README.md

Exllama v2 Quantizations of mistral_7b_norobots at 8.0 bits per weight

Using turboderp's ExLlamaV2 v0.0.8 for quantization.

Conversion was done using wikitext-103-raw-v1-test.parquet as calibration dataset.

Original model: https://huggingface.co/TheBloke/mistral_7b_norobots-fp16

Download instructions

With git:

git clone --single-branch --branch 8.0 https://huggingface.co/bartowski/mistral_7b_norobots-exl2

With huggingface hub (credit to TheBloke for instructions):

pip3 install huggingface-hub

To download from a different branch, add the --revision parameter:

mkdir mistral_7b_norobots-exl2
huggingface-cli download bartowski/mistral_7b_norobots-exl2 --revision 8_0 --local-dir mistral_7b_norobots-exl2 --local-dir-use-symlinks False