Table 5.

Ablation experiments on the GoogLeNet module using different datasets.

Model	Datasets
		Music Emotion Dataset				Multimodal DataSet
		Accuracy	Recall	F1 Score	AUC	Accuracy	Recall	F1 Score	AUC
ResNet	93.75	88.89	87.7	90.39	89.47	86.9	85.74	90.29
VGGNet	94	93.38	84.57	86.38	87.08	87.73	84.8	88.74
DenseNet	95.9	86.73	90.68	84.74	95.07	93.53	86.95	90.8
GoogLeNet	96.98	93.46	92.67	93.22	96.98	93.46	92.67	93.22