This is a package for handling several different models that represent words using real-valued vectors: improved Collobert & Weston (2008, 2011) embeddings and (Reflective) Random Indexing (Cohen et al., 2010).

There is also a battery of tools for evaluating the quality of these word embeddings. The main program is implemented in C/BLAS, since the C&W models take quite a while to train. 

 
GPL 3

This software is licensed under the GNU General Public License (GPL) version 3, protecting among other things the freedom to use, modify and redistribute the software.

Download: Latest snapshot

Contact: Robert Östling