List your company

Image Classification on MetaShift Cat vs. Dog 18 (test)

92.4Test AccuracyOpenCLIP ZS prompt (full)
42.06455.13268.281.268Jun 2, 2026

Evaluation results

MethodLinks
OpenCLIP ZS prompt (full)Backbone=ViT-L-142026.0692.488.8
Full onlyBackbone=ViT-L (300M), Image size=518x5182026.0683.282.4
Full onlyBackbone=ViT-L (300M)2026.068079.2
Full onlyBackbone=ViT-G (1.1B)2026.0678.676.8
A^2 cross-modelBackbone=ViT-S × ViT-G2026.0677.376.6
Center CropBackbone=ViT-G (1.1B)2026.0676.974.8
Attending on Attention (A^2)Backbone=ViT-G (1.1B)2026.0676.173.7
DFRBackbone=ViT-G (1.1B), Requires group labels=true2026.067674.2
Full onlyBackbone=ViT-B (86M), Image size=518x5182026.0675.373.9
Attending on Attention (A^2)Backbone=ViT-B (86M)2026.0673.971.8
DFRBackbone=ViT-L (300M), Requires group labels=true2026.0673.670.4
Attending on Attention (A^2)Backbone=ViT-L (300M)2026.0672.570.6
Attending on Attention (A^2)Backbone=ViT-S (21M)2026.0669.868.3
DFR + A^2Backbone=ViT-S (21M), Requires group labels=true2026.0669.467.7
DFRBackbone=ViT-B (86M), Requires group labels=true2026.0668.665.9
Full onlyBackbone=ViT-B (86M)2026.0668.568.2
Full onlyBackbone=ViT-S (21M), Image size=518x5182026.0665.260.9
Center CropBackbone=ViT-S (21M)2026.066461.3
Full onlyBackbone=ViT-S (21M)2026.0659.856
Scaling on ScalesBackbone=ViT-S (21M), scales=[1, 2], Image size=224x2242026.0659.756.9
DFRBackbone=ViT-S (21M), Requires group labels=true2026.0658.754.5
Random CropsBackbone=ViT-S (21M)2026.064440.6