Kenji8000/Breeze-ASR-25-mlx
013
Breeze-ASR-25-mlx
This model was converted to MLX format from `MediaTek-Research/Breeze-ASR-25`.
Breeze ASR 25 is an advanced ASR model fine-tuned from Whisper-large-v2, optimized for Taiwanese Mandarin and Mandarin-English code-switching scenarios.
Key Features
- Traditional Chinese direct output (no OpenCC post-processing needed)
- Mandarin-English code-switching: 55.9% WER reduction vs Whisper-large-v2
- Taiwanese Mandarin optimized: CommonVoice16-zh-TW WER 7.97 (vs 9.84 baseline)
- MLX native: runs on Apple Silicon GPU via Metal
Use with mlx-whisper
pip install mlx-whisperimport mlx_whisper
result = mlx_whisper.transcribe(
"audio.wav",
path_or_hf_repo="Kenji8000/Breeze-ASR-25-mlx",
language="zh",
)
print(result["text"])Performance
Tested on Mac Studio M2 Ultra 128GB, mlx-whisper 0.4.3, mlx 0.31.1:
Conversion
Converted using mlx-examples/whisper/convert.py with dtype float16.
python convert.py --torch-name-or-path MediaTek-Research/Breeze-ASR-25 \
--mlx-path mlx_models/breeze-asr-25 --dtype float16
# Note: rename output model.safetensors → weights.safetensors for mlx-whisperOriginal Model
- Base: openai/whisper-large-v2
- Fine-tuned by: MediaTek Research
- License: Apache 2.0
- Paper: arXiv:2506.11130
