ActiveVLN: Towards Active Exploration via Multi-Turn RL in Vision-and-Language Navigation

Open in new window