Hi,
I’m having issues with the brms::loo
(https://github.com/paul-buerkner/brms/issues/769), and in the meanwhile I’m looking for a workaround using loo::loo
I see that there is a brms::log_lik
, but it’s not very documented, and it returns a strange object. I’m looking for a way to extact the log likelihood of a brms model in a similar way to loo:extract_log_lik
so that I can feed it to a loo::loo()
function.
Thanks!
Bruno