VGG16 CNN with dense top trained on subset of data The user will have to update train_loc and test_loc with the file path to the train and test folders. The reduced data set is available through Google Drive at: https://drive.google.com/open?id=1MhvONKcaX6F6q0O1VKzht8_n9M4QTvex. It should already be in the correct format, so just download it.
VGG_demo.py: Holds a pretrained VGG16 network that can be used to classify an object, edit line 11 with desired directory
VGG_try1.py: 1st attempt at a fine tuned VGG16 network, froze all convolutional weights and only working with the dense layer on top