Skip to content

πŸ† A ranked list of awesome atomistic machine learning projects βš›οΈπŸ§¬πŸ’Ž.

License

Notifications You must be signed in to change notification settings

roshan2004/best-of-atomistic-machine-learning

Β 
Β 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

Best of Atomistic Machine Learning βš›οΈπŸ§¬πŸ’Ž

πŸ†Β  A ranked list of awesome atomistic machine learning (AML) projects. Updated quarterly.

This curated list contains 290 awesome open-source projects with a total of 120K stars grouped into 23 categories. All projects are ranked by a project-quality score, which is calculated based on various metrics automatically collected from GitHub and different package managers. If you like to add or update projects, feel free to open an issue, submit a pull request, or directly edit the projects.yaml.

The current focus of this list is more on simulation data rather than experimental data, and more on materials rather than drug design. Nevertheless, contributions from other fields are warmly welcome!

πŸ§™β€β™‚οΈ Discover other best-of lists or create your own.

Contents

Explanation

  • πŸ₯‡πŸ₯ˆπŸ₯‰Β  Combined project-quality score
  • ⭐️  Star count from GitHub
  • 🐣  New project (less than 6 months old)
  • πŸ’€Β  Inactive project (6 months no activity)
  • πŸ’€Β  Dead project (12 months no activity)
  • πŸ“ˆπŸ“‰Β  Project is trending up or down
  • βž•Β  Project was recently added
  • πŸ‘¨β€πŸ’»Β  Contributors count from GitHub
  • πŸ”€Β  Fork count from GitHub
  • πŸ“‹Β  Issue count from GitHub
  • ⏱️  Last update timestamp on package manager
  • πŸ“₯Β  Download count from package manager
  • πŸ“¦Β  Number of dependent projects

Active learning

Back to top

Projects that focus on enabling active learning, iterative learning schemes for atomistic ML.

FLARE (πŸ₯‡19 Β· ⭐ 250) - An open-source Python package for creating fast and accurate interatomic potentials. MIT C++ ML-IAP
  • GitHub (πŸ‘¨β€πŸ’» 36 Β· πŸ”€ 56 Β· πŸ“₯ 1 Β· πŸ“¦ 10 Β· πŸ“‹ 190 - 12% open Β· ⏱️ 26.05.2023):

     git clone https://github.com/mir-group/flare
    
Finetuna (πŸ₯ˆ11 Β· ⭐ 26 Β· πŸ’€) - Active Learning for Machine Learning Potentials. MIT
  • GitHub (πŸ‘¨β€πŸ’» 11 Β· πŸ”€ 7 Β· πŸ“‹ 18 - 22% open Β· ⏱️ 13.02.2023):

     git clone https://github.com/ulissigroup/finetuna
    
ACEHAL (πŸ₯‰6 Β· ⭐ 7) - Hyperactive Learning (HAL) Python interface for building Atomic Cluster Expansion potentials. Unlicensed Julia
  • GitHub (πŸ‘¨β€πŸ’» 3 Β· πŸ”€ 2 Β· πŸ“‹ 10 - 40% open Β· ⏱️ 21.09.2023):

     git clone https://github.com/ACEsuit/ACEHAL
    
Show 1 hidden projects...
  • flare++ (πŸ₯‰10 Β· ⭐ 34 Β· πŸ’€) - A many-body extension of the FLARE code. MIT C++ ML-IAP

Biomolecules

Back to top

Projects that focus on biomolecules, protein structure, protein folding, etc. using atomistic ML.

AlphaFold (πŸ₯‡23 Β· ⭐ 11K) - Open source code for AlphaFold. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 19 Β· πŸ”€ 1.9K Β· πŸ“¦ 6 Β· πŸ“‹ 740 - 23% open Β· ⏱️ 10.08.2023):

     git clone https://github.com/deepmind/alphafold
    
Uni-Fold (πŸ₯‰15 Β· ⭐ 290) - An open-source platform for developing protein models beyond AlphaFold. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 7 Β· πŸ”€ 51 Β· πŸ“₯ 2.3K Β· πŸ“‹ 58 - 18% open Β· ⏱️ 18.09.2023):

     git clone https://github.com/dptech-corp/Uni-Fold
    

Community resources

Back to top

Projects that collect atomistic ML resources or foster communication within community.

πŸ”—Β AI for Science Map - Interactive mindmap of the AI4Science research field, including atomistic machine learning, including papers,..

πŸ”—Β Atomic Cluster Expansion - Atomic Cluster Expansion (ACE) community homepage.

πŸ”—Β CrystaLLM - Generate a crystal structure from a composition. LM generative pre-trained

πŸ”—Β matsci.org - A community forum for the discussion of anything materials science, with a focus on computational materials science..

πŸ”—Β Matter Modeling Stack Exchange - Machine Learning - Forum StackExchange, site Matter Modeling, ML-tagged questions.

Best-of Machine Learning with Python (πŸ₯‡22 Β· ⭐ 14K) - A ranked list of awesome machine learning Python libraries. Updated weekly. CC-BY-4.0 general-ml Python
  • GitHub (πŸ‘¨β€πŸ’» 44 Β· πŸ”€ 2.1K Β· πŸ“‹ 48 - 31% open Β· ⏱️ 28.09.2023):

     git clone https://github.com/ml-tooling/best-of-ml-python
    
Graph-based Deep Learning Literature (πŸ₯ˆ19 Β· ⭐ 4.3K Β· πŸ“ˆ) - links to conference publications in graph-based deep learning. MIT general-ml rep-learn
  • GitHub (πŸ‘¨β€πŸ’» 12 Β· πŸ”€ 710 Β· ⏱️ 28.09.2023):

     git clone https://github.com/naganandy/graph-based-deep-learning-literature
    
MatBench (πŸ₯ˆ17 Β· ⭐ 83) - Matbench: Benchmarks for materials science property prediction. MIT datasets benchmarking
  • GitHub (πŸ‘¨β€πŸ’» 22 Β· πŸ”€ 31 Β· πŸ“¦ 10 Β· πŸ“‹ 57 - 54% open Β· ⏱️ 07.08.2023):

     git clone https://github.com/materialsproject/matbench
    
  • PyPi (πŸ“₯ 140 / month):

     pip install matbench
    
MatBench Discovery (πŸ₯‰15 Β· ⭐ 36) - An evaluation framework for machine learning models simulating high-throughput materials discovery. MIT datasets benchmarking
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 5 Β· πŸ“‹ 24 - 25% open Β· ⏱️ 28.09.2023):

     git clone https://github.com/janosh/matbench-discovery
    
  • PyPi (πŸ“₯ 83 / month):

     pip install matbench-discovery
    
AI for Science Resources (πŸ₯‰14 Β· ⭐ 250) - List of resources for AI4Science research, including learning resources. GPL-3.0 license
  • GitHub (πŸ‘¨β€πŸ’» 23 Β· πŸ”€ 30 Β· ⏱️ 31.08.2023):

     git clone https://github.com/divelab/AIRS
    
Awesome Materials Informatics (πŸ₯‰10 Β· ⭐ 300) - Curated list of known efforts in materials informatics = modern materials science. Custom
  • GitHub (πŸ‘¨β€πŸ’» 18 Β· πŸ”€ 72 Β· ⏱️ 21.08.2023):

     git clone https://github.com/tilde-lab/awesome-materials-informatics
    
The Collection of Database and Dataset Resources in Materials Science (πŸ₯‰7 Β· ⭐ 160) - A list of databases, datasets and books/handbooks where you can find materials properties for machine learning.. Unlicensed datasets
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 26 Β· ⏱️ 28.09.2023):

     git clone https://github.com/sedaoturak/data-resources-for-materials-science
    
Show 3 hidden projects...

Datasets

Back to top

Datasets, databases and trained models for atomistic ML.

πŸ”—Β Catalysis Hub - A web-platform for sharing data and software for computational catalysis research!.

πŸ”—Β Citrination Datasets - AI-Powered Materials Data Platform. Open Citrination has been decommissioned.

πŸ”—Β crystals.ai - Curated datasets for reproducible AI in materials science.

πŸ”—Β DeepChem Models - DeepChem models on HuggingFace. pre-trained LM

πŸ”—Β JARVIS-Leaderboard ( ⭐ 41) - This project provides benchmark-performances for materials science applications including Artificial Intelligence.. benchmarking

πŸ”—Β Materials Project - Charge Densities - Materials Project has started offering charge density information available for download via their public API.

πŸ”—Β matterverse.ai - Database of yet-to-be-sythesized materials predicted using state-of-the-art machine learning algorithms.

πŸ”—Β Quantum-Machine.org Datasets - Collection of datasets, including QM7, QM9, etc. MD, DFT. Small organic molecules, mostly.

πŸ”—Β sGDML Datasets - MD17, MD22, DFT datasets.

