This project detects driver drowsiness from images using a Convolutional Neural Network (CNN). It classifies an image as one of four classes: Closed, Open, yawn or no_yawn. A prediction of Closed or ...

The training script is training/train_forest_cover.py. It explicitly drops both target columns from features. wilderness_area and soil_type use one-hot encoding. Data is stratified into 700 training / ...