Text-to-Motion Retrieval on HumanML3D
88.05Recall@3CoAMD
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CoAMD2026.04 | 88.05 | 68.33 | 82.61 | 92.45 | 95.97 | — | |
| LaMP2026.04 | 87.04 | 67.18 | 81.9 | 92 | 95.73 | — | |
| TMR2026.04 | 86.81 | 67.16 | 81.32 | 91.43 | 95.36 | — | |
| UniMotionbackbone=Show-o2 1.5B2026.03 | 84.1 | — | — | — | — | — | |
| T2M2026.04 | 80.65 | 52.48 | 71.05 | 89.66 | 96.58 | — | |
| MG-MotionLLM2026.03 | 80.2 | — | — | — | — | — | |
| MotionGPT2026.03 | 77.8 | — | — | — | — | — | |
| TEMOS2026.04 | 61.14 | 40.49 | 53.52 | 70.96 | 84.15 | — | |
| Ours-LVenue=Proposed, Backbone=ViT-Large, RoBERTa-Large, LLM Augmentation=false2026.03 | 26.22 | 13.76 | 19.18 | 34.89 | 48.08 | 11 | |
| OursVenue=Proposed, LLM Augmentation=false2026.03 | 23.53 | 11.87 | 18.13 | 31.71 | 43.8 | 14 | |
| SGARVenue=NeurIPS’25, LLM Augmentation=true2026.03 | 23.52 | 12.86 | — | 30.75 | 43 | 15 | |
| Lyu et al.Venue=ICLR’25, LLM Augmentation=false2026.03 | 23.25 | 11.8 | 17.11 | 30.81 | 43.36 | 14 | |
| MoPatch-LBackbone=Large, LLM Augmentation=false2026.03 | 22.88 | 12.07 | 16.88 | 30.36 | 42.36 | 15 | |
| PL-TMRVenue=AAAI’25, LLM Augmentation=true2026.03 | 22.18 | 11 | 17.02 | 29.48 | 43.43 | 14 | |
| MoPatchVenue=CVPR’24, LLM Augmentation=false2026.03 | 21.94 | 11.52 | 15.78 | 29.15 | 40.78 | 17 | |
| SGAR-LBackbone=Large, LLM Augmentation=true2026.03 | 21.59 | 11.27 | 16.43 | 28.5 | 41.37 | 16 | |
| KinMoVenue=ICCV’25, LLM Augmentation=true2026.03 | 20.47 | 9.05 | 15.23 | 28.62 | 41.6 | 16 | |
| CARVenue=ECCV’24, LLM Augmentation=false2026.03 | 18.96 | 10.1 | 13.5 | 26.42 | 37.28 | 19.5 | |
| TMRVenue=ICCV’23, LLM Augmentation=false2026.03 | 16.33 | 8.92 | — | 22.06 | 33.37 | 25 |