MLTE
MLTEMLTE (Machine Learning Test and Evaluation) is a framework and Python package for testing ML models against requirements agreed with stakeholders. Teams record requirements, collect evidence from measurements, and assemble it into reports, with an optional web interface and backend that store the artifacts. Its authors describe the framework in a 2023 IEEE/ACM paper.
Tagged self-attested: its reports are generated by the team being evaluated.
Openness
5 medium confidence- license
- MIT(OSI)
- source
- public(Python package, web frontend and backend)
- core features withheld
- no
Package, frontend and backend are all in the MIT-licensed repository, and nothing in its tree sits under other terms.
- https://cdn.jsdelivr.net/gh/mlte-team/mlte@master/LICENSE recorded 2026-09-26
"MIT License Copyright (c) 2022 Kyle Dotterrer"
- https://cdn.jsdelivr.net/gh/mlte-team/mlte@master/README.md recorded 2026-09-26
"MLTE (pronounced \"melt\") is a framework and infrastructure for evaluating machine learning models and systems"; frontend and backend install as optional dependencies
- https://ungh.cc/repos/mlte-team/mlte/files/master recorded 2026-09-26
Full file listing of mlte-team/mlte master, 751 paths; no ee/, enterprise/, commercial/ or proprietary/ directory
Adoption
1 medium confidencePyPI downloads of mlte, the project's package, measure its use.
- https://pypistats.org/api/packages/mlte/recent recorded 2026-09-26
last_month downloads = 157 for mlte
Capability
2 medium confidenceMLTE turns test results into structured reports, but in its own format and produced by the team under evaluation, so a reader takes them on trust. Compliance Trestle, by contrast, writes to a published standard that outside tools can validate.
- https://cdn.jsdelivr.net/gh/mlte-team/mlte@master/README.md recorded 2026-09-26
README: install the package, start the backend and web UI; the default artifact store is in-memory, with relational storage optional
Verified 2026-09-26