Fetching the paper…

EESEN: End-to-End Speech Recognition using Deep RNN Models and WFST-based Decoding · Around