Repository navigation
v2.2.1
In this version, we added a uniform denoiser (shallow diffusion) to further increase the generation quality of both diffusion trained on bad datasets and HiFiSinger. This is the latest stable version of V2.
In this version, we added a uniform denoiser (shallow diffusion) to further increase the generation quality of both diffusion trained on bad datasets and HiFiSinger. This is the latest stable version of V2.