bartowski's picture
Quant for 4.0
6d5bc01
|
raw
history blame
1 kB
---
license: other
language:
- en
base_model: microsoft/Orca-2-13b
datasets:
- HuggingFaceH4/no_robots
quantized_by: bartowski
---
# Exllama v2 Quantizations of Orca-2-13B-no_robots at 4.0 bits per weight
Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.9">turboderp's ExLlamaV2 v0.0.9</a> for quantization.
Conversion was done using wikitext-103-raw-v1-test.parquet as calibration dataset.
Original model: https://huggingface.co/Locutusque/Orca-2-13B-no_robots
## Download instructions
With git:
```shell
git clone --single-branch --branch 4.0 https://huggingface.co/bartowski/Orca-2-13B-no_robots-exl2
```
With huggingface hub (credit to TheBloke for instructions):
```shell
pip3 install huggingface-hub
```
To download from a different branch, add the `--revision` parameter:
```shell
mkdir Orca-2-13B-no_robots-exl2
huggingface-cli download bartowski/Orca-2-13B-no_robots-exl2 --revision 4_0 --local-dir Orca-2-13B-no_robots-exl2 --local-dir-use-symlinks False
```