Action Recognition on UCF101 (10% labels)
60.1AccuracyL2A
Evaluation results
| Method | Links | |
|---|---|---|
| L2ABackbone=3D ResNet-18, Selector Pre-training=true2022.06 | 60.1 | |
| L2ABackbone=3D ResNet-18, Selector Pre-training=false2022.06 | 56.1 | |
| ActorCut+IDConference=Arxiv21, Backbone=3D ResNet-182022.06 | 53 | |
| TCLConference=ICCV21, Backbone=3D ResNet-182022.06 | 52.1 | |
| VideoSSLConference=WACV21, Backbone=3D ResNet-182022.06 | 42 | |
| ActorCutConference=Arxiv21, Backbone=3D ResNet-182022.06 | 40.2 | |
| ST-VideoMixConference=Arxiv21, Backbone=3D ResNet-182022.06 | 35.2 | |
| CutMixConference=ICCV19, Backbone=3D ResNet-182022.06 | 34.6 | |
| MixUpConference=ICLR18, Backbone=3D ResNet-182022.06 | 34.2 | |
| CutOutConference=Arxiv17, Backbone=3D ResNet-182022.06 | 33.9 | |
| S4LConference=ICCV19, Backbone=3D ResNet-182022.06 | 28.1 | |
| PseudoLabelConference=ICMLW13, Backbone=3D ResNet-182022.06 | 24.7 | |
| MeanTeacherConference=Neurips17, Backbone=3D ResNet-182022.06 | 23.8 |