The training set is used to train the machine learning model, while the test set evaluates the model. Keeping the two sets separate is vital because you don’t want to train the model on the test data.

This video is an overall package to understand Dropout in Neural Network and then implement it in Python from scratch.