[mlpack-git] [mlpack/mlpack] SGD performing optimization on same order of visitation at each rotation (#608)

Ryan Curtin notifications at github.com
Sat Apr 9 12:35:33 EDT 2016

Hi there Ozair,

It's actually a bad idea for debugging and reproducibility to set the seed.  Setting the seed (preferably by calling `math::RandomSeed()`, but the Armadillo function you mentioned works too) should be done at a higher level than SGD.  If the seed is set inside of the SGD class, then if I have a bug that I am trying to track down, it becomes very hard if I can't easily set the seed to one specific value (or leave it as the default).  If anything it may be worth adding a bit to the documentation or one of the tutorials (probably a tutorial is the better place) noting that it might be useful to set the random seed at a higher level.

You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160409/3c79218e/attachment.html>

More information about the mlpack-git mailing list