Audio event classification model trained on YouTube AudioSet embeddings, used for tagging sounds in video and real time monitoring applications.