From the course: AI Workshop: Hands-on with GANs with Deep Convolutional Networks
Unlock this course with a free trial
Join today to access over 23,300 courses taught by industry experts.
Training data for discriminator good fakes and real images
From the course: AI Workshop: Hands-on with GANs with Deep Convolutional Networks
Training data for discriminator good fakes and real images
In this demo, we will train the discriminator as a classification model again, but we'll train with fake images that are much better than the fake images that we used earlier. I've called them good_fake_images. Here I am in the same Google Drive folder as before, My Drive, ai_workshop_dcgans, anime_classification. I have two sets of images here. I have the real images that we used in the earlier demo. You've already seen that, so I won't show you those again. And then I have some good_fake_images. Notice the new folder here called good_fake_images. Now if you click through to this good_fake_images folder, you will find that the fakes that have been generated here in this folder are much better than the previous fakes. As you can see, these are fakes from epoch 55 of training a GAN. The generator that created these was trained for a longer time and hence produced better fakes. And you will see this consistently across all of the images. The fakes are much better though in many cases…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Contents
-
-
-
-
Types of convolutional layers4m 11s
-
(Locked)
Training data for discriminator bad fakes and real images4m 34s
-
(Locked)
Loading and transforming training image data4m 55s
-
(Locked)
Understanding the discriminator architecture7m 8s
-
(Locked)
Training a discriminator on bad fakes5m 58s
-
(Locked)
Training data for discriminator good fakes and real images3m 8s
-
(Locked)
Training a discriminator on good fakes4m 56s
-
-
-