Visual Question Answering on MMVP-VLM
26.7Orientation & Direction ScoreSigLIP ViT-SO-14
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| SigLIP ViT-SO-14Image Size=224^2, Params (M)=877.4, zero-shot=true2024.01 | 26.7 | 20 | 53.3 | 40 | 20 | 66.7 | 40 | 20 | 53.3 | 37.8 | |
| SigLIP VIT-SO-14Image Size=384^2, Params (M)=878, zero-shot=true2024.01 | 20 | 26.7 | 60 | 33.3 | 13.3 | 66.7 | 33.3 | 26.7 | 53.3 | 37 | |
| DFN VIT-H-14Image Size=224^2, Params (M)=986.1, zero-shot=true2024.01 | 20 | 26.7 | 73.3 | 26.7 | 26.7 | 66.7 | 46.7 | 13.3 | 53.3 | 39.3 | |
| EVA01 ViT-g-14Image Size=224^2, Params (M)=1136.4, zero-shot=true2024.01 | 16.7 | 26.7 | 40 | 6.7 | 13.3 | 66.7 | 13.3 | 13.3 | 20 | 23 | |
| OpenAI ViT-L-14Image Size=224^2, Params (M)=427.6, zero-shot=true2024.01 | 13.3 | 13.3 | 20 | 20 | 13.3 | 53.3 | 20 | 6.7 | 13.3 | 19.3 | |
| DFN VIT-H-14Image Size=378^2, Params (M)=986.7, zero-shot=true2024.01 | 13.3 | 20 | 53.3 | 33.3 | 26.7 | 66.7 | 40 | 20 | 40 | 34.8 | |
| MetaCLIP VIT-L-14Image Size=224^2, Params (M)=427.6, zero-shot=true2024.01 | 13.3 | 6.7 | 66.7 | 6.7 | 33.3 | 46.7 | 20 | 6.7 | 13.3 | 23.7 | |
| EVA02 ViT-bigE-14+Image Size=224^2, Params (M)=5044.9, zero-shot=true2024.01 | 13.3 | 20 | 66.7 | 26.7 | 26.7 | 66.7 | 26.7 | 20 | 33.3 | 33.3 | |
| MetaCLIP VIT-H-14Image Size=224^2, Params (M)=986.1, zero-shot=true2024.01 | 6.7 | 13.3 | 60 | 13.3 | 6.7 | 53.3 | 26.7 | 13.3 | 33.3 | 25.2 | |
| OpenAI ViT-L-14Image Size=336^2, Params (M)=427.9, zero-shot=true2024.01 | 0 | 20 | 40 | 20 | 6.7 | 20 | 33.3 | 6.7 | 33.3 | 20 |