Aerial Vision-and-Language Navigation on ANDH Unseen 1.0 (val)
18.3SPLHAA-LSTM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| HAA-LSTMhuman attention prediction training=true2022.05 | 18.3 | 2,000 | 54.4 | |
| HAA-Transformerhuman attention prediction training=true2022.05 | 16.5 | 2,040 | 55.2 | |
| E.T.Input Modality=Vision and Language2022.05 | 14.3 | 1,660 | 51.9 | |
| E.T.Input Modality=Language-only2022.05 | 13.9 | 1,700 | 50.6 | |
| LSTM2022.05 | 13.3 | 1,410 | 35.9 | |
| E.T.Input Modality=Vision-only2022.05 | 6.5 | 730 | 0.2 | |
| Random2022.05 | 0.2 | 100 | -81.4 |