Nvidia debuts Nemotron 3.5 Transcribe, an open-weight model for multilingual audio transcription and translation, and Granary v2, its open-source speech dataset (Kyt Dotson/SiliconANGLE) — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Nvidia released Nemotron 3.5 Transcribe, an open-weight model for multilingual audio transcription and translation across 40 languages, in a leaderboard-winning release.
- Granary v2 is the open-source speech dataset Nvidia released alongside Nemotron 3.5 Transcribe, making data availability a second concrete part of the announcement.
Why it matters: Speech-AI developers gain access to a 40-language transcription model and an open-source speech dataset in the same announcement. The practical addition is not only openly available model weights but also an inspectable data resource, giving developers more of the speech pipeline to examine and adapt.
Ask SkimNews
