Continuous Control on Walker2d v4 (Number of interactions)
98Number of Interactions (10^4 steps)PPO
Evaluation Results
| Method | Links | |
|---|---|---|
| PPOCritic Learning Rate=3e-4, Actor Learning Rate=3e-4, Actor and Critic Structure=(64, 64), Batch Size=322025.05 | 98 | |
| SACCritic Learning Rate=3e-4, Actor Learning Rate=3e-4, Actor and Critic Structure=(256, 256), Batch Size=2562025.05 | 90.5 | |
| TD3Critic Learning Rate=3e-4, Actor Learning Rate=3e-4, Actor and Critic Structure=(256, 256), Batch Size=2562025.05 | 81.5 | |
| SD3Critic Learning Rate=1e-3, Actor Learning Rate=1e-3, Actor and Critic Structure=(256, 256), Batch Size=2562025.05 | 65.5 | |
| CSACCritic Learning Rate=3e-4, Actor Learning Rate=3e-4, Actor and Critic Structure=(256, 256), Batch Size=2562025.05 | 45 |