It works.
#1
by javawock7618 - opened
Nice work. I'm using Ref2VA with FL2VA model and Ref-LoRA, and this 3step LoRA works really well. Thanks.
are you using 3 sampling steps?
This comment has been hidden (marked as Off-Topic)
are you using 3 sampling steps?
No, I mainly use 2-stage sampling. For example, I do 3 steps at 480p with this LoRA, followed by another 3 steps with Latent Upscale. I’ve published my workflow, so feel free to give it a try: Hugging Face workflow collection