This code fine tunes the vgg16 model. It's optimized to use the least possible RAM up to my knowledge. The model and the weights used in this code belong to the author of http://www.cs.toronto.edu/~frossard/post/vgg16/
added inceptionv3 from slim research repo