# BachiLi/redner

Differentiable rendering without approximation.

Repository: https://github.com/BachiLi/redner
Canonical: https://ross.abutalabs.com/products/redner
Homepage: https://people.csail.mit.edu/tzumao/diffrt/
Language: NASL
License: MIT
License Family: permissive
Topics: rendering, monte-carlo-ray-tracing, computer-vision, computer-graphics, differentiable-rendering, pytorch, tensorflow
Last push: 2022-08-19T13:09:03+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2853, "days_push": 1475, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1444, forks 145 (observed 2026-08-28T04:04:44.997724+00:00)

## What it is
redner is a differentiable Monte Carlo ray tracer that computes exact gradients of rendered images with respect to arbitrary scene parameters like geometry, materials, lighting, and camera pose. It integrates with PyTorch and TensorFlow, supports physically-based rendering with correct handling of discontinuities via edge sampling, and offers a fast deferred rendering mode for local shading.

## Use cases
- inverse rendering through gradient descent
- compute gradients of rendered images with respect to 3D scene parameters
- optimize scene geometry, materials, and lighting from images
- generate adversarial examples for neural networks
- physically-based rendering with shadows and global illumination gradients
- fast local shading with correct gradient estimation

## When to choose
- you need unbiased, non-approximated rendering gradients including discontinuities
- you want physically-based differentiable rendering with shadows and global illumination
- your pipeline uses PyTorch or TensorFlow and needs a differentiable renderer
- you are doing inverse rendering or 3D reconstruction from images

## When to avoid
- you need a general-purpose production renderer without gradient requirements
- you need active community support or frequent updates
- you need real-time rendering performance
- you need TensorFlow support on Windows

## Facets
- artifact type: library
- maturity: maintenance
- function: graphics, simulation, machine-learning, computer-vision
- domain: computer-vision, graphics, machine-learning, deep-learning
- platform: python, windows
- tags: differentiable-rendering, ray-tracing, inverse-rendering, monte-carlo, pytorch, tensorflow, cuda, linux, macos, gpu

## Member repositories
- BachiLi/redner (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:44.997724+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 2026-08-30T04:36:16.177535+00:00, confidence not recorded.
  - readme: https://github.com/BachiLi/redner (fetched 2026-08-28T04:04:44.997724+00:00, sha 5e09fb3a3b63)
  - homepage: https://people.csail.mit.edu/tzumao/diffrt/ (fetched 2026-08-29T11:46:12.768231+00:00, sha 783b1ea839b0)
- Data as of 2026-08-30T08:39:29.467469+00:00.