Motion Prediction on AMASS
0.3ADESLD
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SLDIn/Out (s)=0.5/2.02026.03 | 0.3 | — | — | 0.45 | — | — | |
| CoMusionIn/Out (s)=0.5/2.02026.03 | 0.31 | — | — | 0.46 | — | — | |
| Ours (wide)In/Out (s)=0.5/2.02026.03 | 0.31 | — | — | 0.45 | — | — | |
| MotionmapIn/Out (s)=0.5/2.02026.03 | 0.32 | — | — | 0.45 | — | — | |
| BeLFusionIn/Out (s)=0.5/2.02026.03 | 0.35 | — | — | 0.48 | — | — | |
| Ours (deep)In/Out (s)=0.5/2.02026.03 | 0.35 | — | — | 0.47 | — | — | |
| EMPMPIn/Out (s)=0.5/2.0, adapted_for_task=true2026.03 | 0.42 | — | — | 0.65 | — | — | |
| DivSampIn/Out (s)=0.5/2.02026.03 | 0.48 | — | — | 0.64 | — | — | |
| SkeletonDiffIn/Out (s)=0.5/2.02026.03 | 0.56 | — | — | 0.71 | — | — | |
| T2PIn/Out (s)=0.5/2.0, adapted_for_task=true2026.03 | 0.63 | — | — | 0.94 | — | — | |
| MotionGPTModel size=Base, Instruction Tuning=true2023.06 | 4.745 | 0.905 | 8.972 | 6.04 | — | — | |
| MDM2023.06 | 5.446 | 6.031 | 7.813 | 8.561 | — | — | |
| BeLFusionType=Stochastic (Latent Diffusion)2026.03 | — | — | — | — | 0.488 | 0.564 | |
| CoMusionType=Stochastic (Motion Diffusion)2026.03 | — | — | — | — | 0.526 | 0.602 | |
| DivSampType=Stochastic (Gumbel-Softmax)2026.03 | — | — | — | — | 0.623 | 0.728 | |
| MotionmapType=Stochastic (Multi-Stage Encoder-Decoder)2026.03 | — | — | — | — | 0.45 | 0.514 | |
| Real2023.06 | — | 0.002 | 9.503 | — | — | — | |
| SimpliHuMoN (deep)Type=Deterministic (K-Proposal)2026.03 | — | — | — | — | 0.521 | 0.571 | |
| SimpliHuMoN (wide)Type=Deterministic (K-Proposal)2026.03 | — | — | — | — | 0.519 | 0.56 | |
| SkeletonDiff*Type=Stochastic (Gaussian Diffusion)2026.03 | — | — | — | — | 0.641 | 0.74 | |
| SLD*Type=Stochastic (State-Space Diffusion)2026.03 | — | — | — | — | 0.482 | 0.551 |