facebookincubator/MCGrad

MCGrad is a scalable and easy-to-use tool for multicalibration. It ensures your ML model predictions are well-calibrated not just globally (across all data), but also across virtually any segment defined by your features

55
/ 100
Established

When your machine learning model makes predictions, it's crucial that these predictions are reliable, not just overall, but also for specific groups within your data, such as users from different countries or with varied content preferences. MCGrad takes your model's initial predictions and the true outcomes, along with descriptive features, and produces adjusted, more accurate predictions. This tool is for data scientists, ML engineers, or anyone deploying predictive models in fields like online advertising, recommendation systems, or fraud detection, who needs their models to be fair and accurate across diverse user segments.

Available on PyPI.

Use this if you need to ensure your machine learning model's predictions are consistently reliable and unbiased across many different user groups or data segments, without manually defining those groups.

Not ideal if your primary concern is overall model accuracy rather than equitable prediction reliability across specific subpopulations.

predictive-modeling algorithmic-fairness model-calibration responsible-AI online-advertising
Maintenance 10 / 25
Adoption 6 / 25
Maturity 24 / 25
Community 15 / 25

How are scores calculated?

Stars

18

Forks

4

Language

Jupyter Notebook

License

MIT

Last pushed

Mar 13, 2026

Commits (30d)

0

Dependencies

12

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/facebookincubator/MCGrad"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.