mljar / mljar-supervised

Python package for AutoML on Tabular Data with Feature Engineering, Hyper-Parameters Tuning, Explanations and Automatic Documentation
https://mljar.com
MIT License
3.05k stars 406 forks source link

Error with golden_features #450

Closed toan01-uet closed 3 years ago

toan01-uet commented 3 years ago

I can't predict target, because the number of features in data(train, test) is not same. Although the number of features in X_train and X_test is same. The fit method used:

from supervised.automl import AutoML
automl_co = AutoML(mode="Compete", 
                   total_time_limit = 12600, 
                   golden_features=True,
                    features_selection=False,
                    stack_models=True,
                    train_ensemble=True,
                    explain_level=0,
                   results_path='./mljar_co',
                    validation_strategy={
                        "validation_type": "kfold",
                        "k_folds": 5,
                        "shuffle": False,
                        "stratify": True,
    }) 
automl_co.fit(X_train, y_train)
y_pred = automl_co.predict_all(X_test)

The error:

LightGBMError: The number of features in data (129) is not the same as it was in training data (130).
pplonski commented 3 years ago

@toan01-uet thank you for reporting. Looks like some bug. I will need code and data to reproduce it.

pplonski commented 3 years ago

@toan01-uet closing the issue, I cant reproduce it. Please provide input data so I can reproduce the problem.