1 paper · 1 filter
Dongya Jia, Zhuo Chen, Jiawei Chen +8
Several recent studies have attempted to autoregressively generate continuous speech representations without discrete speech tokens by combining diffusion and autoregressive models…