SepMamba: State-space models for speaker separation using Mamba
Abstract
Deep learning-based single-channel speaker separation has improved significantly in recent years largely due to the introduction of the transformer-based attention mechanism. However, these improvements come at the expense of intense computational demands, precluding their use in many practical applications. As a computationally efficient alternative with similar modeling capabilities, Mamba was recently introduced. We propose SepMamba, a U-Net-based architecture composed primarily of bidirectional Mamba layers. We find that our approach outperforms similarly-sized prominent models - including transformer-based models - on the WSJ0 2-speaker dataset while enjoying a significant reduction in computational cost, memory usage, and forward pass time. We additionally report strong results for causal variants of SepMamba. Our approach provides a computationally favorable alternative to transformer-based architectures for deep speech separation.
Cite
@article{arxiv.2410.20997,
title = {SepMamba: State-space models for speaker separation using Mamba},
author = {Thor Højhus Avenstrup and Boldizsár Elek and István László Mádi and András Bence Schin and Morten Mørup and Bjørn Sand Jensen and Kenny Falkær Olsen},
journal= {arXiv preprint arXiv:2410.20997},
year = {2024}
}