Developed by Oktay Uğurlu, Koray Kara, Ahmet Deniz Güner.
Firstly, you have to run main.py file to cluster the dataset. After, you can run style_based_cyclegan_vgg19.py file to train CycleGAN model over the clustered dataset. Also, you can reach the FID evaluation metric implementation from here fretched_instance_distance.py
You can reach the dataset below:
- https://people.eecs.berkeley.edu/~taesung_park/CycleGAN/datasets/
- https://www.kaggle.com/ikarus777/best-artworks-of-all-time
You can see our presentation here: https://www.youtube.com/watch?v=v5Lun1viqBc