Video Action Recognition on UCF101 (test)
92.1Top-1 AccCVRL
Evaluation Results
| Method | Links | |
|---|---|---|
| CVRLBackbone=R3D, Resolution=224, Pre-training=Kinetics-400, Evaluation Protocol=Fine-tuning2021.08 | 92.1 | |
| MCNBackbone=R3D, Resolution=128, Pre-training=Kinetics-400, Evaluation Protocol=Fine-tuning2021.08 | 89.7 | |
| MCNBackbone=R(2+1)D, Resolution=128, Pre-training=Kinetics-400, Evaluation Protocol=Fine-tuning2021.08 | 89.2 | |
| CoCLRBackbone=S3D, Resolution=128, Pre-training=Kinetics-400, Evaluation Protocol=Fine-tuning2021.08 | 87.9 | |
| AWTEvaluation Protocol=EP1, Zero-shot=true2024.07 | 85.4 | |
| AWTEvaluation Protocol=EP2, Zero-shot=true2024.07 | 85.2 | |
| FROSTEREvaluation Protocol=EP1, Zero-shot=true2024.07 | 85 | |
| FROSTEREvaluation Protocol=EP2, Zero-shot=true2024.07 | 84.8 | |
| Open-VCLIPEvaluation Protocol=EP1, Zero-shot=true2024.07 | 83.5 | |
| Open-VCLIPEvaluation Protocol=EP2, Zero-shot=true2024.07 | 83.4 | |
| SpeedNetBackbone=S3D-G, Resolution=224, Pre-training=Kinetics-400, Evaluation Protocol=Fine-tuning2021.08 | 81.1 | |
| AdaptFormerEvaluation Protocol=EP1, Zero-shot=true2024.07 | 80.5 | |
| AdaptFormerEvaluation Protocol=EP2, Zero-shot=true2024.07 | 80.3 | |
| Text4VisEvaluation Protocol=EP1, Zero-shot=true2024.07 | 79.6 | |
| AIMEvaluation Protocol=EP2, Zero-shot=true2024.07 | 79.4 | |
| AIMEvaluation Protocol=EP1, Zero-shot=true2024.07 | 79 | |
| ST-AdapterEvaluation Protocol=EP1, Zero-shot=true2024.07 | 77.9 | |
| ST-AdapterEvaluation Protocol=EP2, Zero-shot=true2024.07 | 77.6 | |
| ActionCLIPEvaluation Protocol=EP2, Zero-shot=true2024.07 | 77.5 | |
| ActionCLIPEvaluation Protocol=EP1, Zero-shot=true2024.07 | 77.4 | |
| ViFi-CLIPEvaluation Protocol=EP2, Zero-shot=true2024.07 | 76.8 | |
| PPBackbone=R(2+1)D, Resolution=112, Pre-training=Kinetics-400, Evaluation Protocol=Fine-tuning2021.08 | 75.9 | |
| DPCBackbone=R2D-3D, Resolution=128, Pre-training=Kinetics-400, Evaluation Protocol=Fine-tuning2021.08 | 75.7 | |
| Vita-CLIPEvaluation Protocol=EP2, Zero-shot=true2024.07 | 75 | |
| MCNEvaluation Protocol=Linear probe, Backbone=R3D-182021.08 | 73.1 | |
| MCNEvaluation Protocol=Linear probe, Backbone=R(2+1)D2021.08 | 72.4 | |
| X-CLIPEvaluation Protocol=EP2, Zero-shot=true2024.07 | 72 | |
| Raw dataBackbone=SDANet2025.12 | 71.98 | |
| StegaVARStego Method=LF-VSN2025.12 | 71.66 | |
| StegaVARStego Method=Weng2025.12 | 70.32 | |
| CoCLREvaluation Protocol=Linear probe2021.08 | 70.2 | |
| StegaVARStego Method=HiNet2025.12 | 70.08 | |
| Ju et al.Evaluation Protocol=EP2, Zero-shot=true2024.07 | 69.3 | |
| ST-PuzzleBackbone=R3D, Resolution=224, Pre-training=Kinetics-400, Evaluation Protocol=Fine-tuning2021.08 | 63.9 | |
| 3D-RotNetBackbone=R3D, Resolution=112, Pre-training=Kinetics-400, Evaluation Protocol=Fine-tuning2021.08 | 62.9 | |
| BPAP2025.12 | 62.11 | |
| VITA2025.12 | 62.1 | |
| SPAct2025.12 | 62.03 | |
| Noise-Features2025.12 | 61.9 | |
| Obf-WeakBlur2025.12 | 61.52 | |
| Obf-StrongBlur2025.12 | 55.59 | |
| DownsampleScale=2x2025.12 | 54.11 | |
| MemDPCEvaluation Protocol=Linear probe2021.08 | 54.1 | |
| CBTEvaluation Protocol=Linear probe2021.08 | 54 | |
| Obf-Blackening2025.12 | 53.13 | |
| DownsampleScale=4x2025.12 | 39.65 |