Nvidia Launches Nemotron 3.5 for Speech Recognition and Translation

Nvidia has launched the Nemotron 3.5 Transcribe, an open speech recognition model that supports transcription, speaker recognition, and translation. This model leads several performance benchmarks and uses only a sixth of the active parameters compared to its strongest open rival. The release also includes a dataset with over 800,000 hours of speech, along with tools for fine-tuning. This move enhances Nvidia's competitiveness in the speech recognition model market, focusing on providing integrated solutions for users.