(AI) Yoruna Mishigure #1

Very tough for the (near) last character to train. Spent a good few hours on just this image polishing as much as I could and failing on some (like the sleeve) but it was the best generation I managed out of hundreds before img2img'ing. R-18 stuff may be easier, who knows.

Thanks user/5153840 for providing the dataset for Louis and Yorna. It would've been much harder without it.

Going to be slowing down posts, retraining some characters and putting together a plan for opening commissions. In the meantime I'm going to try picking up photoshop and taking advantage of the fact the AI can learn what you want it to do with tags and editing images manually (for more advanced positions you cannot naturally generate and cleanup).