Description

OpenVoiceOS STT plugin for Faster Whisper

High-performance inference of OpenAI's Whisper automatic speech recognition (ASR) model:

Install

pip install ovos-stt-plugin-fasterwhisper

Configuration

available models are "tiny.en", "tiny", "base.en", "base", "small.en", "small", "medium.en", "medium", "large-v2", "large-v3"

eg, to use Large model with GPU

To use Whisper as STT

  "stt": {
    "module": "ovos-stt-plugin-fasterwhisper",
    "ovos-stt-plugin-fasterwhisper": {
        "model": "large-v3",
        "use_cuda": true,
        "compute_type": "float16",
        "beam_size": 5,
        "cpu_threads": 4
    }
  }

To use Whisper for lang detection (ovos-dinkum-listener only)

  "listener": {
    "audio_transformers": {
        "ovos-audio-transformer-plugin-fasterwhisper": {
            "model": "small"
        }
    }
  }

Models

Models will be auto downloaded by faster whisper on plugin load