shaoanlu/faceswap-GAN

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

3416 stars · 831 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

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

library · maturity maintenance

machine-learning deep-learning image-processing computer-vision deep-learning computer-vision image-processing artificial-intelligence python cross-platform face-swap gan deepfakes autoencoder keras jupyter-notebook adversarial-loss self-attention gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="shaoanlu/faceswap-GAN")