Implementation of DCGAN using PyTorch on CelebA Dataset

This project is an implementation of Deep Convolutional Generative Adversarial Network (DCGAN) using PyTorch on the CelebA dataset. DCGAN is a neural network architecture that is used for generating realistic images. It is based on the Generative Adversarial Network (GAN) architecture, but uses convolutional layers instead of fully connected layers for both the generator and discriminator networks.

Dataset

The CelebA dataset consists of over 200,000 celebrity face images with annotations. In this project, we use all images for training the DCGAN model.

Requirements

Python 3.6 or later PyTorch 1.8.0 or later torchvision 0.9.0 or later NumPy Matplotlib

Credits

This implementation of DCGAN is based on the original paper by Radford et al., "Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks" (https://arxiv.org/abs/1511.06434). The CelebA dataset was created by Z. Liu, P. Luo, X. Wang, and X. Tang and can be downloaded from http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html.