LeRobot Driving School is a dataset for autonomous driving created using the LeRobot framework. It contains 10 episodes and 3,880 frames of data from a KIA Niro EV 2023 robot platform. The dataset was published by yaak-ai in 2025 and is licensed under Apache 2.0.
Use Cases
- Training next-action prediction models for autonomous vehicles based on multimodal sensor data.
- Benchmarking robotic control algorithms using real-world driving episode data.
- Developing and testing multimodal perception systems based on the described video and time-series modalities.
Strengths
- Includes 3,880 frames of data across 10 distinct driving episodes.
- Specifies the robot platform as a KIA Niro EV 2023, providing context for the data source.
- Provides links to a homepage, paper, and visualization tool for further context.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
- Source
- yaak-ai
- Collection Method
- Created using the LeRobot framework.
- Time Range
- null
- Freshness
- Last updated 2025-03-13 10:20:48.
- Geography
- null