A re-trained LORA trained on Yorihime Nao from Yosuga No Sora from 149 images from both the anime and visual novel. It can do both SFW and NSFW images.

It prompts quite well, use yorihime_nao as a trigger word. Tags such as hairclip, large breasts and glasses will help further steer the model to generating her likeness if it doesn't work with just the character name.

Images representing styles (see first 6 image examples in order)

Anime style - use visual_novel in negative tag

Anime + VN blended style - Just prompt as normal.

Visual Novel style - use visual_novel in positive tag

My models take a lot of work to build image datasets by going through frame by frame, building datasets, stitching images and using photoshop to fix images to build the best models possible. I ask no money of this but would appreciate people to take time to rate my models and create your own image posts to show that they are useful to you. I only take model requests if I have time and energy to do so.