Dog Classification with CNNs and Transfer Learning

In this project, we will we will build a pipeline to process real-world, user-supplied images. The goal is to explore different Deep Learning models, using various architecture and techniques (like CNNs and Transfer Learning) and to get a first version of the model with a good performance. We will explore data sets, discuss metrics, present the results of the models as well as some hints on potential improvements. In the end, the final model could be potentially used within a web or mobile app to create an entertaining user experience in dog (or human) classification. To achieve our goals we need to train models using two different data sets: with human faces and with dog images.