mradermacher commited on
Commit
2d03ef0
1 Parent(s): 56a8ad5

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -3
README.md CHANGED
@@ -25,7 +25,7 @@ more details, including on how to concatenate multi-part files.
25
  | Link | Type | Size/GB | Notes |
26
  |:-----|:-----|--------:|:------|
27
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ1_S.gguf) | i1-IQ1_S | 2.6 | for the desperate |
28
- | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ1_M.gguf) | i1-IQ1_M | 2.7 | for the desperate |
29
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 3.2 | |
30
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ2_XS.gguf) | i1-IQ2_XS | 3.5 | |
31
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ2_S.gguf) | i1-IQ2_S | 3.7 | |
@@ -40,14 +40,13 @@ more details, including on how to concatenate multi-part files.
40
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q3_K_L.gguf) | i1-Q3_K_L | 5.9 | IQ3_M probably better |
41
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ4_XS.gguf) | i1-IQ4_XS | 6.0 | |
42
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q4_0.gguf) | i1-Q4_0 | 6.3 | fast, low quality |
43
- | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ4_NL.gguf) | i1-IQ4_NL | 6.4 | slightly worse than Q4_K_S |
44
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q4_K_S.gguf) | i1-Q4_K_S | 6.4 | optimal size/speed/quality |
45
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q4_K_M.gguf) | i1-Q4_K_M | 6.7 | fast, recommended |
46
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q5_K_S.gguf) | i1-Q5_K_S | 7.7 | |
47
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q5_K_M.gguf) | i1-Q5_K_M | 7.9 | |
48
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q6_K.gguf) | i1-Q6_K | 9.1 | practically like static Q6_K |
49
 
50
-
51
  Here is a handy graph by ikawrakow comparing some lower-quality quant
52
  types (lower is better):
53
 
 
25
  | Link | Type | Size/GB | Notes |
26
  |:-----|:-----|--------:|:------|
27
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ1_S.gguf) | i1-IQ1_S | 2.6 | for the desperate |
28
+ | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ1_M.gguf) | i1-IQ1_M | 2.7 | mostly desperate |
29
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 3.2 | |
30
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ2_XS.gguf) | i1-IQ2_XS | 3.5 | |
31
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ2_S.gguf) | i1-IQ2_S | 3.7 | |
 
40
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q3_K_L.gguf) | i1-Q3_K_L | 5.9 | IQ3_M probably better |
41
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ4_XS.gguf) | i1-IQ4_XS | 6.0 | |
42
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q4_0.gguf) | i1-Q4_0 | 6.3 | fast, low quality |
43
+ | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-IQ4_NL.gguf) | i1-IQ4_NL | 6.4 | prefer IQ4_XS |
44
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q4_K_S.gguf) | i1-Q4_K_S | 6.4 | optimal size/speed/quality |
45
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q4_K_M.gguf) | i1-Q4_K_M | 6.7 | fast, recommended |
46
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q5_K_S.gguf) | i1-Q5_K_S | 7.7 | |
47
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q5_K_M.gguf) | i1-Q5_K_M | 7.9 | |
48
  | [GGUF](https://huggingface.co/mradermacher/Fimbulvetr-11B-v2-i1-GGUF/resolve/main/Fimbulvetr-11B-v2.i1-Q6_K.gguf) | i1-Q6_K | 9.1 | practically like static Q6_K |
49
 
 
50
  Here is a handy graph by ikawrakow comparing some lower-quality quant
51
  types (lower is better):
52