This project is meant to get to know Deep Learning methods in Natural Language Processing. The project starts with searching for the two most common DL models for text classification, then implementing and comparing the results. First, we spend the first two weeks to obtain the baseline model (better than random guess) and the following two weeks to improve their accuracy to reach the median score of the competition, roughly 0.97865 at the time the project started.
Reference: https://www.kaggle.com/competitions/jigsaw-toxic-comment-classification-challenge/data