Hands-On Deep Learning with R

Evaluating the model

Finally, we will evaluate the model using the following steps:

1. The last step is to periodically print model diagnostics along with real and generated images for comparison and to track whether the synthetic images are being generated as expected. We print the model iteration and error rates and also save one real and one fake image in the directory that we created earlier using the following code:

if (i %% 100 == 0) {          cat("step:", i, "\n")     cat("discriminator loss:", d_loss, "\n")     cat("adversarial loss:", a_loss, "\n")            image_array_save(       fake_images[1,,,] * 255,       path = file.path(image_directory, paste0("fake_gwb", i, ".png"))     )          image_array_save(       real_images[1,,,] * 255, path = file.path(image_directory, paste0("real_gwb", ...