Since I made this for myself, I can’t really cut corners, can I? So, I went back to the drawing board and started over with selecting the dataset. I’m keeping Version 1 around since the facial expressions and such have changed from it, but the rework has fixed issues like the hair thickness being completely different on the left and right sides, and the background and angles not matching the prompts.

In accordance with the guidelines of Japan’s Agency for Cultural Affairs, the training dataset for this LoRA does not consist solely of artwork by a specific artist. This LoRA cannot replicate the original artwork. Please use it for fan art purposes only.

recommended weight: 1

trigger words: satomura akane, blonde hair, green eyes
serafuku, red neck ribbon, yellow cardigan, red skirt, back bow