NVIDIA Research Projects ORG

Instant neural graphics primitives: lightning fast NeRF and more

NVIDIA research tool that turns a set of photos or 3D data into an interactive scene you can fly through, training a neural radiance field (NeRF) in under 5 seconds using a GPU-accelerated multiresolution hash encoding.

StyleGAN - Official TensorFlow Implementation

NVIDIA's official StyleGAN code: an AI model that generates photorealistic images of faces, cars, bedrooms, and cats that do not exist in real life.

StyleGAN2 - Official TensorFlow Implementation

NVIDIA Research's AI model that generates photorealistic images of faces, cars, and animals from scratch, and can blend the visual style of two generated images together.

SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformer

Semantic Image Synthesis with SPADE

Official PyTorch implementation of StyleGAN3

Official implementation of "Neuralangelo: High-Fidelity Neural Surface Reconstruction" (CVPR 2023)

Lightning fast C++/CUDA neural network framework

StyleGAN2-ADA - Official PyTorch implementation

Flickr-Faces-HQ Dataset (FFHQ)

NVIDIA's Deep Imagination Team's PyTorch Library

VILA is a family of state-of-the-art vision language models (VLMs) for diverse multimodal AI tasks across the edge, data center, and cloud.

Official PyTorch implementation of SegFormer

No description.

[CVPR 2024 Highlight] FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects

cuda-oxide is an experimental Rust-to-CUDA compiler that lets you write (SIMT) GPU kernels in safe(ish), idiomatic Rust. It compiles standard Rust code directly to PTX — no DSLs, no foreign language bindings, just Rust.

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

Multimodal Unsupervised Image-to-Image Translation

Eagle: Frontier Vision-Language Models with Data-Centric Strategies

Welcome to GR00T Whole-Body Control (WBC)! This is a unified platform for developing and deploying advanced humanoid controllers. This includes: Decoupled WBC models used in NVIDIA Isaac-Gr00t, Gr00t N1.5 and N1.6 and GEAR-SONIC

LongLive 2.0: Infra - Long Video Gen

Official code for the CVPR 2022 (oral) paper "Extracting Triangular 3D Models, Materials, and Lighting From Images".

[CVPR 2025] Official PyTorch Implementation of MambaVision: A Hybrid Mamba-Transformer Vision Backbone

Elucidating the Design Space of Diffusion-Based Generative Models (EDM)

No description.

Official repository for "AM-RADIO: Reduce All Domains Into One"

Nvdiffrast - Modular Primitives for High-Performance Differentiable Rendering

StyleGAN2 with adaptive discriminator augmentation (ADA) - Official TensorFlow implementation

Pytorch implementation for few-shot photorealistic video-to-video translation.

ProtoMotions is a GPU-accelerated simulation and learning framework for training physically simulated digital humans and humanoid robots.

InstantSplat: Sparse-view SfM-free Gaussian Splatting in Seconds

CUDA Accelerated Robot Library

Translate images to unseen domains in the test time with few example images.

Noise2Noise: Learning Image Restoration without Clean Data - Official TensorFlow implementation of the ICML 2018 paper

Sionna: An Open-Source Library for Research on Communication Systems

[ICCV 2025] Implementation for Describe Anything: Detailed Localized Image and Video Captioning

[CVPR 2023] BundleSDF: Neural 6-DoF Tracking and 3D Reconstruction of Unknown Objects

The implementation of "Prismer: A Vision-Language Model with Multi-Task Experts".

:couple: Joint Discriminative and Generative Learning for Person Re-identification. CVPR'19 (Oral) :couple:

[CVPR 2026] Fast-FoundationStereo: Real-Time Zero-Shot Stereo Matching

Official PyTorch implementation of VoxFormer [CVPR 2023 Highlight]

Deep Object Pose Estimation (DOPE) – ROS inference (CoRL 2018)

AlpaSim is an open-source autonomous vehicle simulation platform designed for development and testing of end-to-end AV policies

Official implementation of "Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding"

[ICML2024 (Oral)] Official PyTorch implementation of DoRA: Weight-Decomposed Low-Rank Adaptation

[CVPR 2025] Sparse Voxels Rasterization: Real-time High-fidelity Radiance Field Rendering

[ICLR 2024] Official PyTorch implementation of FasterViT: Fast Vision Transformers with Hierarchical Attention

[ICLR 2026 Oral] DiffusionNFT: Online Diffusion Reinforcement with Forward Process

EDM2 and Autoguidance -- Official PyTorch implementation

[CVPR 2026 Best Paper Finalist] Pixel Diffusion Transformers for Image Generation

Efficient Part-level 3D Object Generation via Dual Volume Packing

NVIDIA FastGen: Fast Generation from Diffusion Models

Cosmos Policy

Official PyTorch implementation of FB-BEV & FB-OCC - Forward-backward view transformation for vision-centric autonomous driving perception

[ICCV 2025] LongSplat: Robust Unposed 3D Gaussian Splatting for Casual Long Videos

Official PyTorch implementation of GroupViT: Semantic Segmentation Emerges from Text Supervision, CVPR 2022.

