medalign
importeddata/medalign
MedAlign is a clinician-generated dataset for instruction following with electronic medical records.
Machine-generated from the listed sources and not yet reviewed by a human.
- Category
- Data & Standards
- Subcategory
- unknown
- License
- MIT(osi)
- Status
- maintained
- Maturity
- deployed
- Organization
- som-shahlab
- Country
- unknown
- Homepage
- unknown
- Repository
- github.com/som-shahlab/medalign
- Documentation
- unknown
- Tags
- benchmarking · ehr · electronic-health-records · foundation-models · healthcare · instruction-following · llms
- Regulatory
- unknown
Top contributors by commit count, from the project’s public repository. Avatars are served by their origin, not stored here. To be removed from this list, open an issue.
Computed from shared tags, weighted so a rare tag counts for more than a common one. These are suggestions, not curated relationships.
- ehr-codesetsehr · electronic-health-records · healthcare
Code sets for Electronic Health Record research
- ukbb-ehr-dataehr · electronic-health-records · healthcare
Prepare UK Biobank Electronic Health Record data for research
- meds_readerehr · foundation-models · healthcare
A high-performance database and Python reader interface for MEDS datasets
- femrelectronic-health-records · foundation-models · healthcare
FEMR (Framework for Electronic Medical Records) provides tooling for large-scale, self-supervised learning using electronic health records
- hf_ehrelectronic-health-records · foundation-models · healthcare
Training HuggingFace models on EHR data
- Patient2Vecehr · electronic-health-records · healthcare
Patient2Vec: A Personalized Interpretable Deep Representation of the Longitudinal Electronic Health Record
- api.github.com/repos/som-shahlab/medalignretrieved 2026-08-05 · via github-api
Machine-imported from GitHub search. Last push 2025-05-17, 102 stars, license reported as MIT. Category and schematic were assigned by keyword heuristics and are unreviewed.
Not yet verified by a human. Correct this record →
/v1/entries/50.json→ .entries["medalign"]
Entries are sharded 64 ways by a stable hash of the id, so a consumer can find any record without an index.