The Missing U for Efficient Diffusion Models

arXiv:2310.20092v4 Announce Type: replace
Abstract: Diffusion Probabilistic Models stand as a critical tool in generative modelling, enabling the generation of complex data distributions. This family of generative models yields record-breaking performance in tasks such as image synthesis, video generation, and molecule design. Despite their capabilities, their efficiency, especially in the reverse process, remains a challenge due to slow convergence rates and high computational costs. In this paper, we introduce an approach that leverages continuous dynamical systems to design a novel …

abstract arxiv breaking capabilities cs.cv cs.lg data design diffusion diffusion models efficiency enabling family generative generative models image modelling performance process synthesis tasks tool type video video generation