All Tools
P
OtherFreeOpen Source
PADDLEGAN
GAN toolkit for motion transfer, restoration, and style on PaddlePaddle
Apache-2.0
ABOUT
GAN applications such as talking-head animation, face restoration, and style transfer live in separate paper repos. PaddleGAN puts First-Order, Wav2Lip, GPEN, and photo2cartoon models on PaddlePaddle with shared configs so CV teams train and run generative video and image pipelines from one toolkit.
INSTALL
pip install ppganINTEGRATION GUIDE
1. Animate a still portrait with First-Order motion transfer from a driving video
2. Lip-sync generated faces to speech with Wav2Lip
3. Restore old or low-quality photos with GPEN and inpainting models
4. Convert photos to cartoon or artistic styles with pretrained PaddleGAN configs
TAGS
pythoncomputer-visiongangenerativepaddlepaddlevideoimage-restoration