Open in3omnia opened 5 years ago
Sorry for replying to your question so late. Actually, the non-makeup parameter means the latent vector that pre-trained Glow generate from your non-makeup images. In addition, you can use the Glow model pre-trained on CelebA dataset, which is provided by openai.
How to train the dataset with chainer-glow?
And I don't understand the input
non-makeup
andmakeup
ofmain.py
, is that the output latent vectors.npy
file generated by chainer-glow?Thanks~