Action Recognition on Something-Something V2 (1-shot/5-shot evaluation)
81.2Accuracy (5-shot)SOAP-Net
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SOAP-NetBackbone=ViT-B2024.07 | 81.2 | 66.7 | |
| SOAP-NetBackbone=ResNet-502024.07 | 79.8 | 61.9 | |
| AMFARBackbone=ResNet-502024.07 | 79.5 | 61.7 | |
| CLIP-FSARBackbone=ViT-B2024.07 | 72.1 | 61.9 | |
| MoLoBackbone=ViT-B2024.07 | 71.7 | 61.1 | |
| TRXBackbone=ViT-B2024.07 | 71.4 | 57.2 | |
| HyRSMBackbone=ViT-B2024.07 | 71.3 | 58.8 | |
| MoLoBackbone=ResNet-502024.07 | 70.7 | 56.6 | |
| STRMBackbone=ViT-B2024.07 | 70.2 | — | |
| Lite-MKDBackbone=ResNet-502024.07 | 69.9 | 55.7 | |
| TRXBackbone=ResNet-502024.07 | 69.8 | 55.8 | |
| GgHMBackbone=ResNet-502024.07 | 69.2 | 54.5 | |
| SA-CTBackbone=ResNet-502024.07 | 69.1 | 48.9 | |
| HyRSMBackbone=ResNet-502024.07 | 68.7 | 54.1 | |
| SloshNetBackbone=ResNet-502024.07 | 68.3 | 46.5 | |
| STRMBackbone=ResNet-502024.07 | 68.1 | — | |
| TempCLR2022.12 | 67.7 | 47.8 | |
| LSTCBackbone=ResNet-502024.07 | 66.7 | 47.6 | |
| SA-CTBackbone=ViT-B2024.07 | 66.3 | — | |
| HCLBackbone=ResNet-502024.07 | 64.9 | 47.3 | |
| TRX2022.12 | 64.6 | 42 | |
| ITANetBackbone=ResNet-502024.07 | 62.6 | 49.2 | |
| TA2NBackbone=ResNet-502024.07 | 61 | 47.6 | |
| MTFAN2022.12 | 60.4 | 45.4 | |
| MTFANBackbone=ResNet-502024.07 | 60.4 | 45.7 | |
| OTAM2022.12 | 52.3 | 42.8 | |
| OTAMBackbone=ResNet-502024.07 | 52.3 | 42.8 | |
| RTRN++2022.12 | 48.9 | 38.6 | |
| CMN++2022.12 | 43.8 | 34.4 | |
| TSN++2022.12 | 43 | 33.6 | |
| CMNBackbone=ResNet-502024.07 | 42.8 | 36.2 |