Due to my mistake, I didn't set up trigger words for this model when preparing the training data. However, I added the tag 'JP_Tour' to every image as a trigger word, so prompting 'JP_Tour' might work.

I added several images (mainly derivative works from Deviantart) and updated the tags. It seems a bit better than my Anima model, but not good enough. Though I introduced a few images for Tour X Ruby, don't expect this model to generate them. There is no hope at the moment. Similarly, I've posted the training data for research here: https://drive.google.com/file/d/1yVWkDPzAGLN3CeTsCOQzTKhjMWaOA8wu/view?usp=sharing
New images I trained on can be found here: https://www.deviantart.com/333aaa333aaa333aaa/favourites

Suggested Prompt: JP_Tour, furry, feral, chibi, cat and so on

Suggested Negative Prompt: eyelashes

Interestingly, the model seems capable of generating Ruby even without specific prompts (as seen here: https://tensor.art/images/1045292604071085926?post_id=1045292468779621688 or https://civarchive.com/posts/31078016), though I'm not entirely sure why this happens!