# CSAILVision/gandissect

Pytorch-based tools for visualizing and understanding the neurons of a GAN.  https://gandissect.csail.mit.edu/

Repository: https://github.com/CSAILVision/gandissect
Canonical: https://ross.abutalabs.com/products/gandissect
Language: Python
License: MIT
License Family: permissive
Topics: pytorch, gan, image-manipulation, deep-learning, interactive-visualizations, generative-adversarial-network, interpretable-ml
Last push: 2021-05-23T10:17:34+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": 2844, "days_push": 1928, "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 1765, forks 276 (observed 2026-08-28T04:05:33.452584+00:00)

## What it is
GANDissect is a PyTorch-based toolkit for visualizing and understanding the internal neurons of generative adversarial networks, showing how internal units align with human-interpretable concepts. It produces static dissection summaries and interactive visualizations like the GANPaint demo for editing generated images.

## Use cases
- visualize what individual GAN neurons represent
- understand how a GAN encodes interpretable concepts
- diagnose artifacts in a trained GAN
- remove or insert objects in GAN-generated images
- create interactive visualizations of GAN internals
- research interpretability of generative models

## When to choose
- you need to interpret or debug the internal representations of a PyTorch GAN
- you are doing research on interpretable machine learning for generative models
- you want to reproduce the GAN Dissection paper's analysis or GANPaint-style editing

## When to avoid
- you need a maintained tool for modern GAN or diffusion architectures - the project targets PyTorch 4.1 and has not been updated since 2021
- you want production image editing rather than model analysis
- you lack a CUDA-enabled GPU, which the dissection workflow requires

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, data-visualization, image-processing
- domain: deep-learning, computer-vision, artificial-intelligence, data-visualization
- platform: python
- tags: gan, interpretable-ml, pytorch, model-interpretability, generative-adversarial-network, research, gpu, linux

## Member repositories
- CSAILVision/gandissect (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:33.452584+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 2026-08-30T03:26:36.051431+00:00, confidence not recorded.
  - readme: https://github.com/CSAILVision/gandissect (fetched 2026-08-28T04:05:33.452584+00:00, sha 7276a3674a52)
- Data as of 2026-08-30T08:39:29.467469+00:00.