Description
Streaming machine-learning models can be trained incrementally while testing development changes from River. Data scientists and developers can build classifiers, regressors, anomaly detectors, and evaluation pipelines without retraining on full batches each time. This Git build is mainly for testing unreleased River behavior.