# princeton-vl/RAFT-Stereo

Repository: https://github.com/princeton-vl/RAFT-Stereo
Canonical: https://ross.abutalabs.com/products/raft-stereo
Language: Python
License: MIT
License Family: permissive
Last push: 2026-08-03T19:53:10+00:00

## Health v2 (maintenance only)
Score: 75/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 95, release rhythm 35, longevity 100
- inputs: {"age_days": 1813, "days_push": 30, "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 1119, forks 186 (observed 2026-08-28T04:03:39.502027+00:00)

## What it is
RAFT-Stereo is a PyTorch implementation of a deep learning model for stereo matching that estimates disparity maps from stereo image pairs using multilevel recurrent field transforms. It includes training and evaluation code for standard stereo benchmarks like SceneFlow, Middlebury, ETH3D, and KITTI.

## Use cases
- estimate depth from stereo image pairs
- compute disparity maps for 3D reconstruction
- run stereo matching on KITTI or Middlebury datasets
- train a stereo depth network on SceneFlow
- generate point clouds from stereo images
- benchmark stereo matching models on ETH3D

## When to choose
- you need state-of-the-art stereo disparity estimation in Python
- you want a research-grade model with pretrained checkpoints and benchmark evaluation
- you have a CUDA GPU and stereo image pairs for depth estimation

## When to avoid
- you need monocular depth estimation from a single image
- you need real-time stereo on CPU or embedded hardware without a GPU
- you want a plug-and-play production library rather than research code

## Facets
- artifact type: library
- maturity: stable
- function: computer-vision, image-processing, machine-learning, deep-learning
- domain: computer-vision, machine-learning, robotics, autonomous-vehicles
- platform: python
- tags: stereo-matching, depth-estimation, disparity, pytorch, 3d-vision, research-code, gpu, linux

## Member repositories
- princeton-vl/RAFT-Stereo (main) score 75

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:39.502027+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 2026-08-30T06:41:25.540218+00:00, confidence not recorded.
  - readme: https://github.com/princeton-vl/RAFT-Stereo (fetched 2026-08-28T04:03:39.502027+00:00, sha 857211e1c463)
- Data as of 2026-08-30T08:39:29.467469+00:00.