Welcome to the wiki for the hate speech detector! Pretty unoriginal name, and I'm not sure what all I will end up doing with this project. Mainly this wiki is to document the progress done so far and keep and eye on what all I might add in the future.

If you simply want to run inference, the model has been deployed at https://hate-speeech-detector.herokuapp.com/

To get prediction on a sample piece of text:

https://hate-speeech-detector.herokuapp.com/predict?text=your+text+here

Local Installation

Before running please make sure to:

pip install -r requirements.txt

Current functionality:

• Reading in the Ethos dataset, pre-processing it and storing the train-test splits python main.py split

• Training a binary model to detect if the comment contains hate-speech or not. python main.py train

• Models tested so far: SGD, AdaBoost, Random Forests, DistillBERT

Future functionality:

• An inference mechanism -> Enter comment -> Get output
• Test out more deep learning models, especially attention-based ones
• Create plots for tracked experiments