INFO:
Breaking cross-modal boundaries in multimodal AI: Introducing CoDi, composable diffusion for any-to-any generation - Microsoft Research->