List your company
ResearchFormerly WizWand

Interaction Recognition on ImageNet-1k 2 objects

88Interaction RecognitionMETAGAME
25.641.85874.2May 7, 2026
Updated 5mo ago

Evaluation results

MethodLinks
METAGAMEModel (Size)=CLIP (ViT-B/16), Explanation Basis=MaskCLIP2026.0588
METAGAMEModel (Size)=SigLIP-2 (ViT-L/16), Base Method=MaskSigLIP2026.0588
METAGAMEModel (Size)=SigLIP-2 (ViT-B/32), Base Method=MaskSigLIP2026.0588
METAGAMEModel (Size)=CLIP (ViT-B/16), Explanation Basis=Attention2026.0587
FIxLIPModel (Size)=SigLIP-2 (ViT-B/32)2026.0587
METAGAMEModel (Size)=SigLIP-2 (ViT-B/32), Base Method=Grad-ESigLIP2026.0585
METAGAMEModel (Size)=MetaCLIP-2 (ViT-H/14), Explanation Basis=Attention2026.0583
FIxLIPModel (Size)=SigLIP-2 (ViT-L/16)2026.0583
METAGAMEModel (Size)=SigLIP-2 (ViT-L/16), Base Method=Grad-ESigLIP2026.0580
METAGAMEModel (Size)=CLIP (ViT-B/16), Explanation Basis=Grad-ECLIP2026.0579
FIxLIPModel (Size)=CLIP (ViT-B/16)2026.0577
METAGAMEModel (Size)=MetaCLIP-2 (ViT-H/14), Explanation Basis=Grad-ECLIP2026.0576
MaskSigLIPModel (Size)=SigLIP-2 (ViT-L/16)2026.0573
METAGAMEModel (Size)=SigLIP-2 (ViT-B/32), Base Method=Attention2026.0571
METAGAMEModel (Size)=SigLIP-2 (ViT-L/16), Base Method=Attention2026.0567
FIxLIPModel (Size)=MetaCLIP-2 (ViT-H/14)2026.0566
MaskSigLIPModel (Size)=SigLIP-2 (ViT-B/32)2026.0554
Grad-ESigLIPModel (Size)=SigLIP-2 (ViT-B/32)2026.0548
AttentionModel (Size)=CLIP (ViT-B/16)2026.0546
AttentionModel (Size)=MetaCLIP-2 (ViT-H/14)2026.0545
Grad-ECLIPModel (Size)=CLIP (ViT-B/16)2026.0543
Grad-ECLIPModel (Size)=MetaCLIP-2 (ViT-H/14)2026.0543
Grad-ESigLIPModel (Size)=SigLIP-2 (ViT-L/16)2026.0542
AttentionModel (Size)=SigLIP-2 (ViT-B/32)2026.0539
AttentionModel (Size)=SigLIP-2 (ViT-L/16)2026.0536
MaskCLIPModel (Size)=CLIP (ViT-B/16)2026.0528