Hugging Face's logo

Datasets:
pnnbao-ump
/
MedCrab

Tasks:
Translation
Modalities:
Text
Formats:
parquet
Languages:
Vietnamese
English
Size:
100K - 1M
Libraries:
Datasets
pandas
License:
Dataset card Data Studio Files
xet
Community
1

You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

You agree to not use the dataset to conduct experiments that cause harm to human subjects and to use the dataset for research purposes only.

Log in or Sign Up to review the conditions and access this dataset content.

🦀 MedCrab Dataset

A high-quality English→Vietnamese medical translation dataset containing 260,000 parallel sentence pairs, validated by clinical physicians for training specialized medical translation models.

Dataset Description

MedCrab is a carefully curated dataset designed to support the development of accurate medical translation systems between English and Vietnamese. The dataset covers a wide range of medical domains including clinical documents, research papers, textbooks, and biomedical engineering content.

⚠️ Access Policy: This dataset is not publicly available for download. Access is granted exclusively to:

  • Hospitals and healthcare institutions
  • Academic research organizations
  • Medical universities and research centers

For access requests, please contact the author with details about your institution and intended research use.

Dataset Creation

Source Data

The dataset is compiled from various medical and biomedical sources including:

  • Medical textbooks and reference materials
  • Clinical research papers
  • Biomedical engineering documentation
  • Healthcare guidelines and protocols

Data Validation

All translations have been reviewed and validated by qualified clinical physicians to ensure:

  • Medical terminology accuracy
  • Clinical context preservation
  • Cultural and linguistic appropriateness for Vietnamese medical practice

Usage 

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("pnnbao-ump/MedCrab")

# Access training data
train_data = dataset["train"]

# Example: Get first translation pair
print(train_data[0])

Acknowledgments

Special thanks to the medical professionals who contributed their expertise to validate this dataset:

  • Dr. Trần Phạm Tuyên - Rostov State Medical University, Russia
  • Dr. Trương Quốc Bảo - University of Medicine and Pharmacy at Ho Chi Minh City, Vietnam

Their clinical expertise and dedication to medical education excellence made this dataset possible.

License

This dataset is released under Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0).

Non-commercial use only. For commercial applications, please contact the author directly.

Citation

If you use this dataset in your research, please cite:

@dataset{medcrab2025,
  title={MedCrab: A High-Quality English-Vietnamese Medical Translation Dataset},
  author={Pham Nguyen Ngoc Bao and Tran Pham Tuyen and Truong Quoc Bao},
  year={2025},
  publisher={Hugging Face},
  howpublished={Clinical validation by Dr. Tran Pham Tuyen (Rostov State Medical University) and Dr. Truong Quoc Bao (University of Medicine and Pharmacy at Ho Chi Minh City)},
  url={https://huggingface.co/datasets/pnnbao-ump/MedCrab}
}

Contact

For questions, collaborations, or commercial licensing inquiries:

Built with ❤️ for Vietnamese medical education

Downloads last month
28

Models trained or fine-tuned on pnnbao-ump/MedCrab

Space using pnnbao-ump/MedCrab 1