I updated pymc to the latest version and ran the model in this gist (should run as is). Unexpectedly, the Metropolist step is chosen for word_order_prior, learning_weights, and softmax_alpha: [image] This is despite the fact that they all have a continuous distribution & a previous version of pym…

Latest pymc version autoassigns Metropolis step to continuous variable

ricardoV94 October 19, 2023, 6:58pm 2

What happens if you call model.dlogp()?

Topic		Replies	Views
Switch point Metropolis tuning Questions	25	5000	May 19, 2018
Error in the sampling process, 'sd' not taken in account in Metropolis.py Questions	8	1506	September 29, 2017
AssertionError in PyMC3 code using Deterministic distributions Questions	10	3299	April 18, 2019
Sample method of pymc3 Questions	4	685	July 15, 2018
Sampling crashes with "The derivative of RV [whatever] is zero" when using a custom likelihood defined with pm.Potential Questions	4	4028	February 23, 2019