🔍 Research Series on Classic GNNs

Benchmarking Series: Reassessing Classic GNNs	Paper
Classic GNNs are Strong Baselines: Reassessing GNNs for Node Classification (NeurIPS 2024)	Link
Can Classic GNNs Be Strong Baselines for Graph-level Tasks? (ICML 2025)	Link

Classic GNNs are Strong Baselines: Reassessing GNNs for Node Classification (NeurIPS 2024)

Python environment setup with Conda

Tested with Python 3.7, PyTorch 1.12.1, and PyTorch Geometric 2.3.1, dgl 1.0.2.

pip install pandas
pip install scikit_learn
pip install numpy
pip install scipy
pip install einops
pip install ogb
pip install pyyaml
pip install googledrivedownloader
pip install networkx
pip install gdown
pip install matplotlib

Overview

• ./medium_graph Experiment code on medium graphs.

• ./large_graph Experiment code on large graphs.

Reference

If you find our codes useful, please consider citing our work

@inproceedings{
luo2024classic,
title={Classic {GNN}s are Strong Baselines: Reassessing {GNN}s for Node Classification},
author={Yuankai Luo and Lei Shi and Xiao-Ming Wu},
booktitle={The Thirty-eight Conference on Neural Information Processing Systems Datasets and Benchmarks Track},
year={2024},
url={https://openreview.net/forum?id=xkljKdGe4E}
}

Poster