WARGA

The experiment implementation (PyTorch) for paper Wasserstein Adversarially Regularized Graph Autoencoder.

Requirements

• Pytorch 1.8.1
• Python 3.8
• scikit-learn
• networkx
• munkres
• pickle
• scipy

Run

• WARGA.py for WARGA-WC.
• WARGA GP.py for WARGA-GP.
• WARGA-GP Clustering Notebook.ipynb for WARGA-GP clustering results reproduction.
• WARGA-GP Link Prediction Notebook.ipynb for WARGA-GP link prediction results reproduction.
• WARGA-WC Clustering Notebook.ipynb for WARGA-WC clustering results reproduction.
• WARGA-WC Link Prediction Notebook.ipynb for WARGA-WC link prediction results reproduction.