Vision-Language Navigation on R2R VLN Challenge Leaderboard (test)
1,257.38PLSpeaker-Follower
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Speaker-Followerbeam search=true2018.11 | 1,257.38 | 4.87 | 96 | 53.5 | 1 | |
| Self-MonitoringExploration=None, Trajectories per sample=12019.08 | 18 | 5.67 | — | 48 | 35 | |
| RCM2018.11 | 15.22 | 6.01 | 50.8 | 43.1 | 35 | |
| Speaker-FollowerExploration=None, Trajectories per sample=12019.08 | 14.8 | 6.62 | — | 35 | 28 | |
| The Regretful AgentExploration=None, Trajectories per sample=12019.08 | 13.7 | 5.69 | — | 48 | 40 | |
| Reinforced Cross-ModalExploration=None, Trajectories per sample=12019.08 | 12 | 6.12 | — | 43.1 | 38 | |
| RCMSIL=train2018.11 | 11.97 | 6.12 | 49.5 | 43 | 38 | |
| Speaker-Follower2018.11 | 11.82 | 6.62 | 44 | 35 | 28 | |
| ALTRExploration=None, Trajectories per sample=1, Pre-training=CMA + NVS2019.08 | 10.3 | 5.49 | — | 48 | 45 | |
| Random2018.11 | 9.89 | 9.79 | 18.3 | 13.2 | 12 | |
| RandomExploration=None, Trajectories per sample=12019.08 | 9.89 | 9.79 | — | 13.2 | 12 | |
| RCMSIL=unseen2018.11 | 9.48 | 4.21 | 66.8 | 60.5 | 59 | |
| RPA2018.11 | 9.15 | 7.53 | 32.5 | 25.3 | 23 | |
| Look Before You LeapExploration=None, Trajectories per sample=12019.08 | 9.15 | 7.53 | — | 25.3 | 23 | |
| seq2seq2018.11 | 8.13 | 7.85 | 26.6 | 20.4 | 18 | |
| Seq-to-SeqExploration=None, Trajectories per sample=12019.08 | 8.13 | 7.85 | — | 20.4 | 18 |