kaist-silab/graphsplinenets

[NeurIPS 23] Official Code for "Learning Efficient Surrogate Dynamic Models with Graph Spline Networks"

GraphSplineNets

Repository has work in progress - Arxiv paper, OpenReview, and final touches coming soon!

Description

GraphSplineNets code based on the Lightning Hydra Template

How to run

Clone repository

First, download the repository on Anonymous Github by running this on a terminal:

curl -sSL https://anonymous.4open.science/r/graphsplinenets/src/utils/download_anonymous_github.py | python3 -

or use the downloader script and run it with your favorite Python interpreter. Note that we use the above since Anonymous Github is currently not providing a way to download the repository as a zip file.

Install dependencies

# Automatically install dependencies with light the torch
pip install light-the-torch && python3 -m light_the_torch install --upgrade -r requirements.txt

The above script will automatically install PyTorch with the right GPU version for your system. Alternatively, you can use pip install -r requirements.txt

Quick Start

Run the example notebook:

python run.py experiment=example

Examples

Train model with chosen experiment configuration from configs/experiment/

Train model with default configuration

# train on CPU
python run.py trainer=cpu

# train on GPU
python run.py trainer=gpu

You can override any parameter from command line like this

python run.py trainer.max_epochs=20 datamodule.batch_size=64