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: <a target="_blank" rel="ugc" href="https://drive.google.com/file/d/1yVWkDPzAGLN3CeTsCOQzTKhjMWaOA8wu/view?usp=sharing">https://drive.google.com/file/d/1yVWkDPzAGLN3CeTsCOQzTKhjMWaOA8wu/view?usp=sharing</a> New images I trained on can be found here: <a target="_blank" rel="ugc" href="https://www.deviantart.com/333aaa333aaa333aaa/favourites">https://www.deviantart.com/333aaa333aaa333aaa/favourites</a> Suggested Prompt: JP_Tour, furry, feral, chibi, cat and so on Suggested Negative Prompt: eyelashesInterestingly, the model seems capable of generating Ruby even without specific prompts (as seen here: <a target="_blank" rel="ugc" href="https://tensor.art/images/1045292604071085926?post_id=1045292468779621688">https://tensor.art/images/1045292604071085926?post_id=1045292468779621688</a> or <a target="_blank" rel="ugc" href="https://civarchive.com/posts/31078016">https://civarchive.com/posts/31078016</a>), though I'm not entirely sure why this happens!

Tour(Jewelpet)

It uses the same training dataset as this model: <a target="_blank" rel="ugc" href="https://civitai.red/models/2908769/tourjewelpet?modelVersionId=3289879">https://civitai.red/models/2908769/tourjewelpet?modelVersionId=3289879</a>, but with captions generated by Gemini.

jpTourAnimaV01_v1.safetensors

v0.1 Amina

v0 Amina

It uses the same training dataset as this model: <a target="_blank" rel="ugc" href="https://civarchive.com/models/2908769/tourjewelpet?modelVersionId=3289879">https://civarchive.com/models/2908769/tourjewelpet?modelVersionId=3289879</a>, but with captions generated by Gemini. I'm still not sure why the quality isn't as good.

Files with SHA256: 95598eec69f63e8683eca8bc714f90e022a5fa8ca0977ec90587cb857abd5e94