mvEllipticalSliceSamplingSimple error in "Episodic Diversification Rate Estimation" tutorial

[robinmdbeck]

I attempted to replicate the "Episodic Diversification Rate Estimation" tutorial using the example primates tree here: https://revbayes.github.io/tutorials/divrate/ebd.html. I set up the analysis exactly as specified in the tutorial, then started the MCMC. After 10000 generations, I got the following error message:

Error: mvEllipticalSliceSamplingSimple has iterated 1000 times without success. This should not be possible, something may be wrong.

I am using the Windows executable of RevBayes v.1.2.1 on a Windows 10 laptop. All commands were copied and pasted directly from the tutorial, and I used the primate_tree.nex file available on the tutorial page.

Thanks for your help!

Robin

[hoehna]

Hi Robin, thanks for the info. We have seen this behavior sometimes and apparently still need to fix the problem. In the meantime, just try running the analysis without the slice sampler, i.e. commenting it out. Sebastian

…

On Sun, Sep 17, 2023, 20:49 robinmdbeck ***@***.***> wrote: I attempted to replicate the "Episodic Diversification Rate Estimation" tutorial using the example primates tree here: https://revbayes.github.io/tutorials/divrate/ebd.html. I set up the analysis exactly as specified in the tutorial, then started the MCMC. After 10000 generations, I got the following error message: Error: mvEllipticalSliceSamplingSimple has iterated 1000 times without success. This should not be possible, something may be wrong. I am using the Windows executable of RevBayes v.1.2.1 on a Windows 10 laptop. All commands were copied and pasted directly from the tutorial, and I used the primate_tree.nex file available on the tutorial page. Thanks for your help! Robin — Reply to this email directly, view it on GitHub <#383>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB6P5R37VFGB3MNTD3HLEH3X25A4TANCNFSM6AAAAAA43ZGWTQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[robinmdbeck]

Thank you!