Nvidia Unveils Fugatto AI Model with 2.5 Billion Parameters for Audio Creation

Nov 25, 2024, 03:41 PM

Nvidia has unveiled Fugatto, a groundbreaking AI model designed to transform audio creation. Fugatto, short for Foundational Generative Audio Transformer Opus 1, features 2.5 billion parameters and can generate music, voices, and soundscapes using both text and audio inputs. The model enables users to modify existing audio, such as changing a piano melody into a human voice or altering a trumpet to sound like a saxophone. Leveraging open-source data, Fugatto is positioned as a versatile tool for sound engineering, with potential applications in music production, advertising, and gaming. This innovation highlights Nvidia's continued advancements in generative AI technology.