Qustion in autoencoder.py

AliJalalifar / Character_Animation

A Re-implementation of the paper "A Deep Learning Framework for Character Motion Synthesis and Editing"

114 stars 34 forks source link

Qustion in autoencoder.py #3

Open kfcbla opened 4 years ago

kfcbla commented 4 years ago

If we use relu in deconv part of the autoencoder, we get the result larger than zero. How can we get the same positions as the input which is zero mean.