目录
是谁在说话?谷歌多人语音识别新系统错误率降至2%
Joint Speech Recognition and Speaker Diarization via Sequence Transductio