datafiles/corpora README

This folder contains corpora that have been pre-processed
and saved as pairs of text files. For each pair, one file contains
a bunch of sentences in language A, one sentence per line, with
spaces separating the words, and the other file contains the
same number of sentences in the same format but in language B,
such that line i in the second file contains the translation of
the sentence from line i in the first file. These pairs are meant
to be used by the CorpusFromTextPair class.

There are two copies of each file; one contains all of the sentences from the Proceedings of the European Parliament and the other contains the randomized set of sentences that we used for training.

These files support spanish to english (en-es) and french to english (fr-en) translation.

The folder results contains:
1)the sentences in spanish that we used to test our system
2)their english translations
3)the english translations generated by several different configurations of our system
