This is a LoRA training test using only 22 images as the dataset.
Nazuna_Hiwatashi_moss_image-000045.safetensors