mirror of
https://github.com/QuentinFuxa/WhisperLiveKit.git
synced 2026-03-07 22:33:36 +00:00
a4da246ea5
Pure-MLX implementation of Voxtral Mini 4B Realtime for low-latency speech transcription on Apple Silicon. Avoids the transformers/torch overhead and runs at 0.18-0.32x real-time factor. - voxtral_mlx/model.py: MLX model with spectrogram, encoder, decoder - voxtral_mlx/loader.py: model loading with 6-bit quantized weights - voxtral_mlx/spectrogram.py: mel spectrogram computation in MLX - voxtral_mlx_asr.py: VoxtralASR adapter for the AudioProcessor pipeline
7 lines
188 B
Python
7 lines
188 B
Python
"""Pure-MLX Voxtral Realtime backend for WhisperLiveKit."""
|
|
|
|
from .loader import load_voxtral_model
|
|
from .model import VoxtralMLXModel
|
|
|
|
__all__ = ["load_voxtral_model", "VoxtralMLXModel"]
|