bartowski commited on
Commit
2d62a6f
1 Parent(s): 84b5764

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +15 -15
README.md CHANGED
@@ -15,7 +15,7 @@ quantized_by: bartowski
15
  pipeline_tag: text-generation
16
  ---
17
 
18
- ## Exllama v2 Quantizations of Mistral-quiet-star
19
 
20
  Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.16">turboderp's ExLlamaV2 v0.0.16</a> for quantization.
21
 
@@ -23,22 +23,22 @@ Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.16">turb
23
 
24
  Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
25
 
26
- Original model: https://huggingface.co/liminerity/Mistral-quiet-star
27
 
28
  | Branch | Bits | lm_head bits | VRAM (4k) | VRAM (16k) | VRAM (32k) | Description |
29
  | ----- | ---- | ------- | ------ | ------ | ------ | ------------ |
30
- | [8_0](https://huggingface.co/bartowski/Mistral-quiet-star-exl2/tree/8_0) | 8.0 | 8.0 | 8.4 GB | 9.8 GB | 11.8 GB | Maximum quality that ExLlamaV2 can produce, near unquantized performance. |
31
- | [6_5](https://huggingface.co/bartowski/Mistral-quiet-star-exl2/tree/6_5) | 6.5 | 8.0 | 7.2 GB | 8.6 GB | 10.6 GB | Very similar to 8.0, good tradeoff of size vs performance, **recommended**. |
32
- | [5_0](https://huggingface.co/bartowski/Mistral-quiet-star-exl2/tree/5_0) | 5.0 | 6.0 | 6.0 GB | 7.4 GB | 9.4 GB | Slightly lower quality vs 6.5, but usable on 8GB cards. |
33
- | [4_25](https://huggingface.co/bartowski/Mistral-quiet-star-exl2/tree/4_25) | 4.25 | 6.0 | 5.3 GB | 6.7 GB | 8.7 GB | GPTQ equivalent bits per weight, slightly higher quality. |
34
- | [3_5](https://huggingface.co/bartowski/Mistral-quiet-star-exl2/tree/3_5) | 3.5 | 6.0 | 4.7 GB | 6.1 GB | 8.1 GB | Lower quality, only use if you have to. |
35
 
36
  ## Download instructions
37
 
38
  With git:
39
 
40
  ```shell
41
- git clone --single-branch --branch 6_5 https://huggingface.co/bartowski/Mistral-quiet-star-exl2 Mistral-quiet-star-exl2-6_5
42
  ```
43
 
