Update README.md
Browse files
README.md
CHANGED
@@ -13,7 +13,25 @@ EXL2 quants of [trajkovnikola/MKLLM-7B-Instruct](https://huggingface.co/trajkovn
|
|
13 |
|
14 |
<b>The "main" branch only contains the measurement.json (which can be used for further conversions), download one of the other branches for the model:</b>
|
15 |
|
|
|
16 |
[4.25 bits per weight with 6 lm_head bits](https://huggingface.co/martinkozle/MKLLM-7B-Instruct-exl2/tree/4.25bpw-h6)
|
17 |
-
[5
|
18 |
|
19 |
-
[measurement.json](https://huggingface.co/martinkozle/MKLLM-7B-Instruct-exl2/blob/main/measurement.json)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
13 |
|
14 |
<b>The "main" branch only contains the measurement.json (which can be used for further conversions), download one of the other branches for the model:</b>
|
15 |
|
16 |
+
[5.00 bits per weight with 6 lm_head bits](https://huggingface.co/martinkozle/MKLLM-7B-Instruct-exl2/tree/5.0bpw-h6)
|
17 |
[4.25 bits per weight with 6 lm_head bits](https://huggingface.co/martinkozle/MKLLM-7B-Instruct-exl2/tree/4.25bpw-h6)
|
18 |
+
[3.5 bits per weight with 6 lm_head bits](https://huggingface.co/martinkozle/MKLLM-7B-Instruct-exl2/tree/3.5bpw-h6)
|
19 |
|
20 |
+
[measurement.json](https://huggingface.co/martinkozle/MKLLM-7B-Instruct-exl2/blob/main/measurement.json)
|
21 |
+
|
22 |
+
|
23 |
+
## Download instructions
|
24 |
+
|
25 |
+
With git:
|
26 |
+
|
27 |
+
```shell
|
28 |
+
git clone --single-branch --branch 5.0bpw-h6 https://huggingface.co/martinkozle/MKLLM-7B-Instruct-exl2 MKLLM-7B-Instruct-exl2-5.0bpw-h6
|
29 |
+
```
|
30 |
+
|
31 |
+
With huggingface hub:
|
32 |
+
|
33 |
+
To download a specific branch, use the `--revision` parameter.
|
34 |
+
|
35 |
+
```shell
|
36 |
+
huggingface-cli download martinkozle/MKLLM-7B-Instruct-exl2 --revision 5.0bpw-h6 --local-dir MKLLM-7B-Instruct-exl2-5.0bpw-h6 --local-dir-use-symlinks False
|
37 |
+
```
|