LCV2: An Efficient Pretraining-Free Framework for Grounded Visual Question Answering

MLLM
Re-assign community