Skip to content

Fast, Expressive SE(n) Equivariant Networks through Weight-Sharing in Position-Orientation Space.

Notifications You must be signed in to change notification settings

ebekkers/ponita-torch

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

✨ 🐴 🔥 Official PONITA implementation for Pytorch with minimal library dependencies

See the original github repo for a PyTorch Geometric implementation. The original repo has more features than this one. The current repository is a minimal dependency implementation that currently only implements the fiber bundle method. Moreover, the dataloaders do not rely on PyTorch Geometric, but provide the same type of objects (graphs consisting of the tensors: x, pos, edge_index, batch).

Conda environment

In order to run the code in this repository install the following conda environment

conda create --yes --name ponita-torch python=3.12 numpy
conda activate ponita-torch
conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia -y
pip install tqdm
pip install rdkit
pip install pandas
pip3 install pytorch_lightning
pip3 install wandb

About

Fast, Expressive SE(n) Equivariant Networks through Weight-Sharing in Position-Orientation Space.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages