How to verify performance on datasets with unbiased training samples?

[hongyonghan]

How to verify performance on datasets with unbiased training samples?

[hongyonghan]

when the --biased-sample is False,the code can not run.

[Akanshaaga]

While running main_gnn.py, I am getting the following error:
if len((diagonals)) == 0 or isscalarlike(diagonals[0]):
TypeError: object of type 'int' has no len()

Could you please help me with this. Thanks in advance.
P.S. I am new to python.