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
Could you add args= and kwargs= when initializing the NestedSampler object that would take an argument list and/or a keyword dictionary that get passed to the likelihood function? This would allow for likelihood functions of the type lnlike(v,*args,**kwargs), and would allow for backwards compatibility with EMCEE.
The text was updated successfully, but these errors were encountered:
As referenced in the commit, this adds support for both providing args and kwargs to the loglikelihoodandprior_transform function via some new input arguments using the same scheme as emcee. I did some brief consistency tests and didn't find anything breaking in both the serial and parallel case. Let me know if you find any bugs @pacargile.
New version release coinciding with a reasonable bugfix. Consolidates
the previous few commits (related to #48 and #49) and includes a further
internal change to how evidence weights are computed in `weight_function`
that should make computations much more stable (which led to problems in
some testing).
Could you add args= and kwargs= when initializing the NestedSampler object that would take an argument list and/or a keyword dictionary that get passed to the likelihood function? This would allow for likelihood functions of the type lnlike(v,*args,**kwargs), and would allow for backwards compatibility with EMCEE.
The text was updated successfully, but these errors were encountered: