Jepa Wms: Robotics Trajectories for World Model Training

Meta AI Research, FAIR hosts this collection of robotics trajectories for training JEPA-WM world models. The datasets are available on HuggingFace and include configurations such as 'metaworld' and 'pusht'. The repository was last updated on 2026-01-02.

Use Cases

• Train world models for robotic control based on the provided trajectory data.
• Benchmark world model architectures using the 'metaworld' and 'pusht' dataset configurations.
• Simulate robotic environments and predict future states based on learned trajectory patterns.

Strengths

• Hosted by Meta AI Research, FAIR, indicating a research-focused origin.
• Includes multiple dataset configurations, such as 'metaworld' and 'pusht', suggesting variety.

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 and file formats are unknown, which may limit suitability assessment.

Provenance

Source

Meta AI Research, FAIR (facebook)

Collection Method

Likely collected from robotic simulations or real-world robotic platforms.

Time Range

null

Freshness

Last updated 2026-01-02 16:30:10; freshness should be verified.

Geography

null