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
It does look that the two issues are essentially the same, and I think what @nwu63 recommended (showing all the shared values, and add the major-only where applies) still makes sense. With some luck we should be able to fix this directly within getValues().
I append an example to show the difference between major and minor history entries:
Major --> 74 dict_keys(['xuser', 'funcs', 'fail', 'iter', 'isMajor', 'nMajor', 'nMinor', 'step', 'merit', 'feasibility', 'optimality', 'penalty'])
Minor --> 76 dict_keys(['xuser', 'funcs', 'fail', 'iter', 'isMajor'])
Gradient evaluation --> 77 dict_keys(['xuser', 'funcsSens', 'fail', 'iter', 'isMajor']) (all calls to func or funcSens increase the counter, then the iteration number is "corrected" at a later stage, we discussed a bit of this in #252 and related #182 )
Description
The code incorrectly assumes that
isMajor
is only defined at major iterations.The code makes the check for the
isMajor
flag as part of the check herePossibly related to issue #250
Steps to reproduce issue
histValues = optHist.getValues('isMajor',major=False)
Current behavior
raises the warning
and returns the values of
isMajor
at only major iterationsExpected behavior
returns the value lf
isMajor
at all iterationsCode versions
List versions only if relevant
The text was updated successfully, but these errors were encountered: