Specifically designed to add texture details to the face, try to avoid input images with dark filters or blurriness. No trigger words (such as adding facial details to characters or adding realistic facial textures to characters, English prompts may work better).
Lora strength: 0.7 (adjust according to your needs, maximum 1; the higher the strength, the more texture, and vice versa, the less texture).
Steps: 8 (the higher the steps, the more texture, and vice versa, the less texture).
It also has some effects on anime characters, but the effectiveness on anime characters has not been deeply tested.
There are two versions of Lora in total:
First version: https://www.runninghub.cn/model/public/2069468964757004290
Second version: https://www.runninghub.cn/model/public/2072311076204662785
Detailed information is provided.
Workflow address: https://www.runninghub.cn/post/2072296887910232066