
+ "group_size": 64,
1979
+ "bits": 4
1980
+ },
1981
+ "model.layers.14.mlp.down_proj": {
1982
+ "group_size": 64,
1983
+ "bits": 8
1984
+ },
1985
+ "model.layers.14.mlp.up_proj": {
1986
+ "group_size": 64,
1987
+ "bits": 4
1988
+ },
1989
+ "model.layers.14.input_layernorm": false,
1990
+ "model.layers.14.post_attention_layernorm": false,
1991
+ "model.layers.15.self_attn.q_proj": {
1992
+ "group_size": 64,
1993
+ "bits": 4
1994
+ },
1995
+ "model.layers.15.self_attn.k_proj": {
1996
+ "group_size": 64,
1997
+ "bits": 4
1998
+ },
1999
+ "model.layers.15.self_attn.v_proj": {
2000
+ "group_size": 64,
2001
+ "bits": 4
2002
+ },
2003
+ "model.layers.15.self_attn.o_proj": {
2004
+ "group_size": 64,
2005
+ "bits": 4
2006
+ },
2007
+ "model.layers.15.self_attn.rope": false,
2008
+ "model.layers.15.mlp.gate_proj": {
2009
+ "group_size": 64,
2010
+ "bits": 4
2011
+ },
2012
+ "model.layers.15.mlp.down_proj": {
2013
+ "group_size": 64,
2014
+ "bits": 4
2015
+ },
2016
+ "model.layers.15.mlp.up_proj": {
2017
+ "group_size": 64,
2018
+ "bits": 4
2019
+ },
2020
+ "model.layers.15.input_layernorm": false,
2021
+ "model.layers.15.post_attention_layernorm": false,
2022
+ "model.layers.16.self_attn.q_proj": {
2023
+ "group_size": 64,
2024
+ "bits": 4
2025
+ },
2026
+ "model.layers.16.self_attn.k_proj": {
2027
+ "group_size": 64,
2028
+ "bits": 4
2029
+ },
2030
+ "model.layers.16.self_attn.v_proj": {
2031
+ "group_size": 64,
2032
+ "bits": 4
2033
+ },
2034
+ "model.layers.16.self_attn.o_proj": {
2035
+ "group_size": 64,
2036
+ "bits": 4
2037
+ },
2038
+ "model.layers.16.self_attn.rope": false,
2039
+ "model.layers.16.mlp.gate_proj": {
2040
+ "group_size": 64,
2041
+ "bits": 4
2042
+ },
2043
+ "model.layers.16.mlp.down_proj": {
2044
+ "group_size": 64,
2045
+ "bits": 4
2046
+ },
2047
+ "model.layers.16.mlp.up_proj": {
2048
+ "group_size": 64,
2049
+ "bits": 4
2050
+ },
2051
+ "model.layers.16.input_layernorm": false,
2052
+ "model.layers.16.post_attention_layernorm": false,
2053
+ "model.layers.17.self_attn.q_proj": {
2054
+ "group_size": 64,
2055
+ "bits": 4
2056
+ },
2057
+ "model.layers.17.self_attn.k_proj": {
2058
+ "group_size": 64,
2059
+ "bits": 4
2060
+ },
2061
+ "model.layers.17.self_attn.v_proj": {
2062
+ "group_size": 64,
2063
+ "bits": 8
2064
+ },
2065
+ "model.layers.17.self_attn.o_proj": {
2066
+ "group_size": 64,
2067
+ "bits": 4
2068
+ },
2069
+ "model.layers.17.self_attn.rope": false,
2070
+ "model.layers.17.mlp.gate_proj": {
2071
+ "group_size": 64,
2072
+ "bits": 4
2073
+ },
2074
+ "model.layers.17.mlp.down_proj": {
2075
+ "group_size": 64,
2076
+ "bits": 8
2077
+ },
2078
+ "model.layers.17.mlp.up_proj": {
2079
+ "group_size": 64,
2080
+ "bits": 4
2081
+ },
2082
+ "model.layers.17.input_layernorm": false,
2083
+ "model.layers.17.post_attention_layernorm": false,
2084
+ "model.layers.18.self_attn.q_proj": {
2085
+ "group_size": 64,
2086
+ "bits": 4
2087
+ },
2088
+ "model.layers.18.self_attn.k_proj": {
2089
+ "group_size": 64,
2090
+ "bits": 4
2091
+ },
2092
+ "model.layers.18.self_attn.v_proj": {
2093
+ "group_size": 64,
2094
+ "bits": 4
2095
+ },
2096
+ "model.layers.18.self_attn.o_proj": {
2097
+ "group_size": 64,
2098
+ "bits": 4
2099
+ },
2100
+ "model.layers.18.self_attn.rope": false,
2101
+ "model.layers.18.mlp.gate_proj": {
2102
+ "group_size": 64,
2103
+ "bits": 4
2104
+ },
2105
+ "model.layers.18.mlp.down_proj": {
2106
+ "group_size": 64,
2107
+ "bits": 4
2108
+ },
2109
+ "model.layers.18.mlp.up_proj": {
2110
+ "group_size": 64,
2111
+ "bits": 4
2112
+ },
2113
+ "model.layers.18.input_layernorm": false,
2114
+ "model.layers.18.post_attention_layernorm": false,
2115
+ "model.layers.19.self_attn.q_proj": {
2116
+ "group_size": 64,
2117
+ "bits": 4
2118
+ },
2119
+ "model.layers.19.self_attn.k_proj": {
2120
+ "group_size": 64,
2121
+ "bits": 4
2122
+ },
2123
+ "model.layers.19.self_attn.v_proj": {
2124
+ "group_size": 64,
2125
+ "bits": 4
2126
+ },
2127
+ "model.layers.19.self_attn.o_proj": {
2128
+ "group_size": 64,
2129
+ "bits": 4
2130
+ },
2131
+ "model.layers.19.self_attn.rope": false,
2132
+ "model.layers.19.mlp.gate_proj": {
2133
+ "group_size": 64,
2134
+ "bits": 4
2135
+ },
2136
+ "model.layers.19.mlp.down_proj": {
2137
+ "group_size": 64,
2138
+ "bits": 4
2139
+ },
2140
+ "model.layers.19.mlp.up_proj": {
2141
+ "group_size": 64,
2142
+ "bits": 4
2143
+ },
2144
+ "model.layers.19.input_layernorm": false,
2145
+ "model.layers.19.post_attention_layernorm": false,
2146
+ "model.layers.20.self_attn.q_proj": {
2147
+ "group_size": 64,
2148
+ "bits": 4
2149
+ },
2150
+ "model.layers.20.self_attn.k_proj": {
2151
+ "group_size": 64,
2152
+ "bits": 4
2153
+ },
2154
+ "model.layers.20.self_attn.v_proj": {
2155
+ "group_size": 64,
2156
+ "bits": 8
2157
+ },
2158
+ "model.layers.20.self_attn.o_proj": {
2159
+ "group_size": 64,
2160
+ "bits": 4
2161
+ },
2162
+ "model.layers.20.self_attn.rope": false,
2163
+ "model.layers.20.mlp.gate_proj": {
2164
+ "group_size": 64,
2165
+ "bits": 4
2166
+ },
2167
+ "model.layers.20.mlp.down_proj": {
2168
+ "group_size": 64,
2169
+ "bits": 8
2170
+ },
2171
+ "model.layers.20.mlp.up_proj": {
2172
+ "group_size": 64,
2173
+ "bits": 4
2174
+ },
2175
+ "model.layers.20.input_layernorm": false,
2176
+ "model.layers.20.post_attention_layernorm": false,
2177
+ "model.layers.21.self_attn.q_proj": {
2178
+ "group_size": 64,
2179
+ "bits": 4
2180
+ },
2181
