torch-geometric

Develop and train Graph Neural Networks using PyTorch Geometric.

1|1|Updated Mar 13, 2026
One-click install
npx skills add https://github.com/youyinnn/skills-collection --skill torch-geometric-youyinnn
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: torch-geometric
Source: https://github.com/youyinnn/skills-collection/tree/main/plugins/model-development-and-experiments/skills/torch-geometric
Command: npx skills add https://github.com/youyinnn/skills-collection --skill torch-geometric-youyinnn

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires torch_geometric, torch_scatter, torch_sparse, torch_cluster, torch_spline_conv, 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 you to leverage deep learning on complex graph-structured data.

Core Features & Use Cases

  • GNN Model Development: Build and train GNNs using PyTorch Geometric's extensive layer library (GCN, GAT, GraphSAGE, etc.).
  • Data Handling: Load and process various graph datasets, including citation networks, molecular graphs, and social networks.
  • Advanced Features: Utilize mini-batch processing, heterogeneous graph support, and custom message-passing layers.
  • Use Case: Predict molecular properties for drug discovery by training a GNN on a molecular dataset, or classify nodes in a citation network to categorize research papers.

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 a Graph Neural Network for node classification?

To build a Graph Neural Network for node classification, use PyTorch Geometric's layer library like GCN or GAT. You load citation network datasets, construct the model, and train it to categorize nodes directly within your deep learning pipeline.

Can I train Graph Neural Networks on heterogeneous graph datasets?

Yes, you can train Graph Neural Networks on heterogeneous graph datasets. PyTorch Geometric provides advanced features specifically designed for heterogeneous graph handling, allowing you to process complex graph-structured data with varying node and edge types.

What is the best way to predict molecular properties for drug discovery using deep learning?

The best way to predict molecular properties for drug discovery is training a Graph Neural Network on molecular graph datasets. This Skill enables you to load molecular structures and apply deep learning models to predict properties effectively.

Does PyTorch Geometric support mini-batch processing for large graph datasets?

Yes, PyTorch Geometric supports mini-batch processing for large graph datasets. This allows you to train Graph Neural Networks efficiently on diverse graph datasets without encountering memory constraints during the model training phase.

Do I need to install specific dependencies to use custom message passing layers?

Yes, utilizing custom message passing layers requires installing specific dependencies. You need PyTorch Geometric alongside its associated libraries like torch_scatter, torch_sparse, and torch_cluster to enable advanced graph operations.