Decided to give this a try based on some new Anima style training settings I found. It's a bit hard to get it to work reliably and some random terms seems to break it but it can work and it's cool when it does. You might need to up the LoRA strength when using other LoRAs or certain terms. The dataset is mostly based on the work of Katsushika Hokusai if anyone is curious.Trigger word: ukiyo-e_styleSide note: I do take informal requests(basically a bounty that you don't have to pony up buzz for) if I like the idea(Tips are nice but not required since if it's a good idea then I'll get generation Buzz). Just leave a comment on one of my models or DM me. (Comments are preferred since that let's other people see that I'm actually fulfilling informal requests.) DMs are perfectly acceptable if your request is a bit out there and needs some image examples or something. As a note I will probably pass on Manhwa characters because I generally find their designs to be kind of bland.(No shade to Manhwa creators, I know many of them are going for regular releases and intricate designs are not what you want in the case.)

Files with SHA256: 8ce6520da9e1eae97755328cb17fb666b0ae0176c873367d1feed6f36d978e91