+ "model.layers.21.self_attn.k_proj": {
2182
+ "group_size": 64,
2183
+ "bits": 4
2184
+ },
2185
+ "model.layers.21.self_attn.v_proj": {
2186
+ "group_size": 64,
2187
+ "bits": 4
2188
+ },
2189
+ "model.layers.21.self_attn.o_proj": {
2190
+ "group_size": 64,
2191
+ "bits": 4
2192
+ },
2193
+ "model.layers.21.self_attn.rope": false,
2194
+ "model.layers.21.mlp.gate_proj": {
2195
+ "group_size": 64,
2196
+ "bits": 4
2197
+ },
2198
+ "model.layers.21.mlp.down_proj": {
2199
+ "group_size": 64,
2200
+ "bits": 4
2201
+ },
2202
+ "model.layers.21.mlp.up_proj": {
2203
+ "group_size": 64,
2204
+ "bits": 4
2205
+ },
2206
+ "model.layers.21.input_layernorm": false,
2207
+ "model.layers.21.post_attention_layernorm": false,
2208
+ "model.layers.22.self_attn.q_proj": {
2209
+ "group_size": 64,
2210
+ "bits": 4
2211
+ },
2212
+ "model.layers.22.self_attn.k_proj": {
2213
+ "group_size": 64,
2214
+ "bits": 4
2215
+ },
2216
+ "model.layers.22.self_attn.v_proj": {
2217
+ "group_size": 64,
2218
+ "bits": 4
2219
+ },
2220
+ "model.layers.22.self_attn.o_proj": {
2221
+ "group_size": 64,
2222
+ "bits": 4
2223
+ },
2224
+ "model.layers.22.self_attn.rope": false,
2225
+ "model.layers.22.mlp.gate_proj": {
2226
+ "group_size": 64,
2227
+ "bits": 4
2228
+ },
2229
+ "model.layers.22.mlp.down_proj": {
2230
+ "group_size": 64,
2231
+ "bits": 4
2232
+ },
2233
+ "model.layers.22.mlp.up_proj": {
2234
+ "group_size": 64,
2235
+ "bits": 4
2236
+ },
2237
+ "model.layers.22.input_layernorm": false,
2238
+ "model.layers.22.post_attention_layernorm": false,
2239
+ "model.layers.23.self_attn.q_proj": {
2240
+ "group_size": 64,
2241
+ "bits": 4
2242
+ },
2243
+ "model.layers.23.self_attn.k_proj": {
2244
+ "group_size": 64,
2245
+ "bits": 4
2246
+ },
2247
+ "model.layers.23.self_attn.v_proj": {
2248
+ "group_size": 64,
2249
+ "bits": 8
2250
+ },
2251
+ "model.layers.23.self_attn.o_proj": {
2252
+ "group_size": 64,
2253
+ "bits": 4
2254
+ },
2255
+ "model.layers.23.self_attn.rope": false,
2256
+ "model.layers.23.mlp.gate_proj": {
2257
+ "group_size": 64,
2258
+ "bits": 4
2259
+ },
2260
+ "model.layers.23.mlp.down_proj": {
2261
+ "group_size": 64,
2262
+ "bits": 8
2263
+ },
2264
+ "model.layers.23.mlp.up_proj": {
2265
+ "group_size": 64,
2266
+ "bits": 4
2267
+ },
2268
+ "model.layers.23.input_layernorm": false,
2269
+ "model.layers.23.post_attention_layernorm": false,
2270
+ "model.layers.24.self_attn.q_proj": {
2271
+ "group_size": 64,
2272
+ "bits": 4
2273
+ },
2274
+ "model.layers.24.self_attn.k_proj": {
2275
+ "group_size": 64,
2276
+ "bits": 4
2277
+ },
2278
+ "model.layers.24.self_attn.v_proj": {
2279
+ "group_size": 64,
2280
+ "bits": 4
2281
+ },
2282
+ "model.layers.24.self_attn.o_proj": {
2283
+ "group_size": 64,
2284
+ "bits": 4
2285
+ },
2286
+ "model.layers.24.self_attn.rope": false,
2287
+ "model.layers.24.mlp.gate_proj": {
2288
+ "group_size": 64,
2289
+ "bits": 4
2290
+ },
2291
+ "model.layers.24.mlp.down_proj": {
2292
+ "group_size": 64,
2293
+ "bits": 4
2294
+ },
2295
+ "model.layers.24.mlp.up_proj": {
2296
+ "group_size": 64,
2297
+ "bits": 4
2298
+ },
2299
+ "model.layers.24.input_layernorm": false,
2300
+ "model.layers.24.post_attention_layernorm": false,
2301
+ "model.layers.25.self_attn.q_proj": {
2302
+ "group_size": 64,
2303
+ "bits": 4
2304
+ },
2305
+ "model.layers.25.self_attn.k_proj": {
2306
+ "group_size": 64,
2307
+ "bits": 4
2308
+ },
2309
+ "model.layers.25.self_attn.v_proj": {
2310
+ "group_size": 64,
2311
+ "bits": 4
2312
+ },
2313
+ "model.layers.25.self_attn.o_proj": {
2314
+ "group_size": 64,
2315
+ "bits": 4
2316
+ },
2317
+ "model.layers.25.self_attn.rope": false,
2318
+ "model.layers.25.mlp.gate_proj": {
2319
+ "group_size": 64,
2320
+ "bits": 4
2321
+ },
2322
+ "model.layers.25.mlp.down_proj": {
2323
+ "group_size": 64,
2324
+ "bits": 4
2325
+ },
2326
+ "model.layers.25.mlp.up_proj": {
2327
+ "group_size": 64,
2328
+ "bits": 4
2329
+ },
2330
+ "model.layers.25.input_layernorm": false,
2331
+ "model.layers.25.post_attention_layernorm": false,
2332
+ "model.layers.26.self_attn.q_proj": {
2333
+ "group_size": 64,
2334
+ "bits": 4
2335
+ },
2336
+ "model.layers.26.self_attn.k_proj": {
2337
+ "group_size": 64,
2338
+ "bits": 4
2339
+ },
2340
+ "model.layers.26.self_attn.v_proj": {
2341
+ "group_size": 64,
2342
+ "bits": 8
2343
+ },
2344
+ "model.layers.26.self_attn.o_proj": {
2345
+ "group_size": 64,
2346
+ "bits": 4
2347
+ },
2348
+ "model.layers.26.self_attn.rope": false,
2349
+ "model.layers.26.mlp.gate_proj": {
2350
+ "group_size": 64,
2351
+ "bits": 4
2352
+ },
2353
+ "model.layers.26.mlp.down_proj": {
2354
+ "group_size": 64,
2355
+ "bits": 8
2356
+ },
2357
+ "model.layers.26.mlp.up_proj": {
2358
+ "group_size": 64,
2359
+ "bits": 4
2360
+ },
2361
+ "model.layers.26.input_layernorm": false,
2362
+ "model.layers.26.post_attention_layernorm": false,
2363
+ "model.layers.27.self_attn.q_proj": {
2364
+ "group_size": 64,
2365
+ "bits": 4
2366
+ },
2367
+ "model.layers.27.self_attn.k_proj": {
2368
+ "group_size": 64,
2369
+ "bits": 4
2370
+ },
2371
+ "model.layers.27.self_attn.v_proj": {
2372
+ "group_size": 64,
2373
+ "bits": 4
2374
+ },
2375
+ "model.layers.27.self_attn.o_proj": {
2376
+ "group_size": 64,
2377
+ "bits": 4
2378
+ },
2379
+ "model.layers.27.self_attn.rope": false,
2380
+ "model.layers.27.mlp.gate_proj": {
2381
+ "group_size": 64,
2382
+ "bits": 4
2383
+ },
2384
+ "model.layers.27.mlp.down_proj": {
