1 paper
Haonan Lin, Mengmeng Wang, Yan Chen +6
While large-scale pre-trained text-to-image models can synthesize diverse and high-quality human-centered images, novel challenges arise with a nuanced task of "identity fine editi…