Text-to-Image Alignment in Denoising-Based Models through Step Selection

DiffM
EGVM
Re-assign community