# harskish/ganspace

Discovering Interpretable GAN Controls [NeurIPS 2020]

Repository: https://github.com/harskish/ganspace
Canonical: https://ross.abutalabs.com/products/ganspace
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Topics: generative-adversarial-network, gan, deep-learning, interactive-visualizations, image-generation, pytorch
Last push: 2023-01-20T21:23:23+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2344, "days_push": 1321, "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 1794, forks 263 (observed 2026-08-28T04:05:37.376794+00:00)

## What it is
GANSpace is a research codebase implementing the NeurIPS 2020 paper on discovering interpretable controls for GANs via PCA in activation space. It includes modified BigGAN, StyleGAN, and StyleGAN2 models with interactive exploration and visualization tools for latent edit directions.

## Use cases
- discover interpretable latent directions in a GAN
- edit generated images by changing viewpoint, lighting, or age
- explore StyleGAN2 latent space interactively
- visualize principal components of a GAN's latent space
- apply layer-wise edits to BigGAN in a StyleGAN-like manner
- create videos of GAN component traversals

## When to choose
- you want to analyze or interpret controls of pretrained GANs like StyleGAN2 or BigGAN
- you need a research reference implementation of PCA-based latent space editing
- you want interactive exploration of GAN latent directions

## When to avoid
- you need production image generation rather than research tooling
- you want to train GANs from scratch
- you need actively maintained code with recent dependency support
- you work outside PyTorch or lack a GPU

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, image-processing, data-visualization
- domain: deep-learning, computer-vision, image-processing, artificial-intelligence
- platform: python
- tags: gan, stylegan, biggan, pca, latent-space-exploration, neurips-2020, pytorch, research-code, image-synthesis, linux, macos, gpu

## Member repositories
- harskish/ganspace (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:37.376794+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 2026-08-30T03:22:54.023234+00:00, confidence not recorded.
  - readme: https://github.com/harskish/ganspace (fetched 2026-08-28T04:05:37.376794+00:00, sha 406964a25d15)
- Data as of 2026-08-30T08:39:29.467469+00:00.