litmudoc/Chatterbox-Multilingual-MLX-v2-Q4

This model was converted to MLX format from ResembleAI/chatterbox using mlx-audio version 0.2.10.

Note: This model requires the S3Tokenizer weights from mlx-community/S3TokenizerV2, which will be downloaded automatically.

Use with mlx-audio

pip install -U mlx-audio

Command line

curl -L -o ko.wav https://huggingface.co/litmudoc/Chatterbox-Multilingual-MLX-v2-Q4/resolve/main/ko.wav

mlx_audio.tts.generate \
  --model litmudoc/Chatterbox-Multilingual-MLX-v2-Q4 \
  --text ", μ§€λ‚œλ‹¬ μš°λ¦¬λŠ” 유튜브 μ±„λ„μ—μ„œ 이십얡 μ‘°νšŒμˆ˜λΌλŠ” μƒˆλ‘œμš΄ μ΄μ •ν‘œμ— λ„λ‹¬ν–ˆμŠ΅λ‹ˆλ‹€." \
  --lang_code ko \
  --ref_audio ko.wav \
  --ref_text "μš°λ¦¬λŠ” μ •λ§λ‘œ ν—ˆλ¦„ν•œ ν˜Έν…”μ— λ¬΅μ—ˆμ§€λ§Œ, κ·Έλž˜λ„ ν–‰λ³΅ν–ˆλ‹€." \
  --verbose --play

Python

from mlx_audio.tts.generate import generate_audio

generate_audio(
    text=", μ§€λ‚œλ‹¬ μš°λ¦¬λŠ” 유튜브 μ±„λ„μ—μ„œ 이십얡 μ‘°νšŒμˆ˜λΌλŠ” μƒˆλ‘œμš΄ μ΄μ •ν‘œμ— λ„λ‹¬ν–ˆμŠ΅λ‹ˆλ‹€.",
    model="litmudoc/Chatterbox-Multilingual-MLX-v2-Q4",
    lang_code="ko",
    ref_audio="ko.wav",
    ref_text="μš°λ¦¬λŠ” μ •λ§λ‘œ ν—ˆλ¦„ν•œ ν˜Έν…”μ— λ¬΅μ—ˆμ§€λ§Œ, κ·Έλž˜λ„ ν–‰λ³΅ν–ˆλ‹€.",
    file_prefix="output",
)
Downloads last month
111
Safetensors
Model size
0.3B params
Tensor type
F32
Β·
U32
Β·
MLX
Hardware compatibility
Log In to view the estimation

Quantized

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

Model tree for litmudoc/Chatterbox-Multilingual-MLX-v2-Q4

Finetuned
(27)
this model