OmniVoice in MLX (4bit). For mlx-audio on Apple Silicon.

Usage

pip install mlx-audio
python -m mlx_audio.tts.generate \
  --model commandeaw/OmniVoice-MLX-4bit \
  --text "เธชเธงเธฑเธชเธ”เธตเธ„เนˆเธฐ เธขเธดเธ™เธ”เธตเธ•เน‰เธญเธ™เธฃเธฑเธš" \
  --lang_code th \
  --ref_audio reference.wav

--ref_audio is optional (zero-shot voice cloning); drop it for the default voice.

Size

variant repo vs base (fp32)
bf16 1.6 GB -50%
8bit 1.1 GB -68%
4bit 0.75 GB -77%
Downloads last month
47
Safetensors
Model size
0.6B params
Tensor type
U32
ยท
BF16
ยท
MLX
Hardware compatibility
Log In to add your hardware

Quantized

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for commandeaw/OmniVoice-MLX-4bit

Finetuned
Qwen/Qwen3-0.6B
Finetuned
k2-fsa/OmniVoice
Quantized
(44)
this model