That post you are looking at is pretty old and the implementation is not optimal. You should instead take a look at How to marginalized Markov Chain with categorical?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Sequence of Observed in a loop, how the log-likelihood are estimated? add them all? | 9 | 1812 | November 16, 2018 | |
| How to marginalized Hidden Markov Model with categorical? | 9 | 5788 | December 19, 2018 | |
| Baum-Welch example in PYMC3 | 6 | 1534 | July 15, 2020 | |
| Custom Categorical Distribution - ensure bounded candidates | 13 | 492 | March 31, 2024 | |
| Importing POMDP code in pymc3 to pymc - Code is running with erroneous results | 5 | 74 | April 9, 2025 |