README.md

metadata

language:
  - fr
  - gcf
license: cc-by-4.0
task_categories:
  - translation
size_categories:
  - 1K<n<10K
tags:
  - guadeloupean-creole
  - french
  - low-resource-language
  - caribbean
pretty_name: Guadeloupean Creole-French Translation Dataset

Guadeloupean Creole-French Translation Dataset

Dataset Description

This dataset contains 2,058 translation pairs between French and Guadeloupean Creole (Kreyòl Gwadloup) as part of the POTOMITAN project.

Languages

• French (fr)
• Guadeloupean Creole (gcf)

Dataset Structure

The dataset is in JSON format with the following structure:

[
  {
    "id": 1,
    "fr": "French text",
    "gcf": "Creole text",
    "section": "category",
    "emergency_context": boolean
  }
]

Fields

• id: Unique identifier
• fr: French text
• gcf: Guadeloupean Creole translation
• section: Category (assimil_lesson, medical_emergency, etc.)
• emergency_context: Boolean indicating if this is emergency-related content

Statistics

• Total pairs: 2,058
• Emergency contexts: 9
• Categories: multiple (assimil_lesson, medical_emergency, natural_disaster, basic_needs)

Dataset Sources

• Assimil language lessons
• Emergency phrase translations
• Common conversational expressions

Potential Applications

• Machine translation systems
• Language learning applications
• Emergency response systems
• Cultural preservation initiatives
• Linguistic research

Usage

import json

with open('gcf_fr_translation_dataset.json', 'r', encoding='utf-8') as f:
    data = json.load(f)

for pair in data['pairs']:
    print(f"French: {pair['fr']}")
    print(f"Creole: {pair['gcf']}")

Citation

If you use this dataset, please cite:

@dataset{potomitan_gcf_fr_2024,
  title={Guadeloupean Creole-French Translation Dataset},
  author={POTOMITAN Project},
  year={2024},
  publisher={Hugging Face}
}

License

This dataset is licensed under CC BY 4.0.

Contact

For questions or collaboration, please contact the POTOMITAN project team:

• GitHub: https://github.com/Brdcie/creoles-translation

Acknowledgments

Special thanks to all contributors who helped collect and verify these translations.