Trained from 20 images in OVA video.
Clothes and body shape are not trained, sorry.
SDXL_Melvina_MacGarlen_v1.safetensors