HuggingFace-native framework for training large language models on structured electronic health records using the MEDS (Medical Event Data Standard) format. Ships pretrained autoregressive foundation models — GPT, Llama, Mamba, and Hyena architectures — trained from scratch on 2 billion tokens of de-identified structured EHR data from Stanford Health Care. Designed for researchers already in the HuggingFace ecosystem who want to fine-tune or benchmark EHR foundation models without custom pipelines. Complementary to FEMR: FEMR covers masked/CLMBR-style training; hf_ehr covers autoregressive HuggingFace-native training. FEMR's own documentation recommends hf_ehr for autoregressive CLMBR-style models. Apache 2.0. From the Shah Lab at Stanford Medicine.
Assess — permissive licence (Apache-2.0).
Derived deterministically by the MACH scorer from the signals below — not written prose.
Dormant — last commit ~13 mo ago; has releases
★ 55 stars · 10 forks · 6 open issues
Jupyter Notebook 91% Python 8% Shell 1%
Every value records the source it was harvested from.
| Field | Value | Source |
|---|---|---|
| description | Training HuggingFace models on EHR data | github_api |
| discovered_conda | not_found | discovery |
| discovered_cran | not_found | discovery |
| discovered_doi | not_found | discovery |
| discovered_hf | StanfordShahLab/gpt-base-512-clmbr,StanfordShahLab/gpt-base-1024-clmbr,StanfordShahLab/gpt-base-2048-clmbr,StanfordShahLab/gpt-base-4096-clmbr,StanfordShahLab/llama-base-512-clmbr,StanfordShahLab/llama-base-1024-clmbr,StanfordShahLab/llama-base-2048-clmbr,StanfordShahLab/llama-base-4096-clmbr,StanfordShahLab/mamba-tiny-1024-clmbr,StanfordShahLab/mamba-tiny-4096-clmbr,StanfordShahLab/mamba-tiny-8192-clmbr,StanfordShahLab/mamba-tiny-16384-clmbr,StanfordShahLab/hyena-large-1024-clmbr,StanfordShahLab/hyena-large-4096-clmbr,StanfordShahLab/hyena-large-8192-clmbr,StanfordShahLab/hyena-large-16384-clmbr | discovery |
| discovered_npm | not_found | discovery |
| discovered_pypi | hf_ehr | discovery |
| has CI | False | github_api |
| has docs | True | github_api |
| has releases | True | github_api |
| has tests | False | github_api |
| HuggingFace downloads (monthly) | not_found | huggingface |
| judgement | Assess | mach_scorer |
| judgement reason | Assess — permissive licence (Apache-2.0). | mach_scorer |
| license | Apache-2.0 | github_api |
| PyPI downloads (monthly) | 61 | pypi |
| archived | False | github_api |
| repository created | 2023-10-05T07:36:06Z | github_api |
| GitHub forks | 10 | github_api |
| repo health | Dormant | mach_health |
| repo health reason | last commit ~13 mo ago; has releases | mach_health |
| homepage | https://arxiv.org/abs/2412.16178 | github_api |
| languages | Jupyter Notebook:91,Python:8,Shell:1 | github_api |
| last commit | 2025-11-02T04:11:51Z | github_api |
| open issues | 6 | github_api |
| primary language | Jupyter Notebook | github_api |
| repo_size_kb | 21021 | github_api |
| GitHub stars | 55 | github_api |
| topics | electronic-health-records,foundation-models,healthcare,machine-learning | github_api |
Recorded deviations — when a harvested value changed between runs.
| When | Field | Was | Now | Source |
|---|---|---|---|---|
| 2026-11-20T21:53:45.625404 | repo_health | — | Dormant | mach_health |
| 2026-11-20T21:53:45.625404 | repo_health_reason | — | last commit ~13 mo ago; has releases | mach_health |
| 2026-11-20T21:53:45.625314 | judgement | — | Assess | mach_scorer |
| 2026-11-20T21:53:45.625314 | judgement_reason | — | Assess — permissive licence (Apache-2.0). | mach_scorer |
| 2026-11-20T21:49:06.290050 | hf_downloads | — | — | huggingface |
| 2026-11-20T21:49:06.232270 | pypi_downloads | — | 61 | pypi |
| 2026-11-20T21:49:06.232208 | discovered_conda | — | — | discovery |
| 2026-11-20T21:49:06.232208 | discovered_cran | — | — | discovery |
| 2026-11-20T21:49:06.232208 | discovered_doi | — | — | discovery |
| 2026-11-20T21:49:06.232208 | discovered_hf | — | StanfordShahLab/gpt-base-512-clmbr,StanfordShahLab/gpt-base-1024-clmbr,StanfordShahLab/gpt-base-2048-clmbr,StanfordShahLab/gpt-base-4096-clmbr,StanfordShahLab/llama-base-512-clmbr,StanfordShahLab/llama-base-1024-clmbr,StanfordShahLab/llama-base-2048-clmbr,StanfordShahLab/llama-base-4096-clmbr,StanfordShahLab/mamba-tiny-1024-clmbr,StanfordShahLab/mamba-tiny-4096-clmbr,StanfordShahLab/mamba-tiny-8192-clmbr,StanfordShahLab/mamba-tiny-16384-clmbr,StanfordShahLab/hyena-large-1024-clmbr,StanfordShahLab/hyena-large-4096-clmbr,StanfordShahLab/hyena-large-8192-clmbr,StanfordShahLab/hyena-large-16384-clmbr | discovery |
| 2026-11-20T21:49:06.232208 | discovered_npm | — | — | discovery |
| 2026-11-20T21:49:06.232208 | discovered_pypi | — | hf_ehr | discovery |
| 2026-11-20T21:49:04.499924 | has_ci | — | False | github_api |
| 2026-11-20T21:49:04.499924 | has_docs | — | True | github_api |
| 2026-11-20T21:49:04.499924 | has_releases | — | True | github_api |
| 2026-11-20T21:49:04.499924 | has_tests | — | False | github_api |
| 2026-11-20T21:49:04.499924 | repo_languages | — | Jupyter Notebook:91,Python:8,Shell:1 | github_api |
| 2026-11-20T21:49:04.499850 | description | — | Training HuggingFace models on EHR data | github_api |
| 2026-11-20T21:49:04.499850 | license | — | Apache-2.0 | github_api |
| 2026-11-20T21:49:04.499850 | repo_archived | — | False | github_api |
| 2026-11-20T21:49:04.499850 | repo_created | — | 2023-10-05T07:36:06Z | github_api |
| 2026-11-20T21:49:04.499850 | repo_forks | — | 10 | github_api |
| 2026-11-20T21:49:04.499850 | repo_homepage | — | https://arxiv.org/abs/2412.16178 | github_api |
| 2026-11-20T21:49:04.499850 | repo_last_commit | — | 2025-11-02T04:11:51Z | github_api |
| 2026-11-20T21:49:04.499850 | repo_open_issues | — | 6 | github_api |
| 2026-11-20T21:49:04.499850 | repo_primary_language | — | Jupyter Notebook | github_api |
| 2026-11-20T21:49:04.499850 | repo_size_kb | — | 21021 | github_api |
| 2026-11-20T21:49:04.499850 | repo_stars | — | 55 | github_api |
| 2026-11-20T21:49:04.499850 | repo_topics | — | electronic-health-records,foundation-models,healthcare,machine-learning | github_api |