List your company

Image-to-Text Retrieval on MS COCO short-caption (test)

37.88R@1CLIP + TEVI
32.409633.829835.2536.6702Jun 5, 2026

Evaluation results

MethodLinks
CLIP + TEVIBackbone=ViT-L/142026.0637.8864.3875.16
CLIPBackbone=ViT-L/142026.0636.5262.573.06
SigLIP + TEVIBackbone=ViT-B/162026.0636.2461.272.82
CLIP + TEVIBackbone=ViT-B/162026.0635.6661.8272.94
AlignCLIP + TEVIBackbone=ViT-B/162026.0634.4260.371.48
SigLIPBackbone=ViT-B/162026.0633.8860.370.62
SharedCLIP + TEVIBackbone=ViT-B/162026.0633.8259.8670.9
CLIPBackbone=ViT-B/162026.0632.9859.0270.04
AlignCLIPBackbone=ViT-B/162026.0632.758.9270.46
SharedCLIPBackbone=ViT-B/162026.0632.6258.8869.94