OmniVoice notice

AbstractVoice can optionally use OmniVoice for local omnilingual TTS, voice
design, and prompt-audio voice cloning.

Upstream sources:
  https://github.com/k2-fsa/OmniVoice
  https://huggingface.co/k2-fsa/OmniVoice

The optional OmniVoice Python package, model weights, tokenizer files, and
related assets are not vendored into this repository. They are installed or
downloaded only when users select the `abstractvoice[omnivoice]` extra and run
an explicit prefetch command, or when downloads are otherwise enabled by the
application.

Before deploying or redistributing OmniVoice assets, review the upstream
license and model-card terms for the exact version/revision in use:
  https://github.com/k2-fsa/OmniVoice/blob/master/LICENSE
  https://huggingface.co/k2-fsa/OmniVoice
