# huggingface/distil-whisper

Distilled variant of Whisper for speech recognition. 6x faster, 50% smaller, within 1% word error rate.

Repository: https://github.com/huggingface/distil-whisper
Canonical: https://ross.abutalabs.com/products/distil-whisper
Language: Python
License: MIT
License Family: permissive
Topics: audio, speech-recognition, whisper
Last push: 2025-01-08T10:09:47+00:00

## Health v2 (maintenance only)
Score: 27/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 74
- inputs: {"age_days": 1037, "days_push": 602, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4112, forks 356 (observed 2026-08-28T04:08:35.721094+00:00)

## What it is
Distil-Whisper is a distilled version of OpenAI's Whisper model for English speech recognition, offering 6x faster inference, 49% fewer parameters, and word error rates within 1% of the original. It provides pretrained model checkpoints and training code that run through Hugging Face Transformers (v4.35+).

## Use cases
- transcribe english audio to text faster than whisper
- run speech recognition on resource-constrained or mobile devices
- deploy low-latency asr for long-form audio
- fine-tune or train a distilled whisper model
- reduce memory footprint of whisper transcription pipelines

## When to choose
- you need english-only transcription with near-whisper accuracy at a fraction of the latency
- you want smaller model weights for on-device or memory-limited deployments
- you already use huggingface transformers and want a drop-in whisper replacement

## When to avoid
- you need multilingual speech recognition - use whisper large-v3-turbo instead
- you require the absolute lowest word error rate regardless of speed
- you need a standalone cli or gui app rather than a model library

## Facets
- artifact type: library
- maturity: active
- function: speech-recognition, machine-learning, deep-learning, transformers
- domain: speech-processing, machine-learning, artificial-intelligence
- platform: python, cross-platform
- tags: whisper, distillation, asr, model-weights, huggingface-transformers, english-speech, transcription, audio, gpu

## Member repositories
- huggingface/distil-whisper (main) score 27

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:35.721094+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:23:11.854833+00:00, confidence not recorded.
  - readme: https://github.com/huggingface/distil-whisper (fetched 2026-08-28T04:08:35.721094+00:00, sha 495a8aea5d1e)
- Data as of 2026-08-30T08:39:29.467469+00:00.