accessing model.best_estimator_.feature_importances_ in gridsearchcv Random forest

Question

I'm using cuML from RAPIDS AI. I used a GridSearchCV for finding the best parameters, however i am unable to get the best features (for feature selection purposes).

here's my code:

combined_df=cpd.concat([train_df,evaluate_df])
combined_df=combined_df.astype('float32')
test_fold = [0] * len(train_df) + [1] * len(evaluate_df)

p_test3 = {'n_estimators':[45,50,200,500],'max_depth':[20,25,15,10]}#'max_features':[30,50,60,70,80]

tuning = GridSearchCV(estimator =cuRFr(n_streams=1, min_samples_split=2, min_samples_leaf=1, random_state=0), 
            param_grid = p_test3, scoring='r2', cv=PredefinedSplit(test_fold=test_fold))

tuning.fit(combined_df.iloc[:,2:].to_numpy(dtype='float32'),combined_df['Mcap_w'].to_numpy(dtype='float32'))

i recieved an error when tried tuning.feature_importances_

----> 1 best_features = tuning.best_estimator_.feature_importances_  File base.pyx:330, in cuml.interna ls.base.Base.__getattr__()

AttributeError: feature_importances_