search

genforce / idinvert

[ECCV 2020] In-Domain GAN Inversion for Real Image Editing
https://genforce.github.io/idinvert/
MIT License
461 stars 65 forks source link

Details about training the encoder #37

Open bingxumu opened 3 years ago

bingxumu commented 3 years ago

Thank you for your great work!! That'cool!! I would like to ask, do you mean to train the Encoder with the FFHQ dataset, and then use 7000 real face images for testing? May I share the 7000 real face images you collected？I would appreciate it very much if you could give more details about encoder training. Thanks very much!

ShenYujun commented 3 years ago

65,000 images from FFHQ are used for training, while the remaining 5,000 are used for testing. You can treat the last 5,000 as the test set or randomly select 5,000 images.