Minor Changes

• There are now Python wheels available for torch-scatter and torch-sparse which should make the installation procedure much more user-friendly. Simply run

pip install torch-scatter==latest+${CUDA} torch-sparse==latest+${CUDA} -f https://pytorch-geometric.com/whl/torch-1.4.0.html
pip install torch-geometric

where ${CUDA} should be replaced by either cpu, cu92, cu100 or cu101 depending on your PyTorch installation.

• torch-cluster is now an optional dependency. All methods that rely on torch-cluster will result in an error requesting you to install torch-cluster.
• torch_geometric.data.Dataset can now also be indexed and shuffled:

dataset.shuffle()[:50]

Bugfixes

• Fixed a bug that prevented the user from saving MessagePassing modules.
• Fixed a bug in RGCNConv when using root_weight=False.