Instructions to use mlx-community/Qwen3-ASR-0.6B-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mlx-community/Qwen3-ASR-0.6B-4bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir Qwen3-ASR-0.6B-4bit mlx-community/Qwen3-ASR-0.6B-4bit
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
'audio_path' parameter in example is wrong, mlx-audio's 'audio'
#1
by qhduan - opened
output_path is not exactly too
from mlx_audio.stt.utils import load_model
from mlx_audio.stt.generate import generate_transcription
model = load_model("mlx-community/Qwen3-ASR-0.6B-4bit")
transcription = generate_transcription(
model=model,
audio="path_to_audio.wav", # audio_path change to audio
output_path="path_to_output", # if you want to generate path_to_output.txt output_path should be filename only, output_path + '.' + format
format="txt",
verbose=True,
)
print(transcription.text)