# NVlabs/nvdiffrast

Nvdiffrast - Modular Primitives for High-Performance Differentiable Rendering

Repository: https://github.com/NVlabs/nvdiffrast
Canonical: https://ross.abutalabs.com/products/nvdiffrast
Homepage: https://nvlabs.github.io/nvdiffrast
Language: C++
License: NOASSERTION
License Family: other
Last push: 2025-12-08T13:25:14+00:00

## Health v2 (maintenance only)
Score: 55/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 56, release rhythm 28, longevity 100
- inputs: {"age_days": 2130, "days_push": 268, "days_rel": 268, "gap_med": null, "n_releases_24m": 1}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1905, forks 243 (observed 2026-08-28T04:05:51.903318+00:00)

## What it is
Nvdiffrast is a PyTorch library from NVIDIA providing high-performance, GPU-accelerated primitive operations for rasterization-based differentiable rendering, including rasterization, interpolation, texturing, and antialiasing. It is intentionally low-level, leaving camera models and material/lighting models to the user, and all operations and their gradients run on CUDA.

## Use cases
- differentiable rendering for 3D reconstruction from images
- optimize 3D mesh textures and geometry via gradient descent
- train neural networks that render 3D scenes in PyTorch
- inverse graphics and 3D-aware generative model training
- fast GPU rasterization with gradients for research pipelines

## When to avoid
- you need a high-level renderer with built-in camera, lighting, and material models
- you don't have an NVIDIA GPU or CUDA toolkit
- you want a pure-Python or CPU-only solution

## Facets
- artifact type: library
- maturity: stable
- function: graphics, machine-learning, image-processing
- domain: computer-vision, graphics, deep-learning, machine-learning
- platform: windows, python
- tags: differentiable-rendering, pytorch, cuda, rasterization, 3d-graphics, nvidia, inverse-graphics, gpu-accelerated, linux, gpu

## Member repositories
- NVlabs/nvdiffrast (main) score 55

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:51.903318+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:11:42.407061+00:00, confidence not recorded.
  - readme: https://github.com/NVlabs/nvdiffrast (fetched 2026-08-28T04:05:51.903318+00:00, sha 2c53380264d4)
  - homepage: https://nvlabs.github.io/nvdiffrast (fetched 2026-08-29T10:50:36.494533+00:00, sha f7ab0316efd6)
- Data as of 2026-08-30T08:39:29.467469+00:00.