The Generative Adversarial Network implemented in this project was based on the original paper Generative Adversarial Networks by Goodfellow et al.
It was trained to learn a simple polynomial function
ReLU and Leaky ReLU with Dropout were used in the generator and discriminator respectively along with having one-sided label smoothing.
These optimisations were based on the recommendations from Improved Techniques for Training GANs and Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks.
To run this project
pip install -r requirements.txt
python GAN.py
