If I want to use OpenNMT(py) to translate between English and Chinese?

Besides parallel-data,which part of code should I change?

no changes for the code itself, except some necessary preprocessions for data.

I’m trying to train a model english -> chinese, and i have added spaces between words manually in chinese sentence before running preprocess.lua, but the model doesn’t work. I wondor if some other preprocesstions are needed…Thanks a lot!