This repo contains code to reproduce all experiments in Generative Models as Distributions of Functions.
Requirements for training the models can be installed using pip install -r requirements.txt. All experiments were run using python 3.8.10.
To train a model on CelebAHQ64, run
python main.py configs/config_celebahq64.json
Example configs to reproduce the results in the paper are provided in the configs folder. Note that you will have to provide a path to the data you wish to train on in the config.
The shapenet and ERA5 climate datasets can be downloaded at this link. The CelebAHQ datasets can be downloaded from here.
All trained models can be downloaded from here. The load-trained-model.ipynb notebook shows an example of using a trained model.
The requirements in requirements.txt allow for basic plotting of 3D shapes with matplotlib. However, to properly render 3D models, you will need to install mcubes (for marching cubes) and pytorch3d. Pytorch3D is not directly pip installable (depending on your version of torch), so please follow the install instructions provided in their repo.
Once these requirements have been installed, you can render samples using
python render_mesh.py -sp samples -mr 32 -rs 64 -ns 8 -mp trained-models/shapenet_voxels/model.pt
See render_mesh.py for a full list of rendering options.
By default the samples from the climate data will be visualized as flat projections. Plotting these on a globe (using the viz/plots_globe.py functions) requires the cartopy library. Installation instructions for this library can be found here.
MIT