You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.
yes, when using the Deconvnet for segmentation, the last op is uppooling, but currently, the io part for segmentation is not support.
if the mxneters agreed adding the uppolling op, i will try it.
#422 and #503 have added the deconvolution operator for segmentation. The other missing piece to construct a complete segmentation network is the unpooling layer (#337 (comment)).
Current reference implementation is only in Caffe.
https://github.com/HyeonwooNoh/caffe/blob/new_data_layer/src/caffe/layers/unpooling_layer.cpp
https://github.com/HyeonwooNoh/caffe/blob/new_data_layer/src/caffe/layers/unpooling_layer.cu
The text was updated successfully, but these errors were encountered: