3.0bpw quant?

#2
by narpas - opened

Since you got a few of these up, would you be willing to create and upload a 3.0bpw quant for this? It would fit great on a 120GB rig.
I'd do it myself but the largest GPU I have is 24GB which just isn't big enough to quant this big boy.

I will try when I can! I had to delete the FP16 as it was 600GB and I needed the space.

Thanks! And if someone else comes across this and it doesn't yet exist, I'd appreciate it too.

narpas changed discussion status to closed

Sign up or log in to comment