Hello everyone, I am trying the following, fairly simple experiment with standard pm.sample, but I am confused by the results I see. My script looks like this: import pymc3 as pm from pymc3 import sample, Model, effective_n import numpy as np for i in range(10): with Model(): energy …

Weird effective sample size instability

_eigenfoo July 28, 2018, 3:43pm 3

Perhaps the ordered transform could help you? See #5 on Colin’s blog post.

Topic		Replies	Views
Label switching in multivariate mixtures Questions	0	765	September 28, 2020
Mixture of Linear models Questions	3	1206	July 21, 2018
The number of effective samples is smaller than 25% for some parameters Questions	4	11343	April 5, 2018
How to improve fit of Poisson Mixture Model in PyMC3? Questions	20	5074	May 24, 2018
Sampling does not start or very very slow while attempting a Mixture Model tutorial Questions	4	2267	September 19, 2018