Possible error of relativistic gan

[gregjohnso]

in https://github.com/eriklindernoren/PyTorch-GAN/blob/master/implementations/relativistic_gan/relativistic_gan.py

        if opt.rel_avg_gan:
            g_loss = adversarial_loss(fake_pred - real_pred.mean(0, keepdim=True), valid)
        else:
            g_loss = adversarial_loss(fake_pred - real_pred, valid)

        # Loss measures generator's ability to fool the discriminator
        g_loss = adversarial_loss(discriminator(gen_imgs), valid)

        g_loss.backward()
        optimizer_G.step()

Is this expected? Does it look like g_loss is getting overwritten?

[jbdel]

Up, error is still there.

I believe you can just discard g_loss = adversarial_loss(discriminator(gen_imgs), valid)

[phongnhhn92]

Why dont author merge this ? I think this is a serious error.