Question about train_explorer

[aochaoch]

Thank you for this great work. I'm studing this project and I find in train_explorer.py line 143: data.node_free = torch.diag(data.node_free, 0). This diag function returns main diagonal elements of the node_free which will always be 0 unless there are self loop edges. I don't understand this step and I think this might be unreasonable. Since this is an imortant parameter to the model, I will be very grateful if you can help me with this question. Thank you again for your time and help.