MPContribs (πŸ₯‡23 Β· ⭐ 34) - Platform for materials scientists to contribute and disseminate their materials data through Materials Project. MIT
  • GitHub (πŸ‘¨β€πŸ’» 25 Β· πŸ”€ 20 Β· πŸ“¦ 22 Β· πŸ“‹ 98 - 20% open Β· ⏱️ 25.09.2023):

     git clone https://github.com/materialsproject/MPContribs
    
  • PyPi (πŸ“₯ 2.6K / month):

     pip install mpcontribs-client
    
Open Catalyst datasets (πŸ₯‡18 Β· ⭐ 470) - The datasets of the Open Catalyst project, OC20, OC22. CC-BY-4.0
  • GitHub (πŸ‘¨β€πŸ’» 32 Β· πŸ”€ 170 Β· πŸ“‹ 130 - 11% open Β· ⏱️ 14.09.2023):

     git clone https://github.com/Open-Catalyst-Project/ocp
    
Open Databases Integration for Materials Design (OPTIMADE) (πŸ₯ˆ15 Β· ⭐ 61) - Specification of a common REST API for access to materials databases. CC-BY-4.0
  • GitHub (πŸ‘¨β€πŸ’» 19 Β· πŸ”€ 35 Β· πŸ“‹ 220 - 33% open Β· ⏱️ 22.06.2023):

     git clone https://github.com/Materials-Consortia/OPTIMADE
    
SPICE (πŸ₯ˆ13 Β· ⭐ 93) - A collection of QM data for training potential functions. MIT ML-IAP MD
  • GitHub (πŸ”€ 5 Β· πŸ“₯ 220 Β· πŸ“‹ 47 - 25% open Β· ⏱️ 18.08.2023):

     git clone https://github.com/openmm/spice-dataset
    
SciGlass (πŸ₯‰7 Β· ⭐ 6) - The database contains a vast set of data on the properties of glass materials. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 3 Β· ⏱️ 27.08.2023):

     git clone https://github.com/drcassar/SciGlass
    
3DSC Database (πŸ₯‰5 Β· ⭐ 7) - Repo for the paper publishing the superconductor database with 3D crystal structures. Custom superconductors materials-discovery
  • GitHub (πŸ”€ 2 Β· ⏱️ 21.07.2023):

     git clone https://github.com/aimat-lab/3DSC
    
Show 9 hidden projects...
  • OpenKIM (πŸ₯ˆ11 Β· ⭐ 29 Β· πŸ’€) - The Open Knowledgebase of Interatomic Models (OpenKIM) aims to be an online resource for standardized testing, long-.. LGPL-2.1 knowledge-base pre-trained
  • 2DMD dataset (πŸ₯ˆ9 Β· ⭐ 1) - Code for Kazeev, N., Al-Maeeni, A.R., Romanov, I. et al. Sparse representation for machine learning the properties of.. Apache-2 material-defect
  • ANI-1 Dataset (πŸ₯‰8 Β· ⭐ 89 Β· πŸ’€) - A data set of 20 million calculated off-equilibrium conformations for organic molecules. MIT
  • MoleculeNet Leaderboard (πŸ₯‰8 Β· ⭐ 73 Β· πŸ’€) - MIT benchmarking
  • ANI-1x Datasets (πŸ₯‰6 Β· ⭐ 45 Β· πŸ’€) - The ANI-1ccx and ANI-1x data sets, coupled-cluster and density functional theory properties for organic molecules. MIT
  • COMP6 Benchmark dataset (πŸ₯‰6 Β· ⭐ 36 Β· πŸ’€) - COMP6 Benchmark dataset for ML potentials. MIT
  • GEOM (πŸ₯‰5 Β· ⭐ 140 Β· πŸ’€) - GEOM: Energy-annotated molecular conformations. Unlicensed drug-discovery
  • linear-regression-benchmarks (πŸ₯‰5 Β· ⭐ 1 Β· πŸ’€) - Data sets used for linear regression benchmarks. MIT benchmarking single-paper
  • Visual Graph Datasets (πŸ₯‰3 Β· ⭐ 1 Β· 🐣) - Datasets for the training of graph neural networks (GNNs) and subsequent visualization of attributional explanations.. MIT

Data Structures

Back to top

Projects that focus on providing data structures used in atomistic machine learning.

dpdata (πŸ₯‡22 Β· ⭐ 140) - Manipulating multiple atomic simulation data formats, including DeePMD-kit, VASP, LAMMPS, ABACUS, etc. LGPL-3.0
  • GitHub (πŸ‘¨β€πŸ’» 51 Β· πŸ”€ 100 Β· πŸ“¦ 100 Β· πŸ“‹ 71 - 23% open Β· ⏱️ 08.09.2023):

     git clone https://github.com/deepmodeling/dpdata
    
  • PyPi (πŸ“₯ 6.4K / month):

     pip install dpdata
    
  • Conda (πŸ“₯ 120 Β· ⏱️ 27.09.2023):

     conda install -c deepmodeling dpdata
    
mp-pyrho (πŸ₯‰14 Β· ⭐ 28 Β· πŸ“‰) - Tools for re-griding volumetric quantum chemistry data for machine-learning purposes. Custom ML-DFT
  • GitHub (πŸ‘¨β€πŸ’» 8 Β· πŸ”€ 6 Β· πŸ“¦ 18 Β· πŸ“‹ 4 - 25% open Β· ⏱️ 21.03.2023):

     git clone https://github.com/materialsproject/pyrho
    
  • PyPi (πŸ“₯ 89 / month):

     pip install mp-pyrho
    
Equistore (πŸ₯‰14 Β· ⭐ 25) - Storage format for equivariant atomistic machine learning. BSD-3
  • GitHub (πŸ‘¨β€πŸ’» 17 Β· πŸ”€ 12 Β· πŸ“¦ 3 Β· πŸ“‹ 110 - 33% open Β· ⏱️ 27.09.2023):

     git clone https://github.com/lab-cosmo/equistore
    

Density functional theory (ML-DFT)

Back to top

Projects and models that focus on quantities of DFT, such as density functional approximations (ML-DFA), the charge density, density of states, the Hamiltonian, etc.

DM21 (πŸ₯‡20 Β· ⭐ 12K) - This package provides a PySCF interface to the DM21 (DeepMind 21) family of exchange-correlation functionals described.. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 92 Β· πŸ”€ 2.4K Β· πŸ“‹ 290 - 53% open Β· ⏱️ 02.06.2023):

     git clone https://github.com/deepmind/deepmind-research
    
MALA (πŸ₯‡19 Β· ⭐ 52 Β· πŸ“ˆ) - Materials Learning Algorithms. A framework for machine learning materials properties from first-principles data. BSD-3
  • GitHub (πŸ‘¨β€πŸ’» 41 Β· πŸ”€ 19 Β· πŸ“‹ 240 - 11% open Β· ⏱️ 28.09.2023):

     git clone https://github.com/mala-project/mala
    
DeepH-pack (πŸ₯ˆ12 Β· ⭐ 140) - Deep neural networks for density functional theory Hamiltonian. LGPL-3.0 Julia
  • GitHub (πŸ‘¨β€πŸ’» 6 Β· πŸ”€ 26 Β· πŸ“‹ 35 - 11% open Β· ⏱️ 11.07.2023):

     git clone https://github.com/mzjb/DeepH-pack
    
DeePKS-kit (πŸ₯ˆ10 Β· ⭐ 94) - a package for developing machine learning-based chemically accurate energy and density functional models. LGPL-3.0
  • GitHub (πŸ‘¨β€πŸ’» 6 Β· πŸ”€ 30 Β· πŸ“‹ 10 - 10% open Β· ⏱️ 01.04.2023):

     git clone https://github.com/deepmodeling/deepks-kit
    
SALTED (πŸ₯ˆ9 Β· ⭐ 13) - Symmetry-Adapted Learning of Three-dimensional Electron Densities. GPL-3.0
  • GitHub (πŸ‘¨β€πŸ’» 10 Β· πŸ”€ 2 Β· ⏱️ 25.09.2023):

     git clone https://github.com/andreagrisafi/SALTED
    
ACEhamiltonians (πŸ₯ˆ7 Β· ⭐ 8) - Provides tools for constructing, fitting, and predicting self-consistent Hamiltonian and overlap matrices in solid-.. MIT Julia
  • GitHub (πŸ‘¨β€πŸ’» 4 Β· πŸ”€ 3 Β· πŸ“‹ 4 - 25% open Β· ⏱️ 12.04.2023):

     git clone https://github.com/ACEsuit/ACEhamiltonians.jl
    
DeepDFT (πŸ₯‰6 Β· ⭐ 36 Β· πŸ’€) - Official implementation of DeepDFT model. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 6 Β· ⏱️ 28.02.2023):

     git clone https://github.com/peterbjorgensen/DeepDFT
    
DeepH-E3 (πŸ₯‰6 Β· ⭐ 29 Β· 🐣) - General framework for E(3)-equivariant neural network representation of density functional theory Hamiltonian. MIT magnetism
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 8 Β· πŸ“‹ 8 - 25% open Β· ⏱️ 04.04.2023):

