Jukebox
OpenAI Blog
OpenAI released Jukebox, a neural network that generates music with vocals across multiple genres and artist styles as raw audio files. The model outputs audio directly rather than MIDI, and the team is releasing both the model weights and code along with an exploration tool. This enables researchers and developers to experiment with AI-generated music and potentially build applications using the released model.
Why it matters
We’re introducing Jukebox, a neural net that generates music, including rudimentary singing, as raw audio in a variety of genres and artist styles. We’re releasing the model weights and code, along with a tool to explore the generated samples.
