( Already has a VAE )

Recommended Settings are :

• Steps: 30-64

• CFG scale: 7

• Sampler: DPM++ 2M

• ADetailer can fix face, eyes and some other problem

recommended ADetailer :

yolov8s' ADetailer

How to install :

go to the "Extension" tab then "Available" and check if you have already "https://raw.githubusercontent.com/AUTOMATIC1111/stable-diffusion-webui-extensions/master/index.json" then press "Load from:" and search for "ADetailer", install it and put your file previously installed in webui/models/adetailer and reload your stable diffusion ui. finally use it in the txt2img tab.

"anime source" for anime style (Other are for a more realistic style)

Positive Prompt :

score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up

anime source

Negative prompt is not perfect, it can influence the result's quality. (Use your own negative prompt for better results or use negative embedding).

Negative Prompt :

score_4, score_3, score_2, score_1  bad hands, missing fingers, (censor), monochrome, blurry  3d, source_cartoon  text, signature, watermark, username, artist name