torch-geometric

Name: torch-geometric
Availability: InStock
Author: robinbarvaag

Create and train Graph Neural Networks with PyTorch Geometric for node and graph classification tasks.

Updated Jan 10, 2026

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

Ready to run

npx skills add https://github.com/robinbarvaag/poynt --skill torch-geometric-robinbarvaag

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: torch-geometric
Source: https://github.com/robinbarvaag/poynt/tree/main/.github/skills/torch_geometric
Command: npx skills add https://github.com/robinbarvaag/poynt --skill torch-geometric-robinbarvaag

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill simplifies the development and training of Graph Neural Networks (GNNs) within the PyTorch ecosystem, enabling advanced machine learning on graph-structured data.

Core Features & Use Cases

GNN Model Building: Provides pre-built layers (GCN, GAT, GraphSAGE) and a framework for custom message-passing layers.
Data Handling: Supports loading and processing various graph datasets, including custom and benchmark datasets.
Use Case: Predict molecular properties by training a GNN on chemical graph data, or analyze social networks for community detection.

Quick Start

Install PyTorch Geometric by running 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 a Graph Neural Network in PyTorch?▼

To build a Graph Neural Network in PyTorch, you can use pre-built layers like GCN, GAT, and GraphSAGE, or define custom message-passing layers to train models on graph-structured data.

Can I train models for molecular property prediction using graph data?▼

Yes, molecular property prediction is supported by training Graph Neural Networks on chemical graph data, allowing you to map molecular structures directly to their chemical properties.

Does this framework support heterogeneous graph processing?▼

Heterogeneous graph processing is fully supported, alongside mini-batch processing, to enable efficient geometric deep learning on complex and irregular data structures.

How do I load custom datasets for node classification tasks?▼

Loading custom datasets for node classification is handled by built-in data handling utilities that support processing both benchmark graph datasets and your own custom graph structures.

What is the best way to perform link prediction on social networks?▼

Link prediction on social networks is achieved by training Graph Neural Networks to analyze graph-structured data, facilitating advanced tasks like community detection and relationship forecasting.