The model can create life-like audio from a 15-second clip and a text prompt.
If you’ve been following the publication for a while, you’ll notice I’ve made some…