Object-goal navigation on AI2-THOR Offline Data 1 (test)
74.4CF-SRTwo-Stage DRL training framework
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Two-Stage DRL training frameworkBase Navigation Model=AKGVP2025.02 | 74.4 | 41.6 | |
| AKGVPTraining Strategy=Standard2025.02 | 70.5 | 40.4 | |
| Two-Stage DRL training frameworkBase Navigation Model=L-sTDE2025.02 | 65.3 | 35.9 | |
| Two-Stage DRL training frameworkBase Navigation Model=HOZ2025.02 | 60.2 | 34.1 | |
| L-sTDETraining Strategy=Standard2025.02 | 60 | 34 | |
| Two-Stage DRL training frameworkBase Navigation Model=Baseline2025.02 | 59.3 | 33.1 | |
| HOZTraining Strategy=Standard2025.02 | 58.1 | 33.2 | |
| BaselineTraining Strategy=Standard2025.02 | 55.8 | 32.2 |