Random_state not giving same result after restarting the python program


I have the following example:

    for i in range(10):
        xgb_model = XGBRegressor(random_state=self.seed)
        xgb_model.fit(X_train, y_train)
        xgb_pred = xgb_model.predict(X_test)

        best_features = xgb_model.get_booster().get_score(importance_type='weight')
        best_features = list(best_features.keys())
        print(xgb_pred, best_features)

This code produce the same result.

I stop my python program and start it again, now I will different result from the previous run.

What could be the cause?

I am using xgboost-1.6.1

X wasn’t in the same order and was having different floating point. :frowning: