Researchers at US tech giant Google have created an AI that can generate minute-long musical pieces from text prompts, and can even convert a whistle or hum to other instruments, as written by systems such as DAL-E. The signals generate images, reported The Verge, an American technology news website, via TechCrunch.
According to the outlet, the model is called the Music LM, and while you can’t play with it for yourself, the company uploaded a bunch of samples it created using the model.