Skip to content

stream.train + sklearn fixes

Patrick Godwin requested to merge sklearn_provenance_fixes into master

Three changes:

  • Made it so that only incremental classifiers in stream.train have their models updated at the beginning to continue retraining.
  • Fixed a typo with rank_scaler that should cause issues if you use this type of classifier.
  • Always redefine the model in SupervisedSklearnClassifier.train() before fitting the model, since it will be training a single time anyways.
Edited by Patrick Godwin

Merge request reports