Fetching the paper…

Any2Point: Empowering Any-modality Large Models for Efficient 3D Understanding · Around