NVIDIA releases FoundationPose: 6DoF pose estimation model for robotics

nvidia · hf · 2026-09-15

NVIDIA has released foundationpose on Hugging Face, a foundation model for robotics pipelines supporting 6DoF object pose estimation and tracking. It ships with TensorRT and ONNX exports and integrates with the TAO Toolkit, tagged across computer vision, pose estimation, and robotics.

Related event: NVIDIA open-sources FoundationPose, a 6-DoF pose estimation foundation model(2 posts)→

Original post →

More from Embodied

Embodied channel →