Spatial-Temporal Anticipation on Ego4D STA v1, v2 (val)
55.98Base Performance (B)STAformer++ & AFF(learned)
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| STAformer++ & AFF(learned)Training Data=v22026.02 | 55.98 | 47.63 | 24.82 | 16.4 | 20.77 | 13.86 | 8.42 | 6.77 | |
| STAformer++Training Data=v22026.02 | 55.95 | 47.02 | 24.4 | 16.91 | 20.24 | 14.14 | 8.4 | 6.85 | |
| STAformer++Training Data=v12026.02 | 49.68 | 38.83 | 18.02 | 12.92 | 14 | 10.58 | 5.75 | 4.63 | |
| STAformer++ & AFF(learned)Training Data=v12026.02 | 48.6 | 39.12 | 19.45 | 12.29 | 15.77 | 10.28 | 5.71 | 4.77 | |
| GANO (w/ guided attention)Training Data=v1, Guided Attention=true2026.02 | 45.3 | 25.8 | 10.56 | 10.1 | 10.56 | 5.9 | 2.77 | 1.7 | |
| STAformerTraining Data=v22026.02 | 43.24 | 33.53 | 20.88 | 14.84 | 16.52 | 11.23 | 7.7 | 5.89 | |
| SlowfastTraining Data=v12026.02 | 40.5 | 24.5 | 0.34 | 8.16 | 0.34 | 5 | 0.06 | 0.06 | |
| Slowfast (w/Transformer)Training Data=v1, Backbone=Transformer2026.02 | 40.5 | 24.5 | 8.2 | 7.5 | 8.2 | 4.5 | 1.3 | 0.73 | |
| AVTTraining Data=v12026.02 | 40.5 | 24.5 | 8.45 | 7.12 | 8.45 | 4.39 | 1.15 | 0.71 | |
| ANACTOTraining Data=v12026.02 | 40.5 | 24.5 | 8.9 | 7.47 | 8.9 | 4.55 | 1.54 | 0.91 | |
| MeMVITTraining Data=v12026.02 | 40.5 | 24.5 | 10.05 | 9.27 | 10.04 | 4.95 | 2.11 | 1.34 | |
| GANO (w/o guided attention)Training Data=v1, Guided Attention=false2026.02 | 40.5 | 24.5 | 7.1 | 9.01 | 7.1 | 4.2 | 1.22 | 0.75 | |
| STAformerTraining Data=v12026.02 | 38.38 | 28.36 | 16.66 | 12.27 | 12.66 | 8.89 | 5.47 | 4.06 | |
| StillFastTraining Data=v12026.02 | 27.78 | 17.75 | 10.21 | 7.33 | 7.01 | 4.61 | 2.68 | 1.77 |