shinka nibutani from Chūnibyō Demo Koi ga Shitai!

trained with anything4.5 on 21 images. most were her school uniform. some of the images were doujinshi covers so putting text in the negative prompt should get rid of it.

just using the base trigger word usually makes he generation look like her, but in order to get the different outfits using the prompts from my images should help or describe the outfit. im not sure if if putting the outfit in parentheses helps or not.