This library provides highly optimized implementations of graph cluster algorithms designed to integrate with PyTorch. It includes algorithms such as Graclus for weighted graph cuts, Voxel Grid pooling, Iterative Farthest Point Sampling, k-NN and radius graph generation, and Random Walk sampling. These are commonly used as building blocks in graph neural networks, point cloud processing (e.g., PointNet++), and other geometric deep learning architectures.