Nvidia just released a free speech-to-text model that can transcribe up to 40 languages, but it only works on recordings, not live audio, so it’s a big step for meeting and medical transcription tools while still leaving real-time use to rivals like ElevenLabs.
Voice AI in vogue: Nvidia releases new open transcription model
Nvidia just released a free speech-to-text model that can transcribe up to 40 languages, but it only works on recordings, not live audio, so it’s a big step for meeting and medical transcription tools while still leaving real-time use to rivals like ElevenLabs.
Originally reported by TechCrunch
More like this
All Technology
Reddit is the latest platform to embrace the ‘For You’ feed
The For You feed combines posts from communities users have joined with recommendations tailored to their interests.
AI execs may be too out of touch to build for the general public
Tech leaders are flaunting how out of touch they are when they brag about using AI to connect with their families.
Trump DOE loans $5B to nuclear plants that will power data centers
Constellation Energy will get $5 billion in government loans to modernize nuclear reactors. Most of the new capacity i…