This is an experimental LoRA model, and it was trained with images of the following characters: Lillie, Akari, Dawn, Hilda, Lana, Marnie, May, Misty, Rosa, and Serena. However, May and Misty are not working well for me.
I personally find it funny, sometimes it generates good images, but other times it simply ignores your prompt and generates a different character. For example, I tried to generate an image of Akari, and it generated one of Marnie instead. This happens especially if you don't use enough images to describe the character. I think this is probably because the text encoder wasn't trained.
Important: The triggers are not working well. I recommend using them but also adding as many tags as possible to describe each character. I recommend using this tool for that: https://huggingface.co/spaces/SmilingWolf/wd-tagger
This LoRA model was trained using booru tags at a resolution of 512px.
Commissions:
https://ko-fi.com/bionagato/commissions
Visit my Ko-Fi page if you want to support me:
Alternatively, you can find me on Buy Me a Coffee at: