Chatterbox TTS Support

[skyne98]

Hi llama.cpp team,

I'd like to request support for ChatterBox TTS, a high-quality open-source text-to-speech model from Resemble AI.

Key advantages:

• High quality: Benchmarked against ElevenLabs and preferred in side-by-side evaluations
• Small model size: Only 0.5B parameters with Llama backbone (already supported by llama.cpp)
• Advanced features: Supports emotion exaggeration control and zero-shot voice cloning
• Open source: MIT licensed

I believe ChatterBox has high potential to become an excellent TTS option within the llama.cpp ecosystem, especially given its Llama architecture compatibility.

I'm planning to attempt implementing this myself with AI assistance, but would greatly appreciate any guidance from the community. If anyone has experience with TTS integration or is interested in collaborating, I'd welcome the partnership. Or if someone wants to grab the initiative!

Happy to share progress and contribute back any working implementation.

Thank you for the incredible work on llama.cpp!