SynCD: Generating Multi-Image Synthetic Data for Text-to-Image Customization

Can we generate a training dataset of the same object in different contexts for customization? Check out our work SynCD, which uses Objaverse assets and shared attention in text-to-image models for the same.
cs.cmu.edu/~syncd-proje...
w/ Xi Yin, @junyanz.bsky.social, Ishan Misra, and Samaneh Azadi