Hi, I had the same problem.
You can solve the problem by using "saveCompactModel" and "loadCompactModel".
Given the trained model, you can sace the model ("Mdl") to a file named "Model.mat".
Mdl = fitrgp(X, Y);
saveCompactModel(Mdl, 'Model');
In a Simulink model, for example, you can use MATLAB Function and load the model as follow;
function y = compute_y(x)
Mdl = loadCompactModel('Model');
y = predict(Mdl, x);
end
