Possible bug in L2 norm computation

[lllAlexanderlll]

Hi,

Great paper! I enjoyed reading it.
I'm wondering wether there is a bug in https://github.com/copenlu/xai-benchmark/blob/master/saliency_gen/interpret_grads_occ.py#L35,
since actually the L1 norm is calculated and for the L2 norm p should be set to 2?

Cheers,
Alex

[j-adamczyk]

I have the same question, @apepa could you please confirm this?