I'm calling xgboost via its scikit-learn-style Python interface: model = xgboost.XGBRegressor() %time model.fit(trainX, trainY) testY = model.predict(testX) Some sklearn models tell you which