Instructions to use c1ayvveng/Breeze-ASR-25-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use c1ayvveng/Breeze-ASR-25-mlx with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] hf download c1ayvveng/Breeze-ASR-25-mlx --local-dir Breeze-ASR-25-mlx
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
- Atomic Chat
mlx-breeze-asr-25
MediaTek-Research/Breeze-ASR-25 的 MLX 轉換版本(float16)。
Breeze ASR 25 是基於 Whisper-large-v2 微調的 ASR 模型,針對台灣華語、中英混用情境最佳化,並強化時間戳記對齊。
Use with mlx
pip install mlx-whisper
import mlx_whisper
result = mlx_whisper.transcribe(
"FILE_NAME",
path_or_hf_repo="c1ayvveng/mlx-breeze-asr-25",
)
print(result["text"])
- Downloads last month
- 17
Hardware compatibility
Log In to add your hardware
Quantized
Model tree for c1ayvveng/Breeze-ASR-25-mlx
Base model
openai/whisper-large-v2 Finetuned
MediaTek-Research/Breeze-ASR-25