In this project, a DCGAN (Deep Convolutional Generative Adversarial Network) is defined and trained on a dataset of faces. The goal is to get a generator network to generate new images of faces that look as realistic as possible!
-
Clone the repository and navigate to the downloaded folder.
-
Download the CelebFaces Attributes Dataset (CelebA). Unzip the folder and place it in the repo, at location path/to/face-generation/data processed_celeba_small.
-
Make sure you have already installed the necessary Python packages according to the requirements.txt in the repository.
-
Open a terminal window and navigate to the project folder. Open the notebook using:
jupyter notebook dlnd_face_generation.ipynb
NOTE: If the code is taking too long to run switch to running the code on a GPU.