Trajectory Forecasting on NBA SportVU (test)
0.64Min ADE (20)SVAE
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SVAEconditioning_frames=8, prediction_frames=12, units=feet2025.05 | 0.64 | 1.17 | — | — | |
| LaM-SLidEconditioning_frames=8, prediction_frames=12, units=feet2025.05 | 0.64 | 1.09 | — | — | |
| MoFlow2026.05 | 0.71 | 0.86 | 1.52 | 2.73 | |
| MART2026.05 | 0.72 | 0.9 | 1.52 | 2.77 | |
| Trajectron++conditioning_frames=8, prediction_frames=12, units=feet2025.05 | 0.73 | 1.46 | — | — | |
| LED2026.05 | 0.81 | 1.1 | 1.63 | 2.99 | |
| STFlowconditioning_frames=8, prediction_frames=12, units=feet2025.05 | 0.82 | 1.35 | — | — | |
| STFlow-Transformerconditioning_frames=8, prediction_frames=12, units=feet, Temporal layer=Transformer2025.05 | 0.82 | 1.35 | — | — | |
| U2Diffine2026.05 | 0.84 | 1.11 | 1.47 | 2.65 | |
| U2Diff2026.05 | 0.85 | 1.11 | 1.48 | 2.68 | |
| GroupNetremark=pretrained model from their codebase used2026.05 | 0.94 | 1.22 | 2.12 | 3.72 | |
| MID2026.05 | 0.96 | 1.27 | — | — | |
| MemoNet2026.05 | 1.15 | 1.57 | — | — | |
| AutoBotsremark=trained using their original code2026.05 | 1.19 | 1.55 | 1.75 | 2.73 | |
| NPSN2026.05 | 1.25 | 1.47 | — | — |