A hybrid CNN-RNN model trained on the DEAM dataset for music emotion recognition using the valence-arousal model. Useful for tagging emotional content of music in media or recommendation systems