Show HN: A lightweight ML model to predict music emotion - energy, valence, etc.

Spotify has 7 features for each of their music tracks (acousticness, danceability, energy, instrumentalness, liveness, speechiness, valence) which describe the perceptual/emotional content of the song. I wanted to tag my own offline music library with these features so that I could sort my songs into playlists for different occasions (working out, driving, etc.), but unfortunately Spotify doesn't share how they calculate these features. So, I trained my own lightweight neural network to predict these features!