Baal
github.com
Spec sheet
- Purpose
- Baal is an active learning library for industrial applications and research.Read 14 Nov 2026 · github.com
- Deployment
- The repository provides Docker commands to build and run an experiment.Read 14 Nov 2026 · github.com
- Integrations
- The README links a seminar with Label Studio and describes training with Hugging Face or ModelWrapper.Read 14 Nov 2026 · github.com
- License
- GitHub lists the project license as Apache-2.0.Read 14 Nov 2026 · github.com
- Active learning methods
- The framework supports Monte-Carlo Dropout, MCDropConnect, deep ensembles, and semi-supervised learning.Read 14 Nov 2026 · github.com
- Dataset handling
- ActiveLearningDataset splits a dataset into training and unlabeled pool sets.Read 14 Nov 2026 · github.com
- Framework components
- The framework has four main parts: ActiveLearningDataset, Heuristics, ModelWrapper, and ActiveLearningLoop.Read 14 Nov 2026 · github.com
- Installation
- The README gives pip and Poetry installation methods.Read 14 Nov 2026 · github.com
- Model preparation
- Baal provides modules for preparing models for active learning, including an MCDropoutModule wrapper.Read 14 Nov 2026 · github.com
- Python requirement
- Baal requires Python 3.10 or later.Read 14 Nov 2026 · github.com
- Uncertainty-based selection
- ActiveLearningLoop computes uncertainty and labels the most uncertain items in the pool.Read 14 Nov 2026 · github.com
Where Specifiction places it
- Active Learning Tools#10 of 166.9
