Any plans to integrate Qwen TTS to apple?

[zaidbren]

Hello everyone, I love the WhisperKit package for the speech to text capabilities, and I want to implement the TTS using Qwen models like 0.6B and 1.7B. However, I am not being able to find any reliable package like WhisperKit for that.

Is there any way to integrate this in my app? Are there any future development for this?

[ZachNagengast]

In fact, you're in luck! This PR is dropping very soon #425. Give the sample app a try if you're curious - would love the feedback!

[zaidbren]

@ZachNagengast Thank you so much, I need to implement CustomVoice TTS Qwen today itself, is there any chance to use it today?

[zaidbren]

@ZachNagengast Is there any way to try the "TTSKit Example app"? Can I see the code implementation as well?

[ZachNagengast]

Yes that requires the 1.7b model, the UI will have a text area for instruction when that one is loaded

[zaidbren]

Ohh sure, I missed that part, sorry

Also, may I know tomorrow you mentioned there's gonna be new update which is faster, would that be a big significant change or an optimization in the current architecture. I just want to be sure because I am implementing the library in my app today as its the deadline I have to maintain. Any hints of what changes would be pushed?

[ZachNagengast]

Should be mostly under the hood, any api changes should be obvious and easy to adapt to, but it's still a work in progress so TBD. Model download locations and file management will remain the same.

[zaidbren]

@ZachNagengast Any plans for the Voice Cloneing model "Base" and the Voice Design?

[ZachNagengast]

Yes big plans 👍