In recent years, visually-rich document understanding has attracted increasing attention. Transformer-based pre-trained models have become the mainstream approach, yielding significant performance gains in this field. However, the self-attention mechanism's quadratic computational complexity hinders their efficiency and ability to process long documents. In this paper, we present DocMamba, a novel framework based on the state space model. It is designed to reduce computational complexity to linear while preserving global modeling capabilities. To further enhance its effectiveness in document processing, we introduce the Segment-First Bidirectional Scan (SFBS) to capture contiguous semantic information. Experimental results demonstrate that DocMamba achieves new state-of-the-art results on downstream datasets such as FUNSD, CORD, and SORIE, while significantly improving speed and reducing memory usage. Notably, experiments on the HRDoc confirm DocMamba's potential for length extrapolation.
View on arXiv@article{hu2025_2409.11887,
title={ DocMamba: Efficient Document Pre-training with State Space Model },
author={ Pengfei Hu and Zhenrong Zhang and Jiefeng Ma and Shuhang Liu and Jun Du and Jianshu Zhang },
journal={arXiv preprint arXiv:2409.11887},
year={ 2025 }
}