GNNLens2 is an interactive visualization tool for graph neural networks (GNN). It allows seamless integration with deep graph library (DGL) and can meet your various visualization requirements for presentation, analysis and model explanation. It is an open source version of GNNLens with simplification and extension.

A video demo is available here. Switch the video quality for the best viewing experience.

• PyTorch
• DGL
• Flask-CORS

You can install Flask-CORS with

pip install -U flask-cors

pip install Flask==2.0.3
pip install gnnlens

If you want to try experimental features, you can install from source as follows:

git clone https://github.com/dmlc/GNNLens2.git
cd GNNLens2/python
python setup.py install

Once you have installed the package, you can verify the success of installation with

import gnnlens

print(gnnlens.__version__)
# 0.1.0

We provide a set of tutorials to get you started with the library:

• Tutorial 1: Graph structure
• Tutorial 2: Ground truth and predicted node labels
• Tutorial 3: Edge weights and attention
• Tutorial 4: Weighted subgraphs and explanation methods

HKUST VisLab: Zhihua Jin, Huamin Qu

AWS Shanghai AI Lab: Mufei Li, Wanru Zhao (work done during internship), Jian Zhang, Minjie Wang

SMU: Yong Wang