📝 Post-Editing Evaluation Tool
This project is a web-based evaluation tool that scores machine translation (MT) output against human-edited references using BLEU, CHRF, and COMET metrics. It is packaged as a Gradio interface and deployable via Hugging Face Spaces.
🚀 Features
- 📊 Evaluate MT output with:
- BLEU
- CHRF
- COMET (requires OpenAI API key)
- 🖥️ Simple, interactive web UI via Gradio
- 🐳 Hugging Face Spaces–compatible Docker deployment
🧪 Example Use
Paste or upload:
- Source text
- Machine translation output
- Post-edited reference
Then click "Evaluate" to see automatic quality scores.
📦 Installation (for local development)
git clone https://github.com/yourusername/post_editing_evaluator.git
cd post_editing_evaluator
python -m venv venv
source venv/bin/activate # or .\venv\Scripts\activate on Windows
pip install -r requirements.txt
python interface.py