mradermacher commited on
Commit
9ec31b0
·
verified ·
1 Parent(s): 45512ad

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +8 -0
README.md CHANGED
@@ -49,12 +49,20 @@ more details, including on how to concatenate multi-part files.
49
 
50
  | Link | Type | Size/GB | Notes |
51
  |:-----|:-----|--------:|:------|
 
 
 
 
52
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-Q2_K.gguf) | i1-Q2_K | 4.1 | IQ3_XXS probably better |
53
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 4.2 | lower quality |
 
54
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-IQ3_M.gguf) | i1-IQ3_M | 4.9 | |
55
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-Q3_K_M.gguf) | i1-Q3_K_M | 5.2 | IQ3_S probably better |
 
56
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-IQ4_NL.gguf) | i1-IQ4_NL | 6.0 | prefer IQ4_XS |
57
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-Q4_K_S.gguf) | i1-Q4_K_S | 6.0 | optimal size/speed/quality |
 
 
58
 
59
  Here is a handy graph by ikawrakow comparing some lower-quality quant
60
  types (lower is better):
 
49
 
50
  | Link | Type | Size/GB | Notes |
51
  |:-----|:-----|--------:|:------|
52
+ | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-IQ1_M.gguf) | i1-IQ1_M | 2.8 | mostly desperate |
53
+ | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 3.1 | |
54
+ | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-IQ2_M.gguf) | i1-IQ2_M | 3.8 | |
55
+ | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-Q2_K_S.gguf) | i1-Q2_K_S | 3.8 | very low quality |
56
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-Q2_K.gguf) | i1-Q2_K | 4.1 | IQ3_XXS probably better |
57
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 4.2 | lower quality |
58
+ | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-Q3_K_S.gguf) | i1-Q3_K_S | 4.7 | IQ3_XS probably better |
59
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-IQ3_M.gguf) | i1-IQ3_M | 4.9 | |
60
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-Q3_K_M.gguf) | i1-Q3_K_M | 5.2 | IQ3_S probably better |
61
+ | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-IQ4_XS.gguf) | i1-IQ4_XS | 5.7 | |
62
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-IQ4_NL.gguf) | i1-IQ4_NL | 6.0 | prefer IQ4_XS |
63
  | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-Q4_K_S.gguf) | i1-Q4_K_S | 6.0 | optimal size/speed/quality |
64
+ | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-Q4_K_M.gguf) | i1-Q4_K_M | 6.3 | fast, recommended |
65
+ | [GGUF](https://huggingface.co/mradermacher/INTELLECT-1-Instruct-i1-GGUF/resolve/main/INTELLECT-1-Instruct.i1-Q6_K.gguf) | i1-Q6_K | 8.5 | practically like static Q6_K |
66
 
67
  Here is a handy graph by ikawrakow comparing some lower-quality quant
68
  types (lower is better):