# google/compare_gan

Compare GAN code.

Repository: https://github.com/google/compare_gan
Canonical: https://ross.abutalabs.com/products/compare_gan
Language: Python
License: Apache-2.0
License Family: permissive
Archived: true
Last push: 2021-01-31T18:25:24+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3116, "days_push": 2040, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1815, forks 309 (observed 2026-08-28T04:05:39.903803+00:00)

## What it is
A TensorFlow library from Google Research providing configurable implementations of GAN components: losses, penalties, normalization techniques, architectures (BigGAN, ResNet, DCGAN), and evaluation metrics (FID, Inception Score, precision-recall, KID). It was used to support several published GAN research papers.

## Use cases
- compare GAN losses and architectures in a large-scale study
- compute FID and Inception Score for generative models
- train BigGAN or DCGAN models in TensorFlow
- evaluate precision and recall of generative models
- reproduce GAN research paper experiments
- benchmark GAN training on GPU or TPU

## When to choose
- you need standardized, paper-backed GAN evaluation metrics like FID, KID, or precision-recall
- you want to reproduce or extend Google's GAN comparison studies
- you work in TensorFlow and need configurable GAN training components via Gin

## When to avoid
- you want to train GANs in PyTorch or JAX
- you need actively maintained code with recent updates
- you need production image generation rather than research tooling

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, benchmarking, gpu-computing
- domain: machine-learning, deep-learning
- platform: python
- tags: gan, tensorflow, generative-models, fid, inception-score, biggan, research-code, tpu, research, image-generation, gpu, linux

## Member repositories
- google/compare_gan (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:39.903803+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 2026-08-30T03:20:38.775611+00:00, confidence not recorded.
  - readme: https://github.com/google/compare_gan (fetched 2026-08-28T04:05:39.903803+00:00, sha 7d602c8754a6)
- Data as of 2026-08-30T08:39:29.467469+00:00.