Pickup on BabyAI
33.4Average Pass RatePolychromic PPO
Evaluation Results
| Method | Links | |
|---|---|---|
| Polychromic PPOevaluation_protocol=pass@1, initial-state perturbations=large2025.09 | 33.4 | |
| Poly-PPO w/ UCBevaluation_protocol=pass@1, initial-state perturbations=large2025.09 | 28 | |
| REINFORCE w/ Baselineevaluation_protocol=pass@1, initial-state perturbations=large2025.09 | 22 | |
| REINFORCE w/ UCBevaluation_protocol=pass@1, initial-state perturbations=large2025.09 | 20.5 | |
| Pretrained policyevaluation_protocol=pass@1, initial-state perturbations=large2025.09 | 15.2 | |
| PPOevaluation_protocol=pass@1, initial-state perturbations=large2025.09 | 12.5 | |
| PPO w/ UCBevaluation_protocol=pass@1, initial-state perturbations=large2025.09 | 8.87 |