Autonomous Driving on Waymo Open Motion Dataset (WOMD) 10k Human-replay (test)
97.6Scorespiced self-play
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| spiced self-playHuman demos used=30 hours, regularization=true2026.06 | 97.6 | 1.4 | 0.8 | 1.1 | 99.9 | |
| spiced self-playHuman demos used=3 hours, regularization=true2026.06 | 97.3 | 1.6 | 0.6 | 1.2 | 100 | |
| spiced self-playHuman demos used=30 min, regularization=true2026.06 | 96.8 | 2 | 0.7 | 1.4 | 99.8 | |
| unreg. self-playHuman demos used=—2026.06 | 96.7 | 2.7 | 2.1 | 0.6 | 100 | |
| spiced self-playHuman demos used=10 min, regularization=true2026.06 | 94.1 | 3.9 | 1.4 | 1.4 | 99.6 | |
| SMARTHuman demos used=52 days2026.06 | 65.4 | 4.4 | 1.6 | 1.1 | 88.5 | |
| SMARTHuman demos used=30 hours2026.06 | 60.1 | 6.8 | 3.3 | 1.6 | 85.4 | |
| SMARTHuman demos used=3 hours2026.06 | 51.8 | 11.4 | 6.9 | 4.5 | 81.5 | |
| SMARTHuman demos used=30 min2026.06 | 37.9 | 17.9 | 12.5 | 16.8 | 76.9 | |
| SMARTHuman demos used=10 min2026.06 | 24.6 | 32 | 25 | 18.6 | 57.7 |