Models trained using the Myna framework for contrastive learning of musical representations