Nvidia releases open-weight Nemotron 3.5 Transcribe speech model and Granary v2 dataset
Nvidia released Nemotron 3.5 Transcribe, a free, open-weight, 30-billion-parameter model that transcribes audio in up to 40 languages, along with version 2 of its open-source speech dataset Granary. The company says the model translates 31 languages to and from English, handles recordings up to an hour long, and distinguishes up to eight speakers. Nvidia also claims it beats rivals on benchmarks, including Hugging Face's Open ASR Leaderboard. The tools target developers building meeting and medical transcription apps. The model works only on recorded audio, not live streams. This is product news with no security incident, vulnerability or threat activity described.
Linha do tempo
-
Nvidia releases open-weight Nemotron 3.5 Transcribe speech model and Granary v2 dataset
ler a análise desta fonte ocultar
Nvidia released Nemotron 3.5 Transcribe, a free, open-weight, 30-billion-parameter model that transcribes audio in up to 40 languages, along with version 2 of its open-source speech dataset Granary. The company says the model translates 31 languages to and from English, handles recordings up to an hour long, and distinguishes up to eight speakers. Nvidia also claims it beats rivals on benchmarks, including Hugging Face's Open ASR Leaderboard. The tools target developers building meeting and medical transcription apps. The model works only on recorded audio, not live streams. This is product news with no security incident, vulnerability or threat activity described.