Caffe vs Tensorflow Deconv results

I am converting an autoencoder model from Caffe to TensorFlow. After transferring the weights, the network performs as expected through convolutions until it reaches deconvolutions. Output maps differ from those expected with specified inputs and weights from Caffe. Does anyone know why I am unable to transfer weight from Caffe to TF for deconv layers?

The only thing I could find on the internet is a dead discussion with the same problem: https://groups.google.com/a/tensorflow.org/forum/#!topic/discuss/Dw4PmlRfaaM

Thank!