VocalNet-MDM: Accelerating Streaming Speech LLM via Self-Distilled Masked Diffusion Modeling

TL;DR

VocalNet-MDM introduces a non-autoregressive masked diffusion modeling approach for streaming speech LLMs, significantly improving decoding speed and latency while maintaining high recognition accuracy and naturalness.

Contribution

The paper presents VocalNet-MDM, a novel streaming speech LLM framework using masked diffusion modeling with hierarchical masking and self-distillation, addressing key challenges for efficiency and low latency.

Findings

Achieves 3.7×–10× decoding speedup over autoregressive baselines.

Reduces first-chunk latency by 34%.

Maintains competitive speech recognition accuracy.

Abstract

Recent Speech Large Language Models~(LLMs) have achieved impressive capabilities in end-to-end speech interaction. However, the prevailing autoregressive paradigm imposes strict serial constraints, limiting generation efficiency and introducing exposure bias. In this paper, we investigate Masked Diffusion Modeling~(MDM) as a non-autoregressive paradigm for speech LLMs and introduce VocalNet-MDM. To adapt MDM for streaming speech interaction, we address two critical challenges: training-inference mismatch and iterative overhead. We propose Hierarchical Block-wise Masking to align training objectives with the progressive masked states encountered during block diffusion decoding, and Iterative Self-Distillation to compress multi-step refinement into fewer steps for low-latency inference. Trained on a limited scale of only 6K hours of speech data, VocalNet-MDM achieves a 3.7$\times$--10$\times$ decoding speedup and reduces first-chunk latency by 34\% compared to AR baselines. It maintains competitive recognition accuracy while achieving state-of-the-art text quality and speech naturalness, demonstrating that MDM is a promising and scalable alternative for low-latency, efficient speech LLMs.