The goal of the project was to examine whether incorporating attention mechanism in GAN architecture can improve image quality. Several models were tested for this task:
- Classic DCGAN
- DCGAN with attention mechanism
- Pretrained Progressive GAN
- Classic DCGAN
- "SWINGAN"