1 paper
Kangcong Li, Peng Ye, Lin Zhang +3
Transitioning Multimodal Large Language Models (MLLMs) from offline to online streaming video understanding is essential for continuous perception. However, existing methods lack f…