mradermacher commited on
Commit
c644b18
·
verified ·
1 Parent(s): ea0cca2

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +10 -0
README.md CHANGED
@@ -30,7 +30,17 @@ more details, including on how to concatenate multi-part files.
30
  | Link | Type | Size/GB | Notes |
31
  |:-----|:-----|--------:|:------|
32
  | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q2_K.gguf) | Q2_K | 4.9 | |
 
 
 
 
 
33
  | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q4_K_S.gguf) | Q4_K_S | 7.4 | fast, recommended |
 
 
 
 
 
34
 
35
  Here is a handy graph by ikawrakow comparing some lower-quality quant
36
  types (lower is better):
 
30
  | Link | Type | Size/GB | Notes |
31
  |:-----|:-----|--------:|:------|
32
  | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q2_K.gguf) | Q2_K | 4.9 | |
33
+ | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q3_K_S.gguf) | Q3_K_S | 5.7 | |
34
+ | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q3_K_M.gguf) | Q3_K_M | 6.3 | lower quality |
35
+ | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q3_K_L.gguf) | Q3_K_L | 6.8 | |
36
+ | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.IQ4_XS.gguf) | IQ4_XS | 7.1 | |
37
+ | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q4_0_4_4.gguf) | Q4_0_4_4 | 7.4 | fast on arm, low quality |
38
  | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q4_K_S.gguf) | Q4_K_S | 7.4 | fast, recommended |
39
+ | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q4_K_M.gguf) | Q4_K_M | 7.9 | fast, recommended |
40
+ | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q5_K_S.gguf) | Q5_K_S | 9.0 | |
41
+ | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q5_K_M.gguf) | Q5_K_M | 9.2 | |
42
+ | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q6_K.gguf) | Q6_K | 10.7 | very good quality |
43
+ | [GGUF](https://huggingface.co/mradermacher/Moe-4x7b-math-reason-code-GGUF/resolve/main/Moe-4x7b-math-reason-code.Q8_0.gguf) | Q8_0 | 13.8 | fast, best quality |
44
 
45
  Here is a handy graph by ikawrakow comparing some lower-quality quant
46
  types (lower is better):