Reinforcement Learning on Halfcheetah Medium-Replay
96.71Normalized ReturnROAD
Evaluation results
| Method | Links | |
|---|---|---|
| ROADBackbone=Proto2026.05 | 96.71 | |
| Fixed-0.1Mixing Ratio=0.1, Backbone=Proto2026.05 | 93.4 | |
| Fixed-0.0Mixing Ratio=0.0, Backbone=Proto2026.05 | 91.06 | |
| Fixed-0.3Mixing Ratio=0.3, Backbone=Proto2026.05 | 90.19 | |
| Fixed-0.2Mixing Ratio=0.2, Backbone=Proto2026.05 | 89.74 | |
| Fixed-0.4Mixing Ratio=0.4, Backbone=Proto2026.05 | 87.62 | |
| Fixed-0.5Mixing Ratio=0.5, Backbone=Proto2026.05 | 86.25 |