/498L_Final_Project

DCGANs for image generation

Primary LanguageJupyter Notebook

498L_Final_Project

Generative Adversarial Networks and Their Applications

Getting the Data

Spirits

We have made the data available in this repo itself, so it can be used immediately when cloned into a Jupyter Notebook.

If you want to run the scraping script on your local machine, simply run python spirit_data.py to download the images into a new img/ folder.

CIFAR-10 and STL-10

We use Pytorch's builtin Dataset object: https://pytorch.org/docs/stable/torchvision/datasets.html#cifar
Setting download=True will automatically download the data to your specified root directory.

Notes: we experimented with Celeb-A and LSUN, but ran into issues with Colab crashing.