Open capabilities will be gradually released (word embedding, Chinese-English translation, layout analysis)
