Task1 - The first task is to predict the emotional state of a person while he/she is speaking. I have used ravdess dataset to train the model. It is a huge file so, Iam giving the download link of that dataset here: https://www.kaggle.com/uwrfkaggler/ravdess-emotional-speech-audio. Task-2 - The second task is to predict whether a person is speaking positively or negatively based on the words spoken by him/her. I have used three datsets to train the model. They are uploaded in my repository: Amazon reviews, yelp reviews, IMDB reviews.