     git clone https://github.com/Xiaoxun-Gong/DeepH-E3
    
Show 11 hidden projects...
  • NeuralXC (πŸ₯ˆ10 Β· ⭐ 28 Β· πŸ’€) - Implementation of a machine learned density functional. BSD-3
  • PROPhet (πŸ₯ˆ9 Β· ⭐ 60 Β· πŸ’€) - PROPhet is a code to integrate machine learning techniques with first-principles quantum chemistry approaches. GPL-3.0 ML-IAP MD single-paper C++
  • Libnxc (πŸ₯ˆ7 Β· ⭐ 14 Β· πŸ’€) - A library for using machine-learned exchange-correlation functionals for density-functional theory. MPL-2.0 C++ Fortran
  • gprep (πŸ₯‰4 Β· πŸ’€) - Fitting DFTB repulsive potentials with GPR. MIT single-paper
  • ML-DFT (πŸ₯‰3 Β· ⭐ 18 Β· πŸ’€) - A package for density functional approximation using machine learning. MIT
  • charge-density-models (πŸ₯‰3 Β· ⭐ 2) - Tools to build charge density models using ocpmodels. MIT
  • CSNN (πŸ₯‰3 Β· ⭐ 1 Β· πŸ’€) - Primary codebase of CSNN - Concentric Spherical Neural Network for 3D Representation Learning. BSD-3
  • MALADA (πŸ₯‰3) - MALA Data Acquisition: Helpful tools to build data for MALA. BSD-3
  • xDeepH (πŸ₯‰2 Β· ⭐ 18) - Extended DeepH (xDeepH) method for magnetic materials. LGPL-3.0 magnetism Julia
  • DeepCDP (πŸ₯‰2 Β· ⭐ 2) - DeepCDP: Deep learning Charge Density Prediction. Unlicensed
  • kdft (πŸ₯‰1 Β· ⭐ 2 Β· πŸ’€) - The Kernel Density Functional (KDF) code allows generating ML based DFT functionals. Unlicensed

Educational Resources

Back to top

Tutorials, guides, cookbooks, recipes, etc.

πŸ”—Β Quantum Chemistry in the Age of Machine Learning - Book, 2022.

Deep Learning for Molecules and Materials Book (πŸ₯‡12 Β· ⭐ 530) - Deep learning for molecules and materials book. Custom
  • GitHub (πŸ‘¨β€πŸ’» 19 Β· πŸ”€ 96 Β· πŸ“‹ 150 - 16% open Β· ⏱️ 02.07.2023):

     git clone https://github.com/whitead/dmol-book
    
jarvis-tools-notebooks (πŸ₯‡12 Β· ⭐ 38) - A Google-Colab Notebook Collection for Materials Design: https://jarvis.nist.gov/. NIST
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 21 Β· ⏱️ 19.08.2023):

     git clone https://github.com/JARVIS-Materials-Design/jarvis-tools-notebooks
    
RDKit Tutorials (πŸ₯ˆ8 Β· ⭐ 210) - Tutorials to learn how to work with the RDKit. Custom
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 65 Β· πŸ“‹ 4 - 75% open Β· ⏱️ 19.03.2023):

     git clone https://github.com/rdkit/rdkit-tutorials
    
iam-notebooks (πŸ₯ˆ8 Β· ⭐ 19) - Jupyter notebooks for the lectures of the Introduction to Atomistic Modeling. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 6 Β· πŸ”€ 4 Β· ⏱️ 07.08.2023):

     git clone https://github.com/ceriottm/iam-notebooks
    
AI4Science101 (πŸ₯‰5 Β· ⭐ 73 Β· πŸ’€) - AI for Science. Unlicensed
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 11 Β· ⏱️ 04.09.2022):

     git clone https://github.com/deepmodeling/AI4Science101
    
Data Handling, DoE and Statistical Analysis for Material Chemists (πŸ₯‰5 Β· 🐣) - Notebooks for workshops of DoE course, hosted by the Computational Materials Chemistry group at Uppsala University. GPL-3.0
  • GitHub (πŸ‘¨β€πŸ’» 3 Β· πŸ”€ 12 Β· ⏱️ 26.06.2023):

     git clone https://github.com/Teoroo-CMC/DoE_Course_Material
    
Show 11 hidden projects...
  • DeepLearningLifeSciences (πŸ₯ˆ11 Β· ⭐ 300 Β· πŸ’€) - Example code from the book Deep Learning for the Life Sciences. MIT
  • MAChINE (πŸ₯ˆ8 Β· ⭐ 1 Β· 🐣) - Client-Server Web App to introduce usage of ML in materials science to beginners. MIT
  • BestPractices (πŸ₯ˆ6 Β· ⭐ 140 Β· πŸ’€) - Things that you should (and should not) do in your Materials Informatics research. MIT
  • Applied AI for Materials (πŸ₯ˆ6 Β· ⭐ 47 Β· πŸ’€) - Course materials for Applied AI for Materials Science and Engineering. Unlicensed
  • COSMO Software Cookbook (πŸ₯ˆ6 Β· ⭐ 3 Β· 🐣) - The COSMO cookbook contains recipes for atomic-scale modelling for materials and molecules. BSD-3
  • Machine Learning for Materials Hard and Soft (πŸ₯‰5 Β· ⭐ 32 Β· πŸ’€) - ESI-DCAFM-TACO-VDSP Summer School on Machine Learning for Materials Hard and Soft. Unlicensed
  • ML-in-chemistry-101 (πŸ₯‰4 Β· ⭐ 54 Β· πŸ’€) - The course materials for Machine Learning in Chemistry 101. Unlicensed
  • chemrev-gpr (πŸ₯‰4 Β· ⭐ 5 Β· πŸ’€) - Notebooks accompanying the paper on GPR in materials and molecules in Chemical Reviews 2020. Unlicensed
  • MLDensity_tutorial (πŸ₯‰2 Β· ⭐ 6 Β· πŸ’€) - Tutorial files to work with ML for the charge density in molecules and solids. Unlicensed
  • MALA Tutorial (πŸ₯‰2 Β· ⭐ 2 Β· 🐣) - A full MALA hands-on tutorial. Unlicensed
  • PiNN Lab (πŸ₯‰1 Β· ⭐ 2) - GPL-3.0

Explainable Artificial intelligence (XAI)

Back to top

Projects that focus on explainability and model interpretability in atomistic ML.

exmol (πŸ₯‡19 Β· ⭐ 260) - Explainer for black box models that predict molecule properties. MIT
  • GitHub (πŸ‘¨β€πŸ’» 7 Β· πŸ”€ 39 Β· πŸ“¦ 12 Β· πŸ“‹ 67 - 14% open Β· ⏱️ 19.06.2023):

     git clone https://github.com/ur-whitelab/exmol
    
  • PyPi (πŸ“₯ 470 / month):

     pip install exmol
    
Show 3 hidden projects...

Electronic structure methods (ML-ESM)

Back to top

Projects and models that focus on quantities of electronic structure methods, which do not fit into either of the categories ML-WFT or ML-DFT.

Show 2 hidden projects...
  • QDF for molecule (πŸ₯‡9 Β· ⭐ 170 Β· πŸ’€) - Quantum deep field: data-driven wave function, electron density generation, and energy prediction and extrapolation.. MIT
  • e3psi (πŸ₯‰3 Β· ⭐ 2) - Equivariant machine learning library for learning from electronic structures. LGPL-3.0

General Tools

Back to top

General tools for atomistic machine learning.

DeepChem (πŸ₯‡36 Β· ⭐ 4.6K Β· πŸ“‰) - Democratizing Deep-Learning for Drug Discovery, Quantum Chemistry, Materials Science and Biology. MIT
  • GitHub (πŸ‘¨β€πŸ’» 230 Β· πŸ”€ 1.5K Β· πŸ“¦ 260 Β· πŸ“‹ 1.6K - 25% open Β· ⏱️ 27.09.2023):

     git clone https://github.com/deepchem/deepchem
    
  • PyPi (πŸ“₯ 13K / month):

     pip install deepchem
    
  • Conda (πŸ“₯ 100K Β· ⏱️ 16.06.2023):

     conda install -c conda-forge deepchem
    
  • Docker Hub (πŸ“₯ 7K Β· ⭐ 4 Β· ⏱️ 11.03.2022):

     docker pull deepchemio/deepchem
    
RDKit (πŸ₯‡30 Β· ⭐ 2.2K) - BSD-3 C++
  • GitHub (πŸ‘¨β€πŸ’» 200 Β· πŸ”€ 760 Β· πŸ“₯ 1.4K Β· πŸ“¦ 2 Β· πŸ“‹ 2.9K - 29% open Β· ⏱️ 28.09.2023):

     git clone https://github.com/rdkit/rdkit
    
  • PyPi (πŸ“₯ 210K / month):

     pip install rdkit
    
  • Conda (πŸ“₯ 2.5M Β· ⏱️ 16.06.2023):

     conda install -c rdkit rdkit
    
Matminer (πŸ₯‡28 Β· ⭐ 410) - Data mining for materials science. Custom
  • GitHub (πŸ‘¨β€πŸ’» 49 Β· πŸ”€ 170 Β· πŸ“¦ 230 Β· πŸ“‹ 210 - 11% open Β· ⏱️ 11.09.2023):