No description.

Tackling the Generative Learning Trilemma with Denoising Diffusion GANs https://arxiv.org/abs/2112.07804

HOVER

ToolOrchestra is an end-to-end RL training framework for orchestrating tools and agentic workflows.

No description.

Long-RL: Scaling RL to Long Sequences (NeurIPS 2025)

rCM & Causal-rCM: Leading and Unified Algorithms/Infrastructures for Bidirectional/Autoregressive Video Diffusion Distillation at Scale

Initial public release of code, data, and model weights for FourCastNet

SOMA: Unifying Parametric Human Body Models

No description.

PyTorch Implementation of EmerNeRF: Emergent Spatial-Temporal Scene Decomposition via Self-Supervision

A tool for visualizing and communicating the errors in rendered images.

Our first fully AI generated deep learning system

No description.

cuTile Rust provides a safe, tile-based kernel programming DSL for the Rust programming language. It features a safe host-side API for passing tensors to asynchronously executed kernel functions.

PlaneRCNN detects and reconstructs piece-wise planar surfaces from a single RGB image

No description.

[ICLR 2025] Official PyTorch Implementation of Gated Delta Networks: Improving Mamba2 with Delta Rule

This code corresponds to simulation environments used as part of the MimicGen project.

No description.

Official PyTorch implementation of Dreaming to Distill: Data-free Knowledge Transfer via DeepInversion (CVPR 2020)

[ECCV 2024] Official Repository for DiffiT: Diffusion Vision Transformers for Image Generation

[ICLR 2026]QeRL enables RL for 32B LLMs on a single H100 GPU.

Efficient 6-DoF Grasp Generation in Cluttered Scenes

Timeloop performs modeling, mapping and code-generation for tensor algebra workloads on various accelerator architectures.

Differentiable rasterization applied to 3D model simplification tasks

Official repo for GraspGen: A Diffusion-based Framework for 6-DOF Grasping

Official PyTorch implementation of Fully Attentional Networks

VLA-0: Building State-of-the-Art VLAs with Zero Modification

Official implementation of GDPO: Group reward-Decoupled Normalization Policy Optimization for Multi-reward RL Optimization

Code to accompany the Universal Deep Research paper (https://arxiv.org/abs/2509.00244)

Intrinsic3D - High-Quality 3D Reconstruction by Joint Appearance and Geometry Optimization with Spatially-Varying Lighting (ICCV 2017)

No description.

A unified interface to many trajectory forecasting datasets.

Verilog evaluation benchmark for large language model

PointWorld: Scaling 3D World Models for In-The-Wild Robotic Manipulation

No description.

Web Based Visualizer for Simulation Environments

Official Code for RVT-2 and RVT

No description.

A family of compressed models obtained via pruning and knowledge distillation

Flow Map OPD for AnyStep Video Diffusion

[ICLR 2022] Accelerated Policy Learning with Parallel Differentiable Simulation

[ICLR2025] A PyTorch implementation for STORM: Spatiotemporal Reconstruction Model for Large-Scale Outdoor Scenes

[CoRL 2025] Neural Robot Dynamics

Monocular whole-body 3D human pose estimation using the SOMA body model

Official Code Release of SAGE: Scalable Agentic 3D Scene Generation for Embodied AI

No description.

Devkit and documentation for the NVIDIA Physical AI Autonomous Vehicles Dataset

A digital data-generation pipeline that synthesizes humanoid loco-manipulation data from 3D assets and video priors.

GR00T-VisualSim2Real: Open-source sim-to-real framework for humanoid visual loco-manipulation. Train in simulation, deploy zero-shot on real robots with RGB + proprioception for tasks like pick-and-place and door opening.

No description.

No description.

SystemC/C++ library of commonly-used hardware functions and components for HLS.

No description.

AutoGaze automatically removes redundant patches in a video, reducing #tokens in ViT/MLLM by 4x-100x.

Adaptive FNO transformer - official Pytorch implementation

Parrot is an array fusion GPU library built on NVIDIA's CCCL libaries (Thrust/CUB).

Open-source RTL logic simulator with CUDA acceleration

Implementation of 6-DoF GraspNet with tensorflow and python. This repo has been tested with python 2.7 and tensorflow 1.12.

No description.

No description.

This code corresponds to simulation environments used as part of the DexMimicGen project.

STEP: Spatio-Temporal Progressive Learning for Video Action Detection. CVPR'19 (Oral)

[ICLR 2026] Official PyTorch Implementation of RLP: Reinforcement as a Pretraining Objective

Stochastic Tensor Optimization for Robot Motion - A GPU Robot Motion Toolkit

CGBN: CUDA Accelerated Multiple Precision Arithmetic (Big Num) using Cooperative Groups

SpatialClaw: Rethinking Action Interface for Agentic Spatial Reasoning

[NeurIPS'24] Large Spatial Model: End-to-end Unposed Images to Semantic 3D

No description.

