Saved in:
Bibliographic Details
Main Authors: Wei, Jui-Chiang, Lin, Yi-Cheng, Ritter-Gutierrez, Fabian, Lee, Hung-yi
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.07237
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915336349024256
author Wei, Jui-Chiang
Lin, Yi-Cheng
Ritter-Gutierrez, Fabian
Lee, Hung-yi
author_facet Wei, Jui-Chiang
Lin, Yi-Cheng
Ritter-Gutierrez, Fabian
Lee, Hung-yi
contents Real-world audio often mixes speech and music, yet models typically handle only one domain. This paper introduces a multi-teacher distillation framework that unifies speech and music models into a single one while significantly reducing model size. Our approach leverages the strengths of domain-specific teacher models, such as HuBERT for speech and MERT for music, and explores various strategies to balance both domains. Experiments across diverse tasks demonstrate that our model matches the performance of domain-specific models, showing the effectiveness of cross-domain distillation. Additionally, we conduct few-shot learning experiments, highlighting the need for general models in real-world scenarios where labeled data is limited. Our results show that our model not only performs on par with specialized models but also outperforms them in few-shot scenarios, proving that a cross-domain approach is essential and effective for diverse tasks with limited data.
format Preprint
id arxiv_https___arxiv_org_abs_2506_07237
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Multi-Distillation from Speech and Music Representation Models
Wei, Jui-Chiang
Lin, Yi-Cheng
Ritter-Gutierrez, Fabian
Lee, Hung-yi
Audio and Speech Processing
Real-world audio often mixes speech and music, yet models typically handle only one domain. This paper introduces a multi-teacher distillation framework that unifies speech and music models into a single one while significantly reducing model size. Our approach leverages the strengths of domain-specific teacher models, such as HuBERT for speech and MERT for music, and explores various strategies to balance both domains. Experiments across diverse tasks demonstrate that our model matches the performance of domain-specific models, showing the effectiveness of cross-domain distillation. Additionally, we conduct few-shot learning experiments, highlighting the need for general models in real-world scenarios where labeled data is limited. Our results show that our model not only performs on par with specialized models but also outperforms them in few-shot scenarios, proving that a cross-domain approach is essential and effective for diverse tasks with limited data.
title Multi-Distillation from Speech and Music Representation Models
topic Audio and Speech Processing
url https://arxiv.org/abs/2506.07237