online-ml/river

Health v2 · maintenance only

Adoption not part of the score

5924 stars · 659 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

River is a Python library for online machine learning, allowing models to learn incrementally from streaming data one observation at a time. It is the merger of the creme and scikit-multiflow projects and provides a scikit-learn-like API for streaming pipelines, metrics, and preprocessing.

Use cases

• train machine learning models on streaming data in python
• incrementally update models without retraining from scratch
• handle concept drift in real-time predictions
• build online learning pipelines like sklearn but for streams
• compute streaming statistics and metrics on the fly
• classify or regress data one sample at a time

When to choose

• your data arrives as a stream and doesn't fit in memory
• you need models that update continuously in real time
• you want a scikit-learn-like API for incremental learning
• you need to detect and adapt to concept drift

When to avoid

• you need batch learning on static datasets with GPU acceleration
• you rely on the full scikit-learn ecosystem of estimators
• your project requires deep learning frameworks like PyTorch

Facets

library · maturity active

machine-learning streaming data-science machine-learning data-science python cross-platform online-learning incremental-learning concept-drift stream-processing scikit-learn-alternative real-time

Member repositories

For agents

markdown · JSON · MCP: product_card(name="online-ml/river")