You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I was speaking with Kai about this and agreed it would be good to be able to identify divergence of a particular transition (Sec. 6.2 in here ) allowing us to identify pathologies which may bias the posterior sample.
The text was updated successfully, but these errors were encountered:
Thanks for raising this issue! Sec. 6.2 in fact corresponds to https://github.com/TuringLang/AdvancedHMC.jl/blob/master/src/proposal.jl#L104 in the codebase (not obviously to catch as we are following the alg. description in the original NUTS paper). After the current open PR (#49), in which the interface to return this divergence number is set up, is merged, we plan to introduce a type for this divergence information and hopefully this issue will be solved by then.
I think I misread the section number. This needs some effort to implement.
I was speaking with Kai about this and agreed it would be good to be able to identify divergence of a particular transition (Sec. 6.2 in here ) allowing us to identify pathologies which may bias the posterior sample.
The text was updated successfully, but these errors were encountered: