Please consider adding support in Diffusers for this.

[tin2tin]

As written in the title, it would be great if this project could have support added in Diffusers. It really needs a good project like this.

[Owen718]

Thank you for your interest! Since our team is limited and we are also pushing forward other research projects, we would be very grateful if someone from the community could help implement ComfyUI support and Diffusers support. If anyone can submit a PR, we would deeply appreciate it. Otherwise, I’ll do my best to make time to complete it myself. Thanks!

[tin2tin]

The model is not interchangeable with vanilla Flux, right? In a img2img workflow, I mean.

[Owen718]

Yes—it's not a drop-in replacement. You’d need to adapt the existing Flux components and build a new inference pipeline.