Improve constrained optimization#368
Improve constrained optimization#368bwheelz36 merged 10 commits intobayesian-optimization:masterfrom
Conversation
|
I have a feeling that test may have failed randomly, can you make a tiny change and push to force it to run again? |
Codecov ReportBase: 99.27% // Head: 99.04% // Decreases project coverage by
Additional details and impacted files@@ Coverage Diff @@
## master #368 +/- ##
==========================================
- Coverage 99.27% 99.04% -0.23%
==========================================
Files 8 8
Lines 549 525 -24
Branches 82 87 +5
==========================================
- Hits 545 520 -25
Misses 2 2
- Partials 2 3 +1
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
|
I added a stopgap so that I can make sure loading from logs behaves correctly and replaced the (imho) slightly obscure @bwheelz36 feel free to review at your convenience -- I've addressed all the things I wanted to address. |
941e71c to
8bf1ef8
Compare
ScreenLoggerdoesn't print information related to constraints #363: ScreenLogger now prints "allowed"-nesslb(lower bound) is lower thanubupper boundSomething that I mentioned in the original PR but then forgot to do.
ConstrainedTargetSpaceintoTargetSpaceOriginally I thought there would be less clutter if I subclass the original
TargetSpace, but I think it will make future maintenance much harder since code will need to be changed in multiple places at once (see e.g. feat(bayesian_optimization.py): Allow registering of temporary observations #365).int(presumably the bug mentioned in Constraint Error #361)ScreenLoggerdoesn't print information related to constraints #363 as well)