Vision-and-Language Navigation on map2seq seen (dev)
16SDTWRConcat
Evaluation results
| Method | Links | ||
|---|---|---|---|
| RConcat2022.03 | 16 | 19 | |
| GA2022.03 | 17.2 | 16.5 | |
| VLN Transformer2022.03 | 17.5 | 18.6 | |
| ORAR full modelBackbone=ResNet 4th-to-last2022.03 | 41.1 | 7.2 | |
| ORAR full modelBackbone=ResNet pre-final, Ablation=no heading delta2022.03 | 45.4 | 6.8 | |
| ORAR full modelBackbone=ResNet pre-final2022.03 | 46.7 | 5.9 | |
| ORAR full modelBackbone=ResNet pre-final, Ablation=no junction type2022.03 | 47.2 | 7.6 |