Table of Contents
In this project, three different GANs were implemented. Various components of the GAN network, including the Generator, Discriminator, and loss functions, were built from scratch. Additionally, several techniques were employed to improve the training process and enhance GAN stability. These techniques included One-sided label smoothing, batch normalization, and the addition of noise. The dataset used for training is the Finger Digits 0-5 dataset, accessible via this link.
The programming language, frameworks, and technologies used in the project are listed here:
- Python
- Tensorflow
Some useful links and tutorials about this project can be found here:
Mehrdad Nourbakhsh - mehrdad.nb4@gmail.com