# google-research/multinerf

A Code Release for Mip-NeRF 360, Ref-NeRF, and RawNeRF

Repository: https://github.com/google-research/multinerf
Canonical: https://ross.abutalabs.com/products/multinerf
Language: Python
License: Apache-2.0
License Family: permissive
Topics: nerf, neural-radiance-fields
Archived: true
Last push: 2023-12-08T20:14:39+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 1503, "days_push": 999, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3808, forks 354 (observed 2026-08-28T04:08:20.474868+00:00)

## What it is
Google Research's official code release for three NeRF papers: Mip-NeRF 360, Ref-NeRF, and RawNeRF, written in JAX. It trains neural radiance field models from images to enable novel view synthesis, reflection handling, and reconstruction from raw photos.

## Use cases
- train a NeRF model on my own photos to render novel views
- reproduce Mip-NeRF 360 paper results on unbounded scenes
- render 3D scenes with accurate reflections using Ref-NeRF
- reconstruct scenes from raw camera data with RawNeRF
- compute camera poses from images with COLMAP and train a radiance field
- generate fly-through videos from a trained NeRF checkpoint

## When to choose
- you need reference implementations of Mip-NeRF 360, Ref-NeRF, or RawNeRF for research
- you work in JAX and want GPU/TPU-accelerated NeRF training
- you want to reproduce published CVPR 2022 paper metrics exactly

## When to avoid
- you need a production-ready or user-friendly 3D capture tool
- you prefer PyTorch-based NeRF frameworks with active community support
- you need real-time rendering or interactive viewer features

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, graphics, simulation
- domain: computer-vision, graphics, machine-learning
- platform: python
- tags: nerf, neural-radiance-fields, jax, 3d-reconstruction, novel-view-synthesis, research-code, cvpr-2022, research, gpu, linux

## Member repositories
- google-research/multinerf (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:20.474868+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 2026-08-29T18:26:39.548470+00:00, confidence not recorded.
  - readme: https://github.com/google-research/multinerf (fetched 2026-08-28T04:08:20.474868+00:00, sha e70ebde8161c)
- Data as of 2026-08-30T08:39:29.467469+00:00.