[Insight-users] Problem with lbfgsb optimiser

afe fedorov at bwh.harvard.edu
Thu Sep 19 12:42:33 EDT 2013


Tim,

thank you for investigating this and sharing your experience. I've also had
mixed success with this optimizer, but I never traced it down to see if this
is due to the specifics of a registration tool I use, or to the optimizer
itself. Looks like you made a good progress to localize the issue.

I would be very interested to hear about your experience applying different
optimizers instead of LBFGS. Especially, how well those optimizers work on
datasets where LBFGS works well. I think one advantage of LBFGS is it allows
to set bounds, which is not supported by other optimizers. I am currently
planning to explore the behavior of the registration tool I use (BRAINSFit,
https://github.com/BRAINSia/BRAINSTools) when I switch from LBFGS to
gradient descent for B-spline registration stage.

AF



--
View this message in context: http://itk-insight-users.2283740.n2.nabble.com/Problem-with-lbfgsb-optimiser-tp7583999p7584004.html
Sent from the ITK Insight Users mailing list archive at Nabble.com.


More information about the Insight-users mailing list