Vision-Language Navigation on LH-VLN (test)
244SRFantasyVLN
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| FantasyVLNCoT Modal=unified multimodal2026.01 | 244 | 1,101 | 964 | 899 | |
| Aux-ThinkCoT Modal=Textual2026.01 | 65 | 316 | 204 | 147 | |
| RandomCoT Modal=None/ZS2026.01 | 0 | 0 | 0 | 0 | |
| GLM-4v promptCoT Modal=None/ZS2026.01 | 0 | 0 | 0 | 0 | |
| GPT-4 + NaviLLMCoT Modal=None/ZS2026.01 | 0 | 219 | 145 | 261 | |
| MGDMCoT Modal=None/ZS2026.01 | 0 | 234 | 165 | 291 | |
| CoT-VLACoT Modal=Visual2026.01 | 0 | 0 | 0 | 0 | |
| WorldVLACoT Modal=Visual2026.01 | 0 | 0 | 0 | 0 |