| action-classification-on-kinetics-400 | CoVeR (JFT-300M) | – | Acc@1: 86.3Acc@5: 97.2 |
| action-classification-on-kinetics-400 | CoVeR (JFT-3B) | – | Acc@1: 87.2Acc@5: 97.5 |
| action-classification-on-kinetics-600 | CoVeR (JFT-300M) | – | Top-1 Accuracy: 86.8Top-5 Accuracy: 97.3 |
| action-classification-on-kinetics-600 | CoVeR (JFT-3B) | – | Top-1 Accuracy: 87.9Top-5 Accuracy: 97.8 |
| action-classification-on-kinetics-700 | CoVeR (JFT-300M) | – | Top-1 Accuracy: 78.5Top-5 Accuracy: 94.2 |
| action-classification-on-kinetics-700 | CoVeR (JFT-3B) | – | Top-1 Accuracy: 79.8Top-5 Accuracy: 94.9 |
| action-classification-on-moments-in-time | CoVeR(JFT-300M) | – | Top 1 Accuracy: 45.0Top 5 Accuracy: 73.9 |
| action-classification-on-moments-in-time | CoVeR(JFT-3B) | – | Top 1 Accuracy: 46.1Top 5 Accuracy: 75.4 |
| action-recognition-in-videos-on-something | CoVeR(JFT-300M) | – | Top-1 Accuracy: 69.8Top-5 Accuracy: 91.9 |
| action-recognition-in-videos-on-something | CoVeR(JFT-3B) | – | Top-1 Accuracy: 70.9Top-5 Accuracy: 92.5 |