Great. About the switch point for SGD / ASGD: in the other thread David pointed to http://cs.nyu.edu/~zsx/nips2011/ . And it seems that Moving Averaged SGD might be interesting to implement to.
AFAIK it is not covered by the bounds provided by F. Bach nor by Wei Xu but if it works good in practice, a simple empirical trick like that is good enough for me ;) -- Olivier ------------------------------------------------------------------------------ This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure _______________________________________________ Scikit-learn-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