2385
+ "group_size": 64,
2386
+ "bits": 4
2387
+ },
2388
+ "model.layers.27.mlp.up_proj": {
2389
+ "group_size": 64,
2390
+ "bits": 4
2391
+ },
2392
+ "model.layers.27.input_layernorm": false,
2393
+ "model.layers.27.post_attention_layernorm": false,
2394
+ "model.layers.28.self_attn.q_proj": {
2395
+ "group_size": 64,
2396
+ "bits": 4
2397
+ },
2398
+ "model.layers.28.self_attn.k_proj": {
2399
+ "group_size": 64,
2400
+ "bits": 4
2401
+ },
2402
+ "model.layers.28.self_attn.v_proj": {
2403
+ "group_size": 64,
2404
+ "bits": 4
2405
+ },
2406
+ "model.layers.28.self_attn.o_proj": {
2407
+ "group_size": 64,
2408
+ "bits": 4
2409
+ },
2410
+ "model.layers.28.self_attn.rope": false,
2411
+ "model.layers.28.mlp.gate_proj": {
2412
+ "group_size": 64,
2413
+ "bits": 4
2414
+ },
2415
+ "model.layers.28.mlp.down_proj": {
2416
+ "group_size": 64,
2417
+ "bits": 4
2418
+ },
2419
+ "model.layers.28.mlp.up_proj": {
2420
+ "group_size": 64,
2421
+ "bits": 4
2422
+ },
2423
+ "model.layers.28.input_layernorm": false,
2424
+ "model.layers.28.post_attention_layernorm": false,
2425
+ "model.layers.29.self_attn.q_proj": {
2426
+ "group_size": 64,
2427
+ "bits": 4
2428
+ },
2429
+ "model.layers.29.self_attn.k_proj": {
2430
+ "group_size": 64,
2431
+ "bits": 4
2432
+ },
2433
+ "model.layers.29.self_attn.v_proj": {
2434
+ "group_size": 64,
2435
+ "bits": 8
2436
+ },
2437
+ "model.layers.29.self_attn.o_proj": {
2438
+ "group_size": 64,
2439
+ "bits": 4
2440
+ },
2441
+ "model.layers.29.self_attn.rope": false,
2442
+ "model.layers.29.mlp.gate_proj": {
2443
+ "group_size": 64,
2444
+ "bits": 4
2445
+ },
2446
+ "model.layers.29.mlp.down_proj": {
2447
+ "group_size": 64,
2448
+ "bits": 8
2449
+ },
2450
+ "model.layers.29.mlp.up_proj": {
2451
+ "group_size": 64,
2452
+ "bits": 4
2453
+ },
2454
+ "model.layers.29.input_layernorm": false,
2455
+ "model.layers.29.post_attention_layernorm": false,
2456
+ "model.layers.30.self_attn.q_proj": {
2457
+ "group_size": 64,
2458
+ "bits": 4
2459
+ },
2460
+ "model.layers.30.self_attn.k_proj": {
2461
+ "group_size": 64,
2462
+ "bits": 4
2463
+ },
2464
+ "model.layers.30.self_attn.v_proj": {
2465
+ "group_size": 64,
2466
+ "bits": 4
2467
+ },
2468
+ "model.layers.30.self_attn.o_proj": {
2469
+ "group_size": 64,
2470
+ "bits": 4
2471
+ },
2472
+ "model.layers.30.self_attn.rope": false,
2473
+ "model.layers.30.mlp.gate_proj": {
2474
+ "group_size": 64,
2475
+ "bits": 4
2476
+ },
2477
+ "model.layers.30.mlp.down_proj": {
2478
+ "group_size": 64,
2479
+ "bits": 4
2480
+ },
2481
+ "model.layers.30.mlp.up_proj": {
2482
+ "group_size": 64,
2483
+ "bits": 4
2484
+ },
2485
+ "model.layers.30.input_layernorm": false,
2486
+ "model.layers.30.post_attention_layernorm": false,
2487
+ "model.layers.31.self_attn.q_proj": {
2488
+ "group_size": 64,
2489
+ "bits": 4
2490
+ },
2491
+ "model.layers.31.self_attn.k_proj": {
2492
+ "group_size": 64,
2493
+ "bits": 4
2494
+ },
2495
+ "model.layers.31.self_attn.v_proj": {
2496
+ "group_size": 64,
2497
+ "bits": 4
2498
+ },
2499
+ "model.layers.31.self_attn.o_proj": {
2500
+ "group_size": 64,
2501
+ "bits": 4
2502
+ },
2503
+ "model.layers.31.self_attn.rope": false,
2504
+ "model.layers.31.mlp.gate_proj": {
2505
+ "group_size": 64,
2506
+ "bits": 4
2507
+ },
2508
+ "model.layers.31.mlp.down_proj": {
2509
+ "group_size": 64,
2510
+ "bits": 4
2511
+ },
2512
+ "model.layers.31.mlp.up_proj": {
2513
+ "group_size": 64,
2514
+ "bits": 4
2515
+ },
2516
+ "model.layers.31.input_layernorm": false,
2517
+ "model.layers.31.post_attention_layernorm": false,
2518
+ "model.layers.32.self_attn.q_proj": {
2519
+ "group_size": 64,
2520
+ "bits": 4
2521
+ },
2522
+ "model.layers.32.self_attn.k_proj": {
2523
+ "group_size": 64,
2524
+ "bits": 4
2525
+ },
2526
+ "model.layers.32.self_attn.v_proj": {
2527
+ "group_size": 64,
2528
+ "bits": 8
2529
+ },
2530
+ "model.layers.32.self_attn.o_proj": {
2531
+ "group_size": 64,
2532
+ "bits": 4
2533
+ },
2534
+ "model.layers.32.self_attn.rope": false,
2535
+ "model.layers.32.mlp.gate_proj": {
2536
+ "group_size": 64,
2537
+ "bits": 4
2538
+ },
2539
+ "model.layers.32.mlp.down_proj": {
2540
+ "group_size": 64,
2541
+ "bits": 8
2542
+ },
2543
+ "model.layers.32.mlp.up_proj": {
2544
+ "group_size": 64,
2545
+ "bits": 4
2546
+ },
2547
+ "model.layers.32.input_layernorm": false,
2548
+ "model.layers.32.post_attention_layernorm": false,
2549
+ "model.layers.33.self_attn.q_proj": {
2550
+ "group_size": 64,
2551
+ "bits": 4
2552
+ },
2553
+ "model.layers.33.self_attn.k_proj": {
2554
+ "group_size": 64,
2555
+ "bits": 4
2556
+ },
2557
+ "model.layers.33.self_attn.v_proj": {
2558
+ "group_size": 64,
2559
+ "bits": 4
2560
+ },
2561
+ "model.layers.33.self_attn.o_proj": {
2562
+ "group_size": 64,
2563
+ "bits": 4
2564
+ },
2565
+ "model.layers.33.self_attn.rope": false,
2566
+ "model.layers.33.mlp.gate_proj": {
2567
+ "group_size": 64,
2568
+ "bits": 4
2569
+ },
2570
+ "model.layers.33.mlp.down_proj": {
2571
+ "group_size": 64,
2572
+ "bits": 4
2573
+ },
2574
+ "model.layers.33.mlp.up_proj": {
2575
+ "group_size": 64,
2576
+ "bits": 4
2577
+ },
2578
+ "model.layers.33.input_layernorm": false,
2579
+ "model.layers.33.post_attention_layernorm": false,
2580
+ "model.layers.34.self_attn.q_proj": {
2581
+ "group_size": 64,
2582
+ "bits": 4
2583
+ },
2584
+ "model.layers.34.self_attn.k_proj": {
2585
+ "group_size": 64,
2586
+ "bits": 4
2587
+ },
2588
