Video-to-motion retrieval on HumanML3D (test)
74R@1LAVIMO
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| LAVIMOProtocol=(C) Dissimilar subset, Modality=3-modal2024.03 | 74 | 93 | 98 | 100 | 100 | 1 | |
| MotionSetProtocol=(C) Dissimilar subset2024.03 | 58 | 69 | 81 | 89 | 97 | 2 | |
| MotionCLIPProtocol=(C) Dissimilar subset2024.03 | 42 | 61 | 72 | 80 | 88 | 2 | |
| LAVIMOProtocol=(b) All with threshold, Modality=3-modal2024.03 | 30.42 | 46.97 | 56.82 | 68.77 | 81.41 | 3 | |
| LAVIMOProtocol=(a) All, Modality=3-modal2024.03 | 21.25 | 41.38 | 50.51 | 63.4 | 77.5 | 3 | |
| MotionSetProtocol=(b) All with threshold2024.03 | 15.63 | 31.17 | 40.22 | 52.93 | 65.47 | 8 | |
| MotionSetProtocol=(a) All2024.03 | 8.26 | 18.69 | 35.83 | 48.91 | 56.32 | 16 | |
| MotionCLIPProtocol=(b) All with threshold2024.03 | 7.1 | 12.62 | 16.67 | 22.56 | 32.23 | 29 | |
| MotionCLIPProtocol=(a) All2024.03 | 4.96 | 9.4 | 12.48 | 17.46 | 26.46 | 43 |