video generative model

RA VEN is a novel unconditional video generative model designed to address long-term spatial and temporal dependencies. It uses a hybrid explicit-implicit tri-plane representation to model an entire video sequence….