Why use retain_graph? about pytorch-gnn-meta-attack HOT 2 CLOSED

Hi, you are right. According to pytorch documentation, retain_graph will retain the computation graph for calculating gradient.

retain_graph (bool, optional) – If False, the graph used to compute the grad will be freed. Note that in nearly all cases setting this option to True is not needed and often can be worked around in a much more efficient way. Defaults to the value of create_graph.

For line 216, there is no need to maintain the computation graph since it is in the outer training stage. For other lines containing retain_graph=True, I think we need to keep it.

from pytorch-gnn-meta-attack.

Thanks for your information.

from pytorch-gnn-meta-attack.