VMRNN: Integrating Vision Mamba and LSTM for Efficient and Accurate Spatiotemporal Forecasting

TL;DR

This paper introduces VMRNN, a novel recurrent architecture combining Vision Mamba blocks with LSTM to improve efficiency and accuracy in spatiotemporal forecasting, addressing limitations of CNNs and ViTs.

Contribution

The paper presents the VMRNN cell, integrating Vision Mamba blocks with LSTM, offering a new approach for effective and efficient spatiotemporal prediction.

Findings

Achieves competitive results on various spatiotemporal tasks.

Maintains smaller model size compared to existing methods.

Demonstrates superior long-sequence modeling capabilities.

Abstract

Combining CNNs or ViTs, with RNNs for spatiotemporal forecasting, has yielded unparalleled results in predicting temporal and spatial dynamics. However, modeling extensive global information remains a formidable challenge; CNNs are limited by their narrow receptive fields, and ViTs struggle with the intensive computational demands of their attention mechanisms. The emergence of recent Mamba-based architectures has been met with enthusiasm for their exceptional long-sequence modeling capabilities, surpassing established vision models in efficiency and accuracy, which motivates us to develop an innovative architecture tailored for spatiotemporal forecasting. In this paper, we propose the VMRNN cell, a new recurrent unit that integrates the strengths of Vision Mamba blocks with LSTM. We construct a network centered on VMRNN cells to tackle spatiotemporal prediction tasks effectively. Our extensive evaluations show that our proposed approach secures competitive results on a variety of tasks while maintaining a smaller model size. Our code is available at https://github.com/yyyujintang/VMRNN-PyTorch.