     git clone https://github.com/hackingmaterials/matminer
    
  • PyPi (πŸ“₯ 15K / month):

     pip install matminer
    
  • Conda (πŸ“₯ 50K Β· ⏱️ 27.06.2023):

     conda install -c conda-forge matminer
    
QUIP (πŸ₯ˆ25 Β· ⭐ 300) - libAtoms/QUIP molecular dynamics framework: https://libatoms.github.io. GPL-2.0
  • GitHub (πŸ‘¨β€πŸ’» 80 Β· πŸ”€ 120 Β· πŸ“₯ 350 Β· πŸ“¦ 24 Β· πŸ“‹ 430 - 20% open Β· ⏱️ 29.08.2023):

     git clone https://github.com/libAtoms/QUIP
    
  • PyPi (πŸ“₯ 1.5K / month):

     pip install quippy-ase
    
  • Docker Hub (πŸ“₯ 9.9K Β· ⭐ 4 Β· ⏱️ 24.04.2023):

     docker pull libatomsquip/quip
    
MAML (πŸ₯ˆ24 Β· ⭐ 270) - Python for Materials Machine Learning, Materials Descriptors, Machine Learning Force Fields, Deep Learning, etc. BSD-3
  • GitHub (πŸ‘¨β€πŸ’» 29 Β· πŸ”€ 58 Β· πŸ“¦ 4 Β· πŸ“‹ 64 - 6% open Β· ⏱️ 27.09.2023):

     git clone https://github.com/materialsvirtuallab/maml
    
  • PyPi (πŸ“₯ 270 / month):

     pip install maml
    
JARVIS-Tools (πŸ₯ˆ22 Β· ⭐ 250 Β· πŸ“‰) - JARVIS-Tools: an open-source software package for data-driven atomistic materials design. Publications:.. Custom
  • GitHub (πŸ‘¨β€πŸ’» 15 Β· πŸ”€ 110 Β· πŸ“¦ 64 Β· πŸ“‹ 81 - 46% open Β· ⏱️ 22.09.2023):

     git clone https://github.com/usnistgov/jarvis
    
  • PyPi (πŸ“₯ 4.8K / month):

     pip install jarvis-tools
    
  • Conda (πŸ“₯ 57K Β· ⏱️ 16.06.2023):

     conda install -c conda-forge jarvis-tools
    
Scikit-Matter (πŸ₯ˆ18 Β· ⭐ 61) - A collection of scikit-learn compatible utilities that implement methods born out of the materials science and.. BSD-3 scikit-learn
  • GitHub (πŸ‘¨β€πŸ’» 13 Β· πŸ”€ 15 Β· πŸ“¦ 6 Β· πŸ“‹ 68 - 17% open Β· ⏱️ 22.09.2023):

     git clone https://github.com/scikit-learn-contrib/scikit-matter
    
  • PyPi (πŸ“₯ 1K / month):

     pip install skmatter
    
  • Conda (πŸ“₯ 510 Β· ⏱️ 24.08.2023):

     conda install -c conda-forge skmatter
    
MAST-ML (πŸ₯ˆ17 Β· ⭐ 88) - MAterials Simulation Toolkit for Machine Learning (MAST-ML). MIT
  • GitHub (πŸ‘¨β€πŸ’» 19 Β· πŸ”€ 50 Β· πŸ“₯ 81 Β· πŸ“¦ 6 Β· πŸ“‹ 210 - 10% open Β· ⏱️ 28.07.2023):

     git clone https://github.com/uw-cmg/MAST-ML
    
XenonPy (πŸ₯ˆ16 Β· ⭐ 110) - XenonPy is a Python Software for Materials Informatics. BSD-3
  • GitHub (πŸ‘¨β€πŸ’» 10 Β· πŸ”€ 58 Β· πŸ“₯ 1.2K Β· πŸ“‹ 82 - 19% open Β· ⏱️ 21.05.2023):

     git clone https://github.com/yoshida-lab/XenonPy
    
  • PyPi (πŸ“₯ 220 / month):

     pip install xenonpy
    
Artificial Intelligence for Science (AIRS) (πŸ₯‰14 Β· ⭐ 250) - Artificial Intelligence for Science (AIRS). GPL-3.0 license rep-learn generative ML-IAP MD ML-DFT ML-WFT biomolecules
  • GitHub (πŸ‘¨β€πŸ’» 23 Β· πŸ”€ 30 Β· ⏱️ 31.08.2023):

     git clone https://github.com/divelab/AIRS
    
AMPtorch (πŸ₯‰12 Β· ⭐ 55) - AMPtorch: Atomistic Machine Learning Package (AMP) - PyTorch. GPL-3.0
  • GitHub (πŸ‘¨β€πŸ’» 14 Β· πŸ”€ 31 Β· πŸ“‹ 30 - 16% open Β· ⏱️ 16.07.2023):

     git clone https://github.com/ulissigroup/amptorch
    
Show 11 hidden projects...
  • QML (πŸ₯ˆ16 Β· ⭐ 180 Β· πŸ’€) - QML: Quantum Machine Learning. MIT
  • Automatminer (πŸ₯‰14 Β· ⭐ 130 Β· πŸ’€) - An automatic engine for predicting materials properties. Custom
  • OpenChem (πŸ₯‰10 Β· ⭐ 620 Β· πŸ’€) - OpenChem: Deep Learning toolkit for Computational Chemistry and Drug Design Research. MIT
  • Equisolve (πŸ₯‰8 Β· ⭐ 4) - A package tasked with taking equistore objects and computing machine learning models using them. BSD-3 ML-IAP
  • JAXChem (πŸ₯‰7 Β· ⭐ 74 Β· πŸ’€) - JAXChem is a JAX-based deep learning library for complex and versatile chemical modeling. MIT
  • uncertainty_benchmarking (πŸ₯‰7 Β· ⭐ 34 Β· πŸ’€) - Various code/notebooks to benchmark different ways we could estimate uncertainty in ML predictions. Unlicensed benchmarking probabilistic
  • torchchem (πŸ₯‰7 Β· ⭐ 33 Β· πŸ’€) - An experimental repo for experimenting with PyTorch models. MIT
  • ACEatoms (πŸ₯‰4 Β· ⭐ 2 Β· πŸ’€) - Generic code for modelling atomic properties using ACE. Custom Julia
  • MLatom (πŸ₯‰4) - Machine learning for atomistic simulations. Custom
  • Magpie (πŸ₯‰3) - Materials Agnostic Platform for Informatics and Exploration (Magpie). MIT Java
  • quantum-structure-ml (πŸ₯‰2 Β· ⭐ 1 Β· πŸ’€) - Multi-class classification model for predicting the magnetic order of magnetic structures and a binary classification.. Unlicensed magnetism benchmarking

Generative Models

Back to top

Projects that implement generative models for atomistic ML.

GT4SD (πŸ₯‡18 Β· ⭐ 250) - GT4SD, an open-source library to accelerate hypothesis generation in the scientific discovery process. MIT pre-trained drug-discovery rep-learn
  • GitHub (πŸ‘¨β€πŸ’» 19 Β· πŸ”€ 55 Β· πŸ“‹ 92 - 2% open Β· ⏱️ 15.09.2023):

     git clone https://github.com/GT4SD/gt4sd-core
    
  • PyPi (πŸ“₯ 780 / month):

     pip install gt4sd
    
MoLeR (πŸ₯ˆ16 Β· ⭐ 210) - Implementation of MoLeR: a generative model of molecular graphs which supports scaffold-constrained generation. MIT
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 31 Β· πŸ“‹ 30 - 20% open Β· ⏱️ 09.08.2023):

     git clone https://github.com/microsoft/molecule-generation
    
  • PyPi (πŸ“₯ 220 / month):

     pip install molecule-generation
    
synspace (πŸ₯ˆ12 Β· ⭐ 30) - Synthesis generative model. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 3 Β· πŸ“¦ 5 Β· πŸ“‹ 3 - 66% open Β· ⏱️ 15.04.2023):

     git clone https://github.com/whitead/synspace
    
  • PyPi (πŸ“₯ 370 / month):

     pip install synspace
    
SchNetPack G-SchNet (πŸ₯ˆ11 Β· ⭐ 27) - G-SchNet extension for SchNetPack. MIT
  • GitHub (πŸ‘¨β€πŸ’» 3 Β· πŸ”€ 4 Β· ⏱️ 01.06.2023):

     git clone https://github.com/atomistic-machine-learning/schnetpack-gschnet
    
G-SchNet (πŸ₯‰8 Β· ⭐ 120) - G-SchNet - a generative model for 3d molecular structures. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 22 Β· ⏱️ 24.03.2023):

     git clone https://github.com/atomistic-machine-learning/G-SchNet
    
cG-SchNet (πŸ₯‰8 Β· ⭐ 43) - cG-SchNet - a conditional generative neural network for 3d molecular structures. MIT
  • GitHub (πŸ”€ 13 Β· ⏱️ 24.03.2023):

     git clone https://github.com/atomistic-machine-learning/cG-SchNet
    
bVAE-IM (πŸ₯‰8 Β· ⭐ 8 Β· 🐣) - Implementation of Chemical Design with GPU-based Ising Machine. MIT QML single-paper
  • GitHub (πŸ”€ 2 Β· ⏱️ 11.07.2023):

     git clone https://github.com/tsudalab/bVAE-IM
    
Show 3 hidden projects...
  • EDM (πŸ₯‰9 Β· ⭐ 300 Β· πŸ’€) - E(3) Equivariant Diffusion Model for Molecule Generation in 3D. MIT
  • rxngenerator (πŸ₯‰5 Β· ⭐ 12 Β· πŸ’€) - A generative model for molecular generation via multi-step chemical reactions. MIT
  • MolSLEPA (πŸ₯‰4 Β· ⭐ 3 Β· 🐣) - Interpretable Fragment-based Molecule Design with Self-learning Entropic Population Annealing. MIT XAI

Interatomic Potentials (ML-IAP)

Back to top

Machine learning interatomic potentials (aka ML-IAP, MLIAP, MLIP, MLP) and force fields (ML-FF) for molecular dynamics.

