mradermacher
commited on
Commit
•
a561b15
1
Parent(s):
2858a45
auto-patch README.md
Browse files
README.md
CHANGED
@@ -17,7 +17,6 @@ quantized_by: mradermacher
|
|
17 |
weighted/imatrix quants of https://huggingface.co/Qwen/Qwen2.5-7B
|
18 |
|
19 |
<!-- provided-files -->
|
20 |
-
static quants are available at https://huggingface.co/mradermacher/Qwen2.5-7B-GGUF
|
21 |
## Usage
|
22 |
|
23 |
If you are unsure how to use GGUF files, refer to one of [TheBloke's
|
@@ -36,6 +35,7 @@ more details, including on how to concatenate multi-part files.
|
|
36 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ2_XS.gguf) | i1-IQ2_XS | 2.6 | |
|
37 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ2_S.gguf) | i1-IQ2_S | 2.7 | |
|
38 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ2_M.gguf) | i1-IQ2_M | 2.9 | |
|
|
|
39 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q2_K.gguf) | i1-Q2_K | 3.1 | IQ3_XXS probably better |
|
40 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 3.2 | lower quality |
|
41 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ3_XS.gguf) | i1-IQ3_XS | 3.4 | |
|
@@ -48,11 +48,15 @@ more details, including on how to concatenate multi-part files.
|
|
48 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_0_4_4.gguf) | i1-Q4_0_4_4 | 4.5 | fast on arm, low quality |
|
49 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_0_4_8.gguf) | i1-Q4_0_4_8 | 4.5 | fast on arm+i8mm, low quality |
|
50 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_0_8_8.gguf) | i1-Q4_0_8_8 | 4.5 | fast on arm+sve, low quality |
|
|
|
51 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_0.gguf) | i1-Q4_0 | 4.5 | fast, low quality |
|
52 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_K_S.gguf) | i1-Q4_K_S | 4.6 | optimal size/speed/quality |
|
53 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_K_M.gguf) | i1-Q4_K_M | 4.8 | fast, recommended |
|
|
|
54 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q5_K_S.gguf) | i1-Q5_K_S | 5.4 | |
|
|
|
55 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q5_K_M.gguf) | i1-Q5_K_M | 5.5 | |
|
|
|
56 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q6_K.gguf) | i1-Q6_K | 6.4 | practically like static Q6_K |
|
57 |
|
58 |
Here is a handy graph by ikawrakow comparing some lower-quality quant
|
|
|
17 |
weighted/imatrix quants of https://huggingface.co/Qwen/Qwen2.5-7B
|
18 |
|
19 |
<!-- provided-files -->
|
|
|
20 |
## Usage
|
21 |
|
22 |
If you are unsure how to use GGUF files, refer to one of [TheBloke's
|
|
|
35 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ2_XS.gguf) | i1-IQ2_XS | 2.6 | |
|
36 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ2_S.gguf) | i1-IQ2_S | 2.7 | |
|
37 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ2_M.gguf) | i1-IQ2_M | 2.9 | |
|
38 |
+
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q2_K_S.gguf) | i1-Q2_K_S | 2.9 | very low quality |
|
39 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q2_K.gguf) | i1-Q2_K | 3.1 | IQ3_XXS probably better |
|
40 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 3.2 | lower quality |
|
41 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ3_XS.gguf) | i1-IQ3_XS | 3.4 | |
|
|
|
48 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_0_4_4.gguf) | i1-Q4_0_4_4 | 4.5 | fast on arm, low quality |
|
49 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_0_4_8.gguf) | i1-Q4_0_4_8 | 4.5 | fast on arm+i8mm, low quality |
|
50 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_0_8_8.gguf) | i1-Q4_0_8_8 | 4.5 | fast on arm+sve, low quality |
|
51 |
+
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-IQ4_NL.gguf) | i1-IQ4_NL | 4.5 | prefer IQ4_XS |
|
52 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_0.gguf) | i1-Q4_0 | 4.5 | fast, low quality |
|
53 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_K_S.gguf) | i1-Q4_K_S | 4.6 | optimal size/speed/quality |
|
54 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_K_M.gguf) | i1-Q4_K_M | 4.8 | fast, recommended |
|
55 |
+
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q4_1.gguf) | i1-Q4_1 | 5.0 | |
|
56 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q5_K_S.gguf) | i1-Q5_K_S | 5.4 | |
|
57 |
+
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q5_0.gguf) | i1-Q5_0 | 5.4 | |
|
58 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q5_K_M.gguf) | i1-Q5_K_M | 5.5 | |
|
59 |
+
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q5_1.gguf) | i1-Q5_1 | 5.9 | |
|
60 |
| [GGUF](https://huggingface.co/mradermacher/Qwen2.5-7B-i1-GGUF/resolve/main/Qwen2.5-7B.i1-Q6_K.gguf) | i1-Q6_K | 6.4 | practically like static Q6_K |
|
61 |
|
62 |
Here is a handy graph by ikawrakow comparing some lower-quality quant
|