Face detector currently doesn't work, and is not trained on our input data / generalized inputs. EDA is also missing?. Lets do that today.