DeePMD-kit (πŸ₯‡28 Β· ⭐ 1.2K) - A deep learning package for many-body potential energy representation and molecular dynamics. LGPL-3.0 C++
  • GitHub (πŸ‘¨β€πŸ’» 62 Β· πŸ”€ 430 Β· πŸ“₯ 26K Β· πŸ“¦ 11 Β· πŸ“‹ 460 - 9% open Β· ⏱️ 27.09.2023):

     git clone https://github.com/deepmodeling/deepmd-kit
    
  • PyPi (πŸ“₯ 1.3K / month):

     pip install deepmd-kit
    
  • Conda (πŸ“₯ 1.1K Β· ⏱️ 27.09.2023):

     conda install -c deepmodeling deepmd-kit
    
  • Docker Hub (πŸ“₯ 2K Β· ⭐ 1 Β· ⏱️ 01.09.2023):

     docker pull deepmodeling/deepmd-kit
    
MEGNet (πŸ₯‡21 Β· ⭐ 460) - Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals. BSD-3
  • GitHub (πŸ‘¨β€πŸ’» 13 Β· πŸ”€ 140 Β· πŸ“¦ 69 Β· πŸ“‹ 74 - 22% open Β· ⏱️ 27.04.2023):

     git clone https://github.com/materialsvirtuallab/megnet
    
  • PyPi (πŸ“₯ 340 / month):

     pip install megnet
    
TorchANI (πŸ₯‡21 Β· ⭐ 400) - Accurate Neural Network Potential on PyTorch. MIT
  • GitHub (πŸ‘¨β€πŸ’» 16 Β· πŸ”€ 110 Β· πŸ“¦ 26 Β· πŸ“‹ 160 - 11% open Β· ⏱️ 21.03.2023):

     git clone https://github.com/aiqm/torchani
    
  • PyPi (πŸ“₯ 1.6K / month):

     pip install torchani
    
  • Conda (πŸ“₯ 210K Β· ⏱️ 25.06.2023):

     conda install -c conda-forge torchani
    
DP-GEN (πŸ₯‡21 Β· ⭐ 240) - The deep potential generator to generate a deep-learning based model of interatomic potential energy and force field. LGPL-3.0 workflows
  • GitHub (πŸ‘¨β€πŸ’» 59 Β· πŸ”€ 160 Β· πŸ“₯ 1.5K Β· πŸ“¦ 4 Β· πŸ“‹ 240 - 16% open Β· ⏱️ 21.03.2023):

     git clone https://github.com/deepmodeling/dpgen
    
  • PyPi (πŸ“₯ 280 / month):

     pip install dpgen
    
  • Conda (πŸ“₯ 180 Β· ⏱️ 16.06.2023):

     conda install -c deepmodeling dpgen
    
NequIP (πŸ₯‡20 Β· ⭐ 440) - NequIP is a code for building E(3)-equivariant interatomic potentials. MIT
  • GitHub (πŸ‘¨β€πŸ’» 8 Β· πŸ”€ 96 Β· πŸ“¦ 15 Β· πŸ“‹ 61 - 19% open Β· ⏱️ 26.03.2023):

     git clone https://github.com/mir-group/nequip
    
  • PyPi (πŸ“₯ 510 / month):

     pip install nequip
    
  • Conda (πŸ“₯ 3.4K Β· ⏱️ 18.06.2023):

     conda install -c conda-forge nequip
    
Pre-trained OCP models (πŸ₯‡19 Β· ⭐ 470) - Pre-trained models released as part of the Open Catalyst Project. MIT pre-trained
  • GitHub (πŸ‘¨β€πŸ’» 32 Β· πŸ”€ 170 Β· πŸ“‹ 130 - 11% open Β· ⏱️ 14.09.2023):

     git clone https://github.com/Open-Catalyst-Project/ocp
    
CHGNet (πŸ₯ˆ18 Β· ⭐ 100) - Pretrained universal neural network potential for charge-informed atomistic modeling https://chgnet.lbl.gov. Custom MD pre-trained electrostatics magnetism structure-relaxation
  • GitHub (πŸ‘¨β€πŸ’» 6 Β· πŸ”€ 20 Β· πŸ“¦ 7 Β· πŸ“‹ 18 - 5% open Β· ⏱️ 28.09.2023):

     git clone https://github.com/CederGroupHub/chgnet
    
  • PyPi (πŸ“₯ 5.9K / month):

     pip install chgnet
    
M3GNet (πŸ₯ˆ17 Β· ⭐ 180) - Materials graph network with 3-body interactions featuring a DFT surrogate crystal relaxer and a state-of-the-art.. BSD-3
  • GitHub (πŸ‘¨β€πŸ’» 14 Β· πŸ”€ 51 Β· πŸ“¦ 16 Β· πŸ“‹ 35 - 42% open Β· ⏱️ 06.06.2023):

     git clone https://github.com/materialsvirtuallab/m3gnet
    
  • PyPi (πŸ“₯ 390 / month):

     pip install m3gnet
    
sGDML (πŸ₯ˆ17 Β· ⭐ 130) - sGDML - Reference implementation of the Symmetric Gradient Domain Machine Learning model. MIT
  • GitHub (πŸ‘¨β€πŸ’» 8 Β· πŸ”€ 34 Β· πŸ“¦ 8 Β· πŸ“‹ 16 - 31% open Β· ⏱️ 31.08.2023):

     git clone https://github.com/stefanch/sGDML
    
  • PyPi (πŸ“₯ 180 / month):

     pip install sgdml
    
PyXtalFF (πŸ₯ˆ15 Β· ⭐ 71) - Machine Learning Interatomic Potential Predictions. MIT
  • GitHub (πŸ‘¨β€πŸ’» 8 Β· πŸ”€ 19 Β· πŸ“‹ 60 - 15% open Β· ⏱️ 17.08.2023):

     git clone https://github.com/MaterSim/PyXtal_FF
    
  • PyPi (πŸ“₯ 22 / month):

     pip install pyxtal_ff
    
MACE (πŸ₯ˆ14 Β· ⭐ 190) - MACE - Fast and accurate machine learning interatomic potentials with higher order equivariant message passing. MIT
  • GitHub (πŸ‘¨β€πŸ’» 11 Β· πŸ”€ 67 Β· πŸ“‹ 72 - 31% open Β· ⏱️ 17.08.2023):

     git clone https://github.com/ACEsuit/mace
    
NNPOps (πŸ₯ˆ14 Β· ⭐ 63) - High-performance operations for neural network potentials. MIT MD C++
  • GitHub (πŸ‘¨β€πŸ’» 7 Β· πŸ”€ 14 Β· πŸ“‹ 52 - 38% open Β· ⏱️ 25.07.2023):

     git clone https://github.com/openmm/NNPOps
    
  • Conda (πŸ“₯ 67K Β· ⏱️ 17.08.2023):

     conda install -c conda-forge nnpops
    
Ultra-Fast Force Fields (UF3) (πŸ₯ˆ14 Β· ⭐ 42) - UF3: a python library for generating ultra-fast interatomic potentials. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 8 Β· πŸ”€ 15 Β· πŸ“‹ 31 - 38% open Β· ⏱️ 26.09.2023):

     git clone https://github.com/uf3/uf3
    
  • PyPi (πŸ“₯ 10 / month):

     pip install uf3
    
KLIFF (πŸ₯ˆ14 Β· ⭐ 27) - KIM-based Learning-Integrated Fitting Framework (KLIFF). LGPL-2.1 probabilistic workflows
  • GitHub (πŸ‘¨β€πŸ’» 10 Β· πŸ”€ 17 Β· πŸ“‹ 33 - 51% open Β· ⏱️ 28.08.2023):

     git clone https://github.com/openkim/kliff
    
  • PyPi (πŸ“₯ 46 / month):

     pip install kliff
    
  • Conda (πŸ“₯ 63K Β· ⏱️ 16.06.2023):

