English

CATP: Cross-Attention Token Pruning for Accuracy Preserved Multimodal Model Inference

Computation and Language 2026-02-16 v2 Artificial Intelligence

Abstract

In response to the rising interest in large multimodal models, we introduce Cross-Attention Token Pruning (CATP), a precision-focused token pruning method. Our approach leverages cross-attention layers in multimodal models, exemplified by BLIP-2, to extract valuable information for token importance determination. CATP employs a refined voting strategy across model heads and layers. In evaluations, CATP achieves up to 12.1X higher accuracy compared to existing token pruning methods, addressing the trade-off between computational efficiency and model precision.

Keywords

Cite

@article{arxiv.2404.08567,
  title  = {CATP: Cross-Attention Token Pruning for Accuracy Preserved Multimodal Model Inference},
  author = {Ruqi Liao and Chuqing Zhao and Jin Li and Weiqi Feng and Yi Lyu and Bingxian Chen and Haochen Yang},
  journal= {arXiv preprint arXiv:2404.08567},
  year   = {2026}
}
R2 v1 2026-06-28T15:52:39.718Z