baal-org/baal

Bayesian active learning library for research and industrial usecases.

50
/ 100
Established

This helps machine learning practitioners efficiently train models, especially when labeling data is expensive or time-consuming. You input an existing dataset with some labeled examples and a larger pool of unlabeled data, and it outputs a highly accurate model with significantly fewer labels than traditional training methods. Researchers and MLOps engineers who develop and deploy AI models will find this useful.

923 stars.

Use this if you need to build robust machine learning models but have limited resources for data annotation.

Not ideal if you have abundant, readily available labeled data or are not working with deep learning models.

machine-learning-operations data-annotation model-training deep-learning-research unsupervised-learning
No Package No Dependents
Maintenance 6 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 18 / 25

How are scores calculated?

Stars

923

Forks

89

Language

Python

License

Apache-2.0

Last pushed

Dec 03, 2025

Commits (30d)

0

Get this data via API

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

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