README.md

GANs Using Keras

Training and Prediction of Images using Keras on MNIST dataset.

Introduction

This repository demonstrates the implementation of Generative Adversarial Networks (GANs) using the Keras framework. GANs are a powerful deep learning architecture that consists of two neural networks, a generator and a discriminator, competing against each other. The generator aims to create realistic fake data, such as images, while the discriminator tries to differentiate between real and fake data. Through a process of adversarial training, GANs can generate new data that closely resembles real data, opening up exciting possibilities in image generation, style transfer, and more.

I have used the MNIST dataset for this project. You can use it by just adding a line in the code

 from keras.datasets import mnist

Content

The repository includes examples and tutorials for:

• Vanilla GAN implementation using Keras.
• Conditional GAN (CGAN) for conditional image generation.
• Deep Convolutional GAN (DCGAN) for high-quality image synthesis.
• CycleGAN for domain transfer and style conversion.

References

• Original GAN Paper
• Conditional GAN (CGAN) Paper
• DCGAN Paper
• CycleGAN Paper
• Domain Transfer and Style Conversion Paper

Acknowledgments

This project is inspired by the groundbreaking work on Generative Adversarial Networks (GANs). The references provided above offer in-depth insights into various GAN architectures and applications.

Contact

For questions or feedback, please feel free to reach out:

• Author: Kunal Tilaganji

Project Link: https://github.com/kunaltilaganji/GANs-Using-Keras