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.
Output from an untrained generator and discriminator
From the course: AI Workshop: Hands-on with GANs with Deep Convolutional Networks
Output from an untrained generator and discriminator
Now that we have our generator and discriminator set up, let's take a look at what a sample output from each of these modules looks like. Now z here is the random noise variable that I initialize to feed into the generator network. Observe that I've used a batch size of 64 for the latent noise variable. So we'll have 64 noise variables, each having a dimensionality of 100 that we'll use to generate 64 images. We then pass z in to our untrained generator model, and this will generate a batch of 64 images. Now remember that these 64 images are unlikely to be very good. In fact, they'll be terrible because the generator has not been trained at all. So here is what the tensor of 64 images looks like. Let's take a look at the shape of the generated output. And you can get this by invoking the .shape member variable. There are 64 images in this batch. That's the first dimension. Each image is a multi-channel, three-channel, 64 x 64 image. Let's view one of these generated images to see what…
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
-
-
-
-
-
Generator and discriminator1m 43s
-
(Locked)
Deep convolutional GANs (DCGANs)6m 16s
-
(Locked)
Setting up data for GAN training4m 8s
-
(Locked)
Setting up the generator and discriminator6m 18s
-
(Locked)
Output from an untrained generator and discriminator2m 17s
-
(Locked)
Setting up the GAN training loop7m 9s
-
(Locked)
Viewing GAN training results4m 34s
-
-