Thanks for the impressive idea and high-quality code.
I am trying to experiment on the compositional learning with CelebA dataset. However, I just found the path for CelebA is defined in this line, it seems to be an npy file. I wonder where can I find the corresponding files for training? Thanks a lot.
Thanks for the question! You will need to download the CelebA-HQ dataset. Instead of loading the CelebA numpy files, you can then directly read the images. Feel free to let me know if there are any issue.
Hi,
Thanks for the impressive idea and high-quality code.
I am trying to experiment on the compositional learning with CelebA dataset. However, I just found the path for CelebA is defined in this line, it seems to be an npy file. I wonder where can I find the corresponding files for training? Thanks a lot.