[mlpack-git] [mlpack] FFN serialization (#516)
Ryan Curtin
notifications at github.com
Thu Feb 11 15:45:03 EST 2016
My thoughts here don't apply to this situation in particular (I haven't really looked over this), but I think that in general, serializing user preferences for training is worth doing. For instance, a user may want to load a model, and then train it some more. Any intermediate training variables probably aren't worth saving, but something like a learning rate is worth holding onto, in my opinion. This is what I've tried to do with other serialization code. I hope these thoughts are helpful. :)
---
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/pull/516#issuecomment-183053985
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160211/9b3e8920/attachment.html>
More information about the mlpack-git
mailing list