I know that Replicate supports streaming output for LLM's.
Is there any reason this shouldn't be easy to implement for streaming audio?
One very obvious application would be streaming text-to-speech output in near-realtime.
I can imagine a lot of apps being built with that (and that's my intended use-case) - or streaming music with MusicGen, AudioGen, etc.
I know that Replicate supports streaming output for LLM's.
Is there any reason this shouldn't be easy to implement for streaming audio?
One very obvious application would be streaming text-to-speech output in near-realtime.
I can imagine a lot of apps being built with that (and that's my intended use-case) - or streaming music with MusicGen, AudioGen, etc.