StreamVLN-Trajectory-Data: Vision-and-Language Navigation Paths for Embodied AI

This repository contains trajectory data for the paper "StreamVLN: Streaming Vision-and-Language Navigation via SlowFast Context Modeling." The dataset, created by author cywan, supports research in embodied AI and was last updated on December 31, 2025. It includes versions of the R2R and RxR navigation datasets.

Use Cases

• Training vision-and-language navigation agents based on the described trajectory data.
• Benchmarking agent performance in simulated environments using the included R2R and RxR datasets.
• Researching context modeling for streaming navigation tasks as referenced in the associated paper.

Strengths

• Data is directly associated with a published research paper, providing a clear research context.
• Maintenance is indicated by update logs from August and September 2025.
• Includes multiple dataset versions (R2R v1 and v1-3) and compatibility fixes for RxR.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and license information are unknown.

Provenance

Source

huggingface user cywan

Collection Method

Likely generated or compiled for the StreamVLN research project.

Time Range

null

Freshness

Last updated 2025-12-31 06:41:36; freshness should be verified.

Geography

null