CoolFace
Modelpublic

Suparious/Zebrafish-7B-exl2

sourceHugging Facecc-by-nc-4.0updated 2y agoView on Hugging Face
0likes10downloads
Model Card

Exllama v2 Quantizations of Zebrafish-7B

Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.17">turboderp's ExLlamaV2 v0.0.17</a> for quantization.

<b>The "main" branch only contains the measurement.json, download one of the other branches for the model (see below)</b>

Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.

Original model: https://huggingface.co/mlabonne/Zebrafish-7B

BranchBitslm_head bitsVRAM (4k)VRAM (16k)VRAM (32k)Description
6_56.58.07.2 GB8.6 GB10.6 GBVery similar to 8.0, good tradeoff of size vs performance, recommended.
5_05.06.06.0 GB7.4 GB9.4 GBSlightly lower quality vs 6.5, but usable on 8GB cards.
4_254.256.05.3 GB6.7 GB8.7 GBGPTQ equivalent bits per weight, slightly higher quality.
3_53.56.04.7 GB6.1 GB8.1 GBLower quality, only use if you have to.

Download instructions

With git:

shell
git clone --single-branch --branch 6_5 https://huggingface.co/suparious/Zebrafish-7B-exl2 Zebrafish-7B-exl2-6_5

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 Zebrafish-7B-exl2:

shell
mkdir Zebrafish-7B-exl2
huggingface-cli download suparious/Zebrafish-7B-exl2 --local-dir Zebrafish-7B-exl2 --local-dir-use-symlinks False

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

Linux:

shell
mkdir Zebrafish-7B-exl2-6_5
huggingface-cli download suparious/Zebrafish-7B-exl2 --revision 6_5 --local-dir Zebrafish-7B-exl2-6_5 --local-dir-use-symlinks False

Windows (which apparently doesn't like _ in folders sometimes?):

shell
mkdir Zebrafish-7B-exl2-6.5
huggingface-cli download suparious/Zebrafish-7B-exl2 --revision 6_5 --local-dir Zebrafish-7B-exl2-6.5 --local-dir-use-symlinks False