List your company
ResearchFormerly WizWand

Image-to-Text Retrieval on diagnostic (test-unseen)

85.21Accuracy@50Human
21.156437.785754.41571.0443Nov 29, 2022
Updated 8mo ago

Evaluation results

MethodLinks
Human2022.1185.21
BLIPModel Category=Both, Backbone=ViT-L, Objective Variant=itm2022.1141.94
CLIPModel Category=Contrastive, Backbone=ViT-L2022.1139.61
FLAVAModel Category=Matching, Backbone=ViT-B2022.1138.43
ALBEFModel Category=Both, Backbone=ViT-B2022.1138.32
ViLTModel Category=Matching, Backbone=ViT-B2022.1135.34
OwlVitModel Category=Contrastive, Backbone=ViT-L2022.1132.3
BLIPModel Category=Both, Backbone=ViT-L, Objective Variant=itc2022.1131
Random2022.1123.62