US patent US12802698
Method and apparatus for training speech conversion model, device, and medium
Abstract
This application discloses a speech conversion method performed by a computer device. The method includes: obtaining first accent audio corresponding to a first accent; extracting a first content feature from the first accent audio through the first ASR model, the first content feature being corresponding to the first accent; converting the first content feature into a second content feature through the second conversion model, the second content feature being corresponding to a second accent; and performing audio conversion on the second content feature through the third conversion model to obtain second accent audio.