why we always use img_real to train generator weights in ccGAN？ about generative_adversarial_networks_101 HOT 1 CLOSED

Hello,

Firstly, I want to express my gratitude for your patience and diligence in pointing out this issue 3 years ago. Your insight was invaluable, and I sincerely appreciate your contribution to the project.

I am writing to inform you that the problem you highlighted regarding the input for training the generator weights has been successfully rectified. You're absolutely right in your understanding. The correct input, [masked_imgs, real], is now being utilized, enabling the generator to learn and generate new images effectively. I want to apologize for any confusion or inconvenience caused by the previous oversight.

d_g_loss_batch = d_g.train_on_batch(x=masked_imgs, y=real)

Thank you once again for your understanding and patience.

Best regards,

from generative_adversarial_networks_101.