Alejandro Molina
Alejandro Molina
can we have a skype call to go over this? thanks a lot for the push!
It’s more complicated than that. We had talked about it, basically the problem is that tensorflow doesn’t allow for probability objects of type float64. At least it didn’t back then....
could you provide some code for me to replicate that?
I started writing a multivariate gaussian Leaf. it turns out marginalization there is going to be a bit more painful :D but I believe inference at the node works, and...
sure, It's in the parametric module for now. I believe we have to add something to mark leaves as multivariate - univariate, only then I can modify the current learning...
sorry, try again
I’m working on that :) On Tue 17. Nov 2020 at 22:18, Jesus Nieto wrote: > All right, thank you for that. Do you know if there is any plan...
could you check version 0.0.41?
that would be pretty great
I really like the SKlearn wrapper, but could you elaborate a bit more on the marginal probability? can we decouple those two things?