search

gradientinstitute / aboleth

A bare-bones TensorFlow framework for Bayesian deep learning and Gaussian process approximation
Apache License 2.0
127 stars 10 forks source link

Turn prior variable optimisation *off* by default? #174

Closed dsteinberg closed 6 years ago

dsteinberg commented 6 years ago

since there are no guarantees that ELBO is a lower bound on LML for these...

Also, we can see the effect of this using #173