     conda install -c conda-forge kliff
    
n2p2 (πŸ₯ˆ13 Β· ⭐ 190 Β· πŸ’€) - n2p2 - A Neural Network Potential Package. GPL-3.0 C++
  • GitHub (πŸ‘¨β€πŸ’» 9 Β· πŸ”€ 67 Β· πŸ“‹ 140 - 39% open Β· ⏱️ 05.09.2022):

     git clone https://github.com/CompPhysVienna/n2p2
    
So3krates (MLFF) (πŸ₯ˆ13 Β· ⭐ 37) - Build neural networks for machine learning force fields with JAX. MIT
  • GitHub (πŸ‘¨β€πŸ’» 3 Β· πŸ”€ 6 Β· πŸ“‹ 5 - 20% open Β· ⏱️ 06.09.2023):

     git clone https://github.com/thorben-frank/mlff
    
CCS_fit (πŸ₯ˆ13 Β· ⭐ 5) - Curvature Constrained Splines. GPL-3.0
  • GitHub (πŸ‘¨β€πŸ’» 8 Β· πŸ”€ 9 Β· πŸ“₯ 380 Β· πŸ“‹ 14 - 57% open Β· ⏱️ 25.09.2023):

     git clone https://github.com/Teoroo-CMC/CCS
    
  • PyPi (πŸ“₯ 240 / month):

     pip install ccs_fit
    
DMFF (πŸ₯ˆ12 Β· ⭐ 110 Β· πŸ’€) - DMFF (Differentiable Molecular Force Field) is a Jax-based python package that provides a full differentiable.. LGPL-3.0
  • GitHub (πŸ‘¨β€πŸ’» 7 Β· πŸ”€ 29 Β· πŸ“‹ 17 - 64% open Β· ⏱️ 14.02.2023):

     git clone https://github.com/deepmodeling/DMFF
    
PiNN (πŸ₯ˆ11 Β· ⭐ 96) - A Python library for building atomic neural networks. BSD-3
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 26 Β· πŸ“‹ 6 - 16% open Β· ⏱️ 28.09.2023):

     git clone https://github.com/Teoroo-CMC/PiNN
    
  • Docker Hub (πŸ“₯ 220 Β· ⏱️ 28.09.2023):

     docker pull teoroo/pinn
    
Pacemaker (πŸ₯ˆ11 Β· ⭐ 45 Β· πŸ“‰) - Python package for fitting atomic cluster expansion (ACE) potentials. Custom
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 11 Β· πŸ“‹ 35 - 22% open Β· ⏱️ 17.09.2023):

     git clone https://github.com/ICAMS/python-ace
    
  • PyPi (πŸ“₯ 1 / month):

     pip install python-ace
    
ACEfit (πŸ₯ˆ11 Β· ⭐ 5) - MIT Julia
  • GitHub (πŸ‘¨β€πŸ’» 6 Β· πŸ”€ 3 Β· πŸ“‹ 54 - 40% open Β· ⏱️ 18.08.2023):

     git clone https://github.com/ACEsuit/ACEfit.jl
    
Neural Force Field (πŸ₯‰10 Β· ⭐ 190 Β· πŸ“‰) - Neural Network Force Field based on PyTorch. MIT pre-trained
  • GitHub (πŸ‘¨β€πŸ’» 10 Β· πŸ”€ 41 Β· ⏱️ 25.07.2023):

     git clone https://github.com/learningmatter-mit/NeuralForceField
    
DimeNet (πŸ₯‰9 Β· ⭐ 250) - DimeNet and DimeNet++ models, as proposed in Directional Message Passing for Molecular Graphs (ICLR 2020) and Fast and.. Custom
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 53 Β· πŸ“‹ 30 - 3% open Β· ⏱️ 01.08.2023):

     git clone https://github.com/gasteigerjo/dimenet
    
Allegro (πŸ₯‰9 Β· ⭐ 230) - Allegro is an open-source code for building highly scalable and accurate equivariant deep learning interatomic.. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 33 Β· πŸ“‹ 18 - 33% open Β· ⏱️ 08.05.2023):

     git clone https://github.com/mir-group/allegro
    
ACE.jl (πŸ₯‰9 Β· ⭐ 62) - Parameterisation of Equivariant Properties of Particle Systems. Custom Julia
  • GitHub (πŸ‘¨β€πŸ’» 12 Β· πŸ”€ 14 Β· πŸ“‹ 81 - 28% open Β· ⏱️ 09.06.2023):

     git clone https://github.com/ACEsuit/ACE.jl
    
ACE1.jl (πŸ₯‰9 Β· ⭐ 19) - Atomic Cluster Expansion for Modelling Invariant Atomic Properties. Custom Julia
  • GitHub (πŸ‘¨β€πŸ’» 7 Β· πŸ”€ 4 Β· πŸ“‹ 45 - 46% open Β· ⏱️ 19.09.2023):

     git clone https://github.com/ACEsuit/ACE1.jl
    
wfl (πŸ₯‰9 Β· ⭐ 14) - Workflow is a Python toolkit for building interatomic potential creation and atomistic simulation workflows. Unlicensed workflows HTC
  • GitHub (πŸ‘¨β€πŸ’» 13 Β· πŸ”€ 13 Β· πŸ“‹ 130 - 46% open Β· ⏱️ 22.09.2023):

     git clone https://github.com/libAtoms/workflow
    
NNsforMD (πŸ₯‰9 Β· ⭐ 10 Β· πŸ’€) - Neural network class for molecular dynamics to predict potential energy, forces and non-adiabatic couplings. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 4 Β· ⏱️ 10.11.2022):

     git clone https://github.com/aimat-lab/NNsForMD
    
  • PyPi (πŸ“₯ 16 / month):

     pip install pyNNsMD
    
GemNet (πŸ₯‰8 Β· ⭐ 150) - GemNet model in PyTorch, as proposed in GemNet: Universal Directional Graph Neural Networks for Molecules (NeurIPS.. Custom
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 25 Β· ⏱️ 26.04.2023):

     git clone https://github.com/TUM-DAML/gemnet_pytorch
    
GAP (πŸ₯‰8 Β· ⭐ 32) - Gaussian Approximation Potential (GAP). Custom
  • GitHub (πŸ‘¨β€πŸ’» 12 Β· πŸ”€ 20 Β· ⏱️ 08.06.2023):

     git clone https://github.com/libAtoms/GAP
    
Atomistic Adversarial Attacks (πŸ₯‰8 Β· ⭐ 24 Β· πŸ’€) - Code for performing adversarial attacks on atomistic systems using NN potentials. MIT probabilistic
  • GitHub (πŸ‘¨β€πŸ’» 6 Β· πŸ”€ 6 Β· ⏱️ 03.10.2022):

     git clone https://github.com/learningmatter-mit/Atomistic-Adversarial-Attacks
    
ACE1Pack.jl (πŸ₯‰8 Β· 🐣) - Provides convenience functionality for the usage of ACE1.jl, ACEfit.jl, JuLIP.jl for fitting interatomic potentials.. MIT Julia
  • GitHub (πŸ‘¨β€πŸ’» 11 Β· ⏱️ 21.08.2023):

     git clone https://github.com/ACEsuit/ACE1pack.jl
    
MACE-Jax (πŸ₯‰7 Β· ⭐ 35) - Equivariant machine learning interatomic potentials in JAX. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 1 Β· ⏱️ 20.07.2023):

     git clone https://github.com/ACEsuit/mace-jax
    
ALF (πŸ₯‰7 Β· ⭐ 19) - A framework for performing active learning for training machine-learned interatomic potentials. Custom active-learning
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 7 Β· ⏱️ 04.08.2023):

