Project

Overview

This project classifies toxic comments using deep learning and Python. The dataset leveraged is originally from Kaggle, which provided a large number of Wikipedia comments which have been labeled by human raters for toxic behavior. The types of toxicity are:

• toxic
• severe_toxic
• obscene
• threat
• insult
• identity_hate

A model which predicts the probability of each type of toxicity for each comment is created. Finally, the model is integrated into a Gradio app to be demoed.