We build a VGGnet CNN to build a model that can classify Lexus, Porshce, and Ford cars. We detail the motivation and results of this study in our CNNReport.pdf .
This CNN is created in Python using the following files:
- downloadImage.py downloads images to a sepcified directory given a text file containing 1 URL per line.
- smallervggnet.py: The VGGnet implementation
- train_smallervgg.py Used to train CNN using training set
- predict_smallervgg.py Used to make predictions using an unseen test set. Will output its confidence of prediction directly on image