A Comparison of Modeling Units in Sequence-to-Sequence Speech Recognition with the Transformer on Mandarin Chinese.

Shiyu Zhou, Linhao Dong, Shuang Xu, Bo Xu

Browse the full ICONIP paper archive.