Deep Learning with PyTorch

Inspecting the generated images

So, let's compare the generated images and the real images.

The generated images will be as follows:

The real images will be as follows:

Comparing both sets of images, we can see that our GAN was able to learn how to generate images. Apart from training to generate new images, we also have a discriminator, which can be used for classification problems. The discriminator learns important features about the images which can be used for classification tasks when there is a limited amount of labeled data available. ...