+ "model.layers.34.self_attn.v_proj": {
2589
+ "group_size": 64,
2590
+ "bits": 4
2591
+ },
2592
+ "model.layers.34.self_attn.o_proj": {
2593
+ "group_size": 64,
2594
+ "bits": 4
2595
+ },
2596
+ "model.layers.34.self_attn.rope": false,
2597
+ "model.layers.34.mlp.gate_proj": {
2598
+ "group_size": 64,
2599
+ "bits": 4
2600
+ },
2601
+ "model.layers.34.mlp.down_proj": {
2602
+ "group_size": 64,
2603
+ "bits": 4
2604
+ },
2605
+ "model.layers.34.mlp.up_proj": {
2606
+ "group_size": 64,
2607
+ "bits": 4
2608
+ },
2609
+ "model.layers.34.input_layernorm": false,
2610
+ "model.layers.34.post_attention_layernorm": false,
2611
+ "model.layers.35.self_attn.q_proj": {
2612
+ "group_size": 64,
2613
+ "bits": 4
2614
+ },
2615
+ "model.layers.35.self_attn.k_proj": {
2616
+ "group_size": 64,
2617
+ "bits": 4
2618
+ },
2619
+ "model.layers.35.self_attn.v_proj": {
2620
+ "group_size": 64,
2621
+ "bits": 8
2622
+ },
2623
+ "model.layers.35.self_attn.o_proj": {
2624
+ "group_size": 64,
2625
+ "bits": 4
2626
+ },
2627
+ "model.layers.35.self_attn.rope": false,
2628
+ "model.layers.35.mlp.gate_proj": {
2629
+ "group_size": 64,
2630
+ "bits": 4
2631
+ },
2632
+ "model.layers.35.mlp.down_proj": {
2633
+ "group_size": 64,
2634
+ "bits": 8
2635
+ },
2636
+ "model.layers.35.mlp.up_proj": {
2637
+ "group_size": 64,
2638
+ "bits": 4
2639
+ },
2640
+ "model.layers.35.input_layernorm": false,
2641
+ "model.layers.35.post_attention_layernorm": false,
2642
+ "model.layers.36.self_attn.q_proj": {
2643
+ "group_size": 64,
2644
+ "bits": 4
2645
+ },
2646
+ "model.layers.36.self_attn.k_proj": {
2647
+ "group_size": 64,
2648
+ "bits": 4
2649
+ },
2650
+ "model.layers.36.self_attn.v_proj": {
2651
+ "group_size": 64,
2652
+ "bits": 4
2653
+ },
2654
+ "model.layers.36.self_attn.o_proj": {
2655
+ "group_size": 64,
2656
+ "bits": 4
2657
+ },
2658
+ "model.layers.36.self_attn.rope": false,
2659
+ "model.layers.36.mlp.gate_proj": {
2660
+ "group_size": 64,
2661
+ "bits": 4
2662
+ },
2663
+ "model.layers.36.mlp.down_proj": {
2664
+ "group_size": 64,
2665
+ "bits": 4
2666
+ },
2667
+ "model.layers.36.mlp.up_proj": {
2668
+ "group_size": 64,
2669
+ "bits": 4
2670
+ },
2671
+ "model.layers.36.input_layernorm": false,
2672
+ "model.layers.36.post_attention_layernorm": false,
2673
+ "model.layers.37.self_attn.q_proj": {
2674
+ "group_size": 64,
2675
+ "bits": 4
2676
+ },
2677
+ "model.layers.37.self_attn.k_proj": {
2678
+ "group_size": 64,
2679
+ "bits": 4
2680
+ },
2681
+ "model.layers.37.self_attn.v_proj": {
2682
+ "group_size": 64,
2683
+ "bits": 4
2684
+ },
2685
+ "model.layers.37.self_attn.o_proj": {
2686
+ "group_size": 64,
2687
+ "bits": 4
2688
+ },
2689
+ "model.layers.37.self_attn.rope": false,
2690
+ "model.layers.37.mlp.gate_proj": {
2691
+ "group_size": 64,
2692
+ "bits": 4
2693
+ },
2694
+ "model.layers.37.mlp.down_proj": {
2695
+ "group_size": 64,
2696
+ "bits": 4
2697
+ },
2698
+ "model.layers.37.mlp.up_proj": {
2699
+ "group_size": 64,
2700
+ "bits": 4
2701
+ },
2702
+ "model.layers.37.input_layernorm": false,
2703
+ "model.layers.37.post_attention_layernorm": false,
2704
+ "model.layers.38.self_attn.q_proj": {
2705
+ "group_size": 64,
2706
+ "bits": 4
2707
+ },
2708
+ "model.layers.38.self_attn.k_proj": {
2709
+ "group_size": 64,
2710
+ "bits": 4
2711
+ },
2712
+ "model.layers.38.self_attn.v_proj": {
2713
+ "group_size": 64,
2714
+ "bits": 8
2715
+ },
2716
+ "model.layers.38.self_attn.o_proj": {
2717
+ "group_size": 64,
2718
+ "bits": 4
2719
+ },
2720
+ "model.layers.38.self_attn.rope": false,
2721
+ "model.layers.38.mlp.gate_proj": {
2722
+ "group_size": 64,
2723
+ "bits": 4
2724
+ },
2725
+ "model.layers.38.mlp.down_proj": {
2726
+ "group_size": 64,
2727
+ "bits": 8
2728
+ },
2729
+ "model.layers.38.mlp.up_proj": {
2730
+ "group_size": 64,
2731
+ "bits": 4
2732
+ },
2733
+ "model.layers.38.input_layernorm": false,
2734
+ "model.layers.38.post_attention_layernorm": false,
2735
+ "model.layers.39.self_attn.q_proj": {
2736
+ "group_size": 64,
2737
+ "bits": 4
2738
+ },
2739
+ "model.layers.39.self_attn.k_proj": {
2740
+ "group_size": 64,
2741
+ "bits": 4
2742
+ },
2743
+ "model.layers.39.self_attn.v_proj": {
2744
+ "group_size": 64,
2745
+ "bits": 4
2746
+ },
2747
+ "model.layers.39.self_attn.o_proj": {
2748
+ "group_size": 64,
2749
+ "bits": 4
2750
+ },
2751
+ "model.layers.39.self_attn.rope": false,
2752
+ "model.layers.39.mlp.gate_proj": {
2753
+ "group_size": 64,
2754
+ "bits": 4
2755
+ },
2756
+ "model.layers.39.mlp.down_proj": {
2757
+ "group_size": 64,
2758
+ "bits": 4
2759
+ },
2760
+ "model.layers.39.mlp.up_proj": {
2761
+ "group_size": 64,
2762
+ "bits": 4
2763
+ },
2764
+ "model.layers.39.input_layernorm": false,
2765
+ "model.layers.39.post_attention_layernorm": false,
2766
+ "model.layers.40.self_attn.q_proj": {
2767
+ "group_size": 64,
2768
+ "bits": 4
2769
+ },
2770
+ "model.layers.40.self_attn.k_proj": {
2771
+ "group_size": 64,
2772
+ "bits": 4
2773
+ },
2774
+ "model.layers.40.self_attn.v_proj": {
2775
+ "group_size": 64,
2776
+ "bits": 4
2777
+ },
2778
+ "model.layers.40.self_attn.o_proj": {
2779
+ "group_size": 64,
2780
+ "bits": 4
2781
+ },
2782
+ "model.layers.40.self_attn.rope": false,
2783
+ "model.layers.40.mlp.gate_proj": {
2784
+ "group_size": 64,
2785
+ "bits": 4
2786
+ },
2787
+ "model.layers.40.mlp.down_proj": {
2788
+ "group_size": 64,
2789
+ "bits": 4
2790
+ },
2791
+ "model.layers.40.mlp.up_proj": {
