I’m fairly new to HMC, how would I test that? I assume that it’s per chain.
To address the issue of multi-modality in omega and Omega, I’ve adjusted the uniform priors for these variables to range from 0 to pi instead of 0 to 2*pi. However, despite this adjustment, I’m still seeing the divergences. Here is the pair plot with the parameter ‘divergences=True’.
The reason the covariances for omega, Omega, and Tau is split in half is because the range was split in half.
As far as I can tell most samples are an issue. I believe that it’s most likely because of omega, Omega, and Tau. Maybe it is something easy?