Bayesian calibration on GP model

Thank you for your answer, I will gladly have a look at these notebooks. I will try PyMC v4 when possible, to compare with PyMC3.

Yet in my case, I mostly face this error:

TypeError: Unsupported dtype for TensorType: object

which, in my understanding, is caused by the fact that I put a pymc3 free variable in the X argument.