Emoji Generator

A powerful Generative Adversarial Network (GAN) based on the recommendations outlined in Alec Radford et al.'s influential 2015 paper titled “Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks”. The project’s primary objective was to generate emojis, drawing inspiration primarily from the iOS emojis developed by Apple.

Generator Architecture

I crafted a generator and discriminator for the DCGAN architecture to achieve this. By employing random sampling techniques and implementing architectural tweaks, I ensured the stable generation of emojis. Moreover, additional data processing techniques were employed for stabilizing GANs to enhance the performance and quality of the generated emojis.

Google Colaboratory and Tensorflow have been utilized in training and testing the DCGAN.