List your company

Image-Text Retrieval on MS-COCO 5K (test)

472.1RSum (Composite Score)CDDS
411.052426.901442.75458.599Mar 5, 2026Mar 8, 2026Mar 11, 2026Mar 15, 2026Mar 18, 2026Mar 21, 2026Mar 25, 2026

Evaluation results

MethodLinks
CDDSBackbone=Swin, Resolution=384x384, Number of patches=12x122026.03472.167.989.694.951.180.288.4——
LAPSBackbone=Swin, Resolution=384x384, Number of patches=12x122026.03470.167.188.694.35379.587.6——
CLIPStructure=ViT-L/142026.03469.6————————
CDDSBackbone=Swin, Resolution=224x224, Number of patches=7x72026.03467.764.689.894.851.97987.6——
LAPSBackbone=Swin, Resolution=224x224, Number of patches=7x72026.03465.864.589.294.451.678.987.2——
VSE++Backbone=Swin, Resolution=384x384, Number of patches=12x122026.03462.26388.594.350.178.987.4——
VSE++Backbone=Swin, Resolution=224x224, Number of patches=7x72026.03460.96488.294.249.97886.6——
CHANBackbone=Swin, Resolution=384x384, Number of patches=12x122026.03459.563.488.494.149.277.986.6——
CHANBackbone=Swin, Resolution=224x224, Number of patches=7x72026.0345864.187.993.549.177.386.1——
SGRBackbone=Swin, Resolution=384x384, Number of patches=12x122026.03453.862.88792.948.17786——
SGRBackbone=Swin, Resolution=224x224, Number of patches=7x72026.03453.16186.793.248.677.286.3——
SCANBackbone=Swin, Resolution=384x384, Number of patches=12x122026.03452.461.187.393.347.876.985.9——
SCANBackbone=Swin, Resolution=224x224, Number of patches=7x72026.03451.860.786.693.248.177.186.1——
CMSFStructure=frcnn+bert2026.03445.8————————
CDDSBackbone=ViT, Resolution=384x384, Number of patches=24x242026.03444.457.885.292.846.976.285.5——
LAPSBackbone=ViT, Resolution=384x384, Number of patches=24x242026.03442.257.484.992.546.475.885.2——
CDDSBackbone=ViT, Resolution=224x224, Number of patches=14x142026.03437.857.984.691.44574.884.1——
LAPSBackbone=ViT, Resolution=224x224, Number of patches=14x142026.03434.457.58490.844.57483.6——
CMSFStructure=frcnn+bigru2026.03432————————
CHANBackbone=ViT, Resolution=384x384, Number of patches=24x242026.03431.155.683.891.243.473.683.5——
CHANBackbone=ViT, Resolution=224x224, Number of patches=14x142026.0342856.383.290.14372.682.8——
SGRBackbone=ViT, Resolution=224x224, Number of patches=14x142026.03425.854.982.890.542.872.282.5——
VSE++Backbone=ViT, Resolution=384x384, Number of patches=24x242026.03425.854.982.890.442.472.482.8——
SGRBackbone=ViT, Resolution=384x384, Number of patches=24x242026.03423.653.38189.642.973.183.7——
SCANBackbone=ViT, Resolution=224x224, Number of patches=14x142026.03423.553.981.89042.972.382.5——
SCANBackbone=ViT, Resolution=384x384, Number of patches=24x242026.03423.153.381.89042.672.682.9——
CLIPStructure=ViT-B/322026.03422.6————————
VSE++Backbone=ViT, Resolution=224x224, Number of patches=14x142026.03413.452.480.388.840.670.481.1——
SGD–CLIPBackbone=RN-50 + SBERT, Train time=5121.72 s2026.04———————5.727
SGD–CLIPBackbone=CLIP ViT-B/32, Train time=1066.60 s2026.04———————12.642.1
UniConBackbone=RN-50 + SBERT, Train time=11.11 s2026.04———————11.741.4
UniConBackbone=CLIP ViT-B/32, Train time=11.15 s2026.04———————31.166.5