Updated: Jan 31, 2025
concept
This embedding can be used to get improved results for female faces. The embedding causes stable diffusion to concentrate more on the face. Depending on the checkpoint, a weight between 0.5 and 3 is possible. Like every embedding it will not just change the face, but the whole image when used with text2img.
When using Loras, then often the weight has to be lowered to 0.3 to 1.5 or results will get worse instead of better.
The embedding also works well with img2img.