Closed MatiasSanchezCabrera closed 7 years ago
Python's subprocess.check_call(cmd)
suppresses the original Java exception. So, this sklearn2pmml error indicates that there's something wrong, but not what exactly.
You can see the original Java exception if you re-try the conversion of preserved pickle files using JPMML-SkLearn command-line application:
java -jar target/converter-executable-1.1-SNAPSHOT.jar --pkl-mapper-input <path to mapper pickle file> --pkl-estimator-input <path to estimator pickle file> --pmml-output <pmml file>
Paste the output of the above command here. Alternatively, you can send these preserved pickle files to my e-mail (assuming they do not contain overly sensitive information) and I will conduct the investigation on my own.
I will get in touch with you through email.
Thanks a lot!
Hello Villu,
I was updating my prediction model, and my script started failling were it worked before. I just added a couple of variables and transformations to the mapper, but the pmml generation keeps failing even though I ignore them.
Will generate this result:
Do you have any ideas of what is not working?
Thanks in advance