2792
+ "group_size": 64,
2793
+ "bits": 4
2794
+ },
2795
+ "model.layers.40.input_layernorm": false,
2796
+ "model.layers.40.post_attention_layernorm": false,
2797
+ "model.layers.41.self_attn.q_proj": {
2798
+ "group_size": 64,
2799
+ "bits": 4
2800
+ },
2801
+ "model.layers.41.self_attn.k_proj": {
2802
+ "group_size": 64,
2803
+ "bits": 4
2804
+ },
2805
+ "model.layers.41.self_attn.v_proj": {
2806
+ "group_size": 64,
2807
+ "bits": 8
2808
+ },
2809
+ "model.layers.41.self_attn.o_proj": {
2810
+ "group_size": 64,
2811
+ "bits": 4
2812
+ },
2813
+ "model.layers.41.self_attn.rope": false,
2814
+ "model.layers.41.mlp.gate_proj": {
2815
+ "group_size": 64,
2816
+ "bits": 4
2817
+ },
2818
+ "model.layers.41.mlp.down_proj": {
2819
+ "group_size": 64,
2820
+ "bits": 8
2821
+ },
2822
+ "model.layers.41.mlp.up_proj": {
2823
+ "group_size": 64,
2824
+ "bits": 4
2825
+ },
2826
+ "model.layers.41.input_layernorm": false,
2827
+ "model.layers.41.post_attention_layernorm": false,
2828
+ "model.layers.42.self_attn.q_proj": {
2829
+ "group_size": 64,
2830
+ "bits": 4
2831
+ },
2832
+ "model.layers.42.self_attn.k_proj": {
2833
+ "group_size": 64,
2834
+ "bits": 4
2835
+ },
2836
+ "model.layers.42.self_attn.v_proj": {
2837
+ "group_size": 64,
2838
+ "bits": 8
2839
+ },
2840
+ "model.layers.42.self_attn.o_proj": {
2841
+ "group_size": 64,
2842
+ "bits": 4
2843
+ },
2844
+ "model.layers.42.self_attn.rope": false,
2845
+ "model.layers.42.mlp.gate_proj": {
2846
+ "group_size": 64,
2847
+ "bits": 4
2848
+ },
2849
+ "model.layers.42.mlp.down_proj": {
2850
+ "group_size": 64,
2851
+ "bits": 8
2852
+ },
2853
+ "model.layers.42.mlp.up_proj": {
2854
+ "group_size": 64,
2855
+ "bits": 4
2856
+ },
2857
+ "model.layers.42.input_layernorm": false,
2858
+ "model.layers.42.post_attention_layernorm": false,
2859
+ "model.layers.43.self_attn.q_proj": {
2860
+ "group_size": 64,
2861
+ "bits": 4
2862
+ },
2863
+ "model.layers.43.self_attn.k_proj": {
2864
+ "group_size": 64,
2865
+ "bits": 4
2866
+ },
2867
+ "model.layers.43.self_attn.v_proj": {
2868
+ "group_size": 64,
2869
+ "bits": 8
2870
+ },
2871
+ "model.layers.43.self_attn.o_proj": {
2872
+ "group_size": 64,
2873
+ "bits": 4
2874
+ },
2875
+ "model.layers.43.self_attn.rope": false,
2876
+ "model.layers.43.mlp.gate_proj": {
2877
+ "group_size": 64,
2878
+ "bits": 4
2879
+ },
2880
+ "model.layers.43.mlp.down_proj": {
2881
+ "group_size": 64,
2882
+ "bits": 8
2883
+ },
2884
+ "model.layers.43.mlp.up_proj": {
2885
+ "group_size": 64,
2886
+ "bits": 4
2887
+ },
2888
+ "model.layers.43.input_layernorm": false,
2889
+ "model.layers.43.post_attention_layernorm": false,
2890
+ "model.layers.44.self_attn.q_proj": {
2891
+ "group_size": 64,
2892
+ "bits": 4
2893
+ },
2894
+ "model.layers.44.self_attn.k_proj": {
2895
+ "group_size": 64,
2896
+ "bits": 4
2897
+ },
2898
+ "model.layers.44.self_attn.v_proj": {
2899
+ "group_size": 64,
2900
+ "bits": 8
2901
+ },
2902
+ "model.layers.44.self_attn.o_proj": {
2903
+ "group_size": 64,
2904
+ "bits": 4
2905
+ },
2906
+ "model.layers.44.self_attn.rope": false,
2907
+ "model.layers.44.mlp.gate_proj": {
2908
+ "group_size": 64,
2909
+ "bits": 4
2910
+ },
2911
+ "model.layers.44.mlp.down_proj": {
2912
+ "group_size": 64,
2913
+ "bits": 8
2914
+ },
2915
+ "model.layers.44.mlp.up_proj": {
2916
+ "group_size": 64,
2917
+ "bits": 4
2918
+ },
2919
+ "model.layers.44.input_layernorm": false,
2920
+ "model.layers.44.post_attention_layernorm": false,
2921
+ "model.layers.45.self_attn.q_proj": {
2922
+ "group_size": 64,
2923
+ "bits": 4
2924
+ },
2925
+ "model.layers.45.self_attn.k_proj": {
2926
+ "group_size": 64,
2927
+ "bits": 4
2928
+ },
2929
+ "model.layers.45.self_attn.v_proj": {
2930
+ "group_size": 64,
2931
+ "bits": 8
2932
+ },
2933
+ "model.layers.45.self_attn.o_proj": {
2934
+ "group_size": 64,
2935
+ "bits": 4
2936
+ },
2937
+ "model.layers.45.self_attn.rope": false,
2938
+ "model.layers.45.mlp.gate_proj": {
2939
+ "group_size": 64,
2940
+ "bits": 4
2941
+ },
2942
+ "model.layers.45.mlp.down_proj": {
2943
+ "group_size": 64,
2944
+ "bits": 8
2945
+ },
2946
+ "model.layers.45.mlp.up_proj": {
2947
+ "group_size": 64,
2948
+ "bits": 4
2949
+ },
2950
+ "model.layers.45.input_layernorm": false,
2951
+ "model.layers.45.post_attention_layernorm": false,
2952
+ "model.layers.46.self_attn.q_proj": {
2953
+ "group_size": 64,
2954
+ "bits": 4
2955
+ },
2956
+ "model.layers.46.self_attn.k_proj": {
2957
+ "group_size": 64,
2958
+ "bits": 4
2959
+ },
2960
+ "model.layers.46.self_attn.v_proj": {
2961
+ "group_size": 64,
2962
+ "bits": 8
2963
+ },
2964
+ "model.layers.46.self_attn.o_proj": {
2965
+ "group_size": 64,
2966
+ "bits": 4
2967
+ },
2968
+ "model.layers.46.self_attn.rope": false,
2969
+ "model.layers.46.mlp.gate_proj": {
2970
+ "group_size": 64,
2971
+ "bits": 4
2972
+ },
2973
+ "model.layers.46.mlp.down_proj": {
2974
+ "group_size": 64,
2975
+ "bits": 8
2976
+ },
2977
+ "model.layers.46.mlp.up_proj": {
2978
+ "group_size": 64,
2979
+ "bits": 4
2980
+ },
2981
+ "model.layers.46.input_layernorm": false,
2982
+ "model.layers.46.post_attention_layernorm": false,
2983
+ "model.layers.47.self_attn.q_proj": {
2984
+ "group_size": 64,
2985
+ "bits": 4
2986
+ },
2987
+ "model.layers.47.self_attn.k_proj": {
2988
+ "group_size": 64,
2989
+ "bits": 4
2990
+ },
2991
+ "model.layers.47.self_attn.v_proj": {
2992
+ "group_size": 64,
2993
+ "bits": 8
2994