     git clone https://github.com/lanl/alf
    
Show 17 hidden projects...
  • TensorMol (πŸ₯ˆ12 Β· ⭐ 260 Β· πŸ’€) - Tensorflow + Molecules = TensorMol. GPL-3.0 single-paper
  • ANI-1 (πŸ₯ˆ12 Β· ⭐ 200 Β· πŸ’€) - ANI-1 neural net potential with python interface (ASE). MIT
  • SIMPLE-NN (πŸ₯ˆ11 Β· ⭐ 43 Β· πŸ’€) - SIMPLE-NN(SNU Interatomic Machine-learning PotentiaL packagE version Neural Network). GPL-3.0
  • SchNet (πŸ₯‰9 Β· ⭐ 180 Β· πŸ’€) - SchNet - a deep learning architecture for quantum chemistry. MIT
  • SNAP (πŸ₯‰8 Β· ⭐ 33 Β· πŸ’€) - Repository for spectral neighbor analysis potential (SNAP) model development. BSD-3
  • SIMPLE-NN v2 (πŸ₯‰8 Β· ⭐ 28) - GPL-3.0
  • PhysNet (πŸ₯‰7 Β· ⭐ 78 Β· πŸ’€) - Code for training PhysNet models. MIT electrostatics
  • AIMNet (πŸ₯‰7 Β· ⭐ 76 Β· πŸ’€) - Atoms In Molecules Neural Network Potential. MIT single-paper
  • testing-framework (πŸ₯‰6 Β· ⭐ 11 Β· πŸ’€) - The purpose of this repository is to aid the testing of a large number of interatomic potentials for a variety of.. Unlicensed benchmarking
  • PANNA (πŸ₯‰6 Β· ⭐ 6 Β· πŸ’€) - A package to train and validate all-to-all connected network models for BP[1] and modified-BP[2] type local atomic.. MIT benchmarking
  • Alchemical learning (πŸ₯‰5 Β· ⭐ 2) - Code for the Modeling high-entropy transition metal alloys with alchemical compression article. BSD-3
  • glp (πŸ₯‰4 Β· ⭐ 12 Β· 🐣) - tools for graph-based machine-learning potentials in jax. MIT
  • TensorPotential (πŸ₯‰4 Β· ⭐ 5) - Tensorpotential is a TensorFlow based tool for development, fitting ML interatomic potentials from electronic.. Custom
  • MLIP-3 (πŸ₯‰2 Β· ⭐ 12 Β· 🐣) - MLIP-3: Active learning on atomic environments with Moment Tensor Potentials (MTP). BSD-2 C++
  • SingleNN (πŸ₯‰2 Β· ⭐ 6 Β· πŸ’€) - An efficient package for training and executing neural-network interatomic potentials. Unlicensed C++
  • ACE Workflows (πŸ₯‰2 Β· 🐣) - Workflow Examples for ACE Models. Unlicensed Julia workflows
  • RuNNer (πŸ₯‰2) - The RuNNer Neural Network Energy Representation is a Fortran-based framework for the construction of Behler-.. GPL-3.0 Fortran

Language Models

Back to top

Projects that use (large) language models (LMs, LLMs) or natural language procesing (NLP) techniques for atomistic ML.

paper-qa (πŸ₯‡25 Β· ⭐ 3.1K) - LLM Chain for answering questions from documents with citations. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 12 Β· πŸ”€ 290 Β· πŸ“¦ 32 Β· πŸ“‹ 100 - 39% open Β· ⏱️ 19.09.2023):

     git clone https://github.com/whitead/paper-qa
    
  • PyPi (πŸ“₯ 4.4K / month):

     pip install paper-qa
    
mat2vec (πŸ₯ˆ12 Β· ⭐ 600) - Supplementary Materials for Tshitoyan et al. Unsupervised word embeddings capture latent knowledge from materials.. MIT rep-learn
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 170 Β· πŸ“‹ 23 - 26% open Β· ⏱️ 06.05.2023):

     git clone https://github.com/materialsintelligence/mat2vec
    
gptchem (πŸ₯ˆ12 Β· ⭐ 160) - Use GPT-3 to solve chemistry problems. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 30 Β· πŸ“‹ 8 - 75% open Β· ⏱️ 22.06.2023):

     git clone https://github.com/kjappelbaum/gptchem
    
  • PyPi (πŸ“₯ 80 / month):

     pip install gptchem
    
MolSkill (πŸ₯‰11 Β· ⭐ 81) - Learning chemical intuition from humans in the loop. Supporting code. MIT drug-discovery recommender
  • GitHub (πŸ‘¨β€πŸ’» 4 Β· πŸ”€ 5 Β· πŸ“‹ 5 - 40% open Β· ⏱️ 13.06.2023):

     git clone https://github.com/microsoft/molskill
    
  • Conda (πŸ“₯ 120 Β· ⏱️ 18.06.2023):

     conda install -c msr-ai4science molskill
    
nlcc (πŸ₯‰10 Β· ⭐ 41 Β· πŸ’€) - Natural language computational chemistry command line interface. MIT single-paper
  • GitHub (πŸ‘¨β€πŸ’» 3 Β· πŸ”€ 6 Β· ⏱️ 04.02.2023):

     git clone https://github.com/whitead/nlcc
    
  • PyPi (πŸ“₯ 9 / month):

     pip install nlcc
    
Show 1 hidden projects...
  • BERT-PSIE-TC (πŸ₯‰4 Β· ⭐ 2) - A dataset of Curie temperatures automatically extracted from scientific literature with the use of the BERT-PSIE.. MIT magnetism

Materials Discovery

Back to top

Projects that implement materials discovery methods using atomistic ML.

aviary (πŸ₯‡13 Β· ⭐ 29) - The Wren sits on its Roost in the Aviary. MIT
  • GitHub (πŸ‘¨β€πŸ’» 7 Β· πŸ”€ 7 Β· πŸ“‹ 26 - 15% open Β· ⏱️ 10.08.2023):

     git clone https://github.com/CompRhys/aviary
    
closed-loop-acceleration-benchmarks (πŸ₯‰5) - Data and scripts in support of the publication By how much can closed-loop frameworks accelerate computational.. MIT materials-discovery active-learning single-paper
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 1 Β· ⏱️ 02.05.2023):

     git clone https://github.com/aced-differentiate/closed-loop-acceleration-benchmarks
    
Show 6 hidden projects...

Mathematical tools

Back to top

Projects that implement mathematical objects used in atomistic machine learning.

KFAC-JAX (πŸ₯‡18 Β· ⭐ 160) - Second Order Optimization and Curvature Estimation with K-FAC in JAX. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 11 Β· πŸ”€ 13 Β· πŸ“¦ 6 Β· πŸ“‹ 9 - 22% open Β· ⏱️ 26.09.2023):

     git clone https://github.com/deepmind/kfac-jax
    
  • PyPi (πŸ“₯ 950 / month):

     pip install kfac-jax
    
gpax (πŸ₯ˆ17 Β· ⭐ 120) - Gaussian Processes for Experimental Sciences. MIT probabilistic active-learning
  • GitHub (πŸ”€ 15 Β· πŸ“‹ 8 - 37% open Β· ⏱️ 04.09.2023):

     git clone https://github.com/ziatdinovmax/gpax
    
  • PyPi (πŸ“₯ 350 / month):

     pip install gpax
    
Polynomials4ML.jl (πŸ₯ˆ15 Β· ⭐ 10) - Polynomials for ML: fast evaluation, batching, differentiation. MIT Julia
  • GitHub (πŸ‘¨β€πŸ’» 8 Β· πŸ”€ 4 Β· πŸ“‹ 42 - 35% open Β· ⏱️ 24.09.2023):

     git clone https://github.com/ACEsuit/Polynomials4ML.jl
    
SpheriCart (πŸ₯ˆ13 Β· ⭐ 41) - Multi-language library for the calculation of spherical harmonics in Cartesian coordinates. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 6 Β· πŸ”€ 3 Β· πŸ“₯ 1 Β· πŸ“¦ 1 Β· πŸ“‹ 14 - 57% open Β· ⏱️ 18.08.2023):

     git clone https://github.com/lab-cosmo/sphericart
    
  • PyPi (πŸ“₯ 93 / month):

     pip install sphericart
    
GElib (πŸ₯‰6 Β· ⭐ 16) - C++/CUDA library for SO(3) equivariant operations. MPL-2.0 C++
  • GitHub (πŸ‘¨β€πŸ’» 3 Β· πŸ”€ 2 Β· πŸ“‹ 5 - 80% open Β· ⏱️ 23.07.2023):

     git clone https://github.com/risi-kondor/GElib
    
COSMO Toolbox (πŸ₯‰6 Β· ⭐ 6) - Assorted libraries and utilities for atomistic simulation analysis. Unlicensed C++
  • GitHub (πŸ‘¨β€πŸ’» 9 Β· πŸ”€ 5 Β· ⏱️ 23.06.2023):

     git clone https://github.com/lab-cosmo/toolbox
    
Show 3 hidden projects...
  • EquivariantOperators.jl (πŸ₯‰6 Β· ⭐ 17) - MIT Julia
  • cnine (πŸ₯‰4 Β· ⭐ 2) - Cnine tensor library. Unlicensed C++
  • Wigner Kernels (πŸ₯‰3 Β· ⭐ 1) - Collection of programs to benchmark Wigner kernels. Unlicensed benchmarking

Molecular Dynamics

Back to top

Projects that simplify the integration of molecular dynamics and atomistic machine learning.

