I try to do the dependency parse on some dataset, as I found CoreNLP is useful to do it. However， it seems like I can only use the default tokenize it offered, so in the dow
