# ermongroup/SDEdit

PyTorch implementation for SDEdit: Image Synthesis and Editing with Stochastic Differential Equations

Repository: https://github.com/ermongroup/SDEdit
Canonical: https://ross.abutalabs.com/products/sdedit
Homepage: https://sde-image-editing.github.io/
Language: Python
License: MIT
License Family: permissive
Topics: pytorch, score-matching, image-editing, image-generation, controllable-generation, image-manipulation
Last push: 2023-02-12T07:40:43+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 1946, "days_push": 1298, "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 1168, forks 92 (observed 2026-08-28T04:03:50.874077+00:00)

## What it is
SDEdit is a PyTorch implementation of the ICLR 2022 paper 'Guided Image Synthesis and Editing with Stochastic Differential Equations'. It performs stroke-based image synthesis, editing, and compositing by adding noise to an input image and denoising it with a pre-trained score-based/diffusion model, without task-specific training.

## Use cases
- edit images with hand-drawn colored strokes
- generate realistic images from rough stroke paintings
- composite and manipulate photos using a diffusion prior
- run guided image synthesis with pretrained SDE models on CelebA-HQ and LSUN
- apply img2img-style noise-and-denoise editing to diffusion models
- reproduce the SDEdit paper results in PyTorch

## When to choose
- you want stroke-based or guide-based image editing without task-specific training
- you need a research reference implementation of SDE-based image editing
- you want to plug editing into off-the-shelf pretrained score-based or diffusion models

## When to avoid
- you need a production-ready image editor with a polished UI
- you want text-guided editing out of the box (use Stable Diffusion img2img instead)
- you lack a GPU or cannot download large pretrained checkpoints

## Facets
- artifact type: library
- maturity: maintenance
- function: image-processing, machine-learning, deep-learning
- domain: image-processing, machine-learning, computer-vision
- platform: python
- tags: diffusion-models, score-based-generative-models, image-editing, image-synthesis, pytorch, research-code, iclr-2022, gpu

## Member repositories
- ermongroup/SDEdit (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:50.874077+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 2026-08-30T06:28:46.765530+00:00, confidence not recorded.
  - readme: https://github.com/ermongroup/SDEdit (fetched 2026-08-28T04:03:50.874077+00:00, sha 5fdf9923b0c1)
  - homepage: https://sde-image-editing.github.io/ (fetched 2026-08-29T12:34:53.564041+00:00, sha 07bc0c305033)
- Data as of 2026-08-30T08:39:29.467469+00:00.