[scikit-learn] Gaussian Process Regression with multiple features

Hi, I am trying to predict a target (DeltaNDT) which is dependent on five features (Cu, Ni, P, T, Fluence). Please find attached the ipython notebook and the csv input data. gpr.predict is generating predictions for the train data. However, with test_preds = gpr.predict(X_test) all values test_preds are zeros. Please help me to find the mistake in my approach. Thanks. Apu