torch-geometric

Develop and train Graph Neural Networks with PyTorch Geometric.

2|Updated Feb 2, 2026

Search Tags:#deep learning #pytorch #graph neural networks #graph data #gnn #geometric deep learning

Ready to run

npx skills add https://github.com/Weiwei-Mao/hydrology-skills --skill torch-geometric-weiwei-mao

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: torch-geometric
Source: https://github.com/Weiwei-Mao/hydrology-skills/tree/main/hydrology-skills/torch_geometric
Command: npx skills add https://github.com/Weiwei-Mao/hydrology-skills --skill torch-geometric-weiwei-mao

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires torch_geometric, torch_scatter, torch_sparse, torch_cluster, pyg_lib, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill simplifies the development and training of Graph Neural Networks (GNNs), enabling advanced machine learning on complex, interconnected data.

Core Features & Use Cases

Graph Data Handling: Load, process, and batch graph-structured data efficiently.
GNN Model Building: Implement and train various GNN architectures (GCN, GAT, GraphSAGE, etc.).
Use Case: Analyze social networks for community detection, predict molecular properties for drug discovery, or understand citation networks for research paper classification.

Quick Start

Install the PyTorch Geometric library using the command uv pip install torch_geometric.

Frequently Asked Questions about torch-geometric

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I build and train Graph Neural Networks using PyTorch?▼

To build Graph Neural Networks using PyTorch, you can use this library to implement architectures like GCN and GraphSAGE for tasks including node classification, link prediction, and graph classification on irregular data structures.

Can I process heterogeneous graphs for geometric deep learning with mini-batches?▼

Yes, geometric deep learning on heterogeneous graphs is fully supported, allowing you to efficiently load, process, and batch complex graph-structured data for multi-GPU training.

How do I handle graph data loading and batching for GNN models?▼

Handling graph data for GNN models involves using the library's built-in utilities to efficiently load, process, and batch graph-structured data, enabling seamless mini-batch processing during training.

What are common use cases for graph neural networks on interconnected data?▼

Common use cases for graph neural networks include analyzing social networks for community detection, predicting molecular properties for drug discovery, and understanding citation networks for research paper classification.

Do I need to install additional dependencies to train GNNs with PyTorch?▼

Yes, training GNNs requires installing the PyTorch Geometric library along with dependencies like torch_scatter, torch_sparse, torch_cluster, and pyg_lib to support advanced geometric deep learning operations.