mradermacher commited on
Commit
743b909
1 Parent(s): fb4f0f5

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +23 -0
README.md CHANGED
@@ -32,7 +32,30 @@ more details, including on how to concatenate multi-part files.
32
 
33
  | Link | Type | Size/GB | Notes |
34
  |:-----|:-----|--------:|:------|
 
 
 
 
 
 
35
  | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q2_K.gguf) | i1-Q2_K | 3.3 | IQ3_XXS probably better |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
36
 
37
  Here is a handy graph by ikawrakow comparing some lower-quality quant
38
  types (lower is better):
 
32
 
33
  | Link | Type | Size/GB | Notes |
34
  |:-----|:-----|--------:|:------|
35
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ1_S.gguf) | i1-IQ1_S | 2.1 | for the desperate |
36
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ1_M.gguf) | i1-IQ1_M | 2.3 | mostly desperate |
37
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 2.5 | |
38
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ2_XS.gguf) | i1-IQ2_XS | 2.7 | |
39
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ2_S.gguf) | i1-IQ2_S | 2.9 | |
40
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ2_M.gguf) | i1-IQ2_M | 3.0 | |
41
  | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q2_K.gguf) | i1-Q2_K | 3.3 | IQ3_XXS probably better |
42
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 3.4 | lower quality |
43
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ3_XS.gguf) | i1-IQ3_XS | 3.6 | |
44
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q3_K_S.gguf) | i1-Q3_K_S | 3.8 | IQ3_XS probably better |
45
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ3_S.gguf) | i1-IQ3_S | 3.8 | beats Q3_K* |
46
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ3_M.gguf) | i1-IQ3_M | 3.9 | |
47
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q3_K_M.gguf) | i1-Q3_K_M | 4.1 | IQ3_S probably better |
48
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q3_K_L.gguf) | i1-Q3_K_L | 4.4 | IQ3_M probably better |
49
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-IQ4_XS.gguf) | i1-IQ4_XS | 4.5 | |
50
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q4_0_4_4.gguf) | i1-Q4_0_4_4 | 4.8 | fast on arm, low quality |
51
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q4_0_4_8.gguf) | i1-Q4_0_4_8 | 4.8 | fast on arm+i8mm, low quality |
52
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q4_0_8_8.gguf) | i1-Q4_0_8_8 | 4.8 | fast on arm+sve, low quality |
53
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q4_0.gguf) | i1-Q4_0 | 4.8 | fast, low quality |
54
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q4_K_S.gguf) | i1-Q4_K_S | 4.8 | optimal size/speed/quality |
55
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q4_K_M.gguf) | i1-Q4_K_M | 5.0 | fast, recommended |
56
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q5_K_S.gguf) | i1-Q5_K_S | 5.7 | |
57
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q5_K_M.gguf) | i1-Q5_K_M | 5.8 | |
58
+ | [GGUF](https://huggingface.co/mradermacher/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES-i1-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-HalfAbliterated-TIES.i1-Q6_K.gguf) | i1-Q6_K | 6.7 | practically like static Q6_K |
59
 
60
  Here is a handy graph by ikawrakow comparing some lower-quality quant
61
  types (lower is better):