Detail Perception on MMVP VLM
26.7Orientation and Direction AccuracyDIVA
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| DIVACLIP Backbone=OpenAI ViT-L-14, Resol.=336, #Params=427.9M2026.03 | 26.7 | 20 | 33.3 | 13.3 | 13.3 | 46.7 | 26.7 | 6.7 | 40 | 25.2 | |
| OriginalCLIP Backbone=SigLIP ViT-SO-14, Resol.=224, #Params=877.4M2026.03 | 26.7 | 20 | 53.3 | 40 | 20 | 66.7 | 40 | 20 | 53.3 | 37.8 | |
| DIVACLIP Backbone=SigLIP ViT-SO-14, Resol.=384, #Params=878.0M2026.03 | 26.7 | 33.3 | 53.3 | 26.7 | 13.3 | 80 | 40 | 26.7 | 46.7 | 38.5 | |
| GenHancerCLIP Backbone=SigLIP ViT-SO-14, Resol.=384, #Params=878.0M2026.03 | 26.7 | 20 | 66.7 | 33.3 | 13.3 | 86.7 | 40 | 26.7 | 46.7 | 40 | |
| OursCLIP Backbone=SigLIP ViT-SO-14, Resol.=384, #Params=878.0M2026.03 | 26.7 | 33.3 | 60 | 33.3 | 26.7 | 80 | 40 | 33.3 | 46.7 | 42.2 | |
| OursCLIP Backbone=MetaCLIP ViT-L-14, Resol.=224, #Params=427.6M2026.03 | 20 | 20 | 53.3 | 13.3 | 26.7 | 80 | 33.3 | 13.3 | 33.3 | 32.6 | |
| GenHancerCLIP Backbone=MetaCLIP ViT-H-14, Resol.=224, #Params=986.1M2026.03 | 20 | 20 | 66.7 | 26.7 | 26.7 | 66.7 | 33.3 | 20 | 53.3 | 37 | |
| OursCLIP Backbone=MetaCLIP ViT-H-14, Resol.=224, #Params=986.1M2026.03 | 20 | 20 | 66.7 | 26.7 | 26.7 | 66.7 | 33.3 | 26.7 | 53.3 | 37.8 | |
| GenHancerCLIP Backbone=SigLIP ViT-SO-14, Resol.=224, #Params=877.4M2026.03 | 20 | 20 | 66.7 | 60 | 20 | 86.7 | 40 | 13 | 53.3 | 42.2 | |
| OriginalCLIP Backbone=SigLIP ViT-SO-14, Resol.=384, #Params=878.0M2026.03 | 20 | 26.7 | 60 | 33.3 | 13.3 | 66.7 | 33.3 | 26.7 | 53.3 | 37 | |
| un2CLIPCLIP Backbone=SigLIP ViT-SO-14, Resol.=384, #Params=878.0M2026.03 | 20 | 20 | 60 | 46.7 | 26.7 | 73.3 | 40 | 26.7 | 60 | 41.5 | |
| OriginalCLIP Backbone=OpenAI ViT-L-14, Resol.=224, #Params=427.6M2026.03 | 13.3 | 13.3 | 20 | 20 | 13.3 | 53.3 | 20 | 6.7 | 13.3 | 19.2 | |
| DIVACLIP Backbone=OpenAI ViT-L-14, Resol.=224, #Params=427.6M2026.03 | 13.3 | 20 | 40 | 6.7 | 20 | 53.3 | 46.7 | 20 | 13.3 | 25.9 | |
| GenHancerCLIP Backbone=OpenAI ViT-L-14, Resol.=224, #Params=427.6M2026.03 | 13.3 | 33.3 | 33.3 | 20 | 6.7 | 73.3 | 46.7 | 20 | 40 | 31.8 | |
| OursCLIP Backbone=OpenAI ViT-L-14, Resol.=224, #Params=427.6M2026.03 | 13.3 | 33.3 | 33.3 | 26.7 | 6.7 | 73.3 | 46.7 | 20 | 46.7 | 33.3 | |
| OursCLIP Backbone=OpenAI ViT-L-14, Resol.=336, #Params=427.9M2026.03 | 13.3 | 26.7 | 33.3 | 20 | 6.7 | 73.3 | 46.7 | 33.3 | 26.7 | 31.1 | |
| OriginalCLIP Backbone=MetaCLIP ViT-L-14, Resol.=224, #Params=427.6M2026.03 | 13.3 | 6.7 | 66.7 | 6.7 | 33.3 | 46.7 | 20 | 6.7 | 13.3 | 23.7 | |
| GenHancerCLIP Backbone=MetaCLIP ViT-L-14, Resol.=224, #Params=427.6M2026.03 | 13.3 | 20 | 53.3 | 13.3 | 26.7 | 80 | 33.3 | 13.3 | 33.3 | 31.8 | |
| DIVACLIP Backbone=MetaCLIP ViT-H-14, Resol.=224, #Params=986.1M2026.03 | 13.3 | 20 | 53.3 | 33.3 | 13.3 | 66.7 | 33.3 | 13.3 | 40 | 31.8 | |
| DIVACLIP Backbone=SigLIP ViT-SO-14, Resol.=224, #Params=877.4M2026.03 | 13.3 | 26.7 | 60 | 46.7 | 13.3 | 73.3 | 53.3 | 26.7 | 53.3 | 40.7 | |
| OursCLIP Backbone=SigLIP ViT-SO-14, Resol.=224, #Params=877.4M2026.03 | 13.3 | 26.7 | 73.3 | 60 | 20 | 80 | 40 | 20 | 53.3 | 43 | |
| GenHancerCLIP Backbone=OpenAI ViT-L-14, Resol.=336, #Params=427.9M2026.03 | 6.7 | 20 | 33.3 | 20 | 6.7 | 73.3 | 53.3 | 26.7 | 26.7 | 29.6 | |
| un2CLIPCLIP Backbone=OpenAI ViT-L-14, Resol.=336, #Params=427.9M2026.03 | 6.7 | 33.3 | 46.7 | 13.3 | 13.3 | 80 | 40 | 20 | 20 | 30.4 | |
| DIVACLIP Backbone=MetaCLIP ViT-L-14, Resol.=224, #Params=427.6M2026.03 | 6.7 | 6.7 | 60 | 0 | 26.7 | 66.7 | 20 | 20 | 40 | 27.4 | |
| OriginalCLIP Backbone=MetaCLIP ViT-H-14, Resol.=224, #Params=986.1M2026.03 | 6.7 | 13.3 | 60 | 13.3 | 6.7 | 53.3 | 26.7 | 13.3 | 33.3 | 25.2 | |
| un2CLIPCLIP Backbone=OpenAI ViT-L-14, Resol.=224, #Params=427.6M2026.03 | 0 | 33.3 | 46.7 | 26.7 | 13.3 | 80 | 40 | 20 | 33.3 | 32.6 | |
| OriginalCLIP Backbone=OpenAI ViT-L-14, Resol.=336, #Params=427.9M2026.03 | 0 | 20 | 40 | 20 | 6.7 | 20 | 33.3 | 6.7 | 33.3 | 20 |