Official PyTorch Implementation of Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention

Data Generation Pipeline for Mobility

A Python package that provides evaluation and visualization tools for the DexYCB dataset

Make your wildest 3D ConvNet dream architectures come true

[CVPR 2026] CARI4D: Category Agnostic 4D Reconstruction of Human-Object Interaction

Closed-Loop Supervised Fine-Tuning of Tokenized Traffic Models. CVPR Oral 2025.

Official PyTorch implementation of QUEEN: QUantized Efficient ENcoding of Dynamic Gaussians for Streaming Free-viewpoint Videos (NeurIPS 2024)

No description.

No description.

DEXTRAH

A Video Tokenizer Evaluation Dataset

A Rao-Blackwellized Particle Filter for 6D Object Pose Tracking

X-MOBILITY

No description.

Gradient Boosting Reinforcement Learning (GBRL)

Differentiable Cutting Simulator

Data-Driven Operational Space Control for Adaptive and Robust Robot Manipulation

No description.

Sionna RT: The Ray Tracing Package of Sionna

[ICML 2025 Spotlight] Direct Discriminative Optimization: Reinforcing Diffusion/Autoregressive with GAN Discrimination

[CVPR 2026 Highlight] VideoITG: Multimodal Video Understanding with Instructed Temporal Grounding

cuTAMP: Differentiable GPU-Parallelized Task and Motion Planning

Cross-embOdiment Mobility Policy via ResiduAl RL and Skill Synthesis

No description.

Official code release for CoRL'25 paper: VT-Refine: Learning Bimanual Assembly with Visuo-Tactile Feedback via Simulation Fine-Tuning

FDFO: Finite Difference Flow Optimization

Sionna Research Kit: A GPU-Accelerated Research Platform for AI-RAN

No description.

FABRICS

HANDAL Dataset and Pipeline

No description.

Implementation of Negative-aware Finetuning (NFT) algorithm for "Bridging Supervised Learning and Reinforcement Learning in Math Reasoning"

This code corresponds to transformer training and evaluation code used as part of the OPTIMUS project.

Speed of Light Analysis for ML Model Runtime

M2T2: Multi-Task Masked Transformer for Object-centric Pick and Plac

Code Release for GraspGenX

No description.

This repository contains code to reproduce some of the results from the paper Sionna RT: Differentiable Ray Tracing for Radio Propagation Modeling using the Sionna™ link-level simulator.

[CVPR 2025] Mosaic3D: Foundation Dataset and Model for Open-vocabulary 3D Segmentation

(3DV 2026 Oral) L4P -- a feed-forward foundational model designed for multiple low-level 4D vision perception tasks.

A Vision-Language-Model for Detecting and Reasoning Over Failures in Robotic Manipulation

No description.

Official PyTorch implementation of Data-free Knowledge Distillation for Object Detection, WACV 2021.

No description.

[CVPR 2025] HMAR: Efficient Hierarchical Masked Auto-Regressive Image Generation

A performance analysis tool for VLA models

No description.

Research package for MPC VR teleop with Isaac Sim and cuRobo

SPOT: SE(3) Pose Trajectory Diffusion for Object-Centric Manipulation

No description.

No description.

nvblox Torch

Official PyTorch implementation of BLADE: Single-view Body Mesh Estimation through Accurate Depth Estimation (CVPR 2025). BLADE tackles close-range human mesh recovery where perspective distortion is strongest, and solves for camera pose and focal length in addition to SMPL(-X) parameters.

No description.

[ICLRW'26] EoRA: Fine-tuning-free Compensation for Compressed LLM with Eigenspace Low-Rank Approximation

[CVPR 2025] ZeroMSF: Zero-shot Monocular Scene Flow Estimation in the Wild

No description.

AlpaGym is a reinforcement-learning framework for end-to-end autonomous-driving policies.

[CVPR 2026 (Highlight)] 4D-RGPT: Toward Region-level 4D Understanding via Perceptual Distillation

Temporal Planning with Samplers for GPU-Accelerated Multi-Arm Task and Motion Planning & Scheduling

No description.

No description.

A GPU-oriented coverage-guided fuzzer for userland CUDA applications

GPU accelerated utilities for data on HEALPix grids

SoftMimicGen

Sparse Decoupled Attention for Efficient Long-Context LLM Inference

The code is meant to accompany our paper Point Bridge, which focuses on sim-to-real transfer using 3D key point based representations.

Large-scale simulation of wireless network coverage maps using Sionna RT.

No description.

Code for https://www.arxiv.org/abs/2510.16281

SRSA: Skill Retrieval and Adaptation for Robotic Assembly Tasks

Decoder-only View Synthesis Model

No description.

cuHadron