Perpendicular Plane prediction on 3D physical scene properties (test)
0.86ROC AUCStable Diffusion
Evaluation Results
| Method | Links | |
|---|---|---|
| Stable DiffusionEvaluation protocol=SVM grid search (best time step, layer and C)2023.10 | 0.86 | |
| DINOv2Backbone=ViT-G, Pre-training dataset=LVD-142M, Evaluation protocol=SVM grid search (best layer and C)2023.10 | 0.821 | |
| OpenCLIPBackbone=ViT-G, Pre-training dataset=LAION, Evaluation protocol=SVM grid search (best layer and C)2023.10 | 0.718 | |
| DINOv1Backbone=ViT-B, Pre-training dataset=ImageNet, Evaluation protocol=SVM grid search (best layer and C)2023.10 | 0.713 | |
| VQGANBackbone=48-layer transformer, Pre-training dataset=ImageNet, Evaluation protocol=SVM grid search (best layer and C)2023.10 | 0.62 | |
| RandomDescription=Random classifier2023.10 | 0.5 |