GANs & I2I Translation

Topics covered

This workshop provides at first some theoretical insights into the principles of GANs and image-to-image translation. The theoretical concepts are accompanied by practical (hands on) examples, covering:

• GAN fundamentals
• Deep Convolutional Generative Adversarial Networks (DCGANs)
• Conditional GANs (cGANs)
• CycleGANs (image-to-image translation)

In addition, it provides pratical tips and tricks for training GANs and discusses several potential caveats that might occur in practical application of GANs, e.g. mode collapse.

Course Materials

Please take note of the Prerequisites

Lecture Notes

Lecture Slides

Exercise Material

• Exercises
• Solutions

Prerequisites

Note that there is a certain amount previous knowledge required to fully take advantage of the course. In addition some software has to be installed prior to the course so time is not wasted during the actual exercises.

Required Previous Knowledge

• Python: This course uses Python during the exercises.
• Basics: The basics of deep learning should be known.

Required Software

The same software as for the deep learning basics workshop is required.

In addition to that please make sure that you have installed the following Python packages:

• PyTorch
• Torchvision
• Numpy
• Matplotlib