Microsoft releases VibeVoice-ASR, an open speech-to-text model (opens in new tab)

(github.com)

3 pointsputlake2mo ago1 comments

1 comments

VibeVoice-ASR is a unified speech-to-text model designed to handle 60-minute long-form audio in a single pass, generating structured transcriptions containing Who (Speaker), When (Timestamps), and What (Content), with support for Customized Hotwords.

j / k navigate · click thread line to collapse

Microsoft releases VibeVoice-ASR, an open speech-to-text model (opens in new tab)

(github.com)

3 pointsputlake2mo ago1 comments

1 comments

putlakeOP2mo ago

VibeVoice-ASR is a unified speech-to-text model designed to handle 60-minute long-form audio in a single pass, generating structured transcriptions containing Who (Speaker), When (Timestamps), and What (Content), with support for Customized Hotwords.

j / k navigate · click thread line to collapse