aangelopoulos/conformal_classification

Wrapper for a PyTorch classifier which allows it to output prediction sets. The sets are theoretically guaranteed to contain the true class with high probability (via conformal prediction).

44
/ 100
Emerging

This project helps scientists, researchers, and engineers working with machine learning models by adding a crucial layer of certainty to their classifications. Instead of a single, potentially incorrect prediction, you get a 'prediction set' – a small group of possible classes that is guaranteed to include the true answer a high percentage of the time. This is invaluable when high confidence and error control are critical for image analysis, medical diagnostics, or risk assessment.

255 stars. No commits in the last 6 months.

Use this if you need provable guarantees that your image classifier's predictions will contain the correct answer with a high, user-defined probability, making your model's outputs more reliable and trustworthy.

Not ideal if you simply need the single most likely prediction from your model and do not require formal probabilistic guarantees or prediction sets.

predictive-modeling image-classification model-reliability risk-assessment scientific-research
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 18 / 25

How are scores calculated?

Stars

255

Forks

36

Language

Jupyter Notebook

License

MIT

Last pushed

Jan 23, 2023

Commits (30d)

0

Get this data via API

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

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