Are you working on VC (input: speech of one speaker, output: the same spoken content, but sounds like another speaker) or speaker-adaptive speech synthesis (inp

Are you working on VC (input: speech of one speaker, output: the same spoken content, but sounds like another speaker) or speaker-adaptive speech synthesis (input: text, output: speech)? Also check out ParallelWaveGAN, another high-quality and very fast (on CPU) neural vocoder.