Escapist-007 / GAN

This 'Generative Adversarial Network' project was implemented in grad course CSE-676 : Deep Learning [Fall 2019 @UB_SUNY] Course Instructor : Sargur N. Srihari(https://cedar.buffalo.edu/~srihari/)

Generative Adversarial Network (GAN)

The project is about two types of GAN :

• Deep Convolutional Gan (DC-GAN)
• Self-Attention Gan (SA-GAN)

Project requirements:

• Implement DC-GAN using Binary Cross Entropy loss
• Apply Batch Normalization in DC-GAN
• Self-Attention module implementation for SA-GAN
• Implement wasserstein loss and apply Spectral Normalization in SA-GAN
• Apply Frechet Inception Distance (FID) as an evaluation metric for both DC-GAN and SA-GAN
• Use cifar-10 dataset

                                                                                      Code        Report