How to train a model with Imbalanced classes using SMOTE and Cost sensitive learning

My notebook gives an overview of how to deal with a multi class problem which has imbalanced classes and how we can overcome this issue using synthetic data generation and selective weight distribution while training a model.