Created by: pkuyym

If training texts are end of white space, sentences generated by decoder should be end of white space too. In such situation, we need a post processing logic to re-calculate a LM score after replacing white space by end token.