Nvidia Launches Open Speech Recognition Model Supporting 40 Languages
Nvidia has launched the Nemotron 3.5 Transcribe model for free with open weights. The 30-billion-parameter model transcribes 40 languages and distinguishes the voices of eight speakers.
Nvidia announced the launch of the Nemotron 3.5 Transcribe model, a free and open-weight model featuring 30 billion parameters, as reported by TechCrunch.
The model can transcribe speech in up to 40 languages, translate 31 languages to and from English, and distinguish between up to eight speakers in a single recording.
The model handles recordings up to an hour long, but does not cover live streaming. It runs on a single chip, including Jetson AGX Thor, DGX Spark, and Blackwell hardware platforms.
The company also released the second version of the Granary audio dataset.
What do these terms mean?
Open-weight model: An AI model whose weights and values are made available to everyone for downloading, using, and modifying, rather than being kept within the developing company.
Parameter: Numerical values within an AI model adjusted by training; the higher the count, the greater the model's ability to process complex patterns.
Speech recognition: A technology that converts spoken audio into written text, used in automated transcription, real-time translation, and voice assistants.
Weekly Newsletter
Read between the lines before everyone else. Decode the most important economic, tech, and decision-maker movements in the region.. in 5 minutes every Saturday.
.jpeg?alt=media&token=94befd11-d37f-46fb-9c9d-621a1d92f33e)
