A possible bug worth your attention

[wangywUST]

Dear the authors, here are some concerns I have on your implementation. Thanks for your consideration!

In the batch_utils.py file, you write
pre_indices = indices
but this may not be a copy command, the right one is
pre_indices = indices.copy()
This error bug makes your get_indice_graph() function can only retrieve the first-hop neighborhood. And this leads to the error in Table 2 of your paper, e.g., you gave #Node of Cora being 644 there. This is only the number of the first-hop neighbors.

Is there anything that I missed?

Look forward to your reply!

[zzk0]

you can see this #8 , the author has replied.