Closed gusbeane closed 4 years ago
Yes, I should redefine the internals of loglike
and any other associated functions to actually track the variable names used in fit
. Thanks.
In fit
, the parameter log_initthresh
is passed to _optimize_fit
as init_thresh
, and within _optimize_fit
the action log(init_thresh)
is taken, so this breaks a user-facing parameter, since one would often want to set log_initthresh < 0
. Adding a bug label
Note that it is technically logl
, not log
. So this isn't actually a bug.
There shouldn't be two
wt_thresh
variables running around.