Text-to-video Retrieval on SSv2-Label (test)
47.5R@1SViTT
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SViTTPT=2M, Frames=8, Sparsity=Hybrid, Fine-tuning=true2023.04 | 47.5 | 76.3 | 84.2 | 69.3 | |
| SingularityPT=5M, Frames=4, Fine-tuning=true2023.04 | 44.1 | 73.5 | 82.2 | 66.6 | |
| SViTTPT=2M, Frames=8, Sparsity=Dense, Fine-tuning=true2023.04 | 43.6 | 72.6 | 82.2 | 66.1 | |
| CLIP4ClipPT=400M, Frames=12, Fine-tuning=true2023.04 | 43.1 | 71.4 | 80.7 | 65.1 | |
| SingularityPT=5M, Frames=1, Fine-tuning=true2023.04 | 36.4 | 64.9 | 75.4 | 58.9 |