1 paper · 1 filter
Yaolun Zhang, Ruohui Wang, Jiahao Wang +6
Video understanding with multimodal large language models (MLLMs) remains challenging due to the long token sequences of videos, which contain extensive temporal dependencies and r…