Turning sensor data into objects, poses and maps.
Large-scale benchmark and baseline for general object grasping.
