TraductAL - an offline Neural Machine Translation with NLLB-200

A completely offline, privacy-focused neural machine translation system powered by Meta's NLLB-200 models. All translations happen locally on your machine with no data sent to external servers.

⚠️ Development Status & Disclaimer

USE AT YOUR OWN RISK - This system is currently in active development and not ready for production use.

Known Limitations:

• Incomplete translations: Some words or text segments may be missing from output
• Quality varies: Translation accuracy depends on language pair and text complexity
• No quality guarantees: Results may not be suitable for professional or critical use
• Model limitations: Base NLLB-200 models may require fine-tuning for specific domains

Recommended Use:

• ✅ Development and testing purposes
• ✅ Personal experimentation with offline translation
• ✅ Privacy-focused translation where perfect accuracy isn't critical
• ❌ Production systems requiring reliable translations
• ❌ Professional document translation without human review
• ❌ Critical communications or legal documents

🔒 Privacy & Security

• 100% Offline: No internet connection required after setup
• Zero Data Leakage: All processing happens locally
• Professional Grade: Suitable for confidential document translation
• No Logging: No translation history stored externally

🌍 Language Support

NLLB-200 supports 200+ languages with excellent quality for:

• European languages: English, French, German, Spanish, Italian, Portuguese, Russian
• Asian languages: Chinese, Japanese, Korean, Hindi, Arabic, Turkish
• Many more: Including low-resource languages with good quality

🚀 Quick Start

Installation

# Clone the repository
git clone <repository-url>
cd TraductAL

# Install dependencies
pip install -r requirements.txt

# Download NLLB models
python download_nllb_200.py

Basic Usage

# Simple translation
./translate_enhanced.sh en fr "Hello, how are you?"

# Interactive mode
./translate_enhanced.sh interactive en fr

# List available languages
./translate_enhanced.sh list-languages

# System health check
./translate_enhanced.sh check

Python API

from nllb_translator import EnhancedOfflineTranslator

translator = EnhancedOfflineTranslator()
result = translator.translate("Hello world", "en", "fr")
print(result["translation"])  # "Bonjour ä tous"

📊 Model Options

NLLB-200-1.3B (Recommended)

• Size: ~2.6GB
• Speed: 0.5-1.0 seconds per sentence
• Memory: ~3GB RAM
• Quality: Very High

NLLB-200-3.3B (Maximum Quality)

• Size: ~6.6GB
• Speed: 1.0-2.0 seconds per sentence
• Memory: ~7GB RAM
• Quality: Excellent

🛠️ System Requirements

• Python: 3.8+
• RAM: 4GB minimum (8GB recommended)
• Storage: 3-7GB for models
• OS: Linux, macOS, Windows

📁 Project Structure

TraductAL/
├── README.md                 # This file
├── requirements.txt          # Python dependencies
├── LICENSE                   # MIT License
├── nllb_translator.py        # Main translator class
├── translate_enhanced.sh     # Command-line interface
├── download_nllb_200.py      # Model downloader
├── migrate_to_nllb.py        # Migration utility
├── NLLB_UPGRADE_GUIDE.md     # Detailed upgrade guide
├── QUICK_REFERENCE.md        # Command reference
├── MIGRATION_SUMMARY.md      # Migration documentation
└── models/
    └── deployed_models/
        ├── nllb_200_1.3b/   # NLLB 1.3B model files
        └── nllb_200_3.3b/   # NLLB 3.3B model files

🔧 Advanced Usage

Batch Translation

# Translate multiple lines
echo -e "Hello\nGoodbye\nThank you" | ./translate_enhanced.sh en fr

# Translate from file
./translate_enhanced.sh en fr "$(cat input.txt)"

Model Selection

# Use specific model
translator = EnhancedOfflineTranslator()
result = translator.translate("Hello", "en", "fr", model="nllb_200_3_3b")

Performance Monitoring

result = translator.translate("Hello world", "en", "fr")
print(f"Translation time: {result['time']:.2f}s")
print(f"Model used: {result['method']}")

🌟 Key Features

• High-quality translations with NLLB-200 models
• 200+ language support including low-resource languages
• Bidirectional translation for all language pairs
• Automatic model selection based on availability
• Performance monitoring and optimization
• Professional-grade privacy with offline operation
• Easy integration with Python API and shell commands

🔄 Migration from Other Systems

If you're migrating from MT5 or other translation systems:

# Run the migration script
python migrate_to_nllb.py

# Follow the upgrade guide
cat NLLB_UPGRADE_GUIDE.md

📖 Documentation

• NLLB_UPGRADE_GUIDE.md: Detailed upgrade instructions
• QUICK_REFERENCE.md: Command reference sheet
• MIGRATION_SUMMARY.md: Migration documentation
• LICENSE_INFO.md: Licensing information

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Meta AI for the NLLB-200 models
• Hugging Face for the Transformers library
• PyTorch team for the deep learning framework

---

Ready to translate? Run python download_nllb_200.py to get started!