Branch	Bits	lm_head bits	VRAM (4k)	VRAM (16k)	VRAM (32k)	Description
6_5	6.5	8.0	28.9 GB	31.6 GB	35.6 GB	Near unquantized performance at vastly reduced size, recommended.
4_25	4.25	6.0	19.5 GB	22.2 GB	26.2 GB	GPTQ equivalent bits per weight, slightly higher quality.
3_5	3.5	6.0	16.5 GB	19.2 GB	23.2 GB	Lower quality, only use if you have to.
3_0	3.0	6.0	14.3 GB	17.0 GB	21.0 GB	Very low quality, usable with 16gb of VRAM.

Download instructions

With git:

git clone --single-branch --branch 6_5 https://huggingface.co/bartowski/airoboros-34b-3.2-exl2 airoboros-34b-3.2-exl2-6_5

With huggingface hub (credit to TheBloke for instructions):

pip3 install huggingface-hub

To download the main (only useful if you only care about measurement.json) branch to a folder called airoboros-34b-3.2-exl2:

shell

mkdir airoboros-34b-3.2-exl2
huggingface-cli download bartowski/airoboros-34b-3.2-exl2 --local-dir airoboros-34b-3.2-exl2 --local-dir-use-symlinks False

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

Linux:

shell

mkdir airoboros-34b-3.2-exl2-6_5
huggingface-cli download bartowski/airoboros-34b-3.2-exl2 --revision 6_5 --local-dir airoboros-34b-3.2-exl2-6_5 --local-dir-use-symlinks False

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

shell

mkdir airoboros-34b-3.2-exl2-6.5
huggingface-cli download bartowski/airoboros-34b-3.2-exl2 --revision 6_5 --local-dir airoboros-34b-3.2-exl2-6.5 --local-dir-use-symlinks False

Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski