I am going through the Pytorch Geometric documentation: https://pytorch-geometric.readthedocs.io/en/latest/index.html which is built on Pytorch .Here they mentioned about GNN Cheatsheet: https://pytorch-geometric.readthedocs.io/en/latest/cheatsheet/gnn_cheatsheet.html. Can someone explain what the purpose of this is as it shows a table and there are tick marks if checked. Please explain how to understand the first function with the related attributes and then I can proceed further. Thank you.

  • $\begingroup$ Would put some additional explanation on " to replace the links with text so that people can easily understand."? $\endgroup$
    – Cloud Cho
    Sep 29, 2023 at 0:23
  • $\begingroup$ @Cloud Cho Thanks for your feed back I will look through it please help me in the first question I've asked regaeding GNNsheet. $\endgroup$
    – akhil
    Sep 29, 2023 at 14:47

1 Answer 1


Regarding "what the purpose of this is as it shows a table and there are tick marks if checked.", it sounds like that the tick marks indicating convolution operation support. This means you could train the specific graph with specific convolution method.

  For example (directly from same webpage), "SparseTensor": If checked (✓), supports message passing based on torch_sparse.SparseTensor, e.g., GCNConv(...).forward(x, adj_t).

  In other words, "SparseTensor": If checked (✓), you could run convolution operation using torch_sparse.SparseTensor, e.g., GCNConv(...).forward(x, adj_t)

Here is PyTorch own definition of Message Passing: Generalizing the convolution operator to irregular domains is typically expressed as a neighborhood aggregation or message passing scheme.


Not the answer you're looking for? Browse other questions tagged .