On August 3, global social networking and technology giant Meta (parent company of Facebook, Instagram, etc.) announced the open source text generation music model Audiocraft (open source address: https://github.com/facebookresearch/audiocraft).
It is reported that Audiocraft is a hybrid model, composed of MusicGen, AudioGen and EnCodec. Using only text, you can generate background audio such as bird calls, car horns, footsteps, or more complex music, which is suitable for business scenarios such as game development, social networking, and video dubbing.
MusicGen paper: https://arxiv.org/abs/2306.05284
AudioGen paper: https://arxiv.org/abs/2209.15352
High-fidelity decoder paper: https://arxiv.org/abs/2210.13438