Dissecting Deep Metric Learning Losses for Image-Text Retrieval

Benchmark Model Rank Results
cross-modal-retrieval-on-coco-2014VSE-Gradient#7Text-to-image R@1: 63.6Text-to-image R@5: 86.0
cross-modal-retrieval-on-flickr30kVSE-Gradient#6Image-to-text R@1: 97.0Image-to-text R@5: 99.6Image-to-text R@10: 100