List your company

Zero-shot Image-Text Retrieval on MSCOCO

82.8AccuracySigLIP2-ViT-B/16
73.54475.94778.3580.753Mar 26, 2026

Evaluation results

MethodLinks
SigLIP2-ViT-B/16Backbone=ViT-B/162026.0382.8
C^2LIPBackbone=ViT-B/162026.0382.7
SigLIP ViT-L/16-res256Backbone=ViT-L/16, Input Resolution=256x2562026.0382.3
CLIPS (ViT-L/14)Backbone=ViT-L/142026.0382.1
BLIP-BBackbone=ViT-B2026.0379.1
CLIP-A (ViT-L/14)Backbone=ViT-L/142026.0378.1
FLAVABackbone=ViT-B2026.0373.9