CarilionClinic on Hugging Face

This organization is the home for Carilion Clinic Data Science & AI work on the Hugging Face Hub. We use it to share and collaborate on models, de-identified datasets, and lightweight apps that support clinical, operational, and research use cases.

What lives here

• Models for NLP, imaging, tabular prediction, and tooling like tokenizers
• Datasets that are de-identified, synthetic, or aggregated
• Spaces for interactive demos, evaluations, and stakeholder review
• Docs such as model cards, dataset cards, and usage examples

Privacy and compliance

• Never upload PHI or direct identifiers
• Only de-identified, synthetic, or fully aggregated data
• Private by default, enable gated access when needed
• Document data provenance, intended use, and limitations in the card

Quick start

1. Access, request membership in the CarilionClinic org, private repos are the default
2. Create a repo for a model, dataset, or Space, choose “Private”
3. Describe, complete the model or dataset card template, include eval metrics and known risks
4. Push

   pip install "huggingface_hub>=0.25"
   huggingface-cli login
   git lfs install
   git clone https://huggingface.co/CarilionClinic/<repo>
   # add files, README.md, and card
   git add -A && git commit -m "Initial release" && git push