Sequence Distillation for Purely Sequence Trained Acoustic Models.

Naoyuki Kanda, Yusuke Fujita, Kenji Nagamatsu

Browse the full ICASSP paper archive.