skills.rest
pur3v4d3r
pur3v4d3rCOMMUNITY
Explore all skills by this author →
GitHub

torch-geometric

Build and train graph neural networks with PyG for node classification, graph classification, and link prediction.

6|Updated Dec 30, 2025
Search Tags:#datasets#graph-neural-networks#gnn#node-classification#graph-classification#pytorch-geometric
One-click install
npx skills add https://github.com/pur3v4d3r/pur3-pkb-codebase --skill torch-geometric-pur3v4d3r
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: torch-geometric
Source: https://github.com/pur3v4d3r/pur3-pkb-codebase/tree/main/.claude/skills/__scientific-skills/torch_geometric
Command: npx skills add https://github.com/pur3v4d3r/pur3-pkb-codebase --skill torch-geometric-pur3v4d3r

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires torch, torch_geometric, networkx, matplotlib, numpy, and includes scripts (resource) and references (resource) components.

What problem does it solve?

PyTorch Geometric (PyG) enables researchers and engineers to rapidly construct and train graph neural networks, providing a ready-made toolkit for graph-centric tasks and experiments.

Core Features & Use Cases

  • Rich set of graph layers (GCN, GAT, GraphSAGE, GIN, GCNConv, SAGEConv, GATConv) and data structures (Data, HeteroData) for node and graph classification, link prediction, and regression on graphs.
  • Integrated datasets and utilities for benchmarking and real-world workloads, including Planetoid, TUDataset, MoleculeNet, and OGB-style resources, plus tooling to handle heterogeneity and multi-type graphs.
  • Templates, benchmarks, and visualization utilities through scripts in the repository to accelerate model development, experimentation, and reproducibility.

Quick Start

Install PyG and run a minimal GCN example on a small Planetoid dataset to see graph neural networks in action.

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 for node classification tasks?▼

You can build and train graph neural networks for node classification by using GCN, GAT, and GraphSAGE layers on datasets like Planetoid. The Skill provides templates and scripts to construct end-to-end node classification workflows in PyTorch.

What's the best way to handle heterogeneous graphs with PyTorch Geometric?▼

Handling heterogeneous graphs with PyTorch Geometric utilizes the HeteroData data structure to manage multi-type nodes and edges. This allows you to apply graph neural networks to complex, multi-type graphs for real-world workloads.

Can I use this for molecular property prediction and graph classification?▼

Yes, you can use this for molecular property prediction and graph classification by leveraging integrated datasets like MoleculeNet and TUDataset. It provides the necessary graph layers and benchmarking scripts to train models on molecular structures.

Does this Skill support link prediction and large-scale graph datasets?▼

This Skill supports link prediction and large-scale graph datasets through its core PyG components and OGB-style resources. You can implement link prediction workflows and benchmark performance on large-scale graph datasets using the provided scripts.

Do I need PyTorch and NetworkX installed to run PyTorch Geometric workflows?▼

Yes, you need PyTorch and NetworkX installed along with torch_geometric, matplotlib, and numpy. These dependencies are required to execute the graph neural network training scripts and visualize graph data structures.

Related Skills

Explore More Data & Analytics Skills→
tondevreltondevrel
19

pytorch

Develop and train deep learning models using PyTorch with tensor operations and automatic differentiation.

Community
Advanced
Clay-HHKClay-HHK
2

torch-geometric

Develop Graph Neural Networks with PyTorch Geometric layers and datasets.

Community
Advanced
Hung-3008Hung-3008
1

torch-geometric

Build, train, and deploy graph neural networks with PyTorch Geometric.

Community
Advanced
jaechang-hitsjaechang-hits
298

torchdrug

Train GNN models for molecular property prediction using PyTorch and TorchDrug.

Community
Advanced