Code

torch-graph-components
torch-graph-components holds algorithms for fast graph connected components and graph partitioning, fully implemented in PyTorch with GPU parallelization in mind. In particular, this project contains the graph components and graph partitioning algorithms introduced in EZ-SP.

Code

Superpoint Transformer
Official implementation for Superpoint Transformer, SuperCluster, and EZ-SP. This codebase holds tools for efficient, scalable superpoint-based learning 🧩.

Code

NoRA
A python-populated Notion template to help you keep track of the papers you read 📜, their authors 👤, your notes 📝, and more 🔥. This is my personal framework for getting the best of Zotero but in Notion.

Code

Point Geometric Features
The pgeof library provides utilities for fast, parallelized computing ⚡ of local geometric features for 3D point clouds ☁️ on CPU. It was initially introduced as a dependency for Superpoint Transformer.

Code

DeepViewAgg
Official implementation for DeepViewAgg. This codebase holds tools for multimodal learning on 2D images 🖼️ and 3D point clouds ☁️️. In particular, we developed an efficient and scalable framework for learning to densely map and aggregate multi-view image features to point clouds.

Code