Murmur：基于分层分块的高效长上下文 ASR 推理系统
Murmur: Efficient Long-Context ASR Inference with Hierarchical Chunking
AI 导读
arXiv 论文提出 Murmur，一个基于分层分块的长篇语音识别推理系统：约五分钟的中间块由长上下文模型并行解码，并用 30 秒草稿窗口上的短窗 ASR 做无损结构感知投机解码加速。在两个长上下文 ASR 模型和四个英语、中文会议数据集上，相比单次解码平均降低 2.5x（1.9-2.9x）端到端延迟，tcpWER 平均下降 1.9 点。代码已开源。
正文 · 原文
Abstract:Long-form automatic speech recognition (ASR) requires both high accuracy and low inference latency, but existing systems fall short of handling the trade-off between the two. Chunk-based approaches process audio in parallel windows for low latency, but lose cross-chunk context and need brittle heuristics to align speakers and timestamps at boundaries. Long-context ASR models transcribe everything in a single pass for better accuracy, but run several times slower, and a single repetition loop can corrupt the entire output. In this paper, we first analyze how context length affects the accuracy of long-context ASR models and find that an intermediate chunk size of about five minutes matches single-pass accuracy while avoiding repetition loops. Building on this, we propose Murmur, an inference system for long-form ASR based on hierarchical chunking. The long-context model decodes intermediate target chunks in parallel, and a fast short-window ASR on 30 s draft windows accelerates each chunk with lossless, structure-aware speculative decoding. The short-window ASR drafts the transcript, simple format rules draft the timestamps and speaker tags, and the long-context model verifies every token. On two long-context ASR models and four meeting datasets in English and Mandarin, Murmur cuts end-to-end latency per recording by 2.5x on average (1.9-2.9x) relative to single-pass decoding, including diarization, while lowering tcpWER by 1.9 points on average. Our code is available at this https URL.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Audio and Speech Processing (eess.AS) |
| Cite as: | arXiv:2606.01483 [cs.LG] |
| (or arXiv:2606.01483v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2606.01483 arXiv-issued DOI via DataCite |
Submission history
From: Wei-Tzu Lee [view email]
[v1]
Sun, 31 May 2026 22:54:57 UTC (20,900 KB)
[v2]
Wed, 4 Nov 2026 21:05:30 UTC (859 KB)
来源：arXiv cs.AI · arxiv.org