| Intended Domain: |
Rheo (Isaac for Healthcare simulation workflow focused on preparing for surgical instruments handling tasks in an OR environment). |
| Model Type: |
Robot Vision Language Action (VLA) model |
| Intended Users: |
Isaac For Healthcare users testing operating room environments in simulation. |
| Output: |
Action tensor (next 16 actions) for pick-and-place of a sterilized tray from shelf to cart. |
| Describe how the model works: |
Accepts vision frames, language instruction, and robot observation; encodes multimodal inputs and outputs continuous robot action vectors. |
| Name the adversely impacted groups this has been tested to deliver comparable outcomes regardless of: |
Not Applicable |
| Technical Limitations & Mitigation: |
This model has been trained and tested with simulation data from the Isaac for Healthcare Rheo workflow and is not expected to generalize outside this environment. |
| Verified to have met prescribed NVIDIA quality standards: |
Yes |
| Performance Metrics: |
Latency, Task success rate |
| Potential Known Risks: |
This model may make unexpected movements if deployed in environments outside the Rheo simulation. The model may not fully adhere to protocols or manage unforeseen scenarios when operating outside the simulated environment. |
| Licensing: |
Governing Terms: Your usage of the GR00T-N1.6-Rheo-PickNPlaceTray model is governed by the NVIDIA License. |
