org.clulab.processors.clu.sequences
Counts the bigrams seen in this corpus so we filter out the non-frequent ones
The training corpus