bartowski commited on
Commit
5ab5166
·
1 Parent(s): 8415b99

Main branch

Browse files
Files changed (2) hide show
  1. README.md +43 -0
  2. measurement.json +0 -0
README.md ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ inference: false
4
+ quantized_by: bartowski
5
+ ---
6
+
7
+ ## Exllama v2 Quantizations of MistralLite
8
+
9
+ Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.7">turboderp's ExLlamaV2 v0.0.7</a> for quantization.
10
+
11
+ Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
12
+
13
+ Conversion was done using wikitext-103-raw-v1-test.parquet as calibration dataset.
14
+
15
+ Original model: https://huggingface.co/amazon/MistralLite
16
+
17
+ ## Download instructions
18
+
19
+ With git:
20
+
21
+ ```shell
22
+ git clone --single-branch --branch 4.0 https://huggingface.co/bartowski/MistralLite-exl2
23
+ ```
24
+
25
+ With huggingface hub (credit to TheBloke for instructions):
26
+
27
+ ```shell
28
+ pip3 install huggingface-hub
29
+ ```
30
+
31
+ To download the `main` (only useful if you only care about measurement.json) branch to a folder called `MistralLite-exl2`:
32
+
33
+ ```shell
34
+ mkdir MistralLite-exl2
35
+ huggingface-cli download bartowski/MistralLite-exl2 --local-dir MistralLite-exl2 --local-dir-use-symlinks False
36
+ ```
37
+
38
+ To download from a different branch, add the `--revision` parameter:
39
+
40
+ ```shell
41
+ mkdir MistralLite-exl2
42
+ huggingface-cli download bartowski/MistralLite-exl2 --revision 4.0 --local-dir MistralLite-exl2 --local-dir-use-symlinks False
43
+ ```
measurement.json ADDED
The diff for this file is too large to render. See raw diff