SRGAN Implementation using Pytorch

This repository contains an implementation of the Image Super-Resolution Using Deep Convolutional Networks paper by Christian Ledig, Lucas Theis, Ferenc Huszar, Jose Caballero, et al. The implementation is done using PyTorch and includes a Jupyter notebook demonstrating the training and evaluation of the GAN model for image super-resolution. As a result, here's the average PSNR and SSIM score for my implementation:

For more detail about how SRGAN works, you can check out my blog post series on dev.to here

Contents

• data/: Contains the images used for training and testing
• results/: Contains the result of test images when inferenced with the model
• SRGAN_train.ipynb: Jupyter notebook containing code for training and evaluating the SRCNN model.
• requirements.txt: File listing the required Python packages for running the code.

Usage

1. Clone the repository:

   git clone https://github.com/AdamAzuddin/SRGAN-PyTorch.git
   cd SRGAN-PyTorch

2. Install the required packages:

   pip install -r requirements.txt

3. Open and run SRGAN_train.ipynb in Jupyter Notebook to train the SRGAN model.
4. After training, use the trained model to perform image super-resolution on new images.

References

• Original Paper: Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network
• PyTorch: PyTorch Documentation