Vision-and-Language Navigation on R6R unseen (val)
11.8PLRCM(GOAL)
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| RCM(GOAL)pre-trained with data augmentation=true2020.05 | 11.8 | 9.2 | 18.2 | 14.8 | 31.6 | 25.9 | 7.6 | |
| RCM(GOAL)+trained_on=R4R, data_augmentation=true2020.05 | 12.8 | 9.2 | 19.3 | 15.2 | 31.8 | 23.5 | 7.3 | |
| RCM (Goal)Training Dataset=R2R, Data Augmentation=true2020.05 | 14.2 | 9.7 | 22.4 | 17.7 | 37.1 | 26.6 | 8.2 | |
| RCM (Fidelity)Training Dataset=R2R, Data Augmentation=true2020.05 | 15.7 | 9.8 | 22.7 | 18.3 | 36.4 | 26.1 | 8.4 | |
| RegretfulTraining Dataset=R2R, Data Augmentation=true, Reimplemented or readapted from original authors=true2020.05 | 15.9 | 8.8 | 24.2 | 16.6 | 40.9 | 16.2 | 6.8 | |
| REGRETFUL+*trained_on=R4R, data_augmentation=true, reimplemented=true2020.05 | 19.9 | 9.5 | 18 | 10.6 | 31.7 | 23.5 | 7.5 | |
| BabyWalkTraining Dataset=R2R, Data Augmentation=true2020.05 | 25.9 | 9.8 | 21.7 | 8.8 | 49 | 32.6 | 11.2 | |
| FAST+*trained_on=R4R, data_augmentation=true, reimplemented=true2020.05 | 26.6 | 8.9 | 22.1 | 13.7 | 31.5 | 23 | 7.7 | |
| BABYWALKpre-trained with data augmentation=true2020.05 | 27.2 | 9.3 | 22 | 8.1 | 47.4 | 33.4 | 11.8 | |
| RCM(FIDELITY)pre-trained with data augmentation=true2020.05 | 28 | 9.4 | 20.5 | 7.4 | 39 | 25.8 | 9.5 | |
| BABYWALKpre-trained with data augmentation=false2020.05 | 28.4 | 9.4 | 21.7 | 7.8 | 47.1 | 32.6 | 11.5 | |
| BabyWalk +trained_on=R4R, data_augmentation=true2020.05 | 28.7 | 9.2 | 25.5 | 9.2 | 47.2 | 32.7 | 13.6 | |
| BabyWalkTraining Dataset=R2R2020.05 | 29.1 | 10.1 | 21.4 | 7.9 | 48.4 | 30.8 | 11.2 | |
| FASTTraining Dataset=R2R, Data Augmentation=true, Reimplemented or readapted from original authors=true2020.05 | 32 | 9 | 26 | 16.5 | 37.7 | 21.9 | 8.5 | |
| RCM(FIDELITY)+trained_on=R4R, data_augmentation=true2020.05 | 33 | 9.3 | 20.5 | 8.5 | 38.3 | 23.7 | 7.9 | |
| SEQ2SEQpre-trained with data augmentation=false2020.05 | 34.1 | 9.5 | 18.1 | 9.6 | 23.4 | 19.3 | 6.5 | |
| BabyWalktrained_on=R4R, data_augmentation=false2020.05 | 37 | 8.8 | 26.4 | 8.1 | 44.9 | 30.1 | 13.1 | |
| SF+trained_on=R4R, data_augmentation=true2020.05 | 38.5 | 9.5 | 15.5 | 8.4 | 20.4 | 18.3 | 5.2 | |
| SEQ2SEQTraining Dataset=R2R2020.05 | 39.4 | 9.6 | 20.7 | 11 | 25.9 | 20.5 | 7.7 | |
| SEQ2SEQtrained_on=R4R, data_augmentation=false2020.05 | 40.8 | 9.9 | 14.4 | 6.8 | 17.7 | 16.4 | 4.6 | |
| Speaker-Follower (SF)Training Dataset=R2R, Data Augmentation=true2020.05 | 41.4 | 9.8 | 17.9 | 9.1 | 26.2 | 20.8 | 7.2 | |
| SFpre-trained with data augmentation=true2020.05 | 43.4 | 9.6 | 17.8 | 7.9 | 20.3 | 17.8 | 5.9 |