Ok sweet, I’ll use that instead. Also, this weekend I found out an actual question to what I have been trying to ask in the past. Currently, I use GPy to create the surrogate model of the initial data and then wanted to guide Bayesian analysis on that data.I understand now that I need my mu to be a function that incorporates the GP model from GPy. I do this using gp.predict function from GPy to represent the function for mu, but this function does not take symbolic variables. I found this example that talks about the same issue, but I cannot get it to work in the slightest. Do you have any ideas on creating a Class to transform the symbolic variables into something like a numpy array or something deterministic so it can work with the GPy gp.predict function?