Vision-and-Language Navigation on R2R generalization (unseen)
35.2SRBABYWALK
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| BABYWALKTraining Dataset=R4R, Data Augmentation=false2020.05 | 35.2 | 48.5 | 27.2 | |
| BABYWALKTraining Dataset=R4R, Data Augmentation=true2020.05 | 34.1 | 50.4 | 27.8 | |
| RCM(FIDELITY)Training Dataset=R4R, Data Augmentation=true2020.05 | 29.1 | 34.3 | 18.3 | |
| RCM(GOAL)Training Dataset=R4R, Data Augmentation=true2020.05 | 25.9 | 44.2 | 20.2 | |
| FASTTraining Dataset=R4R, Data Augmentation=true, Reimplemented=true2020.05 | 25.1 | 33.9 | 14.2 | |
| REGRETFULTraining Dataset=R4R, Data Augmentation=true, Reimplemented=true2020.05 | 22.8 | 32.6 | 13.4 | |
| SFTraining Dataset=R4R, Data Augmentation=true2020.05 | 22.5 | 29.5 | 14.8 | |
| SEQ2SEQTraining Dataset=R4R, Data Augmentation=false2020.05 | 16.3 | 27.1 | 10.6 |