##AlexNet trained in Torch7
Using https://github.com/soumith/imagenet-multiGPU.torch
alexnetowtbn, single GPU, default LR/WD policy, everything else also default
Model: https://drive.google.com/open?id=0B7ZgIaKJsQhbbnFWZ0JaRERVRm8
MeanStdCache: https://drive.google.com/open?id=0B7ZgIaKJsQhbeU1qUFJXRENYeVE