Add support for Chatterbox

[brianferri]

Model description

Chatterbox is a multilingual, zero-shot Text-to-Speech (TTS) model designed for flexible voice synthesis across a wide range of languages without requiring task-specific fine-tuning. It is built on a 0.5B parameter Llama-based architecture.

Find some more information on deepwiki

Prerequisites

• The model is supported in Transformers (i.e., listed here)
• The model can be exported to ONNX with Optimum (i.e., listed here)

Additional information

The main repo along with some examples can be found here.
Some attention to it has already been made in the onnx-community, providing onnx exports for the pre-trained model here

Your contribution

With appropriate guidance I can contribute directly to the implementation of this feature.

[xenova]

Hi there 👋 Yes, @VladOS95-cyber and I worked on the ONNX exports (at the links you provided), and chatterbox integration has been completed on my local branch for Transformers.js v4. Expect a demo in the next few days/week.

[xenova]

Chatterbox is now supported in #1382.

We'll be releasing a demo for it soon. Will keep this issue open until V4 is merged.