1 paper
Yichen Han, Xiaoyang Hao, Keming Chen +25
Text-to-speech (TTS) synthesis has seen renewed progress under the discrete modeling paradigm. Existing autoregressive approaches often rely on single-codebook representations, whi…