Graph Neural Network (GNN) Frameworks logopeterindia.net
Knowledge Hub

Graph Neural Network (GNN) Frameworks and Graph Libraries

A directory of libraries for graph deep learning — GNN frameworks for PyTorch, TensorFlow, JAX and Julia, graph analysis and visualisation tools, and benchmarks.

19 topics•4 categories•Updated October 2026

GNN Frameworks

7 entries

Libraries for building graph neural networks.

01PyTorch Geometric (PyG)A library built upon PyTorch to easily write and train Graph Neural Networks (GNNs) for a wide range of applications on structured data. 02Deep Graph Library (DGL)Fast and memory-efficient message passing primitives for training Graph Neural Networks, framework-agnostic across PyTorch, TensorFlow, and MXNet. 03SpektralA Python library for graph deep learning, based on the Keras API and TensorFlow 2, implementing popular GNN and pooling layers. 04GeometricFlux.jlA framework for geometric deep learning in Julia, providing classic graph neural network layers and utility constructs built on Flux. 05Graph NetsDeepMind's library for building graph networks in TensorFlow and Sonnet, with demos for shortest-path, sorting, and physics prediction tasks. The GitHub repository's last commit was in December 2022, so check its activity before adopting it. 06ptgnn (repository archived)A PyTorch GNN library containing code for creating graph neural network models, with sample implementations for PPI, VarMisuse, and more. 07Jraph (repository archived)A lightweight library for working with graph neural networks in JAX, providing a graph data structure and a "zoo" of forkable GNN models.

Graph Analysis & Visualization

3 entries

General graph libraries used alongside GNNs.

08NetworkXA Python package for the creation, manipulation, and study of the structure, dynamics, and functions of complex networks. 09python-igraphigraph is available on the Python Package Index with pre-compiled wheels for most Python distributions and platforms, for fast network analysis. 10Plotly for PythonPlotly's Python graphing library makes interactive, publication-quality graphs — widely used for visualizing graph and network data.

Benchmarks

1 entries

Datasets and leaderboards for graph learning.

11Open Graph BenchmarkBenchmark datasets and tasks for graph machine learning.

Related Pages

8 entries

Related directories on this site.

12Graph Neural NetworksGCN, GAT, GraphSAGE and GIN explained. 13GNN LibrariesLibraries for graph neural networks. 14Understanding GNNsIntroduction to graph neural networks. 15Deep Learning FrameworksDeep learning frameworks. 16Knowledge Graphs & Graph TechnologiesGraph databases and knowledge graphs. 17Graph Data PlatformsGraph analytics platforms. 18Graph and Causal Representation LearningResources on graph and causal representation learning. 19ML FrameworksMachine learning model development frameworks.

Frequently Asked Questions

What is a graph neural network?

A neural network that learns from graph-structured data by passing messages between connected nodes, used for molecules, social networks, recommendations and knowledge graphs.

Which library should I start with?

PyTorch Geometric and Deep Graph Library are the most widely used; Spektral suits TensorFlow users, Jraph and Graph Nets target JAX and TensorFlow.

Are all listed libraries active?

No. The Jraph and ptgnn repositories are archived, so they are marked as such.

Do I need NetworkX?

NetworkX and igraph are not GNN frameworks, but are useful for building, analysing and visualising graphs before training.

Where can I benchmark GNNs?

The Open Graph Benchmark provides standard datasets and tasks.