I am using GPCoregionalized model to predict y1 and y2 out of input data containing 2 features (x1,x2). The code am using for training is as follows:

X_list=[x1,x2]
Y_list=[y1,y2]
m = GPy.models.GPCoregionalizedRegression(X_list, Y_list, kernel=kernel)

Then to make predictions am using the following code snippet:

Xt = np.column_stack((x1t, x2t))
newX = np.hstack([Xt,np.ones_like(Xt)])
noise_dict = {'output_index':newX[:,3:].astype(int)}
preds, sigma = m.predict(newX,Y_metadata=noise_dict)

I assume this is calling for predictions of the first output y1, how can i get predictions for the second output y2?