search

joshspeagle / dynesty

Dynamic Nested Sampling package for computing Bayesian posteriors and evidences
https://dynesty.readthedocs.io/
MIT License
357 stars 77 forks source link

Samples near the edge of the box #261

Closed segasai closed 2 years ago

segasai commented 3 years ago

Over last few days I was trying to debug the issue in sampling slowing down to a crawl, and it was happening with all the samples sitting exactly at the edge of the cube (in u-space) and somehow that was breaking slice sampling. I remember multinest issues a warning in this case. That may be a good idea too. (i.e. if all live points sit within eps of 0 or 1 of the edge)

joshspeagle commented 3 years ago

Adding in warnings for this particular edge case (literally, hah) does sound like a good idea.