mradermacher
commited on
Commit
•
3720d9b
1
Parent(s):
68f867a
auto-patch README.md
Browse files
README.md
CHANGED
@@ -33,7 +33,18 @@ more details, including on how to concatenate multi-part files.
|
|
33 |
|
34 |
| Link | Type | Size/GB | Notes |
|
35 |
|:-----|:-----|--------:|:------|
|
|
|
|
|
|
|
|
|
|
|
|
|
36 |
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-IQ3_M.gguf) | i1-IQ3_M | 35.6 | |
|
|
|
|
|
|
|
|
|
|
|
37 |
|
38 |
Here is a handy graph by ikawrakow comparing some lower-quality quant
|
39 |
types (lower is better):
|
|
|
33 |
|
34 |
| Link | Type | Size/GB | Notes |
|
35 |
|:-----|:-----|--------:|:------|
|
36 |
+
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-IQ1_M.gguf) | i1-IQ1_M | 23.8 | mostly desperate |
|
37 |
+
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-IQ2_M.gguf) | i1-IQ2_M | 29.4 | |
|
38 |
+
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-Q2_K_S.gguf) | i1-Q2_K_S | 29.7 | very low quality |
|
39 |
+
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-Q2_K.gguf) | i1-Q2_K | 29.9 | IQ3_XXS probably better |
|
40 |
+
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 31.9 | lower quality |
|
41 |
+
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-Q3_K_S.gguf) | i1-Q3_K_S | 34.6 | IQ3_XS probably better |
|
42 |
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-IQ3_M.gguf) | i1-IQ3_M | 35.6 | |
|
43 |
+
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-Q3_K_M.gguf) | i1-Q3_K_M | 37.8 | IQ3_S probably better |
|
44 |
+
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-IQ4_XS.gguf) | i1-IQ4_XS | 39.8 | |
|
45 |
+
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-Q4_K_S.gguf) | i1-Q4_K_S | 44.0 | optimal size/speed/quality |
|
46 |
+
| [GGUF](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-Q4_K_M.gguf) | i1-Q4_K_M | 47.5 | fast, recommended |
|
47 |
+
| [PART 1](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-Q6_K.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/QVQ-72B-Preview-i1-GGUF/resolve/main/QVQ-72B-Preview.i1-Q6_K.gguf.part2of2) | i1-Q6_K | 64.4 | practically like static Q6_K |
|
48 |
|
49 |
Here is a handy graph by ikawrakow comparing some lower-quality quant
|
50 |
types (lower is better):
|