Efficient-VLN: A Training-Efficient Vision-Language Navigation Model
By: Duo Zheng , Shijia Huang , Yanyang Li and more
Potential Business Impact:
Teaches robots to navigate using less training.
Multimodal large language models (MLLMs) have shown promising potential in Vision-Language Navigation (VLN). However, their practical development is severely hindered by the substantial training overhead. We recognize two key issues that contribute to the overhead: (1) the quadratic computational burden from processing long-horizon historical observations as massive sequences of tokens, and (2) the exploration-efficiency trade-off in DAgger, i.e., a data aggregation process of collecting agent-explored trajectories. While more exploration yields effective error-recovery trajectories for handling test-time distribution shifts, it comes at the cost of longer trajectory lengths for both training and inference. To address these challenges, we propose Efficient-VLN, a training-efficient VLN model. Specifically, to mitigate the token processing burden, we design two efficient memory mechanisms: a progressive memory that dynamically allocates more tokens to recent observations, and a learnable recursive memory that utilizes the key-value cache of learnable tokens as the memory state. Moreover, we introduce a dynamic mixed policy to balance the exploration-efficiency trade-off. Extensive experiments show that Efficient-VLN achieves state-of-the-art performance on R2R-CE (64.2% SR) and RxR-CE (67.0% SR). Critically, our model consumes merely 282 H800 GPU hours, demonstrating a dramatic reduction in training overhead compared to state-of-the-art methods.
Similar Papers
Harnessing Input-Adaptive Inference for Efficient VLN
CV and Pattern Recognition
Helps robots navigate using less computer power.
FlexVLN: Flexible Adaptation for Diverse Vision-and-Language Navigation Tasks
CV and Pattern Recognition
Helps robots learn new places without retraining.
ActiveVLN: Towards Active Exploration via Multi-Turn RL in Vision-and-Language Navigation
Robotics
Teaches robots to explore new places by themselves.