Action Prediction on ProcTHOR single-object (OOD Systematic)
75AccuracyCDE (ViT-DINO)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CDE (ViT-DINO)Backbone=ViT, Pre-training=DINO2025.08 | 75 | 21 | |
| CDE (ViT-CLIP)Backbone=ViT, Pre-training=CLIP2025.08 | 72 | 25 | |
| CDE (ViT-MAE)Backbone=ViT, Pre-training=MAE2025.08 | 71 | 25 | |
| SMS-RBackbone=ResNet-182025.08 | 54 | 42 | |
| ICM-RBackbone=ResNet-182025.08 | 50 | 45 | |
| ICM-VBackbone=ViT-DINO2025.08 | 49 | 46 | |
| Vanilla-RBackbone=ResNet-182025.08 | 48 | 48 | |
| Vanilla-VBackbone=ViT-DINO2025.08 | 47 | 48 | |
| SMS-VBackbone=ViT-DINO2025.08 | 39 | 56 |