ComfyUI Fp8 loading

#1
by MayensGuds - opened

how can we do that

https://blog.comfy.org/ltxv-day-1-comfyui/

@Symphone
Thanks a lot. I finally can use this model with no OOM.

Can we use it on original inference : https://github.com/Lightricks/LTX-Video/blob/main/inference.py ? (not comfy ui)

Can we use it on original inference : https://github.com/Lightricks/LTX-Video/blob/main/inference.py ? (not comfy ui)

Yes you can with a lot of ram/vram.

Sign up or log in to comment