Hi @NateAM! First, thank you again for your previous feedback.
I am still stuck triny to implement your approach to my problem. Speciffically, when used to estimate in a surrogate model of the form y=f(x1, x2, x3) your definition of wrapperSurrogate doesn’t work (the kernel crashes). Since no feedback is produced, it is very difficult to debug.
I tried to rebuild your line of though by following your previous threads (here) for Guassian Process, but I couldn’t adapt it to my problem. I need to better understand which methods are needed to implement in order to PyMC to be able to sample from any surrogate model from any input shape. Is there some documentation I can read? I’ve seen some feedback from @junpenglao at the old thread. any feedback is welcomed.