Update README.md

README.md

@@ -1,5 +1,5 @@
 ---
-license: mit
+license: cc-by-4.0
 task_categories:
   - robotics
 tags:
@@ -17,44 +17,30 @@ size_categories:
 
 # G1 Locomanipulation Dataset v1
 
-Synthetic dataset of Unitree G1 humanoid robot demonstrations for loco-manipulation — picking up an object at one location, navigating to a second location, and placing it there. Generated using NVIDIA Isaac Lab's Synthetic Data Generation (SDG) pipeline.
+## Dataset Description:
+The G1 Locomanipulation Dataset v1 provides synthetic demonstrations of the Unitree G1 humanoid robot performing loco-manipulation tasks — picking up an object at one location, navigating around obstacles to a second location, and placing it there. Generated using NVIDIA Isaac Lab's Synthetic Data Generation (SDG) pipeline, the dataset extends teleoperated static manipulation recordings with automatically computed navigation segments using occupancy-map-based path planning and PI velocity control. This dataset is an example artifact for the Isaac Lab locomanipulation SDG pipeline — users who have run the data generation pipeline can reproduce it, but it is provided here so that step can be skipped when working through the pipeline examples. This dataset is for demonstration purposes and not for production usage.
 
-## Task Description
+## Dataset Owner(s):
+NVIDIA Corporation <br>
 
-The robot performs a **pick-and-place-with-navigation** task across five state machine phases:
+## Dataset Creation Date:
+04/2026 <br>
 
-1. **Grasp Object** — bimanual grasp of the target object at the pick-up table
-2. **Lift Object** — lift object while stationary
-3. **Navigate** — locomote with object held to approach the drop-off table (occupancy-map path planning + PI control)
-4. **Approach** — fine approach to goal position
-5. **Drop Off** — place object at the goal location
+## License/Terms of Use:
+This dataset is governed by the Creative Commons Attribution 4.0 International License (CC-BY-4.0). <br>
 
-The scene contains two packing tables (pick-up and drop-off) and configurable obstacles (forklifts, boxes) with randomized placement for diversity.
+## Intended Usage:
+This dataset is an example artifact for users working through the [Isaac Lab locomanipulation SDG pipeline](https://github.com/isaac-sim/IsaacLab/tree/main/scripts/imitation_learning). It is provided so users can skip the data generation step and proceed directly to GR00T N1.5 finetuning or rollout examples. Users who wish to generate their own dataset can do so by running the locomanipulation SDG pipeline with the `G1LocomanipulationSDGDataConfig` data config. This dataset is not intended for training production models or for deployment on physical robots. <br>
 
-## Robot
+## Dataset Characterization
+**Data Collection Method:** <br>
+* Synthetic — Generated via NVIDIA Isaac Lab Synthetic Data Generation (SDG) pipeline. The pipeline takes existing teleoperated static manipulation recordings (where the object does not move during collection) and automatically extends them with navigation by replaying manipulation motions while the base locomotes between fixtures. Navigation trajectories are computed via occupancy-map-based path planning with PI velocity control. Only successful episodes are exported. <br>
 
-**Unitree G1** humanoid — 29 degrees of freedom
+**Labeling Method:** <br>
+* Automatic/Sensors — Action and state labels are derived directly from simulation state at each timestep (200 Hz physics simulation). No human annotation was performed. <br>
 
-- Dual manipulation arms with 7-DOF finger joints per hand (14 finger DOF total)
-- Full-body locomotion (base velocity + height control)
-
-## Data Generation
-
-Generated with [`generate_dataset.py`](https://github.com/isaac-sim/IsaacLab/tree/main/scripts/imitation_learning/locomanipulation_sdg) in Isaac Lab. The pipeline:
-
-1. Takes existing **teleoperated static manipulation recordings** (object does not move during data collection)
-2. Automatically extends them with navigation by replaying the manipulation motions while the base locomotes between fixtures
-3. Navigation trajectories are computed via occupancy-map-based path planning with PI velocity control
-
-Only successful episodes are exported.
-
-**Simulation**: 200 Hz physics, ~50 second episodes
-
-## Dataset Format
-
-Stored in **HDF5**, compatible with [LeRobot](https://github.com/huggingface/lerobot) format for GR00T N1.5 training.
-
-### Modalities
+## Dataset Format:
+Video and numerical state/action arrays. Stored in HDF5 format, compatible with the [LeRobot](https://github.com/huggingface/lerobot) data format for GR00T N1.5 training. <br>
 
 | Key | Description | Shape |
 |-----|-------------|-------|
@@ -73,38 +59,20 @@ Stored in **HDF5**, compatible with [LeRobot](https://github.com/huggingface/ler
 | `action.base_velocity` | Base nav command (vx, vy, yaw_rate) | 3D |
 | `action.base_height` | Base height target | 1D |
 
-**Total action dimension**: 32D (28D manipulation + 4D locomotion)
-
-## Usage
-
-This dataset is intended for finetuning [GR00T N1.5](https://huggingface.co/nvidia/GR00T-N1.5-3B) using the `G1LocomanipulationSDGDataConfig` data config. See the [Isaac Lab imitation learning scripts](https://github.com/isaac-sim/IsaacLab/tree/main/scripts/imitation_learning) for conversion to LeRobot format and training instructions.
-
-Recommended GR00T finetuning settings:
-- `max_steps`: 10,000
-- `save_steps`: 1,000
-- Embodiment tag: `new_embodiment`
+**Total action dimension:** 32D (28D manipulation + 4D locomotion) <br>
 
-## Generation Details
+## Dataset Quantification:
+* Record Count: 100K–1M data points (timesteps at 200 Hz across all episodes; ~10,000 timesteps per ~50 s episode) <br>
+* Feature Count: 14 modalities per timestep (8 state inputs + 6 action outputs) <br>
+* Total Data Storage: ~478 MB (compressed) <br>
 
-| Parameter | Value |
-|-----------|-------|
-| Physics frequency | 200 Hz |
-| Episode length | 50 s |
-| Angular control gain | 2.0 |
-| Linear control gain | 1.0 |
-| Max base velocity | 1.0 m/s |
-| Path lookahead distance | 0.6 m |
-| Approach distance | 0.5 m |
+## Reference(s):
+- [Isaac Lab Locomanipulation SDG Pipeline](https://github.com/isaac-sim/IsaacLab/tree/main/scripts/imitation_learning)
 
-## Citation
+## Key Considerations:
+This dataset is provided solely as a demonstration artifact for the Isaac Lab locomanipulation SDG pipeline examples and is not intended for training production models or for use in physical robot deployment. It contains only synthetic simulation data and does not include personal data, biometric information, or copyrighted content. <br>
 
-If you use this dataset, please cite Isaac Lab:
+## 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.  When downloaded or used in accordance with our terms of service, 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. <br>
 
-```bibtex
-@misc{mittal2023orbit,
-  author = {Mittal, Mayank and Yu, Calvin and Yu, Qinxi and Liu, Jingzhou and Rudin, Nikita and Hoeller, David and Yuan, Jia Lin and Singh, Ritvik and Guo, Yunrong and Mazur, Hammad and Mandlekar, Ajay and Babich, Buck and State, Gavriel and Hutter, Marco and Garg, Animesh},
-  title = {Orbit: A Unified Simulation Framework for Interactive Robot Learning Environments},
-  year = {2023},
-  journal = {IEEE Robotics and Automation Letters},
-}
-```
+Please report quality, risk, security vulnerabilities or NVIDIA AI Concerns [here](https://www.nvidia.com/en-us/support/submit-security-vulnerability/). <br>