Multi-label Video Classification on Charades 12 fps setting (test)
66.3mAPTokenLearner
Evaluation results
| Method | Links | |
|---|---|---|
| TokenLearnerInput=RGB, Pre-train=None2021.06 | 66.3 | |
| MoViNetsInput=RGB, Pre-train=None2021.06 | 63.2 | |
| Backbone (X(2+1)D-M)Input=RGB, Pre-train=None2021.06 | 62.7 | |
| AssembleNet-101Input=RGB+Flow, Pre-train=Kinetics2021.06 | 58.6 | |
| AssembleNet++ (w/o obj.)Input=RGB+Flow, Pre-train=None2021.06 | 55 | |
| Multiscale ViTInput=RGB, Pre-train=Kinetics2021.06 | 47.7 | |
| AssembleNet-50Input=RGB+Flow, Pre-train=None2021.06 | 47 | |
| SlowFast-101Input=RGB+RGB, Pre-train=Kinetics2021.06 | 45.2 | |
| SGFB-101Input=RGB, Pre-train=Kinetics2021.06 | 44.3 | |
| LFB-101Input=RGB, Pre-train=Kinetics2021.06 | 42.5 | |
| STRGInput=RGB, Pre-train=Kinetics2021.06 | 39.7 | |
| EvaNetInput=RGB, Pre-train=Kinetics2021.06 | 38.1 | |
| I3D + Non-localInput=RGB, Pre-train=Kinetics2021.06 | 37.5 | |
| I3D (from [32])Input=RGB, Pre-train=Kinetics2021.06 | 35.5 | |
| I3DInput=RGB, Pre-train=Kinetics2021.06 | 32.9 |