https://github.com/shub-garg/fashionmnist-dcgan-generative-adversarial-networks-for-fashion-image-generation

This repository contains an implementation of a Deep Convolutional Generative Adversarial Network (DCGAN) trained on the FashionMNIST dataset. The project aims to generate realistic images of clothing items using a GAN architecture. It includes model definitions, training scripts, and visualizations of generated images at various training stages.
https://github.com/shub-garg/fashionmnist-dcgan-generative-adversarial-networks-for-fashion-image-generation

artificial-intelligence dcgan deeplearning fashionmnist gan generative-adversarial-network pytorch tensorflow

Last synced: 5 days ago
JSON representation

This repository contains an implementation of a Deep Convolutional Generative Adversarial Network (DCGAN) trained on the FashionMNIST dataset. The project aims to generate realistic images of clothing items using a GAN architecture. It includes model definitions, training scripts, and visualizations of generated images at various training stages.

• Host: GitHub
• URL: https://github.com/shub-garg/fashionmnist-dcgan-generative-adversarial-networks-for-fashion-image-generation
• Owner: shub-garg
• License: mit
• Created: 2024-05-29T00:28:27.000Z (9 months ago)
• Default Branch: main
• Last Pushed: 2024-05-31T20:26:31.000Z (9 months ago)
• Last Synced: 2024-06-01T21:15:45.421Z (8 months ago)
• Topics: artificial-intelligence, dcgan, deeplearning, fashionmnist, gan, generative-adversarial-network, pytorch, tensorflow
• Language: Jupyter Notebook
• Homepage:
• Size: 668 KB
• Stars: 4
• Watchers: 3
• Forks: 2
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# FashionMNIST DCGAN

## Project Overview

This project implements a Deep Convolutional Generative Adversarial Network (DCGAN) to generate realistic images of clothing items using the FashionMNIST dataset.

## Folder Structure

- `FashionMNIST_DCGAN.ipynb`: Contains the complete code and implementation for the DCGAN model, including data preprocessing, model architecture, training loop, and result visualization.

- `Images/`: Directory to store sample generated images during training.

## Setup Instructions

### Prerequisites

Ensure you have Python 3 installed. The project also requires several Python packages which can be installed using the following command:

```bash

pip install -r requirements.txt

```

## Running the Notebook

### Clone the Repository:

```bash

git clone https://github.com/yourusername/FashionMNIST_DCGAN.git

cd FashionMNIST_DCGAN

```

## Launch Jupyter Notebook:

```bash

jupyter notebook FashionMNIST_DCGAN.ipynb

```

## Results and Observations

### Discriminator and Generator Losses:

Discriminator losses gradually increase as it becomes harder to distinguish between real and fake images.

Generator losses decrease as it becomes better at generating realistic images.

## Image Quality Progression:

Epoch 10: Blurry and lacking detail.

Epoch 30: Improved clarity and texture.

Epoch 50: Significant enhancements with detailed and realistic images.

## Sample Generated Images



  



## Conclusion

The project successfully demonstrates the ability of a DCGAN to generate realistic images of clothing items from the FashionMNIST dataset. Further training and model refinement can yield even better results.