wrong implementation of denominator in eq.19

[pherrusa7]

Dear @1Konny,

Thanks for your implementation!

I have detected that line 168 in gradcam.py:
alpha_denom = gradients.pow(2).mul(2) + \ activations.mul(gradients.pow(3)).view(b, k, u*v).sum(-1, keepdim=True).view(b, k, 1, 1)

should be:
global_sum = activations.view(b, k, u*v).sum(-1, keepdim=True).view(b, k, 1, 1) alpha_denom = gradients.pow(2).mul(2) + global_sum.mul(gradients.pow(3))

This is because of Eq. 19 in the paper [@adityac94]. If you pay attention, you need first to compute the sum over all {a, b} for each activation map k. Then you have a ponderation for each activation map k that you will use as a multiplier of all gradients {i, j} in the respective kernel k.

In your implementation, you first multiply each cell activation A{a,b,k} by its respective gradient {i, j, k} and then you sum over {i, j}, mixing the indices {a, b} and {i, j}, but they are independent.

I hope it's clear :)

PD: I have fixed this in gradcam.py and added a flag in example.ipynbto automatically detect if Cuda can be used, otherwise, use CPU. I will make you a "Pull request".

Thanks for your time!

[yangfantrinity]

Also, line 172, do we need to do
positive_gradients = F.relu(score.exp()*gradients) # ReLU(dY/dA) == ReLU(exp(S)*dS/dA))?
or it is enough to do just
positive_gradients = F.relu(gradients)?

[devraj89]

@yangfantrinity anything which comes after # is a comment in python.

[yangfantrinity]

@yangfantrinity anything which comes after # is a comment in python.

@devraj89 Sure of course I know anything after # is comment.

May I suggest you read my question one more time?