PyTorch Geometric
A PyTorch library for building and training Graph Neural Networks (GNNs) on structured and irregular data.
What is PyTorch Geometric?
PyTorch Geometric (PyG) is a comprehensive library built upon PyTorch for deep learning on graphs and other irregular structures, a field known as geometric deep learning. It provides a unified framework to easily implement, train, and scale Graph Neural Networks (GNNs) for a wide range of applications involving structured data, from citation networks to 3D point clouds. The library solves the problem of applying neural networks to non-Euclidean data like graphs, meshes, and point clouds.
Machine learning researchers and practitioners, particularly those in fields like computational chemistry, social network analysis, recommendation systems, and computer vision, who need to work with graph-structured or 3D geometric data. It is also suitable for first-time users of GNN toolkits who are already familiar with PyTorch.
Developers choose PyG for its easy-to-use, PyTorch-like API that allows building GNN models in just 10-20 lines of code, its comprehensive collection of state-of-the-art GNN architectures, and its great flexibility for both rapid prototyping and creating custom models. Its unique selling point is being 'PyTorch-on-the-rocks,' offering a familiar experience with extensive support for scalability, heterogeneous graphs, and real-world applications.
Overview
Graph Neural Network Library for PyTorch
Use Cases
Best For
Related Projects
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
