• In this project, the goal is to integrates CycleGAN with Progressive Growing GAN to achieve high-resolution image-to-image translation using unpaired datasets.
• By progressively increasing image resolution during training, this model enhances stability and image quality, beginning with low-resolution images and gradually adding layers as training advances.
• Potential tasks include translating between different seasons (e.g., summer to winter) or styles (e.g., artwork to photo realism).
• The model's performance will be compared against a standard CycleGAN, focusing on improvements in image quality and training stability.