PointNeXt is an official PyTorch implementation of a NeurIPS 2022 research paper that revisits PointNet++ with improved training and scaling strategies. The model processes 3D point clouds for tasks including semantic segmentation on S3DIS and 3D point cloud classification on ScanObjectNN and ModelNet40 datasets. It builds on the PointNet++ architecture and introduces training improvements and scaling strategies to advance performance on point cloud understanding tasks.