This repository provides the official implementation of RandLA-Net, a neural architecture designed for semantic segmentation of large-scale 3D point clouds. It was published at CVPR 2020 (Oral) and IEEE TPAMI 2021. The method processes 3D point cloud data (e.g., from LiDAR scans) to assign semantic labels to each point, enabling scene understanding for applications such as autonomous driving and robotics. It is implemented in TensorFlow.