Vision-and-Language Navigation on R8R unseen (val)
56.4PLSEQ2SEQ
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| SEQ2SEQtrained_on=R4R, data_augmentation=false2020.05 | 56.4 | 10.1 | 20.7 | 10.4 | 15 | 13.4 | 4.7 | |
| SFpre-trained with data augmentation=true2020.05 | 53 | 10.1 | 18.6 | 9.8 | 16.3 | 13.5 | 4.4 | |
| SEQ2SEQTraining Dataset=R2R2020.05 | 52.3 | 10.5 | 16.9 | 6.1 | 22.5 | 17.1 | 4.1 | |
| Speaker-Follower (SF)Training Dataset=R2R, Data Augmentation=true2020.05 | 52.2 | 10.5 | 13.8 | 5.6 | 24.1 | 18.2 | 3.8 | |
| SF+trained_on=R4R, data_augmentation=true2020.05 | 50.8 | 9.5 | 21.6 | 11.8 | 17.2 | 15.1 | 5 | |
| BabyWalktrained_on=R4R, data_augmentation=false2020.05 | 50 | 9.3 | 26.3 | 7.2 | 44.7 | 27.1 | 11.5 | |
| RCM(FIDELITY)pre-trained with data augmentation=true2020.05 | 42.3 | 10.7 | 18.2 | 5.3 | 37.2 | 21.6 | 7.6 | |
| SEQ2SEQpre-trained with data augmentation=false2020.05 | 40 | 9.9 | 20.2 | 12.4 | 19.8 | 15.8 | 5.1 | |
| BabyWalk +trained_on=R4R, data_augmentation=true2020.05 | 39.9 | 10.1 | 23.1 | 7.4 | 46 | 28.2 | 11.1 | |
| BABYWALKpre-trained with data augmentation=true2020.05 | 39.1 | 9.9 | 22 | 7 | 46.4 | 28.3 | 10.1 | |
| RCM(FIDELITY)+trained_on=R4R, data_augmentation=true2020.05 | 38.7 | 9.9 | 20.9 | 9 | 34.6 | 21.7 | 6.1 | |
| BabyWalkTraining Dataset=R2R2020.05 | 38.3 | 11.1 | 19.6 | 6.9 | 48.1 | 26.7 | 9.4 | |
| BABYWALKpre-trained with data augmentation=false2020.05 | 35.6 | 9.6 | 22.3 | 7.3 | 46.4 | 29.6 | 10.4 | |
| FASTTraining Dataset=R2R, Data Augmentation=true, Reimplemented or readapted from original authors=true2020.05 | 34.9 | 10.6 | 11.1 | 6.2 | 33.7 | 14.5 | 2.4 | |
| BabyWalkTraining Dataset=R2R, Data Augmentation=true2020.05 | 34 | 10.5 | 20.7 | 7.8 | 48.7 | 29.1 | 9.8 | |
| FAST+*trained_on=R4R, data_augmentation=true, reimplemented=true2020.05 | 28.2 | 9.1 | 27.7 | 13.7 | 29.6 | 17.7 | 6.9 | |
| REGRETFUL+*trained_on=R4R, data_augmentation=true, reimplemented=true2020.05 | 20.7 | 9.5 | 18.7 | 9.2 | 29.3 | 19 | 5.6 | |
| RCM (Fidelity)Training Dataset=R2R, Data Augmentation=true2020.05 | 16.9 | 11.1 | 12.6 | 7.5 | 30.9 | 23.3 | 4.3 | |
| RegretfulTraining Dataset=R2R, Data Augmentation=true, Reimplemented or readapted from original authors=true2020.05 | 16.6 | 10 | 16.3 | 7.7 | 35.3 | 8.1 | 2.4 | |
| RCM (Goal)Training Dataset=R2R, Data Augmentation=true2020.05 | 15.3 | 11 | 12.4 | 7.4 | 32.4 | 23.9 | 4.3 | |
| RCM(GOAL)+trained_on=R4R, data_augmentation=true2020.05 | 13.9 | 9.5 | 22.8 | 16.9 | 27.6 | 19.5 | 5.1 | |
| RCM(GOAL)pre-trained with data augmentation=true2020.05 | 12.4 | 10.2 | 19.7 | 15.4 | 25.7 | 19.4 | 5.8 |