It's a hybrid of "Illustrious" and "SDXL" (although it's more "Illustrious"), so some lora may not work. It only works properly with the "LCM" sampler. Good results can be achieved using: "euler_a/karras 6-8 Steps, lcm/beta 6-10 steps" (applies to turbo version only). The emphasis was on realism. The settings will be in the picture.
Description
FAQ
Comments (7)
Whatever you're doing, keep it up! I've been in love with Sam Universal and this is just as good.
Great model - it handles realistic LoRAs a lot better than most IL-based checkpoints I've tried. Any chance we could get a non-accelerated version for easier LoRA training?
I think there's scope for releasing a regular version. If necessary, I'll publish it later.
@toya_san I would very much like to see a non-turbo version aswell, just for general use
@gooners_united Okay, but I'll tell you right away that it's not configured and the quality won't be very good.
The "Lora" will have to adjust the quality
@toya_san That was quick! Thank you!