Increased the dataset to 166 images and improved tagging.
Made LORA's style as close to the original style as possible
使用艺术家“Ratatatat74”的图像进行风格训练
为获得更好效果，请使用HiresFix和Adetailer
