Meta has launched its own AI-powered music generator to rival Google.
The music-generating technology created by Meta is called MusicGen, and according to a demo, it can convert a word description into roughly 12 seconds of sounds.
Meta AI researcher, Felix Kreuk, disclosed the development on his Twitter handle.
According to Meta, MusicGen received training on 20,000 hours of music, which included 10,000 “high-quality” licenced music files and 390,000 instrument-only recordings from Pond5 and ShutterStock, two sizable stock music libraries.
“We present MusicGen: A simple and controllable music generation model. MusicGen can be prompted by both text and melody.
“We release code (MIT) and models (CC-BY NC) for open research, reproducibility, and for the music community.” Felix tweeted.
The company has made pre-trained models available that anyone with the necessary hardware, mostly a GPU with about 16GB of memory, may run. However, the company has not published the code it used to train the model.