Hi! Inspired by SuceSuce's LoRA, I trained my own LoRA of Komatsuzaki Rui artstyle, but for SDXL!
This was my first time training LoRA, so I don't know if it turned out good or bad. Needs more testing!
============================
Trained on a base SDXL1.0 model.
Works OK-ish on a base model, works better on other models with better knowledge of human anatomy.
Try putting into positive or negative prompt "thick lines", "hair reflection", "texture"; Those sometimes could give you a look closer to DR1 or DRV3 depending on usage - more of a sketch-y feeling or clearer and sharper picture with thinner lines. I attached a comparison picture above.
Tests and feedback are appreciated!
============================
