You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The main sampling method has a way of writing the sample data into files using the db option. But now there are two additional sampling methods, for the prior predictive and posterior predictive, but they produce results of different types (dictionaries, instead of traces), and don't support the db argument. It would be nice to extend their API to be more orthogonal.
Also, I wonder if it's best to have their return values be simple dictionaries instead of some kind of object, because it means there's no place to hang methods for them. For example, there's no place to put a method that will extract points from these results.
The text was updated successfully, but these errors were encountered:
The main sampling method has a way of writing the sample data into files using the
db
option. But now there are two additional sampling methods, for the prior predictive and posterior predictive, but they produce results of different types (dictionaries, instead of traces), and don't support thedb
argument. It would be nice to extend their API to be more orthogonal.Also, I wonder if it's best to have their return values be simple dictionaries instead of some kind of object, because it means there's no place to hang methods for them. For example, there's no place to put a method that will extract points from these results.
The text was updated successfully, but these errors were encountered: