Aaron Chan
Aaron Chan
Hi @NarineK, thanks for the reply! Below is some toy code illustrating the situation: ``` import torch from torch import nn import torch.nn.functional as F from captum.attr import IntegratedGradients, InputXGradient,...
@NarineK That's a good point! My model is using dropout, so let me check what happens when I turn off dropout in Step 1. However, in eval mode, why does...
Hi @NarineK, apologies for the delayed follow-up, and thanks for your feedback! **Case 1**: When turning off dropout for both Steps 1 and 2, I'm able to get the same...