That way, if you put your classifiers in joblib pipelines, once you're done with fitting steps you can just export your trained classifier with:
joblib.dump(pipe, "trained_classifier.dump")
joblib.load("trained_classifier.dump")
That way, if you put your classifiers in joblib pipelines, once you're done with fitting steps you can just export your trained classifier with:
And resume your work with: Considering this works for any Python object, a lot of heavy lifting can be exported for later (swift) use this way.