About ffhq dataset download

[Freedomcls]

Excuse me, I meet the problem in pre-data because I find the ffhq dataset is so big to download. Then I directly download the thumbnails128x128 from Google drive, but it also seems to need json files when it runs. I also download the json, but I dont know how to merge the thumbnails and the json. Could you give me some suggestions or tell me how to get the data correctly? Looking forward to your reply.

[povolann]

I managed to get the data like this: NVlabs/ffhq-dataset#3 (don't forget to change/download the new download_ffhq.py file so you can use the --pydrive and --cmd_auth flags) and then I changed the line
cmd = "python download_ffhq.py --wilds"
for
cmd = "python3 download_ffhq.py -i --pydrive --cmd_auth"
in runme.py file (depends on you if you want to keep the -i flag or --wilds flag).