Fetching the paper…

BoxDiff: Text-to-Image Synthesis with Training-Free Box-Constrained Diffusion · Around