Pytorch Graph Attention Network

This is a pytorch implementation of the Graph Attention Network (GAT) model presented by Veličković et. al (2017, https://arxiv.org/abs/1710.10903).

The repo has been forked initially from https://github.com/tkipf/pygcn. The official repository for the GAT (Tensorflow) is available in https://github.com/PetarV-/GAT. Therefore, if you make advantage of the pyGAT model in your research, please cite the following:

@article{
  velickovic2018graph,
  title="{Graph Attention Networks}",
  author={Veli{\v{c}}kovi{\'{c}}, Petar and Cucurull, Guillem and Casanova, Arantxa and Romero, Adriana and Li{\`{o}}, Pietro and Bengio, Yoshua},
  journal={International Conference on Learning Representations},
  year={2018},
  url={https://openreview.net/forum?id=rJXMpikCZ},
  note={accepted as poster},
}

Performances

The training of the transductive learning on Cora task on a Titan Xp takes ~0.9 sec per epoch and 10-15 minutes for the whole training (~800 epochs). The final accuracy is between 83.6 and 84.6 (obtained on 3 different runs).

A small note about initial sparse matrix operations of https://github.com/tkipf/pygcn: they have been removed. Therefore, the current model take ~7GB on GRAM.

Requirements

pyGAT relies on Python 3.5 and PyTorch 0.4 (due to torch.where).

Issues/Pull Requests/Feedbacks

Don't hesitate to contact for any feedback or create issues/pull requests.