Fetching the paper…

Localized Text-to-Image Generation for Free via Cross Attention Control · Around