ViPE: Video Pose Engine for Geometric 3D Perception

PyTorch implementation of JiT https://arxiv.org/abs/2511.13720

An open-source, GPU-accelerated physics simulation engine built upon NVIDIA Warp, specifically targeting roboticists and simulation researchers.

SpatialVID: A Large-Scale Video Dataset with Spatial Annotations

Official implementation of "Force Prompting: Video Generation Models Can Learn and Generalize Physics-based Control Signals" (NeurIPS 2025)

[ICCV 2025] PhysTwin: Physics-Informed Reconstruction and Simulation of Deformable Objects from Videos

MapAnything: Universal Feed-Forward Metric 3D Reconstruction

PyTorch code for hierarchical k-means -- a data curation method for self-supervised learning

Cosmos-Reason1 models understand the physical common sense and generate appropriate embodied decisions in natural language through long chain-of-thought reasoning processes.

ICCV 2025 | TesserAct: Learning 4D Embodied World Models

[ICLR 2025 Oral] Official code for "LVSM: A Large View Synthesis Model with Minimal 3D Inductive Bias"

Code for ICCV'2025 (Best student paper honorable mention) "RayZer: A Self-supervised Large View Synthesis Model"

Unified framework for robot learning built on NVIDIA Isaac Sim

PyTorch code and models for VJEPA2 self-supervised learning from video.

A collection of useful functions for 3D vision & graphics research in Python.

🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.

Lets make video diffusion practical!

[CVPR 2025 Best Paper Award] VGGT: Visual Geometry Grounded Transformer

Mobile manipulation research tools for roboticists

[CVPR 2025 Best Paper Nomination] FoundationStereo: Zero-Shot Stereo Matching

[ICCV 2023] MatrixCity: A Large-scale City Dataset for City-scale Neural Rendering and Beyond.

Python optical flow visualization following Baker et al. (ICCV 2007) as used by the MPI-Sintel challenge

Official implementation of Continuous 3D Perception Model with Persistent State

[CVPR 2025 Highlight] Align3R: Aligned Monocular Depth Estimation for Dynamic Videos

Official implementation of ICCV2023 VideoFlow: Exploiting Temporal Cues for Multi-frame Optical Flow Estimation

Official code for "FeatUp: A Model-Agnostic Frameworkfor Features at Any Resolution" ICLR 2024

[CVPR'25 Oral] MoGe: Unlocking Accurate Monocular Geometry Estimation for Open-Domain Images with Optimal Training Supervision

Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene Understanding