Meta has released AudioCraft's AI sound and song generator and opened its source code

Meta has released AudioCraft, an open-source AI tool that generates music and sounds from textual cues.

Here's What We Know

AudioCraft consists of three AI models, each solving different sound generation problems:

• MusicGen creates music. It has been trained on 20,000 hours of music owned by Meta or licensed specifically for this purpose;
• AudioGen generates sounds by imitating barking dogs or footsteps. The model has been trained on recordings collected in public places;
• the EnCodec decoder allows sounds to be created with fewer artefacts.

Meta believes AudioCraft has the potential to spark a new wave of songs, similar to the way synthesisers once changed music. The company has also opened up the source code of the tool to diversify the data used to train it.

You can listen to samples of the music and sounds created by AudioCraft on the project's website. The source code of the tool is available on GitHub.

Source: Meta AI