# shaoanlu/faceswap-GAN

A denoising autoencoder + adversarial losses and attention mechanisms for face swapping.

Repository: https://github.com/shaoanlu/faceswap-GAN
Canonical: https://ross.abutalabs.com/products/faceswap-gan
Language: Jupyter Notebook
License Family: other
Topics: face-swap, generative-adversarial-network, gan, gans, image-manipulation
Last push: 2022-04-17T10:55:49+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": 3175, "days_push": 1599, "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 3416, forks 831 (observed 2026-08-28T04:08:03.740049+00:00)

## What it is
A Jupyter Notebook-based implementation of face swapping using a denoising autoencoder architecture enhanced with adversarial losses, VGGFace perceptual loss, and self-attention mechanisms. It provides notebooks for training models, preprocessing face data with MTCNN detection and alignment, and converting videos.

## Use cases
- swap faces in videos with a GAN model
- train a deepfake face-swapping model
- generate realistic face swaps with adversarial and perceptual losses
- detect and align faces in video for preprocessing
- train face swap models in Google Colab
- experiment with self-attention in GAN face generation

## When to choose
- you want a research-oriented GAN-based face swap implementation with perceptual and adversarial losses
- you prefer notebook-driven workflows for training and video conversion
- you want to experiment with self-attention (SAGAN-style) in face generation models
- you need Colab-based training without local GPU setup

## When to avoid
- you need a maintained, production-ready face swap tool with active support
- you want a simple CLI or GUI face swap application rather than notebooks
- you require modern TensorFlow/Keras versions, since the code depends on old pinned versions
- you need a permissively licensed project, as no license is specified

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, image-processing, computer-vision
- domain: deep-learning, computer-vision, image-processing, artificial-intelligence
- platform: python, cross-platform
- tags: face-swap, gan, deepfakes, autoencoder, keras, jupyter-notebook, adversarial-loss, self-attention, gpu

## Member repositories
- shaoanlu/faceswap-GAN (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:03.740049+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 2026-08-29T18:38:27.995511+00:00, confidence not recorded.
  - readme: https://github.com/shaoanlu/faceswap-GAN (fetched 2026-08-28T04:08:03.740049+00:00, sha 8d88d93a9a79)
- Data as of 2026-08-30T08:39:29.467469+00:00.