Deep Speech 2: End-to-End Speech Recognition in English and Mandarin