mahou shoujo ai style

使用动漫截图训练，由于年代较早清晰度一般。素材主要来自于动漫中的女孩，有些过拟合可以适当降低强度。建议Negative prompt使用text,english text,pillarboxed。示例图片使用了 ADetailer。可以和其他艺术家tag一起使用来增加多样性。

The English part uses Google Translate. The training is done with anime screenshots, which are not very clear due to their age. The material mainly comes from girls in anime, which is a bit overfitted and can be reduced appropriately. It is recommended to use text, english text, pillarboxed for Negative prompt. The sample image uses ADetailer. It can be used with other artist tags to increase diversity.