Graph Convolutional Networks (GCN)

A PyTorch implementation of the paper https://arxiv.org/abs/1609.02907

The folder contains a GCN implementation based on DGL.

Three training scripts are available to run on Cora, Citeseer, and Pubmed.

Run with following:

python train_cora.py

* Cora: ~0.823 (paper: 0.815)
* Citeseer: ~0.705 (paper: 0.703)
* Pubmed: ~0.800 (paper: 0.790)

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
__pycache__		__pycache__
dgl_example		dgl_example
GCN.py		GCN.py
LICENSE		LICENSE
README.md		README.md
train_citeseer.py		train_citeseer.py
train_cora.py		train_cora.py
train_pubmed.py		train_pubmed.py

Provide feedback