Attention-Based Graph Convolutional Network for Zero-Shot Learning with Pre-Training

Table 4

Top-k accuracy of different models on the ImageNet dataset.

	Method	Accuracy (%)
		1	2	5	10	20
OL	Con	8.3	12.9	21.8	30.9	41.7
		10.5	17.7	28.6	40.1	52.0
		12.5	19.5	32.3	43.7	55.2
		19.8	33.3	53.2	65.4	74.6
		24.5	37.8	57.1	69.7	79.5
		24.6	37.4	56.7	69.5	79.1
	AGCNZ	22.2	34.8	53.6	66.5	77.5
	ADGPZ	20.0	31.6	49.4	63.1	74.3
ML	GCN	24.8	38.4	57.5	69.7	79.4
	DGP	24.7	38.1	57.1	69.6	79.3
	AGCNZ	22.3	34.9	53.6	66.7	77.8
	ADGPZ	20.2	31.7	49.3	63.2	74.8