nvidia/GR00T-N1.7-AppleToPlate

Dataset Description:

The GR00T-N1.7-AppleToPlate dataset is a multimodal collection of trajectories collected on a Unitree G1 humanoid robot. It supports a humanoid (G1) static loco-manipulation task in which the robot picks up an apple and places it on a plate. Each entry provides the full context (state, vision, language, action) needed to train and evaluate generalist robot policies for an apple pick-and-place task.

Dataset Name	# Trajectories
G1 Static AppleToPlate Task	402

This dataset is ideal for behavior cloning, policy learning, and generalist robotic manipulation research. It has been used for post-training the GR00T N1.7 model.

This dataset is ready for commercial or non-commercial uses.

Dataset Owner

NVIDIA Corporation

Dataset Creation Date:

05/20/2026

Version:

v1.0
Previous Version(s): No previous version
Relationship to Previous Version(s): N/A

License/Terms of Use:

This dataset is governed by the Creative Commons Attribution 4.0 International License (CC-BY-4.0).

Intended Usage:

This dataset is intended for:

• Training robot manipulation policies using behavior cloning.
• Research in generalist robotics and task-conditioned agents.
• Real-world manipulation and imitation learning research.

Dataset Characterization:

Data Collection Method

• Manually-Collected

All demonstrations were manually collected through human teleoperation on a physical G1 robot using an XR headset. Each demo was recorded at 30 Hz.

Labeling Method

Manually-Labeled

Dataset Format:

The dataset is provided in LeRobot format, organized into the following directories:

• data/ contains the time-indexed trajectories as parquet files.
• videos/ contains the recorded RGB videos in mp4 format.
• meta/ contains the dataset metadata.

Each demo in GR00T-Lerobot datasets consists of a time-indexed sequence of the following modalities:

Actions

• action (FP32): joint desired positions for all body joints (43 DoF)

Observations

• observation.state (FP32): joint positions for all body joints (43 DoF)

Task-specific

• timestamp (FP32): time in seconds of each recorded data entry.
• episode_index (INT64): index indicating the order of each demo
• task_index (INT64): index used in multi-task data loader. Not applicable to GR00T-N1 post training, always set to 0.

Videos

• 640 x 480 RGB videos in mp4 format from an egocentric (ego-view) camera

In addition, a set of metadata describing the following is provided in the meta/ directory,

• episodes.jsonl contains a list of all the episodes in the entire dataset. Each episode contains a list of tasks and the length of the episode.
• tasks.jsonl contains a list of all the tasks in the entire dataset.
• modality.json contains the modality configuration.
• info.json contains the dataset information.

Dataset Quantification:

Record Count

G1 Static AppleToPlate Task

• Number of demonstrations/trajectories: 402
• Number of RGB videos: 402

Total Storage

1.02 GB

Ethical Considerations:

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. Developers should work with their internal developer teams to ensure this dataset meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

Please report quality, risk, security vulnerabilities or NVIDIA AI Concerns here.