NVIDIA Releases Cross-Embodiment Foundation Model for Humanoid Robots
nvidia/gr00t17-lerobot-libero_object-640
NVIDIA has released the 'gr00t17-lerobot-libero_object-640' model, a cross-embodiment foundation model for humanoid robot reasoning and skills. It uses a Cosmos-Reason2/Qwen3-VL backbone and a flow-matching action transformer, trained with LeRobot and available under the Apache 2.0 license. The model supports training and deployment through LeRobot, with detailed instructions provided.
Full article
You are reading the complete RoboSignal summary. The publisher’s full article is available at the original source.
Read full article at sourcehuggingface.co · Opens in a new tab; source language may differ.
NVIDIA has released a cross-embodiment foundation model for humanoid robots, leveraging a Cosmos-Reason2/Qwen3-VL backbone and a flow-matching action transformer. The model is trained using LeRobot and is available under the Apache 2.0 license, with instructions for training and deployment provided in the documentation.
Key findings, evidence and limits
Published automatically after robotics and source-evidence checks; no manual editorial approval is recorded. Source assertions are not independently verified. Missing information remains not reported.
- Environment:
- Not reported
- Control:
- Not reported
- Data origin:
- Not reported
- dataset: Not reported
Source excerpts and review record
No manual editorial approval recorded.
Original source quotation: “Model Card for groot GR00T N1.7 is an open, cross-embodiment foundation model from NVIDIA for generalized humanoid robot reasoning and skills.”
Source E1
Original source quotation: “It uses a Cosmos-Reason2/Qwen3-VL backbone and a flow-matching action transformer to predict actions conditioned on vision, language, and proprioception.”
Source E2
Original source quotation: “This policy has been trained and pushed to the Hub using LeRobot .”
Source E3
Original source quotation: “Model Details License: apache-2.0 Inputs & Outputs The policy consumes these observation features and produces these action features.”
Source E4
Original source quotation: “Training steps 20000 Batch size 320 Optimizer adamw Learning rate 0.0001 Seed 42 LeRobot version 0.6.1”
Source E5
Original source quotation: “How to Get Started with the Model New to LeRobot? These guides cover the full workflow: Install LeRobot — set up the lerobot package.”
Source E6
Original source quotation: “The short version to run and train this policy: Run the policy on your robot lerobot-rollout \ --strategy.type=base \ --robot.type=<your_robot_type> \ --robot.port=<your_robot_port> \ --robot.cameras= "{ <camera_1>: {type: opencv, index_or_path: <index_or_path>, width: 640, height: 480, fps: 30}, <camera_2>: {type: opencv, index_or_path: <index_or_path>, width: 640, height: 480, fps: 30}}" \ --policy.path=nvidia/gr00t17-lerobot-libero_object-640 \ --task= "<your_task_description>" \ --duration=60”
Source E7
Original source quotation: “Replace the remaining <...> placeholders with your own values: --robot.port and the camera names/indices are specific to your machine, and the camera names must match the observation keys this policy was trained on.”
Source E8
Original source quotation: “When --strategy.type=base is used the script doesn't record the episodes. Skipping duration will make the policy run indefinitely.”
Source E9
Original source quotation: “For more information look at rollout documentation . Train your own policy lerobot-train \ --dataset.repo_id= ${HF_USER} /<dataset> \ --policy.type=groot \ --output_dir=outputs/train/<policy_repo_id> \ --job_name=lerobot_training \ --policy.device=cuda \ --policy.repo_id= ${HF_USER} /<policy_repo_id> \ --wandb.enable= true Writes checkpoints to outputs/train/<policy_repo_id>/checkpoints/ .”
Source E10
Original source quotation: “Evaluation No evaluation results have been provided for this policy yet.”
Source E11
Original source quotation: “Citation If you use this policy, please cite the method linked in the description above, along with LeRobot: @misc{cadene2024lerobot, author = {Cadene, Remi and Alibert, Simon and Soare, Alexander and Gallouedec, Quentin and Zouitine, Adil and Palma, Steven and Kooijmans, Pepijn and Aractingi, Michel and Shukor, Mustafa and Aubakirova, Dana and Russi, Martino and Capuano, Francesco and Pascal, Caroline and Choghari, Jade and Moss, Jess and Wolf, Thomas}, title = {LeRobot: State-of-the-art Machine Learning for Real-World Robotics in Pytorch}, howpublished = "\url{https://github.com/huggingface/lerobot}", year = {2024} }”
Source E12
Implications for data suppliers
RoboSignal interpretation and collection questions, not statements of buyer demand.
- Confirm the required data type and collection setting with the buyer; this source does not establish a complete collection specification.
- Validate demand and acceptance criteria with a buyer before scaling. Publication, popularity and a research result do not establish a purchase commitment.
Source：NVIDIA — Robotics model repository updates · huggingface.co