denru

/

Tess-3-Mistral-Large-2-123B-5_0bpw-h8-exl2-pippa

like 0

Thanks for these quants. These don't fit on 48G VRAM. I'm limited in bandwidth so it's not feasible for me to quant and share unfortunately.
If the size is the same as the original Mistral Large Instruct, then 2.75bpw would be perfect for 48G VRAM and about 12k context. The quality of the original version has been really good for me despite being such a low quant, but I'm curious about this one.

Thanks for considering it :)