lukasmasuch/best-of-ml-python
A weekly ranked census of Python’s sprawling ML zoo
It auto-ranks nearly a thousand Python ML libraries by health metrics so you don't have to guess what's alive and what's abandonware.
★23.7k stars Learning
Not currently ranked — collecting fresh signals.
star history
What it does This repository is a living directory of 920 open-source Python libraries for machine learning, bucketed into 34 categories from frameworks to vector search. A bot recalculates a project-quality score each week using public GitHub stats and package-manager data, then re-sorts the list. It is essentially a periodically refreshed card catalog for PyPI’s ML aisle.
The interesting bit The ranking is not just stars-and-prayers. It folds in contributors, forks, issues, download counts, dependents, and last-update timestamps into a single score, and marks dead or inactive projects with skull and sleeping icons. That makes it mildly useful for spotting abandonware before you vendor it into production.
Key highlights
- 920 projects tracked across 34 categories, totaling 5.1M GitHub stars
- Auto-generated quality scores derived from GitHub and package-manager metrics
- Visual status icons flag new, inactive, dead, trending, and risky-license projects
- Covers niches often missed by generic awesome-lists: geospatial, medical, adversarial robustness, ANN search
- Updated weekly via automated releases
Caveats
- The README does not fully expose the scoring formula, so take the medals with a grain of salt
- Some categories are thin (tabular data has six projects, database clients has one), so coverage is broad rather than deep everywhere
Verdict Worth bookmarking if you are exploring tools for a new ML pipeline or auditing dependencies for bit-rot. Skip it if you already know your stack by heart and distrust algorithmic popularity contests.
Frequently asked
- What is lukasmasuch/best-of-ml-python?
- It auto-ranks nearly a thousand Python ML libraries by health metrics so you don't have to guess what's alive and what's abandonware.
- Is best-of-ml-python open source?
- Yes — lukasmasuch/best-of-ml-python is open source, released under the CC-BY-SA-4.0 license.
- How popular is best-of-ml-python?
- lukasmasuch/best-of-ml-python has 23.7k stars on GitHub.
- Where can I find best-of-ml-python?
- lukasmasuch/best-of-ml-python is on GitHub at https://github.com/lukasmasuch/best-of-ml-python.