Current models perform poorly in processing Chinese speech, with a word recognition accuracy of less than 60%. Many Chinese words are incorrectly identified as other Chinese words, highlighting the need for a more intelligent and accurate Chinese speech model.
