底膜推荐：AnythingV5V3_v5v3(或首选动漫模型）
迭代步数：20
采样方法：Euler a
训练尺寸：512*768
高清修复：R-ESRGAN 4x+ Anime6B
推荐权重：1
提示词举例：
正：xiantiaohua, solo, book, glasses, 1boy, male focus, smile, hood, sitting, holding book, holding, hoodie, pants, looking at viewer, crossed legs, white background, short hair
反：low quality,bad anatomy,extra fingers,mutated hands,blurry,watermark,text,lowres,ugly,deformed,