search

THUDM / GraphMAE

GraphMAE: Self-Supervised Masked Graph Autoencoders in KDD'22
467 stars 74 forks source link

keyword type error in GraphMAE torch_geometric #28

Closed voladorlu closed 1 year ago

voladorlu commented 1 year ago

Hi, I just get a keyword error when running the code based on torch_geometric. The keyword "residul" seems to be type error in the gat.py, from line 56 to 70. Please double check it out.

voladorlu commented 1 year ago

In the evaluation.py, the "linear_probing_for_transductive_node_classiifcation" function, the parameters for the model forward step should be model(x, graph.edge_index), not model(graph, x).

THINK2TRY commented 1 year ago

@voladorlu Thanks for pointing out the problems. We will fix them as soon as possible.