RLDS Format: What It Is and How nxted Delivers It
What is the RLDS dataset format?
RLDS (Reinforcement Learning Datasets) is a TensorFlow-Datasets format for storing episodic robot and RL data. It is the format behind the cross-embodiment Open X-Embodiment / RT-X collection, so it is the natural choice if you train on or alongside that data.
What an RLDS dataset contains
Episodes of per-timestep steps (observation, action, reward, discount) stored as sharded TFDS records, with dataset features and metadata.
- ›TFDS-sharded episodes and steps
- ›Observation / action / reward / discount per step
- ›Feature specs and dataset metadata
- ›Compatible with Open X-Embodiment tooling
When to choose RLDS
Choose RLDS if you train alongside Open X-Embodiment data or in a TFDS pipeline. See our format comparison for trade-offs.
How nxted delivers RLDS data
nxted Capture datasets can ship in RLDS, alongside LeRobot and HDF5, with metadata, a dataset card and a Data Trust Pack.
FAQ
What is RLDS?
A TensorFlow-Datasets format for episodic robot and RL data, used by the Open X-Embodiment / RT-X collection.
When should I use RLDS?
When you train on or alongside Open X-Embodiment data, or use a TFDS-based pipeline.
Does nxted deliver in RLDS?
Yes - RLDS, LeRobot and HDF5 are all available with metadata, a dataset card and a QA report.