VLN-R1: Vision-Language Navigation via Reinforcement Fine-Tuning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Qi, Zhangyang, Zhang, Zhixiong, Yu, Yizhou, Wang, Jiaqi, Zhao, Hengshuang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911020962807808
author Qi, Zhangyang
Zhang, Zhixiong
Yu, Yizhou
Wang, Jiaqi
Zhao, Hengshuang
author_facet Qi, Zhangyang
Zhang, Zhixiong
Yu, Yizhou
Wang, Jiaqi
Zhao, Hengshuang
contents Vision-Language Navigation (VLN) is a core challenge in embodied AI, requiring agents to navigate real-world environments using natural language instructions. Current language model-based navigation systems operate on discrete topological graphs, limiting path planning to predefined node connections. We propose VLN-R1, an end-to-end framework that leverages Large Vision-Language Models (LVLM) to directly translate egocentric video streams into continuous navigation actions, adopting GRPO-based training inspired by DeepSeek-R1. To enable effective training, we first construct the VLN-Ego dataset using a 3D simulator, Habitat, and propose Long-Short Memory Sampling to balance historical and current observations. While large language models can supervise complete textual instructions, they lack fine-grained action-level control. Our framework employs a two-stage training approach: a) Supervised fine-tuning (SFT) to align the model's action sequence text predictions with expert demonstrations, followed by b) Reinforcement fine-tuning (RFT) enhanced with a Time-Decayed Reward (TDR) mechanism that strategically weights multi-step future actions. Experimental results show VLN-R1 achieves strong performance on VLN-CE benchmark. VLN-R1 proves LVLMs can drive embodied navigation and enhance task-specific reasoning through data-efficient, reward-driven post-training.
format Preprint
id arxiv_https___arxiv_org_abs_2506_17221
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle VLN-R1: Vision-Language Navigation via Reinforcement Fine-Tuning
Qi, Zhangyang
Zhang, Zhixiong
Yu, Yizhou
Wang, Jiaqi
Zhao, Hengshuang
Computer Vision and Pattern Recognition
Vision-Language Navigation (VLN) is a core challenge in embodied AI, requiring agents to navigate real-world environments using natural language instructions. Current language model-based navigation systems operate on discrete topological graphs, limiting path planning to predefined node connections. We propose VLN-R1, an end-to-end framework that leverages Large Vision-Language Models (LVLM) to directly translate egocentric video streams into continuous navigation actions, adopting GRPO-based training inspired by DeepSeek-R1. To enable effective training, we first construct the VLN-Ego dataset using a 3D simulator, Habitat, and propose Long-Short Memory Sampling to balance historical and current observations. While large language models can supervise complete textual instructions, they lack fine-grained action-level control. Our framework employs a two-stage training approach: a) Supervised fine-tuning (SFT) to align the model's action sequence text predictions with expert demonstrations, followed by b) Reinforcement fine-tuning (RFT) enhanced with a Time-Decayed Reward (TDR) mechanism that strategically weights multi-step future actions. Experimental results show VLN-R1 achieves strong performance on VLN-CE benchmark. VLN-R1 proves LVLMs can drive embodied navigation and enhance task-specific reasoning through data-efficient, reward-driven post-training.
title VLN-R1: Vision-Language Navigation via Reinforcement Fine-Tuning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.17221