Offline-to-online Reinforcement Learning on pen
5.3RegretSMAC
Evaluation results
| Method | Links | |
|---|---|---|
| SMACOffline Algorithm=SMAC, Fine-tuning algorithm=SAC2026.02 | 5.3 | |
| SMACFine-tuning algorithm=TD3+BC, Offline Algorithm=SMAC2026.02 | 7.7 | |
| CalQL/CQLOffline Algorithm=CalQL/CQL, Fine-tuning algorithm=SAC2026.02 | 8 | |
| IQLOffline Algorithm=IQL, Fine-tuning algorithm=SAC2026.02 | 10.3 | |
| SMACOffline Algorithm=SMAC, Fine-tuning algorithm=AWR2026.02 | 13.3 | |
| IQLFine-tuning algorithm=TD3+BC, Offline Algorithm=IQL2026.02 | 16.1 | |
| CalQL/CQLFine-tuning algorithm=TD3+BC, Offline Algorithm=CalQL/CQL2026.02 | 17.8 | |
| IQLOffline Algorithm=IQL, Fine-tuning algorithm=AWR2026.02 | 18 | |
| CalQL/CQLOffline Algorithm=CalQL/CQL, Fine-tuning algorithm=AWR2026.02 | 28.8 | |
| TD3+BCFine-tuning algorithm=TD3+BC, Offline Algorithm=TD3+BC2026.02 | 31.8 | |
| TD3+BCOffline Algorithm=TD3+BC, Fine-tuning algorithm=AWR2026.02 | 32.9 | |
| TD3+BCOffline Algorithm=TD3+BC, Fine-tuning algorithm=SAC2026.02 | 55.2 |