echo # tra #02/07/2016 Corpus 462 and corpus 45 is cleaned and created again. No errors in the dataset. Word2vec model is trained on wikipedia data and also on TRA data Perform a grid search.