Multi-Label Classification on Visual Genome VG256 (test) benchmark leaderboard · SOTA2 Research

Multi-Label Classification on Visual Genome VG256 (test)

50.9mAPML-VPT

Evaluation Results

Method	Links
ML-VPT2025.04		50.9	51.7	64.1
VPT2025.04		49.3	50.3	62.9
E2VPT2025.04		49.2	50.3	62.9
GateVPT2025.04		48.6	49.7	62.4
ML-VPT2025.04		47.7	48.7	61.9
ML-VPT2025.04		47.6	48.6	61.2
VPT2025.04		46	47.4	60.2
E2VPT2025.04		46	47.4	60.1
GateVPT2025.04		45.4	46.8	59.5
VPT2025.04		45.2	46.7	60.2
E2VPT2025.04		45.1	46.7	60.1
GateVPT2025.04		44.3	45.9	59.3
ML-VPT2025.04		44.2	45.8	58.8
VPT2025.04		42.2	44	57.5
E2VPT2025.04		42.2	44.2	57.6
GateVPT2025.04		41.1	43	56.5
ML-VPT2025.04		41	43.1	57.1
ML-VPT2025.04		40.5	42.7	57.3
VPT2025.04		40.4	42.6	56.5
E2VPT2025.04		40.4	42.5	56.5
GateVPT2025.04		39.6	41.8	55.8
VPT2025.04		37.9	40.3	55.2
E2VPT2025.04		37.6	40.2	55
GateVPT2025.04		35.3	38.2	53.2