Distil-Whisper
Distilled version of Whisper that is 6x faster with minimal accuracy loss.
About
Distil-Whisper by Hugging Face is a distilled version of OpenAI Whisper for English speech recognition that runs about six times faster and is roughly half the size while staying within one percent word error rate on out-of-distribution audio. The distil-large-v3 checkpoint is the recommended option, and training code is provided. It suits latency-sensitive production transcription. Released under the MIT license.
Reviews (0)
Leave a Review
No reviews yet. Be the first to review!
Details
- Price
- Free
- Platform
- Local/Desktop
- Difficulty
- Easy (2/5)
- License
- MIT
- Added
- Apr 3, 2026
Related Tools
Multilingual ASR model by NVIDIA supporting 4 languages with translation.
CLI tool that transcribes audio 10x faster using pipeline optimizations.
Self-supervised speech representation model by Meta for ASR.
High-performance C/C++ port of Whisper for CPU-based speech recognition.
Non-autoregressive ASR model by Alibaba achieving fast parallel transcription.
JAX-based Whisper implementation optimized for TPU/GPU with 70x+ speedup.