Description
This lora is suitable for use in scenes such as decorative paintings, posters, picture books, mobile phone wallpaper, and gongbi art
After testing and comparison, Lora effectively increased two-dimensional facial details. If some model faces are not good, it can be used together with this model. When the weight is 0.4 or less, the facial details improve significantly when the weight does not affect the original model's style too much.
Recommended parameters:
Bottom mold suggestion: basic algorithm SDXL
Wanxiang XL Advanced Edition_V70 Foliage-SDXL
Sampling method: Eulera a number of iteration steps: 20
is used to show the image
CFG:7 Weights:
The performance is good between 0.4-0.8 (0.8 is recommended) Picture size: Recommended 512*768
(Recommended resolution repair, 20 iteration steps, 0.3-0.5 redraw amplitude
)Welcome back and follow, we will continue to update in the future
.





