Domain Adaptation on Office-31: Finetune Alexnet

An AlexNet finetune report in Tensorflow
Yes, I still play with ancient NNs

OMG, how could implementation of Padding and Lrn be different between Tensorflow and Caffe

In short, code is here