search

zsyzzsoft / co-mod-gan

[ICLR 2021, Spotlight] Large Scale Image Completion via Co-Modulated Generative Adversarial Networks
Other
444 stars 67 forks source link

How to make custom pix2pix dataset and training #24

Closed SwordBearFire closed 3 years ago

SwordBearFire commented 3 years ago

Hi,

Thanks for your impressive code.

I meet a problem about pix2pix when I try to use co-mod-gan to train my dataset. Could you tell me how to make custom pix2pix dataset that can be used in the program. Thank you very much.

sincerely yours Bell

zsyzzsoft commented 3 years ago

Maybe I will publish the pix2pix code in the next week. Basically you can make each image contain 6 channels, first three channels of the output image followed by three channels of the condition image.

SwordBearFire commented 3 years ago

Maybe I will publish the pix2pix code in the next week. Basically you can make each image contain 6 channels, first three channels of the output image followed by three channels of the condition image.

Got it! thank you so much, I have fixed the code, and is training now, thank you again!