torch-geometric - Develop Graph Neural Networks with PyG
Supports graph neural network development with PyTorch Geometric for graph learning, geometric data, molecular prediction, and heterogeneous graphs.
Tags
Updated: 2026-10-08Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Create graph data
- Load benchmark datasets
- Build GNN models
- Implement message passing layers
- Batch graph data
- Train graph models
- Handle heterogeneous graphs
- Process molecular graphs
Inputs
- Graph data
- Node features
- Edge features
- Task labels
- Dataset files
- Learning task
Outputs
- Graph predictions
- Node predictions
- Link predictions
- Trained GNN models
- Graph embeddings
- Processed graph datasets
Requirements
- Python environment
- PyTorch
- torch_geometric package