# fastgs/FastGS

[CVPR 2026 Highlight, the CVPR Compute Gold Star] Offical code for "FastGS: Training 3D Gaussian Splatting in 100 Seconds"

Repository: https://github.com/fastgs/FastGS
Canonical: https://ross.abutalabs.com/products/fastgs
Homepage: https://fastgs.github.io/
Language: Python
License: MIT
License Family: permissive
Topics: 3dgs, 3dgs-acceleration, acceleration, gaussian-splatting
Last push: 2026-03-23T14:19:38+00:00

## Health v2 (maintenance only)
Score: 49/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 73, release rhythm 35, longevity 21
- inputs: {"age_days": 300, "days_push": 163, "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 1221, forks 127 (observed 2026-08-28T04:04:02.050584+00:00)

## What it is
FastGS is a general acceleration framework for 3D Gaussian Splatting that trains scenes in roughly 100 seconds using multi-view consistent densification and targeted pruning. It integrates with multiple 3DGS backbones and supports tasks including dynamic scenes, surface reconstruction, sparse-view, large-scale reconstruction, and SLAM.

## Use cases
- accelerate 3d gaussian splatting training
- reconstruct 3d scenes from photos quickly
- speed up novel view synthesis training
- train 3dgs on limited gpu memory
- sparse-view 3d reconstruction
- surface reconstruction from multi-view images
- speed up dynamic scene reconstruction

## When to choose
- you need fast 3DGS training without sacrificing rendering quality
- you want to accelerate an existing 3DGS backbone like Mip-Splatting or Scaffold-GS
- you work on dynamic scenes, sparse-view, or SLAM tasks needing faster Gaussian splatting

## When to avoid
- you need non-Gaussian-splatting rendering methods like NeRF or mesh-based pipelines
- you require a turnkey GUI application rather than a research codebase
- your project depends on a license other than MIT

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, graphics, image-processing, gpu-computing
- domain: computer-vision, graphics, machine-learning, simulation
- platform: python, cross-platform
- tags: 3d-gaussian-splatting, 3dgs, neural-rendering, scene-reconstruction, training-acceleration, cvpr-2026, novel-view-synthesis, gpu, linux

## Member repositories
- fastgs/FastGS (main) score 49

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:02.050584+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 2026-08-30T06:15:41.172052+00:00, confidence not recorded.
  - readme: https://github.com/fastgs/FastGS (fetched 2026-08-28T04:04:02.050584+00:00, sha f9ba38481465)
  - homepage: https://fastgs.github.io/ (fetched 2026-08-29T12:24:19.779042+00:00, sha 8e0805af3375)
- Data as of 2026-08-30T08:39:29.467469+00:00.