Action Classification on HMDB51 1.0 (fine-tuned)
72.8AccuracyGDT
Evaluation Results
| Method | Links | |
|---|---|---|
| GDTNxW/H=32x112, Backbone=R21D, Pre-training=Different pre-training data2023.06 | 72.8 | |
| CVRLNxW/H=32x224, Backbone=R21D-18, Pre-training=K4002023.06 | 67.9 | |
| VIMPACNxW/H=10x256, Backbone=ViT-L, Pre-training=Different pre-training data2023.06 | 65.9 | |
| VideoMAENxW/H=16x224, Backbone=ViT-B, Pre-training=K4002023.06 | 62.6 | |
| TransRankNxW/H=16x112, Backbone=R21D-18, Pre-training=K4002023.06 | 60.1 | |
| TCLRNxW/H=16x112, Backbone=R21D-18, Pre-training=K4002023.06 | 60 | |
| AVTSNxW/H=25x224, Backbone=I3D, Pre-training=K4002023.06 | 53 | |
| RSPNet (Ours)NxW/H=16x112, Backbone=R21D-18, Pre-training=30k subset of K4002023.06 | 51.5 | |
| TempTransNxW/H=16x112, Backbone=R3D-18, Pre-training=K4002023.06 | 49.8 | |
| VideoMoCoNxW/H=16x112, Backbone=R21D-18, Pre-training=K4002023.06 | 49.2 | |
| XDCNxW/H=32x224, Backbone=R21D, Pre-training=K4002023.06 | 47.1 | |
| TaCoNxW/H=16x224, Backbone=R21D-18, Pre-training=K4002023.06 | 46 | |
| VideoMAE *NxW/H=16x112, Backbone=R21D-18, Pre-training=K4002023.06 | 45.4 | |
| RSPNetNxW/H=16x112, Backbone=R21D-18, Pre-training=K4002023.06 | 44.6 | |
| STSNxW/H=16x112, Backbone=R21D-18, Pre-training=K4002023.06 | 40.5 | |
| PacePredNxW/H=16x112, Backbone=R21D-18, Pre-training=K4002023.06 | 36.6 |