Action Recognition on ActivityNet v1.3 (test)
96.1mAPBIKE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| BIKEBackbone=ViT-L2022.12 | 96.1 | 94.7 | |
| NSNet2022.12 | 94.3 | 90.2 | |
| TSQNet2022.12 | 93.7 | 88.7 | |
| CESensemble=TSN (Inception V3, ResNet-152) + audio models2017.05 | 93.23 | 88.14 | |
| QCIS2017.05 | 92.41 | 87.79 | |
| MSRA2017.05 | 91.94 | 86.69 | |
| DSANet2022.12 | 90.5 | — | |
| MARL2022.12 | 90.1 | 85.7 | |
| ListenToLook2022.12 | 89.9 | — | |
| UTS2017.05 | 87.16 | 84.9 | |
| Tokyo Univ.2017.05 | 86.46 | 80.43 | |
| AFNetRS=0.4, RT=0.8, GFLOPs=24.62022.11 | 75.6 | — | |
| AdaFocusGFLOPs=26.62022.11 | 75 | — | |
| VideoIQGFLOPs=28.12022.11 | 74.8 | — | |
| AR-NetGFLOPs=33.52022.11 | 73.8 | — | |
| SCSamplerGFLOPs=422022.11 | 72.9 | — | |
| LiteEvalGFLOPs=95.12022.11 | 72.7 | — | |
| ListenToLookGFLOPs=81.42022.11 | 72.3 | — | |
| AdaFrameGFLOPs=792022.11 | 71.5 | — |