Detoxify

A python library for hate speech detection

0 followers

A python library for hate speech detection

0 followers

Pretrained transformer-based models to predict toxicity in online comments, based on the Jigsaw Kaggle challenges. It includes a multilingual model trained on seven different languages.

Overview
Reviews
Team
More

Launch tags:Developer Tools•Artificial Intelligence•Tech

Launch Team

Framer AI AgentsDesign and publish professional sites with AI

Promoted

Maker

This is something we've been working on at Unitary, spearheaded by @laura_hanu. It is a simple user-friendly python library for detecting toxic comments, and gives three solid reference implementations of the Kaggle/Jigsaw toxic comment classification challenges, which are aimed at detecting toxicity and hate speech in online communities, along with extensions to minimise forms of unintended bias and classify multiple languages. It is based on @huggingface_ph transformers and Pytorch Lightning.

Report

6yr ago

Reviews