Image Classification on UCF101 middle frame
81.6AccuracyCLIP
Evaluation Results
| Method | Links | |
|---|---|---|
| CLIPBackbone=ResNet-50, Evaluation Protocol=Linear probe2021.06 | 81.6 | |
| EsViTBackbone=Swin-T, Evaluation Protocol=Linear probe2021.06 | 81.1 | |
| SupervisedBackbone=Swin-T, Evaluation Protocol=Linear probe2021.06 | 79 | |
| SupervisedBackbone=ResNet-50, Evaluation Protocol=Linear probe, Reproduced=true2021.06 | 72.1 | |
| SupervisedBackbone=ResNet-50, Evaluation Protocol=Linear probe2021.06 | 71.2 |