KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#developer-tools #typescript #dsh #dsh-plugin #ai-agent #codex #ai-agents #deepseek-harness #claude-code #cli #react #mcp

synthetic-manipulation-motion-generation

★ 229 Open GitHub ↗

Reference workflow for generating large amounts of synthetic motion trajectories for robot manipulation from a few human demonstrations.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# NVIDIA Omniverse Blueprint: Synthetic Manipulation Motion Generation for Robotics

The NVIDIA Isaac GR00T blueprint for synthetic manipulation motion generation is the ideal place to start. This is a reference workflow for creating exponentially large amounts of synthetic motion trajectories for robot manipulation from a small number of human demonstrations, built on [NVIDIA Omniverse™](https://developer.nvidia.com/isaac/sim) and [NVIDIA Cosmos™](https://www.nvidia.com/en-us/ai/cosmos/).

![image](https://github.com/user-attachments/assets/f3621fcc-91c3-4f4d-a516-c9c9c7f0d339)

# Deploy On Local Workstation

## Prerequisites
**Requirements for local deployment:**
* Ubuntu 22.04 Operating System
* NVIDIA GPU (RTX A 6000 with 48GB VRAM)
* [NVIDIA GPU Driver](https://www.nvidia.com/en-us/drivers/unix/) (recommended version 535.129.03)
* [Docker](https://docs.docker.com/engine/install/ubuntu/)
* [NVIDIA Container Toolkit](https://github.com/NVIDIA/nvidia-container-toolkit) (minimum version 1.17.0)

**Requirements for NVIDIA Cosmos:**
* NVIDIA GPU (H100 or higher) with 80GB of VRAM.
  * NVIDIA H100 GPU is available on AWS in a P5 EC2 instance, GCP in a A3 machine type VM and Azure in a ND H100 v5 series VM
* Visit the [Cosmos Hugging Face Model](https://huggingface.co/nvidia/Cosmos-Transfer1-7B) for specific details
>[!NOTE]
NVIDIA Cosmos must be run on a node separate from the Isaac Lab simulation due to differing hardware requirements.

## Launch a Jupyter Notebook

Steps:

1. Clone this repository to your local workstation and navigate to this repository.

git clone https://github.com/NVIDIA-Omniverse-blueprints/synthetic-manipulation-motion-generation.git
       cd synthetic-manipulation-motion-generation

2. Enable X11 forwarding for a local workstation user.

xhost +local:

3. Deploy the Jupyter Notebook with the Blueprint container.

docker compose -f docker-compose.yml up -d
       
4. Access the Jupyter Notebook from a browser at http://localhost:8888/lab/tree/generate_dataset.ipynb.

5. Follow the instructions inside of the Jupyter Notebook.

6. Run the command below to stop the Jupyter Notebook and end the demo.

docker compose -f docker-compose.yml down

>[!NOTE]
The Blueprint container includes a pre-installed version of Isaac Lab 2.0.2 and Isaac Sim 4.5.0.

# Licenses

By running the docker compose command, you accept the terms and conditions of all the licenses below:

- [Isaac Sim](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-software-license-agreement/)
- [Isaac Lab](https://github.com/isaac-sim/IsaacLab/blob/main/LICENSE)
- [Isaac Lab mimic](https://github.com/isaac-sim/IsaacLab/blob/main/LICENSE-mimic)
- [Cosmos NVIDIA Open Model License Agreement](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/)

synthetic-manipulation-motion-generation

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README