The padding of I3D model should be symmetrical

[fjchange]

The Module of MaxPool3dTFPadding with kernel_size=(1,3,3), stride(1,2,2) can lead to asymmetrical padding. It would influence the output feature map, as the bottom right would be usually higher than other part of the feature map.

When I try to input a all zeros tensor into I3D model pretrained on Kinetics-400, someting strange happen, I average pooling the C and T dim and min-max norm to get a picture as below. The bottom right is much higher than other parts.

By checking each layer output, I find out bottom right is usually have higher activation value than other part but not obvious, until mixed_5b block.

I don't know whether it hurt the model's performance, but at least it hurt the Interpretability.

[hassony2]

Thank you for raising this issue.

Interesting ! I am not sure whether this is influencing the numerical results as I converted the models by checking that the results were consistent on different images. I did observe a slight discrepancy with the original model (as reported in the readme) maybe this is related !

Best,

Yana