+ },
2995
+ "model.layers.47.self_attn.o_proj": {
2996
+ "group_size": 64,
2997
+ "bits": 4
2998
+ },
2999
+ "model.layers.47.self_attn.rope": false,
3000
+ "model.layers.47.mlp.gate_proj": {
3001
+ "group_size": 64,
3002
+ "bits": 4
3003
+ },
3004
+ "model.layers.47.mlp.down_proj": {
3005
+ "group_size": 64,
3006
+ "bits": 8
3007
+ },
3008
+ "model.layers.47.mlp.up_proj": {
3009
+ "group_size": 64,
3010
+ "bits": 4
3011
+ },
3012
+ "model.layers.47.input_layernorm": false,
3013
+ "model.layers.47.post_attention_layernorm": false,
3014
+ "model.norm": false,
3015
+ "lm_head": {
3016
+ "group_size": 64,
3017
+ "bits": 8
3018
+ }
3019
+ },
3020
+ "rms_norm_eps": 1e-05,
3021
+ "rope_theta": 1000000.0,
3022
+ "sliding_window": 131072,
3023
+ "tie_word_embeddings": false,
3024
+ "torch_dtype": "bfloat16",
3025
+ "transformers_version": "4.43.1",
3026
+ "use_cache": true,
3027
+ "use_sliding_window": false,
3028
+ "vocab_size": 152064
3029
+ }
DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f27108a8c6ada421df46a5f048996f810e852e6bd98b946592b923725b18a1b9
3
+ size 5343568565
DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03bd582135369ce6d1bd6d3ad9d1c11cab93ea4574fb16a00f42cf9b1b597ecc
3
+ size 4267470721
DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/special_tokens_map.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<｜begin▁of▁sentence｜>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<｜end▁of▁sentence｜>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<｜end▁of▁sentence｜>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ }
23
+ }
DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e20ddafc659ba90242154b55275402edeca0715e5dbb30f56815a4ce081f4893
3
+ size 11422778
DeepSeek-R1-Distill-Qwen-14B-4,8_mixed/tokenizer_config.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "151643": {
7
+ "content": "<｜end▁of▁sentence｜>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "151644": {
15
+ "content": "<｜User｜>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": false
21
+ },
22
+ "151645": {
23
+ "content": "<｜Assistant｜>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": false
29
+ },
30
+ "151646": {
31
+ "content": "<｜begin▁of▁sentence｜>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false,
36
+ "special": true
37
+ },
38
+ "151647": {
39
+ "content": "<|EOT|>",
40
+ "lstrip": false,
41
+ "normalized": false,
42
+ "rstrip": false,
43
+ "single_word": false,
44
+ "special": false
45
+ },
46
+ "151648": {
47
+ "content": "<think>",
48
+ "lstrip": false,
49
+ "normalized": false,
50
+ "rstrip": false,
51
+ "single_word": false,
52
+ "special": false
53
+ },
54
+ "151649": {
55
+ "content": "</think>",
56
+ "lstrip": false,
57
+ "normalized": false,
58
+ "rstrip": false,
59
+ "single_word": false,
60
+ "special": false
61
+ },
62
+ "151650": {
63
+ "content": "<|quad_start|>",
64
+ "lstrip": false,
65
+ "normalized": false,
66
+ "rstrip": false,
67
+ "single_word": false,
68
+ "special": true
69
+ },
70
+ "151651": {
71
+ "content": "<|quad_end|>",
72
+ "lstrip": false,
73
+ "normalized": false,
74
+ "rstrip": false,
75
+ "single_word": false,
76
+ "special": true
77
+ },
78
+ "151652": {
79
+ "content": "<|vision_start|>",
80
+ "lstrip": false,
81
+ "normalized": false,
82
+ "rstrip": false,
83
+ "single_word": false,
84
+ "special": true
85
+ },
86
+ "151653": {
87
+ "content": "<|vision_end|>",
88
+ "lstrip": false,
89
+ "normalized": false,
90
+ "rstrip": false,
91
+ "single_word": false,
92
+ "special": true
93
+ },
94
+ "151654": {
95
+ "content": "<|vision_pad|>",
96
+ "lstrip": false,
97
+ "normalized": false,
98
+ "rstrip": false,
99
+ "single_word": false,
100
+ "special": true
101
+ },
102
+ "151655": {
103
+ "content": "<|image_pad|>",
104
+ "lstrip": false,
105
+ "normalized": false,
106
+ "rstrip": false,
107
+ "single_word": false,
108
+ "special": true
109
+ },
110
+ "151656": {
111
+ "content": "<|video_pad|>",
112
+ "lstrip": false,
113
+ "normalized": false,
114
+ "rstrip": false,
115
+ "single_word": false,
116
+ "special": true
117
+ },
118
+ "151657": {
119
+ "content": "<tool_call>",
120
+ "lstrip": false,
121
+ "normalized": false,
122
+ "rstrip": false,
123
+ "single_word": false,
124
+ "special": false
125
+ },
126
+ "151658": {
127
+ "content": "</tool_call>",
128
+ "lstrip": false,
129
+ "normalized": false,
130
+ "rstrip": false,
131
+ "single_word": false,
132
+ "special": false
133
+ },
134
+ "151659": {
135
+ "content": "<|fim_prefix|>",
136
+ "lstrip": false,
137
+ "normalized": false,
138
+ "rstrip": false,
139
+ "single_word": false,
140
+ "special": false
141
+ },
142
+ "151660": {
143
+ "content": "<|fim_middle|>",
144
+ "lstrip": false,
145
+ "normalized": false,
146
+ "rstrip": false,
147
+ "single_word": false,
148
+ "special": false
149
+ },
150
+ "151661": {
151
+ "content": "<|fim_suffix|>",
152
+ "lstrip": false,
153
+ "normalized": false,
154
+ "rstrip": false,
155
+ "single_word": false,
156
+ "special": false
157
+ },
158
+ "151662": {
159
+ "content": "<|fim_pad|>",
160
+ "lstrip": false,
161
+ "normalized": false,
162
+ "rstrip": false,
163
+ "single_word": false,
164
+ "special": false
165
+ },
166
+ "151663": {
167
+ "content": "<|repo_name|>",
168
+ "lstrip": false,
169
+ "normalized": false,
170
+ "rstrip": false,
171
+ "single_word": false,
172
+ "special": false
173
+ },
174
+ "151664": {
175
