Hugging Face's logo

mradermacher
/
Gemma2-9b-WangchanLIONv2-instruct-GGUF

Transformers
GGUF
Thai
English
Inference Endpoints
conversational
Model card Files Community
mradermacher commited on
Commit
b215355
·
verified ·
1 Parent(s): 1066e27

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -0
README.md CHANGED
@@ -38,11 +38,15 @@ more details, including on how to concatenate multi-part files.
38
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q3_K_S.gguf) | Q3_K_S | 4.4 | |
39
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q3_K_M.gguf) | Q3_K_M | 4.9 | lower quality |
40
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q3_K_L.gguf) | Q3_K_L | 5.2 | |
 
 
41
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q4_K_S.gguf) | Q4_K_S | 5.6 | fast, recommended |
42
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q4_K_M.gguf) | Q4_K_M | 5.9 | fast, recommended |
43
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q5_K_S.gguf) | Q5_K_S | 6.6 | |
 
44
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q6_K.gguf) | Q6_K | 7.7 | very good quality |
45
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q8_0.gguf) | Q8_0 | 9.9 | fast, best quality |
 
46
 
47
  Here is a handy graph by ikawrakow comparing some lower-quality quant
48
  types (lower is better):
 
38
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q3_K_S.gguf) | Q3_K_S | 4.4 | |
39
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q3_K_M.gguf) | Q3_K_M | 4.9 | lower quality |
40
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q3_K_L.gguf) | Q3_K_L | 5.2 | |
41
+ | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.IQ4_XS.gguf) | IQ4_XS | 5.3 | |
42
+ | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q4_0_4_4.gguf) | Q4_0_4_4 | 5.5 | fast on arm, low quality |
43
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q4_K_S.gguf) | Q4_K_S | 5.6 | fast, recommended |
44
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q4_K_M.gguf) | Q4_K_M | 5.9 | fast, recommended |
45
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q5_K_S.gguf) | Q5_K_S | 6.6 | |
46
+ | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q5_K_M.gguf) | Q5_K_M | 6.7 | |
47
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q6_K.gguf) | Q6_K | 7.7 | very good quality |
48
  | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.Q8_0.gguf) | Q8_0 | 9.9 | fast, best quality |
49
+ | [GGUF](https://huggingface.co/mradermacher/Gemma2-9b-WangchanLIONv2-instruct-GGUF/resolve/main/Gemma2-9b-WangchanLIONv2-instruct.f16.gguf) | f16 | 18.6 | 16 bpw, overkill |
50
 
51
  Here is a handy graph by ikawrakow comparing some lower-quality quant
52
  types (lower is better):