Implementation and derivation of "Variational Bayesian inference for a nonlinear forward model." [Chappell et al. 2008] for arbitrary, user-defined model errors.
MIT License
2
stars
1
forks
source link
Add vb.scaling_eps to control when to apply scale #48
Previously, the scaling was only applied to parameters with mean > 1. This 1 is now replaced by scaling_eps = 1.e-20 (and adjustable via kwargs) to also affect parameters with numerically small means.
Generally, this is quite hacky. An alternative can be discussed here #47
Previously, the scaling was only applied to parameters with mean > 1. This
1
is now replaced byscaling_eps = 1.e-20
(and adjustable via kwargs) to also affect parameters with numerically small means. Generally, this is quite hacky. An alternative can be discussed here #47