VibeVoice-ASR-HF

Home/Models/microsoft/VibeVoice-ASR-HF

microsoft/VibeVoice-ASR-HF

kind: original
quant: bf16 (16-bit)
family: misc
modality: audio-text-to-text
root model: microsoft/VibeVoice-ASR-HF
provider: microsoft
license: mit
language: en, zh, es, pt, de, ja, ko, fr, ru, id, sv, it
context: —
library: transformers
vRAM (GB): —
size on disk: —
downloads: 198,817
likes: 169
created: —
last modified: 2026-03-09T03:11:35+00:00 (2026)
huggingface: https://huggingface.co/microsoft/VibeVoice-ASR-HF
tags: transformerssafetensorsvibevoice_asrautomatic-speech-recognitionASRDiarizationSpeech-to-TextTranscriptionaudio-text-to-textenzhesptdejakofrruidsvithenlpl

benchmark	category	score	context	conf	evidence
WER	speech-asr	17.2		low	`\| ami_test \| 17.20 \|`
WER Earnings-22	speech-asr	13.17	earnings22_test	medium	`\| earnings22_test \| 13.17 \|`
WER GigaSpeech	speech-asr	9.67	gigaspeech_test	medium	`\| gigaspeech_test \| 9.67 \|`
WER VoxPopuli	speech-asr	8.01	voxpopuli_test	medium	`\| voxpopuli_test \| 8.01 \|`
WER LibriSpeech	speech-asr	5.51	librispeech_test.other	medium	`\| librispeech_test.other \| 5.51 \|`
WER SPGISpeech	speech-asr	3.8	spgispeech_test	medium	`\| spgispeech_test \| 3.80 \|`
WER TED-LIUM	speech-asr	2.57	tedlium_test	medium	`\| tedlium_test \| 2.57 \|`
WER LibriSpeech	speech-asr	2.2	librispeech_test.clean	medium	`\| librispeech_test.clean \| 2.20 \|`

none scraped.