How to specify that a kernel hyperarameter is always fixed in scikit-learn GP kernels? I am trying to implement the following simple custom GP kernel in scikit-learn: