TensorX
返回文献探索

Paper · arXiv 2602.04804

OmniSIFT: Modality-Asymmetric Token Compression for Efficient Omni-modal Large Language Models

Yue Ding, Yiyan Ji, Jungang Li, Xuyang Liu, Xinlong Chen, Junfei Wu, Bozhou Li, Bohan Zeng, Yang Shi, Yushuo Guan, Yuanxing Zhang, Jiaheng Liu, Qiang Liu, Pengfei Wan, Liang Wang

50 upvotesFebruary 4, 2026arXiv 预印本
AI 摘要

OmniSIFT is a modality-asymmetric token compression framework for Omni-LLMs that reduces computational overhead through spatio-temporal video pruning and vision-guided audio selection while maintaining superior performance.

Omni-modal Large Language Modelstoken compressionspatio-temporal video pruningvision-guided audio selectiondifferentiable straight-through estimatorend-to-end optimizationmultimodal token sequences

Abstract

Omni-modal Large Language Models (Omni-LLMs) have demonstrated strong capabilities in audio-video understanding tasks. However, their reliance on long multimodal token sequences leads to substantial computational overhead. Despite this challenge, token compression methods designed for Omni-LLMs remain limited. To bridge this gap, we propose OmniSIFT (Omni-modal Spatio-temporal Informed Fine-grained Token compression), a modality-asymmetric token compression framework tailored for Omni-LLMs. Specifically, OmniSIFT adopts a two-stage compression strategy: (i) a spatio-temporal video pruning module that removes video redundancy arising from both intra-frame structure and inter-frame overlap, and (ii) a vision-guided audio selection module that filters audio tokens. The entire framework is optimized end-to-end via a differentiable straight-through estimator. Extensive experiments on five representative benchmarks demonstrate the efficacy and robustness of OmniSIFT. Notably, for Qwen2.5-Omni-7B, OmniSIFT introduces only 4.85M parameters while maintaining lower latency than training-free baselines such as OmniZip. With merely 25% of the original token context, OmniSIFT consistently outperforms all compression baselines and even surpasses the performance of the full-token model on several tasks.

北京市昌平区探索星信息技术及软件开发工作室

京ICP备2026059466号
OmniSIFT: Modality-Asymmetric Token Compression for Efficient Omni-modal Large Language Models | TensorX