Image Classification On CelebA

1. Classified the CelebFaces Attributes Dataset (CelebA) dataset images into celebrities wearing and not wearing glasses.
2. Tuned the hyper parameters such as dropout rates, the number of layers, the number of nodes in each layer, kernel size.
3. Increased the size of training set using data augmentation by rotating, scaling, injecting noise into input.

Report contains details about the structure of the neural network and results.