This model encodes audio files into vectors of 100 dimensions. It was trained on a million Spotify playlists and tracks. The details can be found here.

To encode an audio first install the package with

pip install audiodiffusion

and then run

from audiodiffusion.audio_encoder import AudioEncoder

audio_encoder = AudioEncoder.from_pretrained("teticio/audio-encoder")
audio_encoder.encode(<list of audio files>)