The *WikiArt* dataset is used which can be downloaded [here](https://github.com/cs-chan/ArtGAN/tree/master/WikiArt%20Dataset). It is expected to be at the location `data` by default, but the path can easily be changed at the top of `main.py` (for instance on Euler you might want to change that to your scratch directory). The first time the script is run the images will be transformed and saved in *npz* format at the same location. Our smaller variation of the dataset is already contained in the repository and can be pulled from the `data` folder in the branch `full-data`. Make sure you have at least 1GB of space available before you do so.
The *WikiArt* dataset is used which can be downloaded [here](https://github.com/cs-chan/ArtGAN/tree/master/WikiArt%20Dataset). It is expected to be at the location `data` by default, but the path can easily be changed at the top of `main.py` (for instance on Euler you might want to change that to your scratch directory). The first time the script is run the images will be transformed and saved in *npz* format at the same location. Our smaller variation of the dataset is already contained in the `data` folder and will be loaded automatically.
