Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/LLNL/graphite

A repository for implementing graph network models based on atomic structures.
https://github.com/LLNL/graphite

Last synced: about 2 months ago
JSON representation

A repository for implementing graph network models based on atomic structures.

Awesome Lists containing this project

README 

        
# graphite



A repository for implementing graph network models based on atomic structures.



## Implemented or replicated works



- ALIGNN

    - [*Atomistic Line Graph Neural Network for improved materials property predictions*][ALIGNN paper]

- **ALIGNN-d** (our work, see [demo](notebooks/alignn/demo.ipynb))

    - [*Efficient and interpretable graph network representation for angle-dependent properties applied to optical spectroscopy*][ALIGNN-d paper]

- Gated GCN

    - [*Benchmarking Graph Neural Networks*][Gated GCN paper]

- NequIP (implementation adopted from `e3nn`; requires `e3nn` installation)

    - [*E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials*][NequIP paper]

- **Atomic Structure denoiser** (our work, see [demo](notebooks/denoiser/demo.ipynb))

    - [*Score-based denoising for atomic structure identification*][Denoiser paper]

- MeshGraphNets

    - [*Learning Mesh-Based Simulation with Graph Networks*][MGN paper]

- **Score dynamics** (our work, see [demo](notebooks/score-dynamics/demo.ipynb))

    - [*Score dynamics: scaling molecular dynamics with picoseconds time steps via conditional diffusion model*][SD paper]

- **Spectroscopy-guided generation of amorphous structures** (our work, see [demo](notebooks/amorph-gen/amorph-gen.ipynb))

    - [*Spectroscopy-guided discovery of three-dimensional structures of disordered materials with diffusion models*][a-C paper]

- Equivariant transformer

    - [*Equivariant pretrained transformer for unified geometric learning on multi-domain 3D molecules*][ET paper]

- Graphormer

    - [*Towards predicting equilibrium distributions for molecular systems with deep learning*][Graphormer paper]



## Gallery



### Spectroscopy-guided amorphous material generation



Unconditional generation of amorphous carbons via diffusion model. Color is meant to help give you a sense of depth.



![](/media/amorph-gen/a-C-denoise-traj-d15.gif)



Conditional generation of amorphous carbons based on a given XANES spectrum.



![](/media/amorph-gen/a-C_generation-with_plot.gif)



Generation of multi-element, amorphous carbon nitrides.



![](/media/amorph-gen/a-C-N-denoise.gif)



### Score dynamics



Generative rollouts of molecular dynamics over picosecond timesteps via conditional diffusion model.



![](/media/score-dynamics/ala-dipep.gif)



### Atomic structure denoiser



Simple and effective atomic denoiser for structure characterization.








## Installation



The installation time is typically less than 10 minutes on a normal local machine.



Installation dependencies:

- `pytorch>=2.0.1`

- `torch_geometric`

- `torch-scatter`

- `torch-cluster`



Reasons: many model implementations in this repo are written mostly based on PyTorch, but some operations such as `scatter` and graph pooling are from PyTorch Geometric (PyG), which offers very optimized CUDA implementations. Additionally, most models in this repo treat atomic/molecular data as graphs following the PyG data format (`torch_geometric.data.Data`).



Also, this repo has scattering (via `Torch.scatter_reduce`) and clustering (e.g., `radius_graph` and `knn_graph`) codes such that `torch-scatter` and `torch-cluster` are not strictly required. Still, it is recommended to install `torch-scatter` and `torch-cluster` if you favor CUDA-optimized compute speed.



Lastly, for development purposes, you may want to install packages such as `ase`, `MDAnalysis`, `rdkit`, `lightning`, etc.



Example installation process:

```bash

conda create -n graphite

conda activate graphite

conda install pytorch pytorch-cuda=12.1 -c pytorch -c nvidia

pip install torch_geometric

pip install torch_scatter torch_cluster -f https://data.pyg.org/whl/torch-2.3.0+cu121.html



# For development depending on your use case

pip install ase jupyterlab ipywidgets seaborn lightning tensorboard

```



Then, to install `graphite`, clone this repo and run:

```bash

pip install -e /path/to/the/repo

```



To uninstall:

```bash

pip uninstall graphite

```



## How to use



`graphite` is intended for research and prototyping. It is a general collection of simple codes containing helper functions, custom graph convolutions, model templates, and so on. Full-scale production codes for specific applications and deployments should be hosted elsewhere.



- The `src` folder contains the source code.

- The `notebooks` folder contains Jupyter notebooks that demonstrate running or training models.

    - Some demos require additional packages (e.g., PyTorch Lightning for automated training).

    - The demo notebooks are not always up-to-date. We will try to update the notebooks after every major change to the source code.

- The `media` folder contains media files.



## Release



LLNL-CODE-836648



[ALIGNN paper]: https://www.nature.com/articles/s41524-021-00650-1

[ALIGNN-d paper]: https://www.nature.com/articles/s41524-022-00841-4

[Gated GCN paper]: https://arxiv.org/abs/2003.00982

[e3nn basic conv doc]: https://docs.e3nn.org/en/stable/guide/convolution.html

[NequIP paper]: https://www.nature.com/articles/s41467-022-29939-5

[e3nn transformer doc]: https://docs.e3nn.org/en/stable/guide/transformer.html

[Denoiser paper]: https://arxiv.org/abs/2212.02421

[MGN paper]: https://arxiv.org/abs/2010.03409v4

[ET paper]: https://arxiv.org/abs/2402.12714

[Graphormer paper]: https://arxiv.org/abs/2306.05445

[SD paper]: https://pubs.acs.org/doi/10.1021/acs.jctc.3c01361

[a-C paper]: https://arxiv.org/abs/2312.05472