List your company

Zero-shot Image-Text Retrieval on Flickr30k

97.7Accuracy (Zero-shot)CLIPS (ViT-L/14)
91.56493.15794.7596.343Mar 26, 2026

Evaluation results

MethodLinks
CLIPS (ViT-L/14)Backbone=ViT-L/142026.0397.7
BLIP-BBackbone=ViT-B2026.0397.2
C^2LIPBackbone=ViT-B/162026.0397
SigLIP ViT-L/16-res256Backbone=ViT-L/16, Input Resolution=256x2562026.0396.7
SigLIP2-ViT-B/16Backbone=ViT-B/162026.0396.5
CLIP-A (ViT-L/14)Backbone=ViT-L/142026.0395.2
FLAVABackbone=ViT-B2026.0391.8