Bibliographic Details
| Title: |
Soundwave: Less is More for Speech-Text Alignment in LLMs |
| Authors: |
Zhang, Yuhao, Liu, Zhiheng, Bu, Fan, Zhang, Ruiyu, Wang, Benyou, Li, Haizhou |
| Publication Year: |
2025 |
| Collection: |
Computer Science |
| Subject Terms: |
Computer Science - Computation and Language, Computer Science - Artificial Intelligence, Computer Science - Sound |
| More Details: |
Existing end-to-end speech large language models (LLMs) usually rely on large-scale annotated data for training, while data-efficient training has not been discussed in depth. We focus on two fundamental problems between speech and text: the representation space gap and sequence length inconsistency. We propose Soundwave, which utilizes an efficient training strategy and a novel architecture to address these issues. Results show that Soundwave outperforms the advanced Qwen2-Audio in speech translation and AIR-Bench speech tasks, using only one-fiftieth of the training data. Further analysis shows that Soundwave still retains its intelligence during conversation. The project is available at https://github.com/FreedomIntelligence/Soundwave. |
| Document Type: |
Working Paper |
| Access URL: |
http://arxiv.org/abs/2502.12900 |
| Accession Number: |
edsarx.2502.12900 |
| Database: |
arXiv |
