https://github.com/kushrm2803/mnist_gan

Generative Adversarial Network (GAN) to generate handwritten digits similar to those in the MNIST dataset
https://github.com/kushrm2803/mnist_gan

ann cnn deep-learning keras mnist tensorflow

Last synced: 15 days ago
JSON representation

Generative Adversarial Network (GAN) to generate handwritten digits similar to those in the MNIST dataset

• Host: GitHub
• URL: https://github.com/kushrm2803/mnist_gan
• Owner: kushrm2803
• Created: 2024-10-16T16:20:03.000Z (22 days ago)
• Default Branch: master
• Last Pushed: 2024-10-16T17:46:53.000Z (22 days ago)
• Last Synced: 2024-10-18T16:04:50.455Z (20 days ago)
• Topics: ann, cnn, deep-learning, keras, mnist, tensorflow
• Language: Jupyter Notebook
• Homepage:
• Size: 4.68 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# MNIST GAN (Generative Adversarial Network)

This project implements a Generative Adversarial Network (GAN) to generate handwritten digits similar to those in the MNIST dataset. The GAN comprises two main components: a generator that creates new images and a discriminator that evaluates their authenticity. The objective is for the generator to produce images that are indistinguishable from real handwritten digits.

## Dataset

- The project uses the MNIST dataset, which is built into TensorFlow.

## Steps

The process involves the following steps:

1. **Loading the MNIST dataset** and preprocessing the images.

2. **Defining the generator and discriminator models.**

3. **Utilizing a custom training loop** to train the GAN over a specified number of epochs.

4. During training, **the generator and discriminator compete against each other**, leading to improved performance over time.

5. **Generating digits** using the trained model.

## Results

Here is an example of an image generated after 500 epochs:

![Image Generated after 500 epochs](image500.png)

Increasing the number of epochs may yield more realistic digits.