Vision-and-Language Navigation on R2R Discrete (val-unseen)
2.09Navigation Error (NE)DUET+ScaleVLN
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DUET+ScaleVLNSetting=Supervised Learning2026.01 | 2.09 | 14.09 | 88 | 81 | 70 | |
| HAMTSetting=Supervised Learning2026.01 | 2.29 | 11.46 | 73 | 66 | 61 | |
| DUETSetting=Supervised Learning2026.01 | 3.31 | 13.94 | 81 | 72 | 60 | |
| VLN-BERTSetting=Supervised Learning2026.01 | 3.93 | 12.01 | 69 | 63 | 57 | |
| SpatialNavSetting=Zero-Shot, spatial annotations=ground truth2026.01 | 4.4 | 13.8 | 70.8 | 59.3 | 48 | |
| SpatialNavSetting=Zero-Shot2026.01 | 4.54 | 13.8 | 68.2 | 57.7 | 47.8 | |
| PREVALENTSetting=Supervised Learning2026.01 | 4.71 | 10.19 | — | 58 | 53 | |
| MC-GPTSetting=Zero-Shot2026.01 | 5.42 | — | 68.8 | 32.1 | — | |
| SpatialGPTSetting=Zero-Shot2026.01 | 5.56 | — | 70.8 | 48.4 | 36.1 | |
| MapGPTSetting=Zero-Shot2026.01 | 5.63 | — | 57.6 | 43.7 | 34.8 | |
| NavCoTSetting=Supervised Learning2026.01 | 6.36 | 9.95 | 48 | 40 | 37 | |
| NavGPTSetting=Zero-Shot2026.01 | 6.46 | 11.45 | 42 | 34 | 29 |