THUMNLab/AutoGL

Health v2 · maintenance only

Adoption not part of the score

1140 stars · 125 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

AutoGL is an autoML framework and toolkit for machine learning on graphs, built on PyTorch with PyTorch Geometric and DGL backends. It provides automated feature engineering, model selection, neural architecture search, hyperparameter optimization, and ensembling for graph tasks like node classification, link prediction, and graph classification.

Use cases

• automatically train GNNs on graph datasets without manual tuning
• run neural architecture search for graph neural networks
• hyperparameter optimization for node classification on Cora
• automated link prediction on graph data
• benchmark graph NAS algorithms with NAS-Bench-Graph
• apply robust or self-supervised graph learning automatically

When to choose

• you need automated graph ML pipelines with minimal manual configuration
• you want NAS or HPO specifically for graph neural networks
• you work with PyTorch Geometric or DGL datasets and want AutoML on top

When to avoid

• you need a lightweight, fully manual GNN training loop
• your data is not graph-structured
• you need non-PyTorch frameworks like TensorFlow

Facets

framework · maturity active

machine-learning deep-learning llm-training machine-learning graph-processing artificial-intelligence python automl graph-neural-networks neural-architecture-search hyperparameter-optimization pytorch-geometric dgl gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="THUMNLab/AutoGL")