Exploring architectures, data and units for streaming end-to-end speech recognition with RNN-transducer.

Kanishka Rao, Hasim Sak, Rohit Prabhavalkar

Browse the full ASRU paper archive.