Skip to content

srm-mic/swapping-autoencoder-pytorch

 
 

Repository files navigation

Swapping Autoencoder Interface

An interactive interface for swapping autoencoder

license MIT

💡 Project Description

Our project builds on the paper Swapping Autoencoder for Deep Image Manipulation by Taesung Park, Jun-Yan Zhu, Oliver Wang, Jingwan Lu, Eli Shechtman, Alexei A. Efros, Richard Zhang. Our goal with this project was to make it easier for artists to use it as a tool. In that effort, we have introduced 3 interfaces to interact with a pre-trained model and edit images.

📺 Preview

Screenshot

Some images we genarated with the streamlit inerface

📌 Prerequisites

💻 System requirement :

  1. Nvidia GPU + CUDA.
  2. Operating System : Any (Windows / Linux / Mac).

💿 Software requirement :

  1. python 3.8
  2. poetry (Check out poetry here)

🔧 Installation

Step One - install python dependencies

$ poetry install

Step Two - Download pretrained models

Head over to the Testing and Evaluation section of the official implementation of the paper and download the pretrained models and unzip them, put the checkpoints at ./checkpoints/, you can change this location by specifying it at api/const.py:7

🏁 Quick Start

Streamlit Interface

$ streamlit run streamlit_interface.py

📦 Inside the box

Checkout our wiki for more details

📜 License

saxenabhishek/swapping-autoencoder-pytorch is available under the MIT license. See the LICENSE file for more info.

🤝 Contributing

Please read Contributing.md for details on our code of conduct, and the process for submitting pull requests to us.

⚙️ Maintainers

Abhishek Saxena
Abhishek Saxena

💥 Contributors

Contributors

About

Interface built to use the Swapping Autoencoder model.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 92.6%
  • Cuda 6.4%
  • C++ 1.0%