Sklearn2pmml raises an error on 'classes_' parameter

Question

I'm trying to create a PMML from a model, using this way :

from sklearn.preprocessing import LabelEncoder
y_h_train = LabelEncoder().fit_transform(y_train.copy(deep=True))

modele_label_encoded = model.fit(X_train, y_h_train)



import sklearn2pmml as skpmml
from sklearn2pmml import PMMLPipeline, sklearn2pmml

cols_used = list(X_train.columns)
pmml_model = skpmml.make_pmml_pipeline(modele_label_encoded)
pmml_model.active_fields = cols_used
pmml_model.target_fields = ['Traget']
skpmml.sklearn2pmml(pmml_model, os.path.join('test.pmml'), debug=True)

And the PMML seems not to understand the classes_ :

Exception in thread "main" java.lang.IllegalArgumentException: The value of 'xgboost.sklearn.XGBClassifier.classes_' attribute (null) is not a supported array type

I don't understand the error since my classes_ has an array type, as requested :

modele_label_encoded.classes_
Output : array([0, 1])

I found no post about this error, does someone know how to solve it?

Case also opened here : github.com/jpmml/sklearn2pmml/issues/466 — Adept
– Adept, Commented Nov 12 at 11:04

Adept · Accepted Answer · 2025-11-17 14:17:04Z

0

Question answered here : https://github.com/jpmml/sklearn2pmml/issues/466

TLDR : Version of sklearn2pmml is too old. Find in their github the file requirements.txt of your version of sklearn2pmml, and adapt all other libraries to the ones on the file

answered Nov 17 at 14:17

Adept

5643 silver badges18 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Sklearn2pmml raises an error on 'classes_' parameter

1 Answer 1

Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Related