Code repository of our research paper on AI-generated music detection - D. Afchar, G. Meseguer Brocal, R. Hennequin (2024).
Examples of audio reconstructions may be found in the audio_examples folder or on the demo page: research.deezer.com/deepfake-detector/.
The FMA dataset is available at github.com/mdeff/fma.
Most of our experiment code is available for the review. We will make the trained weights open source for the publication.