# gradslam/gradslam

gradslam is an open source differentiable dense SLAM library for PyTorch

Repository: https://github.com/gradslam/gradslam
Canonical: https://ross.abutalabs.com/products/gradslam
Homepage: https://gradslam.github.io/
Language: Python
License: MIT
License Family: permissive
Topics: slam, deep-learning, machine-learning, robotics, 3d-reconstruction
Last push: 2023-09-02T16:35:54+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2126, "days_push": 1096, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1422, forks 166 (observed 2026-08-28T04:04:41.152102+00:00)

## What it is
gradslam is a fully differentiable dense SLAM library built on PyTorch, providing differentiable building blocks such as nonlinear least squares solvers, ICP techniques, raycasting, and mapping/fusion modules. It enables end-to-end gradient flow from 3D maps and trajectories back to raw color/depth inputs for gradient-based learning in SLAM systems.

## Use cases
- build a differentiable dense SLAM system in PyTorch
- backpropagate from 3D maps to 2D pixels for learned SLAM representations
- run differentiable ICP for point cloud registration
- fuse RGB-D frames into differentiable point cloud maps
- learn camera intrinsics or depth via gradient-based SLAM
- prototype neural SLAM research with differentiable raycasting

## When to choose
- you need gradients to flow through SLAM components for deep learning research
- you work with RGB-D data and PyTorch
- you want differentiable ICP, raycasting, or TSDF-style fusion blocks

## When to avoid
- you need a production real-time SLAM system without learning components
- you don't use PyTorch
- you need active maintenance or support for recent PyTorch versions

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, simulation, graphics, computer-vision
- domain: robotics, deep-learning, machine-learning, simulation
- platform: python, cross-platform
- tags: slam, pytorch, differentiable-programming, 3d-reconstruction, icp, raycasting, rgbd, pointclouds, gpu

## Member repositories
- gradslam/gradslam (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:41.152102+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 2026-08-30T04:37:41.815365+00:00, confidence not recorded.
  - readme: https://github.com/gradslam/gradslam (fetched 2026-08-28T04:04:41.152102+00:00, sha 2672ac526a67)
  - homepage: https://gradslam.github.io/ (fetched 2026-08-29T11:49:46.980122+00:00, sha fa1802c62d0c)
- Data as of 2026-08-30T08:39:29.467469+00:00.