1 paper
Yohai Mazuz, Janna Bruner, Lior Wolf
In text-to-image models, consistent character generation is the task of achieving text alignment while maintaining the subject's appearance across different prompts. However, since…