Human Activity Recognition on Breakfast
90.7AccuracyS5 + LSMCL
Evaluation results
| Method | Links | |
|---|---|---|
| S5 + LSMCLP.T. Dataset=Kinetics-600, P.T. Samples=495K2023.03 | 90.7 | |
| S5P.T. Dataset=Kinetics-600, P.T. Samples=495K2023.03 | 90.14 | |
| D-Sprv.P.T. Dataset=HowTo100M, P.T. Samples=136M2023.03 | 89.9 | |
| ViS4merP.T. Dataset=Kinetics-600, P.T. Samples=495K2023.03 | 85.1 | |
| Br-PromptTraining=fine-tuned on Breakfast2022.03 | 80 | |
| GHRMP.T. Dataset=Kinetics-400, P.T. Samples=306K2023.03 | 75.5 | |
| GHRMPre-training=Kinetics2022.03 | 75.49 | |
| I3DTraining=fine-tuned on Breakfast2022.03 | 74.83 | |
| TimeceptionP.T. Dataset=Kinetics-400, P.T. Samples=306K2023.03 | 71.3 | |
| VideoGraphP.T. Dataset=Kinetics-400, P.T. Samples=306K2023.03 | 69.5 | |
| VideoGraphPre-training=Kinetics2022.03 | 69.45 | |
| TimeceptionPre-training=Kinetics2022.03 | 67.07 | |
| ActionVLADPre-training=Kinetics2022.03 | 65.48 | |
| I3DPre-training=Kinetics2022.03 | 58.61 |