+ "content": "<|file_sep|>",
176
+ "lstrip": false,
177
+ "normalized": false,
178
+ "rstrip": false,
179
+ "single_word": false,
180
+ "special": false
181
+ }
182
+ },
183
+ "bos_token": "<｜begin▁of▁sentence｜>",
184
+ "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='') %}{%- for message in messages %}{%- if message['role'] == 'system' %}{% set ns.system_prompt = message['content'] %}{%- endif %}{%- endfor %}{{bos_token}}{{ns.system_prompt}}{%- for message in messages %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{{'<｜User｜>' + message['content']}}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is none %}{%- set ns.is_tool = false -%}{%- for tool in message['tool_calls']%}{%- if not ns.is_first %}{{'<｜Assistant｜><｜tool▁calls▁begin｜><｜tool▁call▁begin��>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{%- set ns.is_first = true -%}{%- else %}{{'\\n' + '<｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{{'<｜tool▁calls▁end｜><｜end▁of▁sentence｜>'}}{%- endif %}{%- endfor %}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is not none %}{%- if ns.is_tool %}{{'<｜tool▁outputs▁end｜>' + message['content'] + '<｜end▁of▁sentence｜>'}}{%- set ns.is_tool = false -%}{%- else %}{% set content = message['content'] %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{% endif %}{{'<｜Assistant｜>' + content + '<｜end▁of▁sentence｜>'}}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<｜tool▁outputs▁begin｜><｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- set ns.is_output_first = false %}{%- else %}{{'\\n<｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<｜tool▁outputs▁end｜>'}}{% endif %}{% if add_generation_prompt and not ns.is_tool %}{{'<｜Assistant｜><think>\\n'}}{% endif %}",
185
+ "clean_up_tokenization_spaces": false,
186
+ "eos_token": "<｜end▁of▁sentence｜>",
187
+ "extra_special_tokens": {},
188
+ "legacy": true,
189
+ "model_max_length": 16384,
190
+ "pad_token": "<｜end▁of▁sentence｜>",
191
+ "sp_model_kwargs": {},
192
+ "tokenizer_class": "LlamaTokenizerFast",
193
+ "unk_token": null,
194
+ "use_default_system_prompt": false
195
+ }
DeepSeek-R1-Distill-Qwen-14B-4bit/config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151643,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 5120,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 13824,
12
+ "max_position_embeddings": 131072,
13
+ "max_window_layers": 48,
14
+ "model_type": "qwen2",
15
+ "num_attention_heads": 40,
16
+ "num_hidden_layers": 48,
17
+ "num_key_value_heads": 8,
18
+ "quantization": {
19
+ "group_size": 64,
20
+ "bits": 4
21
+ },
22
+ "quantization_config": {
23
+ "group_size": 64,
24
+ "bits": 4
25
+ },
26
+ "rms_norm_eps": 1e-05,
27
+ "rope_theta": 1000000.0,
28
+ "sliding_window": 131072,
29
+ "tie_word_embeddings": false,
30
+ "torch_dtype": "bfloat16",
31
+ "transformers_version": "4.43.1",
32
+ "use_cache": true,
33
+ "use_sliding_window": false,
34
+ "vocab_size": 152064
35
+ }
DeepSeek-R1-Distill-Qwen-14B-4bit/model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:561eacdd7d4d652063a099fb0b6b5340f9e3f82263f7dca70fcf9e026f5860e9
3
+ size 5353840339
DeepSeek-R1-Distill-Qwen-14B-4bit/model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09ec75fe3ad8aa351fce70d1184625b275d4ba578212951e8e4f1883fcc92ed3
3
+ size 2955654022
DeepSeek-R1-Distill-Qwen-14B-4bit/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
DeepSeek-R1-Distill-Qwen-14B-4bit/special_tokens_map.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<｜begin▁of▁sentence｜>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<｜end▁of▁sentence｜>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<｜end▁of▁sentence｜>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ }
23
+ }
DeepSeek-R1-Distill-Qwen-14B-4bit/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e20ddafc659ba90242154b55275402edeca0715e5dbb30f56815a4ce081f4893
3
+ size 11422778
DeepSeek-R1-Distill-Qwen-14B-4bit/tokenizer_config.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "add_prefix_space": null,
5
+ "added_tokens_decoder": {
6
+ "151643": {
7
+ "content": "<｜end▁of▁sentence｜>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false,
12
+ "special": true
13
+ },
14
+ "151644": {
15
+ "content": "<｜User｜>",
16
+ "lstrip": false,
17
+ "normalized": false,
18
+ "rstrip": false,
19
+ "single_word": false,
20
+ "special": false
21
+ },
22
+ "151645": {
23
+ "content": "<｜Assistant｜>",
24
+ "lstrip": false,
25
+ "normalized": false,
26
+ "rstrip": false,
27
+ "single_word": false,
28
+ "special": false
29
+ },
30
+ "151646": {
31
+ "content": "<｜begin▁of▁sentence｜>",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false,
36
+ "special": true
37
+ },
38
+ "151647": {
39
+ "content": "<|EOT|>",
40
+ "lstrip": false,
41
+ "normalized": false,
42
+ "rstrip": false,
43
+ "single_word": false,
44
+ "special": false
45
+ },
46
+ "151648": {
47
+ "content": "<think>",
48
+ "lstrip": false,
49
+ "normalized": false,
50
+ "rstrip": false,
51
+ "single_word": false,
52
+ "special": false
53
+ },
54
+ "151649": {
55
+ "content": "</think>",
56
+ "lstrip": false,
57
+ "normalized": false,
58
+ "rstrip": false,
59
+ "single_word": false,
60
+ "special": false
61
+ },
62
+ "151650": {
63
+ "content": "<|quad_start|>",
64
+ "lstrip": false,
65
+ "normalized": false,
66
+ "rstrip": false,
67
+ "single_word": false,
68
+ "special": true
69
+ },
70
+ "151651": {
71
