metadata
license: apache-2.0
base_model: MediaTek-Research/Breeze-ASR-26
language:
- zh
- nan
pipeline_tag: automatic-speech-recognition
library_name: ctranslate2
tags:
- whisper
- taiwanese-hokkien
- taigi
- ctranslate2
- faster-whisper
- int8
- real-time
- speech-recognition
Breeze-ASR-26 — CTranslate2 / faster-whisper (INT8)
The fastest edge build: RTF 0.21 on CPU, real-time with headroom. Use this on servers, 8 GB+ hosts, or GPU. For 4 GB hosts or desktop apps use the GGML repo; for mobile/WASM use ONNX.
Part of the Breeze-ASR-26 edge family — the same MediaTek model in every runtime, pick by your constraint:
Repo Runtime RSS RTF (CPU 4-thread) Best for Breeze-ASR-26-ct2 CTranslate2 / faster-whisper ~2.9 GB 0.21 servers, 8 GB+ hosts, GPU Breeze-ASR-26-GGML whisper.cpp / MacWhisper 1.85 GB 0.40 4 GB hosts, desktop apps Breeze-ASR-26-ONNX sherpa-onnx / onnxruntime — 1.3 Android / iOS / WASM All Apache-2.0, derived from MediaTek-Research/Breeze-ASR-26. Measured on real multi-speaker Mandarin meeting audio. Mandarin does not regress; Taigi is transcribed as Mandarin meaning (not verbatim Taigi characters).
Usage
from faster_whisper import WhisperModel
m = WhisperModel("weemed/Breeze-ASR-26-ct2", device="cpu", compute_type="int8", cpu_threads=4)
segments, _ = m.transcribe("meeting.wav", language="zh", beam_size=1)
print("".join(s.text for s in segments))
INT8 vs FP32: CER 4.69% (function-word/segmentation jitter only), 2.4x faster, quarter the size.