Example images are normal txt2img 896x1152 without the use of negatives, refiner or other improvements.
HOW TO USE:
Refiner is unnecessary and VAE is included (sdxl_vae.safetensors)
MAIN SAMPLER (Recommended): dpmpp_3m_sde + exponential
OTHER SAMPLERS: dpmpp_3m_sde + karras \ dpmpp_sde + Karras
FOR MORE RALISM: CFG: 2 or 3 | STEPS: 12 to 24
FOR BALANCING (Recommended): CFG: 4 | STEPS: 12 to 40
FOR MORE CREATIVITY: CFG: 5 or 6 | STEPS: 28 to 40
Description
FAQ
Comments (6)
I am getting a "size mismatch" error.
With only this information I can't help you. What interface do you use to generate? In what situations does it give you this error? Thanks for the feedback, if you provide me with some extra information I will try to help you.
H! I am sorry for not answering earlier but I was away from the site. I am afraid that I can't provide more information as I got other issues due to software updates and me having an AMD card. I had to restore an old SD directory to have things work.
can it do lineart like Franco Grignani? I haven't found a single model that can do it yet..
He is able to do line art illustrations, but I don't think he can replicate Franco Grignani's style. I will try to implement this in future versions as it is a very interesting design style. Thanks for your advice. Have a good time.
@ALIENHAZEÂ yeah, but be warned . Franco grignani, is what i call "Hard lineart". It's almost as hard to generate as text. But maybe you will find a trick to achieve it, keep us updated. Have a good day you too friend! :D