This repository contains two Jupyter notebooks for detecting driver drowsiness. There are two ways of detecting: one is by detecting occurence of yawning, which triggers an alarm after some threshold period; the other one is by detecting the state of eyes of a driver. It has three classes: open, half-closed and closed while the first detector has classes awake and drowsy. Detection is implemented using YOLOv5 object detection and is demonstrated on web camera.