fit a model using mapie regressor

[sraghavan0610]

Hi I am trying to use a mapie regressor on top of a trained xgboost model trained using mljar.
This is what I am doing

automl = AutoML(results_path ="AutoML_1) # this will load the best performing model which is xgboost
mapie_re = MapieRegressor(automl, cv=10, random_state=2)
mapie_re.fit(x_train, y_train) 
mapie.predict(x_test)
 

This approach seems to not work so I would like to save the original best performing model as a .pkl or .joblib file. Please advice!

Thanks

[pplonski]

What is the error from MapieRegressor when using AutoML? The good approach might be to use AutoML to find good hyperparameters set and then create model manually which can be used with MapieRegressor.