Toxic Comment Classifier with BERT

This is a pet project that I created with the intention of enhancing my knowledge in Machine Learning, Natural Language Processing, and TensorFlow.

Introduction

A TensorFlow-based neural network model which uses the BERT (Bidirectional Encoder Representations from Transformers) model to classify comments as toxic or non-toxic.

Dataset

The dataset used for training is the cleaned-toxic-comments dataset available on Kaggle.

Requirements

• TensorFlow 2.x
• TensorFlow Hub
• TensorFlow Text
• Pandas
• Seaborn(optional)
• Scikit-learn
• Kaggle API

Example usage

This classifier could be useful in various applications such as pre-moderation of comments, automatic flagging of inappropriate content, and more.