44
  With huggingface hub (credit to TheBloke for instructions):
@@ -47,11 +47,11 @@ With huggingface hub (credit to TheBloke for instructions):
47
  pip3 install huggingface-hub
48
  ```
49
 
50
- To download the `main` (only useful if you only care about measurement.json) branch to a folder called `Mistral-quiet-star-exl2`:
51
 
52
  ```shell
53
- mkdir Mistral-quiet-star-exl2
54
- huggingface-cli download bartowski/Mistral-quiet-star-exl2 --local-dir Mistral-quiet-star-exl2 --local-dir-use-symlinks False
55
  ```
56
 
57
  To download from a different branch, add the `--revision` parameter:
@@ -59,15 +59,15 @@ To download from a different branch, add the `--revision` parameter:
59
  Linux:
60
 
61
  ```shell
62
- mkdir Mistral-quiet-star-exl2-6_5
63
- huggingface-cli download bartowski/Mistral-quiet-star-exl2 --revision 6_5 --local-dir Mistral-quiet-star-exl2-6_5 --local-dir-use-symlinks False
64
  ```
65
 
66
  Windows (which apparently doesn't like _ in folders sometimes?):
67
 
68
  ```shell
69
- mkdir Mistral-quiet-star-exl2-6.5
70
- huggingface-cli download bartowski/Mistral-quiet-star-exl2 --revision 6_5 --local-dir Mistral-quiet-star-exl2-6.5 --local-dir-use-symlinks False
71
  ```
72
 
73
  Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
 
15
  pipeline_tag: text-generation
16
  ---
17
 
18
+ ## Exllama v2 Quantizations of Mistral-quiet-star-demo
19
 
20
  Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.16">turboderp's ExLlamaV2 v0.0.16</a> for quantization.
21
 
 
23
 
24
  Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
25
 
26
+ Original model: https://huggingface.co/liminerity/Mistral-quiet-star-demo
27
 
28
  | Branch | Bits | lm_head bits | VRAM (4k) | VRAM (16k) | VRAM (32k) | Description |
29
  | ----- | ---- | ------- | ------ | ------ | ------ | ------------ |
30
+ | [8_0](https://huggingface.co/bartowski/Mistral-quiet-star-demo-exl2/tree/8_0) | 8.0 | 8.0 | 8.4 GB | 9.8 GB | 11.8 GB | Maximum quality that ExLlamaV2 can produce, near unquantized performance. |
31
+ | [6_5](https://huggingface.co/bartowski/Mistral-quiet-star-demo-exl2/tree/6_5) | 6.5 | 8.0 | 7.2 GB | 8.6 GB | 10.6 GB | Very similar to 8.0, good tradeoff of size vs performance, **recommended**. |
32
+ | [5_0](https://huggingface.co/bartowski/Mistral-quiet-star-demo-exl2/tree/5_0) | 5.0 | 6.0 | 6.0 GB | 7.4 GB | 9.4 GB | Slightly lower quality vs 6.5, but usable on 8GB cards. |
33
+ | [4_25](https://huggingface.co/bartowski/Mistral-quiet-star-demo-exl2/tree/4_25) | 4.25 | 6.0 | 5.3 GB | 6.7 GB | 8.7 GB | GPTQ equivalent bits per weight, slightly higher quality. |
34
+ | [3_5](https://huggingface.co/bartowski/Mistral-quiet-star-demo-exl2/tree/3_5) | 3.5 | 6.0 | 4.7 GB | 6.1 GB | 8.1 GB | Lower quality, only use if you have to. |
35
 
36
  ## Download instructions
37
 
38
  With git:
39
 
40
  ```shell
41
+ git clone --single-branch --branch 6_5 https://huggingface.co/bartowski/Mistral-quiet-star-demo-exl2 Mistral-quiet-star-demo-exl2-6_5
42
  ```
43
 
44
  With huggingface hub (credit to TheBloke for instructions):
 
47
  pip3 install huggingface-hub
48
  ```
49
 
50
+ To download the `main` (only useful if you only care about measurement.json) branch to a folder called `Mistral-quiet-star-demo-exl2`:
51
 
52
  ```shell
53
+ mkdir Mistral-quiet-star-demo-exl2
54
+ huggingface-cli download bartowski/Mistral-quiet-star-demo-exl2 --local-dir Mistral-quiet-star-demo-exl2 --local-dir-use-symlinks False
55
  ```
56
 
57
  To download from a different branch, add the `--revision` parameter:
 
59
  Linux:
60
 
61
  ```shell
62
+ mkdir Mistral-quiet-star-demo-exl2-6_5
63
+ huggingface-cli download bartowski/Mistral-quiet-star-demo-exl2 --revision 6_5 --local-dir Mistral-quiet-star-demo-exl2-6_5 --local-dir-use-symlinks False
64
  ```
65
 
66
  Windows (which apparently doesn't like _ in folders sometimes?):
67
 
68
  ```shell
69
+ mkdir Mistral-quiet-star-demo-exl2-6.5
70
+ huggingface-cli download bartowski/Mistral-quiet-star-demo-exl2 --revision 6_5 --local-dir Mistral-quiet-star-demo-exl2-6.5 --local-dir-use-symlinks False
71
  ```
72
 
73
  Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski