Can explain models be added to edge_attr？

[tanyingjia]

Recently I was using the gnn module to predict the classification task, and I wanted to use the interpretation model to explain the predicted data, but the model had edge_attr during use, so when using PGExplainer, the parameters were wrong. So I checked the code in the explain module and found that the model parameter does not have edge_attr, so I wanted to ask if I could use the gnn model with edge_attr.

Errors Encountered
for batch in testloader:
21 loss = explainer.algorithm.train(
---> 22 epoch, model, batch.x, batch.edge_index, batch.edge_attr,target=batch.y)
23
24 # Generate the explanation for a particular graph:

TypeError: train() takes 5 positional arguments but 6 positional arguments (and 1 keyword-only argument) were given