This is the WSJ parsing model for the BLLIP parser. It is trained from
PTB version 2, sections 2-21 and tuned on section 24.  Unlike the model
typically distributed with the BLLIP parser, it will not produce any
non-standard tags (e.g., AUX).

-- David McClosky
   dmcc@cs.stanford.edu
   1.29.2014
