You can click "show more" to view the english introduction, it follows the chinese introduction.

此lora献给最美的绫华大小姐~

权重建议0.5~0.85，想换衣服的话就把权重调低

CSG Scale~3

使用高分重绘生成图像，下面是示例

建议DPM++ SDE Karras采样，用别的采样方法可能不太好换动作和衣服什么的

服装词：short blue hair/short hair, pillbox hat/yellow pillbox hat, beautiful blue eyes, european style long dress, long white sleeves,white legwear, brwon shoes/brown boots. 如果哪个不明显就增加哪个的权重

全身图都采用了SD Upscale进行了二次加工，所以脸部细节才会这么丰富，SD Upscale在图生图的脚本里开启，下面是示例：(重绘幅度越高效果越好，但要在0.4以下)

此lora应该适用于大部分动漫模型，在2.5D真人模型里表现也不错，比如perfectworld、yorrrlmix

真人图的模型是yorrrlmix, 不是chilloutmix

Set this lora's weight from 0.5~0.85, give it less weights if you want to change her clothes and postures.

CFG Scale~3

I recommend using hires. fix to generate images, the following pic is for reference

Better use DPM++ SDE Karras, it might be a little difficult to change her postures if you use other samplers.

Clothes words：short blue hair/short hair, pillbox hat/yellow pillbox hat, beautiful blue eyes, european style long dress, long white sleeves,white legwear, brwon shoes/brown boots. If the clothing features are not obvious enough, increase their weights appropriately

All full-body iamges have been processed by SD upcale, that's the reason why the faces in full-body images are so rich in detail, but don't set denoising above 0.4

I used yorrrlmix to get the real person example, not chilloutmix