Image-goal navigation on HM3D (test)
80Success RateHRNav
Evaluation results
| Method | Links | ||
|---|---|---|---|
| HRNavtraining_dataset=Gibson, evaluation_mode=Cross-domain2026.04 | 80 | 0.493 | |
| FGPrompt-MFBackbone=Resnet92023.10 | 76.1 | 49.6 | |
| FGPrompt-EFtraining_dataset=Gibson, evaluation_mode=Cross-domain2026.04 | 75.2 | 0.421 | |
| REGNavtraining_dataset=Gibson, evaluation_mode=Cross-domain2026.04 | 75.2 | 0.44 | |
| RFSGtraining_dataset=Gibson, evaluation_mode=Cross-domain2026.04 | 73.4 | 0.427 | |
| Zero-Shot Experience Learning2022.02 | 9.6 | 6.3 | |
| ZERBackbone=Resnet92023.10 | 9.6 | 6.3 | |
| ZERtraining_dataset=Gibson, evaluation_mode=Cross-domain2026.04 | 9.6 | 0.063 | |
| Hahn et al.2022.02 | 6.6 | 4.3 | |
| NRNSBackbone=Resnet182023.10 | 6.6 | 4.3 | |
| DTG-RL2022.02 | 5.5 | 3.7 | |
| Zhu et al.2022.02 | 4.4 | 2.7 | |
| Mezghani et al.Field of View=90°2022.02 | 3.5 | 1.9 | |
| Mem-Augtraining_dataset=Gibson, evaluation_mode=Cross-domain2026.04 | 3.5 | 0.019 | |
| Imitation Learning2022.02 | 2 | 1.9 | |
| Mem-AugBackbone=Resnet182023.10 | 1.9 | 3.5 |