Hugging Face's logo

mradermacher
/
Arabic-llama3-GGUF

Transformers
GGUF
English
Arabic
text-generation-inference
unsloth
llama
trl
sft
Inference Endpoints
Model card Files Community
mradermacher commited on
Commit
b465bb0
·
verified ·
1 Parent(s): 8d1d35d

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +6 -0
README.md CHANGED
@@ -43,8 +43,14 @@ more details, including on how to concatenate multi-part files.
43
  | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q3_K_S.gguf) | Q3_K_S | 3.8 | |
44
  | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q3_K_M.gguf) | Q3_K_M | 4.1 | lower quality |
45
  | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q3_K_L.gguf) | Q3_K_L | 4.4 | |
 
46
  | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q4_K_S.gguf) | Q4_K_S | 4.8 | fast, recommended |
 
 
 
47
  | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q6_K.gguf) | Q6_K | 6.7 | very good quality |
 
 
48
 
49
  Here is a handy graph by ikawrakow comparing some lower-quality quant
50
  types (lower is better):
 
43
  | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q3_K_S.gguf) | Q3_K_S | 3.8 | |
44
  | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q3_K_M.gguf) | Q3_K_M | 4.1 | lower quality |
45
  | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q3_K_L.gguf) | Q3_K_L | 4.4 | |
46
+ | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.IQ4_XS.gguf) | IQ4_XS | 4.6 | |
47
  | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q4_K_S.gguf) | Q4_K_S | 4.8 | fast, recommended |
48
+ | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q4_K_M.gguf) | Q4_K_M | 5.0 | fast, recommended |
49
+ | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q5_K_S.gguf) | Q5_K_S | 5.7 | |
50
+ | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q5_K_M.gguf) | Q5_K_M | 5.8 | |
51
  | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q6_K.gguf) | Q6_K | 6.7 | very good quality |
52
+ | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.Q8_0.gguf) | Q8_0 | 8.6 | fast, best quality |
53
+ | [GGUF](https://huggingface.co/mradermacher/Arabic-llama3-GGUF/resolve/main/Arabic-llama3.f16.gguf) | f16 | 16.2 | 16 bpw, overkill |
54
 
55
  Here is a handy graph by ikawrakow comparing some lower-quality quant
56
  types (lower is better):