Vision and Language Navigation on Room-to-Room (R2R) unseen Leaderboard (test)
8.13NLSeq-to-Seq
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Seq-to-SeqExperimental Setup=Single Run2019.04 | 8.13 | 20.4 | 0.18 | |
| Look Before You LeapExperimental Setup=Single Run2019.04 | 9.15 | 25.3 | 0.23 | |
| Reinforced Cross-ModalExperimental Setup=Pre-Explore2019.04 | 9.48 | 60.5 | 0.59 | |
| EnvDropExperimental Setup=Pre-Explore2019.04 | 9.79 | 63.9 | 0.61 | |
| RandomExperimental Setup=Single Run2019.04 | 9.89 | 13.2 | 0.12 | |
| EnvDropExperimental Setup=Single Run2019.04 | 11.7 | 51.5 | 0.47 | |
| Reinforced Cross-ModalExperimental Setup=Single Run2019.04 | 12 | 43.1 | 0.38 | |
| Speaker-FollowerExperimental Setup=Single Run2019.04 | 14.8 | 35 | 0.28 | |
| Self-MonitoringExperimental Setup=Single Run2019.04 | 18 | 48 | 0.35 | |
| Reinforced Cross-ModalExperimental Setup=Beam Search2019.04 | 358 | 63 | 0.02 | |
| Self-MonitoringExperimental Setup=Beam Search2019.04 | 373 | 61 | 0.02 | |
| EnvDropExperimental Setup=Beam Search2019.04 | 687 | 68.9 | 0.01 | |
| Speaker-FollowerExperimental Setup=Beam Search2019.04 | 1,257 | 53.5 | 0.01 |