Download a checkpoint and place it in weight/ using the filename below. Each VMGGA checkpoint contains the complete inference model state; a separate DINOv3 initialization checkpoint is not required ...
Cameras and LiDARs are currently two types of sensors commonly used for 3D mapping. Vision-based methods are susceptible to textureless regions and lighting, and LiDAR-based methods easily degenerate ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results