Closed vgunjal closed 3 years ago
@vgunjal It needs a script update it seems. For now you can use pre-built docker image, I'll update as soon as I can.
Okay. Thank you, I'll do that.
If you would be so kind, can you please explain how the data is used here? Why do we generate multiple resolutions of the same image? How does that help in the training process?
You can refer to Appendix C of Progressive Growing of GANs paper. It is implemented in the following line: https://github.com/suvojit-0x55aa/celebA-HQ-dataset-download/blob/23e78a0d2affb5447784c4a620f3dc976a5267fe/create_celeba_HQ.py#L174
If you have any other doubt feel free to ask.
I get stuck at the same error whether I try the Docker approach or follow the instructions myself.
Am I doing something wrong? Or does the script need to be udpated?