Image Classification on ImageNet-1k (Zero-shot)
84.4Zero-shot AccuracyDFN ViT-H-14
Evaluation results
| Method | Links | |
|---|---|---|
| DFN ViT-H-14Image Size=378^2, Params (M)=986.72024.01 | 84.4 | |
| DFN ViT-H-14Image Size=224^2, Params (M)=986.12024.01 | 83.4 | |
| SigLIP ViT-SO-14Image Size=384^2, Params (M)=8782024.01 | 83.1 | |
| SigLIP ViT-SO-14Image Size=224^2, Params (M)=877.42024.01 | 82 | |
| EVA02 ViT-bigE-14+Image Size=224^2, Params (M)=5044.92024.01 | 82 | |
| MetaCLIP ViT-H-14Image Size=224^2, Params (M)=986.12024.01 | 80.6 | |
| MetaCLIP ViT-L-14Image Size=224^2, Params (M)=427.62024.01 | 79.2 | |
| EVA01 ViT-g-14Image Size=224^2, Params (M)=1136.42024.01 | 78.5 | |
| OpenAI ViT-L-14Image Size=336^2, Params (M)=427.92024.01 | 76.6 | |
| OpenAI ViT-L-14Image Size=224^2, Params (M)=427.62024.01 | 75.5 |