the eyes seem to stay consistent with more training and bigger training set
5:46 PM · Mar 29, 2019
3
2
1
49
What kind of alignment and scaling strategy are you using? I figure that you can't use face markers for that?
1
0
0
2
I was planning to release this later but here is the kaggle kernel I did the whole thing in kaggle.com/ahsenk/stylegan-p… the only thing I changed in the dataset is resizing them and save as jpg
0
1
0
7
Why is the background both black and white? Is this an inconsistency in the original dataset?
1
0
0
1
