Live Demo -> https://senti-talk.herokuapp.com/
Download the model files from here --> https://www.kaggle.com/himanshu2606/comparison-between-different-models
But here is link to the colab notebook you can try --> https://colab.research.google.com/drive/1FMADqssRV18D4uVLSwkdPXAeV4AHmq6h
To classify comments posted on online blogs and other platforms such as product reviews as being toxic or not. Main goal is to develop a model that will help reduce the need for manually micromanaging online feeds and automate the entire process with considerable accuracy.
Jigsaw Multilingual Toxic Comment Classification Dataset Built by taking content from various websites including Wikipedia Talk Pages and Civil Comments Dataset size : 2.12 M training examples with toxic/non-toxic labels