In this project, I used DCGAN mode for Face Generator using CelebFaces Attributes (CelebA) Dataset https://www.kaggle.com/datasets/jessicali9530/celeba-dataset.
A DCGAN consists of two neural networks:
Generator: This network takes a random noise vector as input and generates images that resemble the real training data. Its goal is to "fool" the discriminator into thinking its fake images are real.
Discriminator: This is the discriminative network. It’s a convolutional neural network (CNN) trained to classify inputs as either real (from the training data) or fake (generated by the generator). The discriminator serves as a binary classifier, distinguishing between actual images from the dataset and the synthetic ones created by the generator.
Output of DCGAN model after training 15 epochs:
