Motion-to-text retrieval on KIT-ML (test)
60.23R@1Ours(3-modal)
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Ours(3-modal)Protocol=Small batches2024.03 | 60.23 | 77.52 | 86.44 | 93.22 | 95.87 | 1.2 | |
| Ours(2-modal)Protocol=Small batches2024.03 | 58.58 | 75.05 | 81.84 | 89.68 | 93.86 | 1.23 | |
| PSTProtocol=Small batches2026.01 | 57.14 | 74.53 | 80.76 | 90.15 | 97.58 | 1.26 | |
| MotionPatchProtocol=Small batches2026.01 | 54.54 | 72.15 | 79.68 | 89.35 | 96.11 | 1.31 | |
| TMRProtocol=Small batches2026.01 | 51.21 | 69.53 | 78.64 | 89 | 95.31 | 1.5 | |
| TMRProtocol=Small batches2024.03 | 50.12 | 67.12 | 76.88 | 88.88 | 94.75 | 1.53 | |
| Ours(3-modal)Protocol=Dissimilar subset2024.03 | 48 | 60 | 66 | 76 | 82 | 2 | |
| Ours(2-modal)Protocol=Dissimilar subset2024.03 | 43 | 59 | 67 | 73 | 83 | 2 | |
| TEMOSProtocol=Small batches2024.03 | 41.88 | 55.88 | 65.62 | 75.25 | 85.75 | 2.25 | |
| TEMOSProtocol=Small batches2026.01 | 41.88 | 55.88 | 65.62 | 75.25 | 85.75 | 2.25 | |
| Guo et al.Protocol=Small batches2024.03 | 39.75 | 62.75 | 73.62 | 86.88 | 95.88 | 1.95 | |
| T2MProtocol=Small batches2026.01 | 39.75 | 62.75 | 73.62 | 86.88 | 95.88 | 1.95 | |
| MotionCLIPProtocol=Small batches2024.03 | 39.55 | 52.07 | 68.13 | 77.94 | 90.85 | 2.16 | |
| TMRProtocol=Dissimilar subset2024.03 | 34 | 45 | 60 | 69 | 82 | 3.5 | |
| TEMOSProtocol=Dissimilar subset2024.03 | 33 | 39 | 45 | 49 | 64 | 6.5 | |
| MotionCLIPProtocol=Dissimilar subset2024.03 | 28 | 36 | 43 | 48 | 65 | 7 | |
| Ours(3-modal)Protocol=All with threshold2024.03 | 25.98 | 31.25 | 38.28 | 45.7 | 63.09 | 6.5 | |
| Guo et al.Protocol=Dissimilar subset2024.03 | 24 | 29 | 36 | 46 | 66 | 7 | |
| Ours(2-modal)Protocol=All with threshold2024.03 | 21.68 | 27.93 | 34.18 | 42.77 | 57.42 | 8 | |
| TMRProtocol=All with threshold2024.03 | 19.64 | 23.73 | 32.53 | 41.2 | 53.01 | 9.5 | |
| TEMOSProtocol=All with threshold2024.03 | 17.71 | 22.41 | 28.8 | 35.42 | 47.11 | 13.25 | |
| PSTProtocol=All2026.01 | 16.76 | 20.97 | 32.15 | 37.54 | 50.17 | 8 | |
| Motion PatchesTraining Dataset=Transferred2024.05 | 16.51 | — | — | 35.78 | 47.47 | 11 | |
| Ours(3-modal)Protocol=All2024.03 | 15.43 | 20.12 | 26.95 | 34.57 | 53.32 | 10 | |
| Lyu et al.Protocol=All2026.01 | 15.01 | 19.47 | 30.06 | 35.63 | 47.53 | 10.5 | |
| Motion PatchesTraining Dataset=KIT-ML2024.05 | 13.61 | — | — | 33.33 | 44.77 | 13 | |
| MotionPatchProtocol=All2026.01 | 13.61 | 17.26 | 27.54 | 33.33 | 44.77 | 13 | |
| MotionCLIPProtocol=All with threshold2024.03 | 13.24 | 22.11 | 29.53 | 38.06 | 50.23 | 10 | |
| TMRTraining Dataset=KIT-ML2024.05 | 11.83 | — | — | 29.39 | 38.55 | 16 | |
| TMRProtocol=All2026.01 | 11.83 | 13.74 | 22.14 | 29.39 | 38.55 | 16 | |
| Ours(2-modal)Protocol=All2024.03 | 11.72 | 17.19 | 23.63 | 32.81 | 48.83 | 12 | |
| TEMOSProtocol=All2024.03 | 11.69 | 15.3 | 20.12 | 26.63 | 36.39 | 26.5 | |
| TEMOSTraining Dataset=KIT-ML2024.05 | 11.69 | — | — | 26.63 | 36.39 | 26.5 | |
| TEMOSProtocol=All2026.01 | 11.69 | 15.3 | 20.12 | 26.63 | 36.39 | 26.5 | |
| TMRProtocol=All2024.03 | 11.2 | 13.86 | 20.12 | 28.07 | 38.55 | 18 | |
| Guo et al.Protocol=All with threshold2024.03 | 10.48 | 13.98 | 20.48 | 27.95 | 38.55 | 17.25 | |
| Motion PatchesTraining Dataset=HumanML3D2024.05 | 7.9 | — | — | 17.71 | 25.88 | 31 | |
| MotionCLIPProtocol=All2024.03 | 6.55 | 11.28 | 17.12 | 25.48 | 34.97 | 23 | |
| Guo et al.Protocol=All2024.03 | 4.94 | 6.51 | 10.72 | 16.14 | 25.3 | 28.5 | |
| T2MTraining Dataset=KIT-ML2024.05 | 4.94 | — | — | 16.14 | 25.3 | 28.5 | |
| T2MProtocol=All2026.01 | 4.94 | 6.51 | 10.72 | 16.14 | 25.3 | 28.5 |