---
title: HateBERT Moderation Prototype
colorFrom: gray
colorTo: blue
sdk: gradio
python_version: "3.11"
app_file: app.py
pinned: false
short_description: Live harmful-content moderation demo
---

# HateBERT Moderation Prototype

This Space runs a Gradio demo for binary harmful-content moderation using the bundled fine-tuned HateBERT model.

## Files

- `app.py` - Gradio application and inference code.
- `fine_tuned_hatebert_model/` - local model directory loaded by the app.
- `requirements.txt` - Python dependencies for the Space.

The app loads the local model folder with `local_files_only=True`, so the model files must stay inside this repository.