This is a simple (and parallelizable) tuning method for cdec
which enables training weights of very many (sparse) features
on the full training set.

Please cite as:
>  "Joint Feature Selection in Distributed Stochastic
>   Learning for Large-Scale Discriminative Training in
>   SMT" (Simianer, Riezler, Dyer; ACL 2012)
>

Building
--------
Builds when building cdec, see ../../BUILDING .

Running
-------
Download runnable examples for all use cases from [1] and extract here.

Legal
-----
Copyright (c) 2012-2015 by Patrick Simianer <p@simianer.de>

See the file LICENSE.txt in the root folder for the licensing terms that this software is
released under.


[1] http://simianer.de/dtrain-example.tar.gz