GGUF?
#2
by
MoonRide
- opened
CFT looks interesting, but it would be nice to have some GGUFs for it, so we can do quick local evaluations. IQ3_XS+ quants are quite usable, and allow to run ~30B models on a 16 GB VRAM GPUs (with reduced context size). @bartowski ?
@bartowski Thank you very much! <3
MoonRide
changed discussion status to
closed