Aerial Vision-and-Language Navigation on ANDH Full Unseen 1.0 (val)
340SPLHAA-LSTM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| HAA-LSTMhuman attention prediction training=true2022.05 | 340 | 370 | 56.1 | |
| HAA-Transformerhuman attention prediction training=true2022.05 | 320 | 470 | 50.9 | |
| LSTM2022.05 | 320 | 370 | 48.7 | |
| E.T.Input Modality=Vision and Language2022.05 | 250 | 370 | 48.9 | |
| E.T.Input Modality=Language-only2022.05 | 230 | 370 | 54.5 | |
| E.T.Input Modality=Vision-only2022.05 | 20 | 20 | -5.2 | |
| Random2022.05 | 0 | 0 | -162.4 |