【AI繪畫】新手訓練圖片 看一下就好

下面是我個人主要的一些參數在stable stable diffusion中用到

主要模型:

AC_H-3, BY_A_RE-1

Sampler:

DPM++ 2M Karras, UniPC

其中一部分的Tag:

"1girl, solo, flat chest, long hair, dress, white dress, smile, blush, hair(cyan:0.2, white:0.8), blue eyes, very long hair, bare shoulders, closed mouth, sleeveless, frills, sleeveless dress, star (sky), floating hair"

Negative prompt:

"EasyNegative,extra fingers,fewer fingers,watermark,Invisible watermark,username,Signature,jpeg artifacts,Bad feet,Bad hands,Extra legs,Three Legs,Three feet,(NSFW),Multiple people,More than one person,2girl,More than two legs,"

尺寸：

768*1024

提示詞引導係數 (CFG Scale):

4.5 / 7.5

重繪幅度: 

0.7