← Back to catalog

EchoJEPA Assess

Latent predictive foundation model for echocardiography trained on 18 million cardiac ultrasound images using joint-embedding predictive architecture (JEPA). Extracts anatomically meaningful cardiac features while filtering ultrasound speckle noise. Outperforms baselines by ~20% on left ventricular ejection fraction (LVEF) estimation and achieves 79% accuracy with only 1% of labelled training data. Demonstrates strong robustness under acoustic perturbation — degrades only 2% vs 17% for competing approaches. Apache-2.0.

Judgement: Assess

Assess — last commit ~5 months ago; 357 stars; permissive licence (Apache-2.0).

Derived deterministically by the MACH scorer from the signals below — not written prose.

Repository

Maintained — last commit ~5 mo ago; has tests

★ 357 stars · 58 forks · 10 open issues

Python 70% Jupyter Notebook 29% Shell 1%

Signals & lineage

Every value records the source it was harvested from.

FieldValueSource
descriptionEchoJEPA: A Latent Predictive Foundation Model for Echocardiographygithub_api
discovered_condanot_founddiscovery
discovered_crannot_founddiscovery
discovered_doinot_founddiscovery
discovered_hfnot_founddiscovery
discovered_npmnot_founddiscovery
discovered_pypinot_founddiscovery
has CIFalsegithub_api
has docsTruegithub_api
has releasesFalsegithub_api
has testsTruegithub_api
HuggingFace downloads (monthly)not_foundhuggingface
judgementAssessmach_scorer
judgement reasonAssess — last commit ~5 months ago; 357 stars; permissive licence (Apache-2.0).mach_scorer
licenseApache-2.0github_api
PyPI downloads (monthly)rate_limitedpypi
archivedFalsegithub_api
repository created2026-02-02T01:08:45Zgithub_api
GitHub forks58github_api
repo healthMaintainedmach_health
repo health reasonlast commit ~5 mo ago; has testsmach_health
homepagehttps://echojepa.comgithub_api
languagesPython:70,Jupyter Notebook:29,Shell:1github_api
last commit2026-06-18T20:06:15Zgithub_api
open issues10github_api
primary languagePythongithub_api
repo_size_kb14254github_api
GitHub stars357github_api
topicsokgithub_api

Change history

Recorded deviations — when a harvested value changed between runs.

WhenFieldWasNowSource
2026-11-20T21:53:45.013109repo_health—Maintainedmach_health
2026-11-20T21:53:45.013109repo_health_reason—last commit ~5 mo ago; has testsmach_health
2026-11-20T21:53:45.013011judgement—Assessmach_scorer
2026-11-20T21:53:45.013011judgement_reason—Assess — last commit ~5 months ago; 357 stars; permissive licence (Apache-2.0).mach_scorer
2026-11-20T21:46:32.451481hf_downloads——huggingface
2026-11-20T21:46:32.396159pypi_downloads——pypi
2026-11-20T21:46:32.396106discovered_conda——discovery
2026-11-20T21:46:32.396106discovered_cran——discovery
2026-11-20T21:46:32.396106discovered_doi——discovery
2026-11-20T21:46:32.396106discovered_hf——discovery
2026-11-20T21:46:32.396106discovered_npm——discovery
2026-11-20T21:46:32.396106discovered_pypi——discovery
2026-11-20T21:46:30.593649has_ci—Falsegithub_api
2026-11-20T21:46:30.593649has_docs—Truegithub_api
2026-11-20T21:46:30.593649has_releases—Falsegithub_api
2026-11-20T21:46:30.593649has_tests—Truegithub_api
2026-11-20T21:46:30.593649repo_languages—Python:70,Jupyter Notebook:29,Shell:1github_api
2026-11-20T21:46:30.593577description—EchoJEPA: A Latent Predictive Foundation Model for Echocardiographygithub_api
2026-11-20T21:46:30.593577license—Apache-2.0github_api
2026-11-20T21:46:30.593577repo_archived—Falsegithub_api
2026-11-20T21:46:30.593577repo_created—2026-02-02T01:08:45Zgithub_api
2026-11-20T21:46:30.593577repo_forks—58github_api
2026-11-20T21:46:30.593577repo_homepage—https://echojepa.comgithub_api
2026-11-20T21:46:30.593577repo_last_commit—2026-06-18T20:06:15Zgithub_api
2026-11-20T21:46:30.593577repo_open_issues—10github_api
2026-11-20T21:46:30.593577repo_primary_language—Pythongithub_api
2026-11-20T21:46:30.593577repo_size_kb—14254github_api
2026-11-20T21:46:30.593577repo_stars—357github_api
2026-11-20T21:46:30.593577repo_topics——github_api