I have build a custom estimator in sklearn, according to the documentation. Inside the fit method, I have a print statement, which I use for debugging. When I create an instance of this model and call the fit method, everything is working as expected. However, when calling cross_val_score, the statements are not printed anymore, although as far as I know, cross_val_score calls the fit method at each fold.
Is there any way to force the print statement inside cross-validation?
Asked
Active
Viewed 219 times
1
desertnaut
- 57,590
- 26
- 140
- 166
Radu Gh
- 33
- 3
-
2Could parallel processing be swallowing the output? Try with n_jobs=1? – Ben Reiniger May 26 '20 at 20:50
-
@BenReiniger yes, that was indeed the case. Thank your for your answer. So, is there any way I could prevent parallel processing from doing so? – Radu Gh May 27 '20 at 21:48
-
are you running this in a Jupyter notebook, or ...? – Ben Reiniger May 28 '20 at 16:23
-
yes, in a Jupyter Notebook – Radu Gh May 29 '20 at 12:13
1 Answers
1
Parallel processing in sklearn is handled by joblib. When running cross_val_score with n_jobs != 1, evidently that doesn't play nicely with Jupyter output:
Printed output not displayed when using joblib in jupyter notebook
Ben Reiniger
- 10,517
- 3
- 16
- 29