keras-squeeze-excite-network icon indicating copy to clipboard operation
keras-squeeze-excite-network copied to clipboard

Published 20 hours ago • verified publisher icon titu1994

Update se_densenet.py

Open ItsCRC opened this issue 5 years ago • 1 comments

Added ZeroPadding2D at the start before Conv2D and in subsample_initial_block to make densenet variants parameters same as original ones.

ItsCRC avatar Mar 10 '20 10:03 ItsCRC

Same padding already does this, and adding a zero padding does not increase parameter count. Where was this stated in the paper?

titu1994 avatar Mar 10 '20 16:03 titu1994