torch_geometric

Build and train Graph Neural Networks for node classification by implementing architectures like GCN and GraphSAGE using the PyTorch Geometric library. This Skill facilitates loading citation networks and processing graph-structured data to classify nodes based on their graph neighborhood.

Predict molecular properties for drug discovery by loading molecular graphs and training Graph Neural Networks. The Skill enables graph classification tasks on diverse graph datasets, mapping molecular structures to property predictions using deep learning.

PyTorch Geometric supports advanced Graph Neural Network concepts including heterogeneous graphs and message passing. You can process interconnected information across diverse node and edge types to analyze complex graph-structured data relationships.

PyTorch Geometric supports link prediction and provides explainability tools for graph analysis. You can train GNN models to predict connections between nodes and utilize built-in tools to interpret the model's decision-making process.

Install PyTorch Geometric by running `uv pip install torch_geometric` to set up the environment. This enables the development and training of various GNN architectures for deep learning on graph-structured data.