# NVlabs/edm

Elucidating the Design Space of Diffusion-Based Generative Models (EDM)

Repository: https://github.com/NVlabs/edm
Canonical: https://ross.abutalabs.com/products/edm
Language: Python
License: NOASSERTION
License Family: other
Last push: 2024-03-16T09:02:18+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": 1423, "days_push": 900, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1991, forks 213 (observed 2026-08-28T04:06:03.105015+00:00)

## What it is
Official PyTorch implementation of the NeurIPS 2022 paper 'Elucidating the Design Space of Diffusion-Based Generative Models' (EDM) by NVIDIA Research. It provides training and sampling code for diffusion-based image generators, achieving state-of-the-art FID on CIFAR-10, FFHQ, AFHQ, and ImageNet-64 with fast sampling.

## Use cases
- generate images with a diffusion model
- reproduce state-of-the-art CIFAR-10 FID results
- train a diffusion model on my own dataset
- improve sampling efficiency of pre-trained score networks
- study design choices of diffusion generative models
- sample images from pre-trained FFHQ or ImageNet-64 models

## When to choose
- you need a research-grade diffusion model implementation with strong FID results
- you want fast deterministic sampling with few network evaluations
- you are experimenting with diffusion training, sampling, or score-network preconditioning
- you have NVIDIA GPUs and want to reproduce paper results

## When to avoid
- you need a production-ready image generation service or API
- you lack high-end NVIDIA GPUs, especially for training
- you need a maintained general-purpose diffusion library with broad model support
- you need permissive licensing for commercial use without checking NVIDIA's license

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, image-processing
- domain: deep-learning, machine-learning, artificial-intelligence, image-processing
- platform: windows, python
- tags: diffusion-models, generative-models, pytorch, research-code, image-generation, score-networks, neurips-2022, linux, gpu, docker

## Member repositories
- NVlabs/edm (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:03.105015+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 2026-08-30T03:02:40.398855+00:00, confidence not recorded.
  - readme: https://github.com/NVlabs/edm (fetched 2026-08-28T04:06:03.105015+00:00, sha 9e12015d1ab2)
- Data as of 2026-08-30T08:39:29.467469+00:00.