Corpora consisting of approximately 2 million words of annotated and parsed text from Chinese newswire, government documents, magazine articles, various broadcast news and broadcast conversation programs, web newsgroups, weblogs, discussion forums, chat messages and transcribed conversational telephone speech.
