VoxServe: Streaming-Centric Serving System for Speech Language Models

TL;DR

VoxServe is a unified streaming system for Speech Language Models that significantly improves throughput and efficiency by decoupling model architecture from system optimizations, supporting diverse models with low latency.

Contribution

It introduces a model-execution abstraction and streaming-aware scheduling to enhance streaming performance and support diverse SpeechLM architectures within a single system.

Findings

Achieves 10-20x higher throughput than existing systems.

Maintains high streaming viability with low latency.

Supports diverse SpeechLM architectures efficiently.

Abstract

Deploying modern Speech Language Models (SpeechLMs) in streaming settings requires systems that provide low latency, high throughput, and strong guarantees of streamability. Existing systems fall short of supporting diverse models flexibly and efficiently. We present VoxServe, a unified serving system for SpeechLMs that optimizes streaming performance. VoxServe introduces a model-execution abstraction that decouples model architecture from system-level optimizations, thereby enabling support for diverse SpeechLM architectures within a single framework. Building on this abstraction, VoxServe implements streaming-aware scheduling and an asynchronous inference pipeline to improve end-to-end efficiency. Evaluations across multiple modern SpeechLMs show that VoxServe achieves 10-20x higher throughput than existing implementations at comparable latency while maintaining high streaming viability. The code of VoxServe is available at https://github.com/vox-serve/vox-serve.