+ "content": "<|quad_end|>",
72
+ "lstrip": false,
73
+ "normalized": false,
74
+ "rstrip": false,
75
+ "single_word": false,
76
+ "special": true
77
+ },
78
+ "151652": {
79
+ "content": "<|vision_start|>",
80
+ "lstrip": false,
81
+ "normalized": false,
82
+ "rstrip": false,
83
+ "single_word": false,
84
+ "special": true
85
+ },
86
+ "151653": {
87
+ "content": "<|vision_end|>",
88
+ "lstrip": false,
89
+ "normalized": false,
90
+ "rstrip": false,
91
+ "single_word": false,
92
+ "special": true
93
+ },
94
+ "151654": {
95
+ "content": "<|vision_pad|>",
96
+ "lstrip": false,
97
+ "normalized": false,
98
+ "rstrip": false,
99
+ "single_word": false,
100
+ "special": true
101
+ },
102
+ "151655": {
103
+ "content": "<|image_pad|>",
104
+ "lstrip": false,
105
+ "normalized": false,
106
+ "rstrip": false,
107
+ "single_word": false,
108
+ "special": true
109
+ },
110
+ "151656": {
111
+ "content": "<|video_pad|>",
112
+ "lstrip": false,
113
+ "normalized": false,
114
+ "rstrip": false,
115
+ "single_word": false,
116
+ "special": true
117
+ },
118
+ "151657": {
119
+ "content": "<tool_call>",
120
+ "lstrip": false,
121
+ "normalized": false,
122
+ "rstrip": false,
123
+ "single_word": false,
124
+ "special": false
125
+ },
126
+ "151658": {
127
+ "content": "</tool_call>",
128
+ "lstrip": false,
129
+ "normalized": false,
130
+ "rstrip": false,
131
+ "single_word": false,
132
+ "special": false
133
+ },
134
+ "151659": {
135
+ "content": "<|fim_prefix|>",
136
+ "lstrip": false,
137
+ "normalized": false,
138
+ "rstrip": false,
139
+ "single_word": false,
140
+ "special": false
141
+ },
142
+ "151660": {
143
+ "content": "<|fim_middle|>",
144
+ "lstrip": false,
145
+ "normalized": false,
146
+ "rstrip": false,
147
+ "single_word": false,
148
+ "special": false
149
+ },
150
+ "151661": {
151
+ "content": "<|fim_suffix|>",
152
+ "lstrip": false,
153
+ "normalized": false,
154
+ "rstrip": false,
155
+ "single_word": false,
156
+ "special": false
157
+ },
158
+ "151662": {
159
+ "content": "<|fim_pad|>",
160
+ "lstrip": false,
161
+ "normalized": false,
162
+ "rstrip": false,
163
+ "single_word": false,
164
+ "special": false
165
+ },
166
+ "151663": {
167
+ "content": "<|repo_name|>",
168
+ "lstrip": false,
169
+ "normalized": false,
170
+ "rstrip": false,
171
+ "single_word": false,
172
+ "special": false
173
+ },
174
+ "151664": {
175
+ "content": "<|file_sep|>",
176
+ "lstrip": false,
177
+ "normalized": false,
178
+ "rstrip": false,
179
+ "single_word": false,
180
+ "special": false
181
+ }
182
+ },
183
+ "bos_token": "<｜begin▁of▁sentence｜>",
184
+ "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='') %}{%- for message in messages %}{%- if message['role'] == 'system' %}{% set ns.system_prompt = message['content'] %}{%- endif %}{%- endfor %}{{bos_token}}{{ns.system_prompt}}{%- for message in messages %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{{'<｜User｜>' + message['content']}}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is none %}{%- set ns.is_tool = false -%}{%- for tool in message['tool_calls']%}{%- if not ns.is_first %}{{'<｜Assistant｜><｜tool▁calls▁begin｜><｜tool▁call▁begin��>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{%- set ns.is_first = true -%}{%- else %}{{'\\n' + '<｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\\n' + '```json' + '\\n' + tool['function']['arguments'] + '\\n' + '```' + '<｜tool▁call▁end｜>'}}{{'<｜tool▁calls▁end｜><｜end▁of▁sentence｜>'}}{%- endif %}{%- endfor %}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is not none %}{%- if ns.is_tool %}{{'<｜tool▁outputs▁end｜>' + message['content'] + '<｜end▁of▁sentence｜>'}}{%- set ns.is_tool = false -%}{%- else %}{% set content = message['content'] %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{% endif %}{{'<｜Assistant｜>' + content + '<｜end▁of▁sentence｜>'}}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<｜tool▁outputs▁begin｜><｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- set ns.is_output_first = false %}{%- else %}{{'\\n<｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<｜tool▁outputs▁end｜>'}}{% endif %}{% if add_generation_prompt and not ns.is_tool %}{{'<｜Assistant｜><think>\\n'}}{% endif %}",
185
+ "clean_up_tokenization_spaces": false,
186
+ "eos_token": "<｜end▁of▁sentence｜>",
187
+ "extra_special_tokens": {},
188
+ "legacy": true,
189
+ "model_max_length": 16384,
190
+ "pad_token": "<｜end▁of▁sentence｜>",
191
+ "sp_model_kwargs": {},
192
+ "tokenizer_class": "LlamaTokenizerFast",
193
+ "unk_token": null,
194
+ "use_default_system_prompt": false
195
+ }
DeepSeek-R1-Distill-Qwen-14B-6bit/config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151643,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 5120,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 13824,
12
+ "max_position_embeddings": 131072,
13
+ "max_window_layers": 48,
14
+ "model_type": "qwen2",
15
+ "num_attention_heads": 40,
16
+ "num_hidden_layers": 48,
17
+ "num_key_value_heads": 8,
18
+ "quantization": {
19
+ "group_size": 64,
20
+ "bits": 6
21
+ },
22
+ "quantization_config": {
23
+ "group_size": 64,
24
+ "bits": 6
25
+ },
26
+ "rms_norm_eps": 1e-05,
27
+ "rope_theta": 1000000.0,
28
+ "sliding_window": 131072,
29
+ "tie_word_embeddings": false,
30
+ "torch_dtype": "bfloat16",
31
+ "transformers_version": "4.43.1",
32
+ "use_cache": true,
33
+ "use_sliding_window": false,
34
+ "vocab_size": 152064
35
+ }