Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Avenstrup, Thor Højhus, Elek, Boldizsár, Mádi, István László, Schin, András Bence, Mørup, Morten, Jensen, Bjørn Sand, Olsen, Kenny Falkær
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2410.20997
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917819363360768
author Avenstrup, Thor Højhus
Elek, Boldizsár
Mádi, István László
Schin, András Bence
Mørup, Morten
Jensen, Bjørn Sand
Olsen, Kenny Falkær
author_facet Avenstrup, Thor Højhus
Elek, Boldizsár
Mádi, István László
Schin, András Bence
Mørup, Morten
Jensen, Bjørn Sand
Olsen, Kenny Falkær
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2410_20997
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SepMamba: State-space models for speaker separation using Mamba
Avenstrup, Thor Højhus
Elek, Boldizsár
Mádi, István László
Schin, András Bence
Mørup, Morten
Jensen, Bjørn Sand
Olsen, Kenny Falkær
Sound
Machine Learning
Audio and Speech Processing
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.
title SepMamba: State-space models for speaker separation using Mamba
topic Sound
Machine Learning
Audio and Speech Processing
url https://arxiv.org/abs/2410.20997