hatespeech / README.md
sandbox338's picture
Update README.md
fad5d56 verified

A newer version of the Gradio SDK is available: 5.35.0

Upgrade
metadata
title: Hatespeech
emoji: πŸ’¬
colorFrom: yellow
colorTo: purple
sdk: gradio
sdk_version: 5.29.0
app_file: app.py
pinned: false
license: mit

Swahili Hate Speech Classifier

This Space is a demo of a fine-tuned BERT model for classifying Swahili text into three categories

  • Non-hate speech
  • Political hate speech
  • Offensive language

🧠 Model

The model was trained using a labeled Swahili dataset for hate speech detection. It is based on a BERT architecture and fine-tuned specifically for this 3-class classification task.

Model repo: sandbox338/hatespeech

πŸ’¬ How to Use

  1. Enter a Swahili sentence or paragraph into the textbox.
  2. Click Submit.
  3. The model will return one of the three class labels.

πŸ” Example Inputs

  • Hii ni ujumbe wa kawaida bila matusi. β†’ Non-hate speech
  • Wanasiasa hawa ni wabaya na lazima waondoke! β†’ Political hate speech
  • Unasema upuuzi na wewe ni mjinga kabisa! β†’ Offensive language

πŸ‘© Usability Testing

This tool is part of an ongoing usability evaluation. Please try it out and share feedback on:

  • Clarity of results
  • Ease of use
  • Suggestions for improvement