The model that I have used for this project is based on Multi-Task Learning. The model is based on the idea of Additive Combination, which is one of the traditional approaches for multi-modal classification. This model proved to be really good as it boosted the F1 scores wrt to the other models.