paper-with-me

홈 › Papers

Towards Audio Token Compression in Large Audio Language Models

2025-11-26 · Saurabhchand Bhati, Samuel Thomas, Hilde Kuehne, Rogerio Feris, James Glass arxiv

Large Audio Language Models (LALMs) deliver strong performance across speech and audio tasks, but their audio encoders generate high-rate token sequences (e.g., 25 tokens/s), making attention computation costly and limiting scalability. In this paper, we explore techniques such as unsupervised segmentation, uniform average pooling, etc., to reduce the number of audio tokens before they are consumed by the LLM decoder. To mitigate potential performance degradation, we employ low-rank adapters during finetuning. We evaluate our proposed models on two tasks, automatic speech recognition and speech-to-speech translation tasks, that are dependent on effectively uncovering the underlying lexical content of the input signal, and study the effect of downsampling on these tasks. Experimental results show that compressed LALMs can achieve performance closer to frame-level LALMs while reducing the input audio token count up to three times before the LLM backbone.

📄 PDF Abstract BibTeX arXiv:2511.20973

Code (0)

등록된 구현이 없습니다.

Tasks

Speech-to-Speech TranslationSpeech Recognition

Similar Papers 제목 키워드 기반

OmniZip: Audio-Guided Dynamic Token Compression for Fast Omnimodal Large Language Models

2025-11-18 · Keda Tao, Kele Shao, Bohan Yu, Weiqiang Wang 외 arxiv

Omnimodal large language models (OmniLLMs) have attracted increasing research attention of late towards unified audio-video understanding. However, the high computational cost of processing longer joint audio-video token…

OmniFocus: Query-Guided Modality-Balanced Token Compression for Omni-Modal Large Language Models

2026-07-03 · Shijie Cao, Qingyu Zhang, Boxi Yu, Yuzhong Zhang 외 arxiv

Omni modal large language models (OmniLLMs) have attracted wide attention for their ability to jointly process audio and video, but they generate large token sequences under audio-visual inputs, leading to substantial in…

Locality Matters for Training-Free Audio Token Compression in Audio-Language Models

2026-05-24 · Jiale Luo, Xiaoyu Liang, Haoji Hu arxiv

Audio-language models (ALMs) are increasingly used for audio captioning, question answering, and open-ended audio understanding, but their inference cost remains high when audio inputs are represented as long prefix-toke…

Question AnsweringAudio captioning

HeadRouter: Dynamic Head-Weight Routing for Task-Adaptive Audio Token Pruning in Large Audio Language Models

2026-04-26 · Peize He, Yaodi Luo, Xiaoqian Liu, Xuyang Liu 외 arxiv

Recent large audio language models (LALMs) demonstrate remarkable capabilities in processing extended multi-modal sequences, yet incur high inference costs. Token compression is an effective method that directly reduces …

OmniRefine: Alignment-Aware Cooperative Compression for Efficient Omnimodal Large Language Models

2026-05-12 · Yuchen Deng, Zidang Cai, Hai-Tao Zheng, Jie Wang 외 arxiv

Omnimodal large language models (Omni-LLMs) show strong capability in audio-video understanding, but their practical deployment remains limited by high inference cost of long video streams and dense audio sequences. Desp…