Version 1.2
-----------

-Fixed an issue where C code of GMM.learn would hang when encountering numerical instability in high-dimensional datasets
-No longer duplicating inputs to .C calls

Version 1.1
-----------

-Changed debug.output levels for several functions
-Fixed documentation incosistencies
-Added additional argument verification to GMM.learn
-Added covariance matrix regularization to LDA.train and QDA.train
