Rename `prior_model` into `datareg`
The regularizer might be not squared, so the term is properly a data vector for the regularization.
prior_model
could be a initial guess for the inversion (and I strongly suggest to use initial_guess
for the variable).