QG-VTC: Question-Guided Visual Token Compression in MLLMs for Efficient VQA

MQ
Re-assign community