Abstract: We propose a lightweight neural front-end framework for on-device speech generation and highlight its benefits towards low-resource language scaling. While data-driven models have shown ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback