Cross-modal Active Complementary Learning with Self-refining Correspondence

Benchmark Model Rank Results
cross-modal-retrieval-with-noisy-1CRCL#4R-Sum: 373.7Image-to-text R@1: 41.8Image-to-text R@5: 67.4
cross-modal-retrieval-with-noisy-2CRCL#2R-Sum: 507.8Image-to-text R@1: 77.9Image-to-text R@5: 95.4
cross-modal-retrieval-with-noisy-3CRCL#4R-Sum: 525.6Image-to-text R@1: 79.6Image-to-text R@5: 96.1