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 Tom Kerridge's Best Ever Dishes - "I am Tom Kerridge, not Tom Daley", Tom Kerridge's Best Ever Dishes - "That's my first fish!", Tom Kerridge's Best Ever Dishes - "Oh my god, that is amazing!"