GNN and Continuous Convolutions based N-body simulations

This project aims to accelerate simulations of complex systems by integrating Graph Neural Networks (GNNs) and Continuous Convolutions with traditional numerical methods. It provides tools for generating initial conditions and running simulation experiments.

Project Overview

• Simulator & Initial Condition Generation:
  The core simulation environment, including initial condition generation, is implemented in the src/galaxify directory.

• Experiment Modules:
  Two main experiments are supported:

  • Continuous Convolution Experiment: Focuses on training and evaluating continuous convolutions models for simulation tasks.
  • GNN Experiment: Focuses on training and evaluating GNN models for simulation tasks.

  Results are stored in the results/ directory, separated into contconv and gnn subfolders, while pre-trained weights reside in contconv_weights and gnn_weights.

Requirements

• Python 3.10
• Key libraries include:
  • pandas
  • PyTorch and torch-geometric (with all related dependencies)
  • numpy
  • tqdm

Install all required packages with:

pip install -r requirements.txt