videos in AudioSet
A list of about a million ten-second YouTube clips, each labelled with the sounds in it, that Google published in 2017. Labs use it to train and test general audio and music models.
Including Graham Nash - Chicago - Acoustic Version w/ Saxophone, Driver's Seat - Sniff'n' The Tears - Acoustic Guitar Cover Unplugged, Au clair de la lune - berceuse - chanson pour enfants en francais comptine dodo