Text-to-Image
Diffusers
Safetensors
flux
flow-matching
distillation

Will there be ComfyUI support for the Flux.2 model?

#2
by CCP6 - opened

Hey!

I absolutely love your method of distillation, and Flux.2 is a model that even w/ super good consumer hardware is heavy and time consuming to generate an image with. I was super stoked seeing your repo update come across my github feed the other day, but then I figured out that only the location implementation had been upgraded to support it.

Will there be a new node update/workflow to use this model w/in ComfyUI?

Thanks so much for any info!!

(oh and I definitely second the Kandinsky5 request, but I'd be cool w/ any Pi-"Video" distillation tbh and that one seems like it'd be the hardest to pull off due to the massive size 20b - buuuut it's fantastic quality if you haven't used it).

Happy holidays!

-D

Thanks for your interest in our work. Yea there will be a ComfyUI update. I just need some time.

It's done! Checkout the ComfyUI-piFlow update

Sign up or log in to comment