In order to use in languages other than English (Chinese or Japanese with tokenized), it is now possible to select words to be separated by word boundaries instead of non-wordboundary when separating words .
Coverage decreased (-2.3%) to 97.049% when pulling 0ff4e77aec656e9b61a08abce94f999daee165fb on mkt3:master into 50c45f1f4a394572381249681046f57e2bf5a591 on vi3k6i5:master.
In order to use in languages other than English (Chinese or Japanese with tokenized), it is now possible to select words to be separated by word boundaries instead of non-wordboundary when separating words .