A model for profane text analysis

Something like this
Flag the comments for admin to review

Datasets to train the model

• Kaggle: Toxic Comment Classification Challenge
• Kaggle: Jigsaw Unintended Bias in Toxicity Classification
• Kaggle: Jigsaw Multilingual Toxic Comment Classification

Model

• BERT based multi-class classification
  • BERT 101 by HuggingFace
  • BERT implementation in TensorFlow
  • DistilBERT