Skip to content

Feature Request: Support Microsoft VibeVoice-ASR-BitNet #45

Description

@hossein-mohseni

Hi! 👋

I'd like to request support for Microsoft's new VibeVoice-ASR-BitNet model:

https://huggingface.co/microsoft/VibeVoice-ASR-BitNet

Why it would be valuable
🚀 Optimized for fast CPU inference using BitNet quantization.
💾 Much lower memory usage than traditional ASR models.
🎙️ High-quality multilingual speech recognition with support for many languages.
🇮🇷 Please consider adding Persian (Farsi) language support as well, as it would be very useful for Persian-speaking users.
👥 Built-in speaker diarization support.
⏱️ Timestamp generation.
💻 Makes local/offline transcription much more accessible for users without high-end GPUs.

Since VoiceFlow already supports multiple ASR backends, adding VibeVoice-ASR-BitNet would provide users with another excellent option, especially for efficient local CPU-based transcription.

Adding Persian (Farsi) support would make the project more accessible to a wider community and would be a great addition.

I think it would be a great improvement for the project. Thanks for considering it, and keep up the awesome work! ❤️

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions