In Convolutional Neural Networks, do all filters of the same convolutional layer need to have the same dimensions and stride?

If they don't, then it would seem the channel produced by each filter would have different sizes. Or is there some way to get around that?

  • 1
    $\begingroup$ You could pick the stride and dimensions accordingly so even though they are different, the output of each filter has the same dimensions, although as to the benefits of this I'm not sure. $\endgroup$ – Recessive Aug 13 at 1:56

It seems that a similar question has been raised here: https://stackoverflow.com/questions/57438922/different-size-filters-in-the-same-layer-with-tensorflow-2-0

Like answered in the link above, you could combine severall Conv2D ops with different kernel sizes on the same input. You would have to adapt each output with padding, or cropping, so that you could concatenate all of them.

Hope this helps!

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.