Hi devs and community,
I’m getting very different n_eff
values in NUTS mcmc.print_summary()
(although the parameter estimates are same) when changing name of the sample site for some random variables in my NumPyro model (in numpyro.sample('some_name', dist)
)
I was not able to reproduce this with a sample data and I won’t be able to share my data.
Is this a known issue? Could you please suggest what could be happening?