Orca-2-7B-16k-exl2 / README.md
bartowski's picture
Add 4.0 link
1401b95
---
pipeline_tag: text-generation
tags:
- orca
- orca2
- microsoft
license: other
license_name: microsoft-research-license
license_link: LICENSE
quantized_by: bartowski
---
## Exllama v2 Quantizations of Orca-2-7B-16k
Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.9">turboderp's ExLlamaV2 v0.0.9</a> for quantization.
Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
Conversion was done using wikitext-103-raw-v1-test.parquet as calibration dataset.
Default arguments used except when the bits per weight is above 6.0, at that point the lm_head layer is quantized at 8 bits per weight instead of the default 6.
Original model: https://huggingface.co/NurtureAI/Orca-2-7B-16k
<a href="https://huggingface.co/bartowski/Orca-2-7B-16k-exl2/tree/4_0">4.0 bits per weight</a>
<a href="https://huggingface.co/bartowski/Orca-2-7B-16k-exl2/tree/5_0">5.0 bits per weight</a>
<a href="https://huggingface.co/bartowski/Orca-2-7B-16k-exl2/tree/6_0">6.0 bits per weight</a>
<a href="https://huggingface.co/bartowski/Orca-2-7B-16k-exl2/tree/8_0">8.0 bits per weight</a>
## Download instructions
With git:
```shell
git clone --single-branch --branch 4_0 https://huggingface.co/bartowski/Orca-2-7B-16k-exl2
```
With huggingface hub (credit to TheBloke for instructions):
```shell
pip3 install huggingface-hub
```
To download the `main` (only useful if you only care about measurement.json) branch to a folder called `Orca-2-7B-16k-exl2`:
```shell
mkdir Orca-2-7B-16k-exl2
huggingface-cli download bartowski/Orca-2-7B-16k-exl2 --local-dir Orca-2-7B-16k-exl2 --local-dir-use-symlinks False
```
To download from a different branch, add the `--revision` parameter:
```shell
mkdir Orca-2-7B-16k-exl2
huggingface-cli download bartowski/Orca-2-7B-16k-exl2 --revision 4_0 --local-dir Orca-2-7B-16k-exl2 --local-dir-use-symlinks False
```