SLAM-Omni: Timbre-Controllable Voice Interaction System with Single-Stage Training
Abstract
Recent advancements highlight the potential of end-to-end real-time spoken dialogue systems, showcasing their low latency and high quality. In this paper, we introduce SLAM-Omni, a timbre-controllable, end-to-end voice interaction system with single-stage training. SLAM-Omni achieves zero-shot timbre control by modeling spoken language with semantic tokens and decoupling speaker information to a vocoder. By predicting grouped speech semantic tokens at each step, our method significantly reduces the sequence length of audio tokens, accelerating both training and inference. Additionally, we propose historical text prompting to compress dialogue history, facilitating efficient multi-round interactions. Comprehensive evaluations reveal that SLAM-Omni outperforms prior models of similar scale, requiring only 15 hours of training on 4 GPUs with limited data. Notably, it is the first spoken dialogue system to achieve competitive performance with a single-stage training approach, eliminating the need for pre-training on TTS or ASR tasks. Further experiments validate its multilingual and multi-turn dialogue capabilities on larger datasets.
Cite
@article{arxiv.2412.15649,
title = {SLAM-Omni: Timbre-Controllable Voice Interaction System with Single-Stage Training},
author = {Wenxi Chen and Ziyang Ma and Ruiqi Yan and Yuzhe Liang and Xiquan Li and Ruiyang Xu and Zhikang Niu and Yanqiao Zhu and Yifan Yang and Zhanxun Liu and Kai Yu and Yuxuan Hu and Jinyu Li and Yan Lu and Shujie Liu and Xie Chen},
journal= {arXiv preprint arXiv:2412.15649},
year = {2024}
}