iduta / pyconv

Pyramidal Convolution: Rethinking Convolutional Neural Networks for Visual Recognition (https://arxiv.org/pdf/2006.11538.pdf)
MIT License
328 stars 53 forks source link

Did you try to add SE Block? #16

Open hegc opened 3 years ago

hegc commented 3 years ago

If adding se layer, will there be any prompt?

iduta commented 3 years ago

Yes, the SE block can bring additional improvement.

hegc commented 3 years ago

Thanks!