TypeError: 'NoneType' object is not iterable

[sagarshankar]

I was running a searcher for the best pipelines on of my datasets and for some reason received the following error, which would result in no best pipeline or output for the dataset. The error is coming from the python file brute_force_search.py when its iterating through pipeline_result.error which is giving me a NoneType object for some reason which cannot be iterated through. What could be a reason for this occurring?

[lhenry15]

This is because the "debugging" code we have are trying to capture potential pipeline failure and return the error message to users. But, it seems like the error is due to there is no error. It used to work well without having this error, we will take investigation about this and fix it. As a temporary solution, you can simply comment line 60-67 in brute_force_search.py to get rid of this error message and get the pipeline searching result.

[sagarshankar]

I did comment out those lines, and that error went away, but it seems that even by doing that my search returns no best pipeline output. There does not appear to be any error message, but it is telling me that the Step 6 for pipeline failed. Do you know where this output message might be coming from or could it be due to the dataset that I am using?

[LSC2204]

I did comment out those lines, and that error went away, but it seems that even by doing that my search returns no best pipeline output. There does not appear to be any error message, but it is telling me that the Step 6 for pipeline failed. Do you know where this output message might be coming from or could it be due to the dataset that I am using?

I also encountered this error.Step 4 for pipeline failed. And I checked that the cause of my error was probably the version of scikit-learn, so I tried pip install scikit-learn==0.22.2.post1 and the problem was solved.