Fetching the paper…

Improving Compositional Text-to-image Generation with Large Vision-Language Models · Around