FaceRank is a deep learning project that trains a convolutional neural network to rank faces by appearance quotient. The model is built on TensorFlow with a Keras version available. Users provide images which the model scores on a scale, with training data consisting of tagged web images at 128x128 resolution.