VisualSparta: An Embarrassingly Simple Approach to Large-scale Text-to-Image Search with Weighted Bag-of-words

Benchmark Model Rank Results
cross-modal-retrieval-on-coco-2014VisualSparta#21Text-to-image R@1: 44.4Text-to-image R@5: 72.8
image-retrieval-on-cocoVisualSparta#2recall@1: 68.2recall@5: 91.8Recall@10: 96.3QPS: 451.4
image-retrieval-on-flickr30kVisualSparta#9Recall@10: 88.1Recall@5: 82.0Recall@1: 57.4QPS: 451.4
image-retrieval-on-flickr30k-1k-testVisualSparta#4R@1: 57.4R@5: 82.0R@10: 88.1