The Readme says FluidAudio's Pocket TTS supports voice cloning, but the hugging face repo says the cloning weights are not available from Fluid because the Pocket TTS model is gated. What's the reality?
If FluidAudio can't do voice cloning in PocketTTS, please update the docs.
If it can, please update the model page with info on how to do so.
Best of all, see if Kyutai will let you contribute a CoreML version of the gated model to their official repo!
The Readme says FluidAudio's Pocket TTS supports voice cloning, but the hugging face repo says the cloning weights are not available from Fluid because the Pocket TTS model is gated. What's the reality?
If FluidAudio can't do voice cloning in PocketTTS, please update the docs.
If it can, please update the model page with info on how to do so.
Best of all, see if Kyutai will let you contribute a CoreML version of the gated model to their official repo!