AI RESEARCH
HeadRouter: Dynamic Head-Weight Routing for Task-Adaptive Audio Token Pruning in Large Audio Language Models
arXiv CS.CL
•
ArXi:2604.23717v1 Announce Type: cross Recent large audio language models (LALMs) nstrate remarkable capabilities in processing extended multi-modal sequences, yet incur high inference costs. Token compression is an effective method that directly reduces redundant tokens in the sequence. Existing compression methods usually assume that all attention heads in LALMs contribute equally to various audio tasks and calculate token importance by averaging scores across all heads. However, our analysis nstrates that attention heads exhibit distinct behaviors across diverse audio domains.