Boosting Multimodal Large Language Models with Visual Tokens Withdrawal for Rapid Inference.

Zhihang Lin, Mingbao Lin, Luxi Lin, Rongrong Ji

Browse the full AAAI paper archive.