!!! LoRA can be difficult to use !!!

LoRA is trained on the artworks of the wonderful illustrator Yun Ling.

• Recommended weight:

• For NoobAI XL: is 0.8-1. 0.9 is good spot to start.

• For WAI-NSFW-illustrious-SDXL v8: 0.6-0.9. 0.7 is good spot to start.

• Hires fix provides quality.

• Although LoRA doesn't require trigger word, I recommend using the artist tag “Yun Ling” for better stability of the style. Especially for WAI.

• The following tags may be used optionally: “wide shot, cinematic, perspective, vanishing point, full body, reflection, sunlight”

• Because the artist draws often a character with a bow and arrow, they may appear from time to time. In this case, just add “(bow \(weapon\):1.1)” to the negative prompt.

Have fun!

描述:

Trained on NoobAI-XL Epsilon-pred 1.1-Version.

训练词语: