distil-large-v3.5

Home/Models/distil-whisper/distil-large-v3.5

distil-whisper/distil-large-v3.5

kind: original
quant: bf16 (16-bit)
family: misc
modality: automatic-speech-recognition
root model: distil-whisper/distil-large-v3.5
provider: distil-whisper
license: mit
language: en
context: —
library: transformers
vRAM (GB): —
size on disk: —
downloads: 6,869
likes: 93
created: —
last modified: 2026-04-13T20:14:38+00:00 (2026)
huggingface: https://huggingface.co/distil-whisper/distil-large-v3.5
tags: transformerstensorboardsafetensorswhisperautomatic-speech-recognitionaudioenarxiv:2311.00430arxiv:2106.05237arxiv:1904.08779arxiv:1910.13267license:miteval-resultsendpoints_compatibleregion:us

benchmark	category	score	context	conf	evidence
WER Earnings-22	speech-asr	39.08	earnings22	medium	`\| earnings22 \| 39.08 \| 42.09 \| 57.28 \| 58.8 \|`
WER GigaSpeech	speech-asr	35.36	Gigaspeech	medium	`\| Gigaspeech \| 35.36 \| 10.02 \| 10.14 \| 10.08 \| 9.84 \|`
WER AMI	speech-asr	8.68	AMI	medium	`\| AMI \| 8.68 \| 15.95 \| 16.13 \| 15.16 \| 14.63 \|`
WER	speech-asr	7.3		low	`\| [large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) \| 809 \| 1.0 \| 7.30 \| 10.25 \|`
WER TED-LIUM	speech-asr	2.61	Tedlium	medium	`\| Tedlium \| 2.61 \| 3.86 \| 3.57 \| 3.86 \| 3.64 \|`
WER TED-LIUM	speech-asr	2.47	tedlium-long-form	medium	`\| tedlium-long-form \| 2.47 \| 3.07 \| 9.66 \| 3.9 \| 4.63 \|`

none scraped.