Comment by diyer22

Comment by diyer22 2 days ago

0 replies

I believe DDN is capable of handling TTS (text-to-speech) tasks, because with the text condition, the generation space is significantly reduced.

And it's recommended to combine it with an autoregressive model (GPT) for more powerful modeling capabilities.