Hello, Currently, I am trying to implement PyMC (5.10.4) and I’m noticing that nearly every sample is divergent. When I make a corner plot of the trace, I see that my posteriors look correct. From what I have found, I see that you can either re-parameterize or turn up the acceptance probability. Si…

Correct posteriors but many divergences

AstroParrotKing March 13, 2024, 7:57pm 10

Since I still have divergences, the model looks fine, and the gradients are relatively okay. Does this mean that manual parameterization is needed?

Topic		Replies	Views
There were 6 divergences after tuning. Increase `target_accept` or reparameterize Questions	5	4316	August 19, 2021
Overcoming divergence in MvNormal covariance model? Questions	1	481	January 28, 2019
Fitting a complex model with a large dataset: chain contains only diverging samples Questions	5	1276	December 12, 2019
Reducing divergences : Implementation challenges Questions	7	2213	August 3, 2021
Advice with hierarchical model Questions	9	2258	October 8, 2018