no code implementations • 4 Oct 2022 • Haoyu Pan, Hao Wu, Tan Yang
In this paper, a modular design is proposed, which decomposes spatiotemporal sequence model into two modules: a spatial encoder-decoder and a predictor.
