Pretty difficult to use. Please start to use this LoRA from the weight 0.5 or around.
Trained with very blurred 33 images in 830 steps.
SDXL_Amachi_Hiromi_v1.safetensors