Please check :doc:`recipes` and :doc:`overview` first. NNSVS v0.0.3 and later supports neural vocoders. This page summarizes how to train neural vocoders. The current implementation of NNSVS does not ...
DisCoder is a neural vocoder that leverages a generative adversarial encoder-decoder architecture informed by a neural audio codec to reconstruct high-fidelity 44.1 kHz audio from mel spectrograms.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback