# MURMUR: An Efficient Inference System for Long-Form ASR

arXiv: https://arxiv.org/abs/2606.01483 · HTML: https://mlsanity.com/paper/2606.01483/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Audio and omni-modal token compression, one of 12 directions in Efficient LLM inference and serving.
- Builds on 21 of its 33 references that are in the map.
- First posted on arXiv 31 May 2026.

## Placed in

- Efficient LLM inference and serving / [Audio and omni-modal token compression](https://mlsanity.com/inference/audio-and-omni-modal-token-compression/)

## Builds on (most cited first, 21 of 21)

- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [Lost in the Middle: How Language Models Use Long Contexts](https://mlsanity.com/paper/2307.03172/) (2023)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [SnapKV: LLM Knows What You are Looking for Before Generation](https://mlsanity.com/paper/2404.14469/) (2024)
- [Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference](https://mlsanity.com/paper/2406.10774/) (2024)
- [GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10,000 Hours of Transcribed Audio](https://mlsanity.com/paper/2106.06909/) (2021)
- [WhisperX: Time-Accurate Speech Transcription of Long-Form Audio](https://mlsanity.com/paper/2303.00747/) (2023)
- [pyannote.audio: neural building blocks for speaker diarization](https://mlsanity.com/paper/1911.01255/) (2019)
- [CHiME-6 Challenge:Tackling Multispeaker Speech Recognition for Unsegmented Recordings](https://mlsanity.com/paper/2004.09249/) (2020)
- [Distil-Whisper: Robust Knowledge Distillation via Large-Scale Pseudo Labelling](https://mlsanity.com/paper/2311.00430/) (2023)
- [Earnings-21: A Practical Benchmark for ASR in the Wild](https://mlsanity.com/paper/2104.11348/) (2021)
- [DQ-Whisper: Joint Distillation and Quantization for Efficient Multilingual Speech Recognition](https://mlsanity.com/paper/2305.10788/) (2023)
- [Updated Corpora and Benchmarks for Long-Form Speech Recognition](https://mlsanity.com/paper/2309.15013/) (2023)
- [TagSpeech: End-to-End Multi-Speaker ASR and Diarization with Fine-Grained Temporal Grounding](https://mlsanity.com/paper/2601.06896/) (2026)
- [VoxServe: Streaming-Centric Serving System for Speech Language Models](https://mlsanity.com/paper/2602.00269/) (2026)
- [Whisper-MLA: Reducing GPU Memory Consumption of ASR Models based on MHA2MLA Conversion](https://mlsanity.com/paper/2603.00563/) (2026)
- [Whisper-CD: Accurate Long-Form Speech Recognition using Multi-Negative Contrastive Decoding](https://mlsanity.com/paper/2603.06193/) (2026)
- [Context-Driven Dynamic Pruning for Large Speech Foundation Models](https://mlsanity.com/paper/2505.18860/) (2025)
- [Early Attentive Sparsification Accelerates Neural Speech Transcription](https://mlsanity.com/paper/2506.15912/) (2025)
- [Beyond the Utterance: An Empirical Study of Very Long Context Speech Recognition](https://mlsanity.com/paper/2602.09044/) (2026)

## Abstract (arXiv)

Long-form automatic speech recognition (ASR) requires both high accuracy and low latency, but existing systems force a trade-off between the two. Chunk-based pipelines 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 resolve everything in a single pass for better accuracy, but are an order of magnitude slower. We propose Murmur, an inference system that overcomes this trade-off by operating at two levels. At the inter-chunk level, we revisit the chunk-based pipeline for modern long-context ASR, treating chunk size as a tunable hyperparameter, and show that intermediate chunk sizes strike a good balance of accuracy and latency. At the intra-chunk level, we exploit attention sparsity through a sliding window KV cache eviction policy applied to both output and speech tokens. On AMI-IHM, Murmur matches single-pass accuracy while reducing latency by 4.2x, with further gains from token eviction at less than 1% relative tcpWER degradation. The code of Murmur is available at https://github.com/uw-syfi/Murmur.