The model I think is fairly self explanatory.
I'm not currently prefectly aware of how to generate images effectively with Illustrious, seems to run a bit different than my usual Autismmix Confetti settings, but I managed to generate some acceptable previews, the settings you can see on the image generation data. Works fine at a strength of 1, previews made with NoobAI-XL (NAI-XL) 1.0.
Edited for 2.0 version, I guess I figured it out in the end. 2.0 version previews made with NovafurryV40
Description
FAQ
Comments (2)
Quality of the v4 is better but it seems to have severely deviated from the artist.
Was trying out a different optimizer, and I think the problem is that I used way too few repeats to really bake the style in, I guess 5 is out, 10 doesn't seem entirely reliable. 15 seems usually good enough for regular hentai with my regular settings. Not sure for furry. I just tried to do a model for a Doxy comic at 20 repeats with my usual settings, but maybe it's because it has too many closeups and extreme positions but it didn't really work at all. Trying to figure out furry stuff, sometimes it works well, but I guess I have to keep in mind how clear the art is too from how much action is going on. I'll put a pause on tails for a bit since I'd want to come back to a v5 after I've sorted that out and not experiment too much like I did for Magnetar a few months ago, haha