The code on how to train a Convolutional Neural Network model has also been uploaded with the help of Google Colab. The program takes input image and imports the weight files from the already trained model. The code predicts the crowd density present inside the bus and classifies into five crowd levels: very low, low, moderate, high and very high. The predicted crowd density is updated at the real-time server as well. A concise program on how to make the code work on desktop has also been uploaded.