Reinforcement Learning on PROCGEN LEAPER 1.0 (test)
57.64AccuracyLAOF-Action
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LAOF-ActionAction Supervision=true2025.11 | 57.64 | 0.79 | |
| LAOFAction Supervision=false2025.11 | 53.23 | 0.74 | |
| CoMo w/ OFOptical Flow (OF) constraints=true2025.11 | 50.38 | 0.71 | |
| CoMo2025.11 | 44.29 | 0.56 | |
| LAOM-ActionAction Supervision=true2025.11 | 41.85 | 0.33 | |
| LAPO2025.11 | 40.09 | 0.55 |