SLAM-Omni: Timbre-Controllable Voice Interaction System with Single-Stage Training

TL;DR

SLAM-Omni is a novel end-to-end voice interaction system that enables zero-shot timbre control and efficient multi-turn dialogue with single-stage training, outperforming prior models with limited data and training time.

Contribution

It introduces a single-stage training approach for a timbre-controllable spoken dialogue system, eliminating the need for pre-training on TTS or ASR tasks.

Findings

Outperforms prior models of similar scale.

Requires only 15 hours of training on 4 GPUs.

Achieves competitive performance with limited data.

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.