mradermacher commited on
Commit
cf26d33
·
verified ·
1 Parent(s): bf6d580

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +18 -0
README.md CHANGED
@@ -54,11 +54,29 @@ more details, including on how to concatenate multi-part files.
54
 
55
  | Link | Type | Size/GB | Notes |
56
  |:-----|:-----|--------:|:------|
 
 
 
 
 
 
 
57
  | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q2_K.gguf) | i1-Q2_K | 12.4 | IQ3_XXS probably better |
58
  | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 12.9 | lower quality |
 
 
 
59
  | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ3_M.gguf) | i1-IQ3_M | 14.9 | |
60
  | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q3_K_M.gguf) | i1-Q3_K_M | 16.0 | IQ3_S probably better |
 
 
 
61
  | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q4_K_S.gguf) | i1-Q4_K_S | 18.9 | optimal size/speed/quality |
 
 
 
 
 
62
 
63
  Here is a handy graph by ikawrakow comparing some lower-quality quant
64
  types (lower is better):
 
54
 
55
  | Link | Type | Size/GB | Notes |
56
  |:-----|:-----|--------:|:------|
57
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ1_S.gguf) | i1-IQ1_S | 7.4 | for the desperate |
58
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ1_M.gguf) | i1-IQ1_M | 8.0 | mostly desperate |
59
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 9.1 | |
60
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ2_XS.gguf) | i1-IQ2_XS | 10.1 | |
61
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ2_S.gguf) | i1-IQ2_S | 10.5 | |
62
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ2_M.gguf) | i1-IQ2_M | 11.4 | |
63
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q2_K_S.gguf) | i1-Q2_K_S | 11.6 | very low quality |
64
  | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q2_K.gguf) | i1-Q2_K | 12.4 | IQ3_XXS probably better |
65
  | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 12.9 | lower quality |
66
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ3_XS.gguf) | i1-IQ3_XS | 13.8 | |
67
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q3_K_S.gguf) | i1-Q3_K_S | 14.5 | IQ3_XS probably better |
68
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ3_S.gguf) | i1-IQ3_S | 14.5 | beats Q3_K* |
69
  | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ3_M.gguf) | i1-IQ3_M | 14.9 | |
70
  | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q3_K_M.gguf) | i1-Q3_K_M | 16.0 | IQ3_S probably better |
71
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q3_K_L.gguf) | i1-Q3_K_L | 17.3 | IQ3_M probably better |
72
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-IQ4_XS.gguf) | i1-IQ4_XS | 17.8 | |
73
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q4_0.gguf) | i1-Q4_0 | 18.8 | fast, low quality |
74
  | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q4_K_S.gguf) | i1-Q4_K_S | 18.9 | optimal size/speed/quality |
75
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q4_K_M.gguf) | i1-Q4_K_M | 20.0 | fast, recommended |
76
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q4_1.gguf) | i1-Q4_1 | 20.7 | |
77
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q5_K_S.gguf) | i1-Q5_K_S | 22.7 | |
78
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q5_K_M.gguf) | i1-Q5_K_M | 23.4 | |
79
+ | [GGUF](https://huggingface.co/mradermacher/Hamanasu-QwQ-V1.5-Instruct-i1-GGUF/resolve/main/Hamanasu-QwQ-V1.5-Instruct.i1-Q6_K.gguf) | i1-Q6_K | 27.0 | practically like static Q6_K |
80
 
81
  Here is a handy graph by ikawrakow comparing some lower-quality quant
82
  types (lower is better):