Trained on around 70 screenshots from DOA5LR (2015). Not trained on all characters, and not all were tagged completely, but it can help to reinforce a few characters or designs Anima doesn't have a full grasp of, like Mila \(DOA\), Helena Douglas, and Tina Armstrong.
No trigger word, weight's usually good around 1. For characters outside of DOA, adding "3d" and upping the weight to 1.1 can help reinforce the style.
Description
FAQ
Comments (3)
sooooooo great!❤️
Instant classic! I'm sure you used DOAHDM pics lol right?
Yeah, I did. I don't think this LoRA would be capable of generating characters not in a fighting stance or action shot if I didn't have access to a better posing/photomode tool like Gravure Studio and Autolink. Made it more fun to get the screenshots too