56 3D Models Integrated

LIDARLearn is a unified PyTorch library for 3D point cloud deep learning.

• It provides 56 ready-to-run configurations, bringing together supervised, self-supervised, and parameter-efficient fine-tuning methods in one place.
• Training can be launched with a single YAML file and one command.
• After training, it automatically generates a publication-ready LaTeX PDF, handling table formatting, best result highlighting, statistical testing, and diagram generation.
• Along with ModelNet40, ShapeNet, and S3DIS, it also includes remote sensing dataset benchmarks such as STPCTLS and HELIALS.
• STPCTLS is already preprocessed and ready to use immediately.
• It is licensed under MIT, and welcomes contributions and benchmark participation in the fields of research 3D computer vision and remote sensing.