DEff-GAN: Diverse Attribute Transfer for Few-Shot Image Synthesis
Generating diverse, high-quality images from minimal training data

Project Overview
This project implemented DEff-GAN, a progressive generative adversarial network (GAN) designed to synthesize realistic and diverse images using only a small number of input samples. Unlike conventional GANs that require large datasets, DEff-GAN enables few-shot and multi-class synthesis, addressing issues such as mode collapse, overfitting, and lack of diversity.