yamada ryou (sdxl)

Used chatgpt made prompts and turned off gradient checkpointing. Removed and added some pictures and had a dataset of 72 images. Older model does better at anatomy but was off quite a bit with style and the character being there. Captioned the normal clothes of the character as well so you would have to prompt them out:

-black v-neck shirt

-white collared shirt underneath

-black bow

-skirt

Choose whichever model you like