How to exit a nested sampling run that hangs

[amunozj]

Hi friends!

I am using nested sampling to find the evidence associated with a model fit. Under certain conditions, everything runs nicely, under others, it simply hangs after displaying the following message:

Number of parallel Markov-chains set to: 50

I have played around with all the input parameters to DefaultNestedSampler, as well as working in 64-bits, to no avail. Is there a way to diagnose what the problem is? and or killing/stopping the call if it takes too long?

thanks a lot!

[amunozj]

My bad friends. My calculation of log_prob inside my model was returning -inf in some cases and that was leading to the hang. So if you run into this problem, check your log_prob :)

[Joshuaalbert]

Just in case others stumble on this there is a very helpful method for the model ```model.saneity_check(key, 100)```On Mar 20, 2024 00:17, Andres Muñoz-Jaramillo ***@***.***> wrote: Closed #155 as resolved. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>