@tcapretto This might be a silly question, but what do I do in Bambi to use the find_map() function in PYMC? I am trying to essentially compare the frequentist likelihood vs Bayesian approach. I know that I could do a .build and then run PYMC but wondering if I can skip that and just run it directly out of Bambi.
1 Like
Hey @zweli! Thanks for the question. It’s not available for now. In principle, I don’t think it would be hard to add it, but I’m unaware of any potential rough edges.
1 Like
Great! I am trying to educate people about Bayesian methods in my space and it is helpful to compare methods on the same likelihood.
1 Like
We can now just use model.build() + Bayeux’s bx.Model.from_pymc(model.backend.model) and an optimize function from Bayeux. More info here: Introducing bayeux - #5 by zweli and bayeux.
1 Like
You should also be able to just call find_map with Bambi backend model object
1 Like