I’m also suffering from having set up a project that used the reliable, consistent style of the old art model, and I’m struggling hard to re-produce it with the new model.
Has anyone seen success with reliably reproducing the style of “Anime” and “Drawn Anime” in the old style? Any prompts and seeds to share?
I know of tools that can generate prompts out of images, but even with very refined prompting, I rarely get what I need and have no consistency in the art style with the new model.
What I want is to reliably reproduce the visuals of the old “Anime” images. Also, is there information available on what exactly the old model was so I might be able to set it up myself? I specific stable diffusion model?
Thank you for sharing. The results appear a bit more consistent, but still too far away from my previous material created with the old “Anime”-style. I’ll tryprompting more on that one and see how far I get.
Good luck! I’m still refining the art styles and getting better results as time goes on. Hopefully the model continues to be updated as the devs promised. Should improve things further…
Thank you, I’ll keep on trying.