Implementation and derivation of "Variational Bayesian inference for a nonlinear forward model." [Chappell et al. 2008] for arbitrary, user-defined model errors.
Instead of calling the logpdf individually for each term, we now
concatenate all terms and call logpdf once. Fewer individual numpy
evaluations is always a good idea!
Instead of calling the logpdf individually for each term, we now concatenate all terms and call logpdf once. Fewer individual numpy evaluations is always a good idea!