Video-to-Text Retrieval on Charades
16.5R@5TrajViT-2
Evaluation Results
| Method | Links | |
|---|---|---|
| TrajViT-2mode=Zero-shot2026.02 | 16.5 | |
| TrajViTEvaluation Protocol=Zero-shot, Training Samples=4M2025.05 | 14.8 | |
| TrajViTmode=Zero-shot2026.02 | 14.6 | |
| ViT3Dmode=Zero-shot2026.02 | 12.9 | |
| TokenLearnerEvaluation Protocol=Zero-shot, Training Samples=4M2025.05 | 12.8 | |
| RLTEvaluation Protocol=Zero-shot, Training Samples=4M2025.05 | 12.8 | |
| ViT3DEvaluation Protocol=Zero-shot, Training Samples=4M2025.05 | 12.6 | |
| TokenLearnermode=Zero-shot2026.02 | 12.5 | |
| ViViTEvaluation Protocol=Zero-shot, Training Samples=4M2025.05 | 12.3 | |
| RLTmode=Zero-shot2026.02 | 12.3 | |
| ViViTmode=Zero-shot2026.02 | 11.8 | |
| AutoMergeEvaluation Protocol=Zero-shot, Training Samples=4M2025.05 | 11.4 | |
| ToMeEvaluation Protocol=Zero-shot, Training Samples=4M2025.05 | 10.5 |