JAX-MD (πŸ₯‡23 Β· ⭐ 990) - Differentiable, Hardware Accelerated, Molecular Dynamics. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 28 Β· πŸ”€ 160 Β· πŸ“¦ 33 Β· πŸ“‹ 130 - 45% open Β· ⏱️ 29.08.2023):

     git clone https://github.com/jax-md/jax-md
    
  • PyPi (πŸ“₯ 1.6K / month):

     pip install jax-md
    
FitSNAP (πŸ₯ˆ19 Β· ⭐ 120) - Software for generating SNAP machine-learning interatomic potentials. GPL-2.0
  • GitHub (πŸ‘¨β€πŸ’» 24 Β· πŸ”€ 43 Β· πŸ“₯ 7 Β· πŸ“‹ 62 - 12% open Β· ⏱️ 15.09.2023):

     git clone https://github.com/FitSNAP/FitSNAP
    
  • Conda (πŸ“₯ 4.6K Β· ⏱️ 16.06.2023):

     conda install -c conda-forge fitsnap3
    
openmm-torch (πŸ₯ˆ15 Β· ⭐ 130) - OpenMM plugin to define forces with neural networks. Custom ML-IAP C++
  • GitHub (πŸ‘¨β€πŸ’» 7 Β· πŸ”€ 24 Β· πŸ“‹ 68 - 25% open Β· ⏱️ 27.09.2023):

     git clone https://github.com/openmm/openmm-torch
    
  • Conda (πŸ“₯ 210K Β· ⏱️ 27.09.2023):

     conda install -c conda-forge openmm-torch
    
OpenMM-ML (πŸ₯‰10 Β· ⭐ 52) - High level API for using machine learning models in OpenMM simulations. MIT ML-IAP
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 15 Β· πŸ“‹ 39 - 53% open Β· ⏱️ 21.08.2023):

     git clone https://github.com/openmm/openmm-ml
    
  • Conda (πŸ“₯ 1.9K Β· ⏱️ 21.08.2023):

     conda install -c conda-forge openmm-ml
    
PACE (πŸ₯‰9 Β· ⭐ 21 Β· πŸ’€) - The LAMMPS ML-IAP `pair_style pace`, aka Atomic Cluster Expansion (ACE), aka ML-PACE,.. Custom
  • GitHub (πŸ‘¨β€πŸ’» 6 Β· πŸ”€ 10 Β· πŸ“‹ 6 - 16% open Β· ⏱️ 31.01.2023):

     git clone https://github.com/ICAMS/lammps-user-pace
    
Show 1 hidden projects...

Probabilistic ML

Back to top

Projects that focus on probabilistic, Bayesian, Gaussian process and adversarial methods for atomistic ML, for optimization, uncertainty quantification (UQ), etc.


Reinforcement Learning

Back to top

Projects that focus on reinforcement learning for atomistic ML.

Show 2 hidden projects...
  • ReLeaSE (πŸ₯‡11 Β· ⭐ 310 Β· πŸ’€) - Deep Reinforcement Learning for de-novo Drug Design. MIT drug-discovery
  • CatGym (πŸ₯‰6 Β· ⭐ 10 Β· πŸ’€) - Surface segregation using Deep Reinforcement Learning. GPL

Representation Engineering

Back to top

Projects that offer implementations of representations aka descriptors, fingerprints of atomistic systems, and models built with them, aka feature engineering.

cdk (πŸ₯‡25 Β· ⭐ 430) - The Chemistry Development Kit. LGPL-2.1 cheminformatics Java
  • GitHub (πŸ‘¨β€πŸ’» 160 Β· πŸ”€ 150 Β· πŸ“₯ 16K Β· πŸ“¦ 18 Β· πŸ“‹ 260 - 8% open Β· ⏱️ 11.09.2023):

     git clone https://github.com/cdk/cdk
    
  • Maven:

     <dependency>
     	<groupId>org.openscience.cdk</groupId>
     	<artifactId>cdk-bundle</artifactId>
     	<version>[VERSION]</version>
     </dependency>
    
DScribe (πŸ₯‡23 Β· ⭐ 340) - DScribe is a python package for creating machine learning descriptors for atomistic systems. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 18 Β· πŸ”€ 79 Β· πŸ“¦ 150 Β· πŸ“‹ 85 - 9% open Β· ⏱️ 05.09.2023):

     git clone https://github.com/SINGROUP/dscribe
    
  • PyPi (πŸ“₯ 25K / month):

     pip install dscribe
    
  • Conda (πŸ“₯ 78K Β· ⏱️ 06.09.2023):

     conda install -c conda-forge dscribe
    
CatLearn (πŸ₯‡16 Β· ⭐ 93 Β· πŸ’€) - GPL-3.0 surface-science
  • GitHub (πŸ‘¨β€πŸ’» 22 Β· πŸ”€ 51 Β· πŸ“¦ 4 Β· πŸ“‹ 25 - 36% open Β· ⏱️ 07.02.2023):

     git clone https://github.com/SUNCAT-Center/CatLearn
    
  • PyPi (πŸ“₯ 100 / month):

     pip install catlearn
    
MODNet (πŸ₯‡16 Β· ⭐ 58) - MODNet: a framework for machine learning materials properties. MIT pre-trained small-data transfer-learning
  • GitHub (πŸ‘¨β€πŸ’» 7 Β· πŸ”€ 26 Β· πŸ“¦ 3 Β· πŸ“‹ 35 - 37% open Β· ⏱️ 29.07.2023):

     git clone https://github.com/ppdebreuck/modnet
    
SISSO (πŸ₯ˆ14 Β· ⭐ 180) - A data-driven method combining symbolic regression and compressed sensing for accurate & interpretable models. Apache-2 Fortran
  • GitHub (πŸ‘¨β€πŸ’» 3 Β· πŸ”€ 61 Β· πŸ“‹ 51 - 1% open Β· ⏱️ 12.09.2023):

     git clone https://github.com/rouyang2017/SISSO
    
GlassPy (πŸ₯ˆ14 Β· ⭐ 16) - Python module for scientists working with glass materials. GPL-3.0
  • GitHub (πŸ”€ 6 Β· πŸ“¦ 2 Β· πŸ“‹ 4 - 25% open Β· ⏱️ 27.08.2023):

     git clone https://github.com/drcassar/glasspy
    
  • PyPi (πŸ“₯ 260 / month):

     pip install glasspy
    
Librascal (πŸ₯ˆ13 Β· ⭐ 71) - A scalable and versatile library to generate representations for atomic-scale learning. LGPL-2.1
  • GitHub (πŸ‘¨β€πŸ’» 29 Β· πŸ”€ 18 Β· πŸ“‹ 230 - 43% open Β· ⏱️ 06.06.2023):

     git clone https://github.com/lab-cosmo/librascal
    
Rascaline (πŸ₯ˆ12 Β· ⭐ 20) - Computing representations for atomistic machine learning. BSD-3 Rust C++
  • GitHub (πŸ‘¨β€πŸ’» 11 Β· πŸ”€ 11 Β· πŸ“‹ 40 - 47% open Β· ⏱️ 21.09.2023):

     git clone https://github.com/Luthaf/rascaline
    
BenchML (πŸ₯‰7 Β· ⭐ 13) - ML benchmarking and pipeling framework. Apache-2 benchmarking
  • GitHub (πŸ‘¨β€πŸ’» 9 Β· πŸ”€ 2 Β· πŸ“‹ 13 - 23% open Β· ⏱️ 24.05.2023):

     git clone https://github.com/capoe/benchml
    
  • PyPi (πŸ“₯ 17 / month):

     pip install benchml
    
milad (πŸ₯‰6 Β· ⭐ 27 Β· πŸ’€) - Moment Invariants Local Atomic Descriptor. GPL-3.0 generative
  • GitHub (πŸ”€ 1 Β· πŸ“¦ 1 Β· ⏱️ 03.12.2022):

     git clone https://github.com/muhrin/milad
    
SA-GPR (πŸ₯‰6 Β· ⭐ 14 Β· πŸ’€) - Public repository for symmetry-adapted Gaussian Process Regression (SA-GPR). LGPL-3.0 C-lang
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 9 Β· πŸ“‹ 7 - 28% open Β· ⏱️ 29.09.2022):

     git clone https://github.com/dilkins/TENSOAP
    
NICE (πŸ₯‰6 Β· ⭐ 10) - NICE (N-body Iteratively Contracted Equivariants) is a set of tools designed for the calculation of invariant and.. MIT
  • GitHub (πŸ‘¨β€πŸ’» 4 Β· πŸ”€ 2 Β· πŸ“‹ 3 - 66% open Β· ⏱️ 01.05.2023):

     git clone https://github.com/lab-cosmo/nice
    
Show 10 hidden projects...
  • cmlkit (πŸ₯ˆ10 Β· ⭐ 31 Β· πŸ’€) - tools for machine learning in condensed matter physics and quantum chemistry. MIT benchmarking
  • SkipAtom (πŸ₯‰8 Β· ⭐ 22 Β· πŸ’€) - Distributed representations of atoms, inspired by the Skip-gram model. MIT
  • CBFV (πŸ₯‰7 Β· ⭐ 14 Β· πŸ’€) - Tool to quickly create a composition-based feature vector. Unlicensed
  • pyLODE (πŸ₯‰7 Β· ⭐ 2) - Pythonic implementation of LOng Distance Equivariants. Apache-2 electrostatics
  • fplib (πŸ₯‰6 Β· ⭐ 7 Β· πŸ’€) - a fingerprint library. MIT C-lang single-paper
  • SOAPxx (πŸ₯‰6 Β· ⭐ 7 Β· πŸ’€) - A SOAP implementation. GPL-2.0 C++
  • magnetism-prediction (πŸ₯‰4 Β· ⭐ 1) - DFT-aided Machine Learning Search for Magnetism in Fe-based Bimetallic Chalcogenides. Apache-2 magnetism single-paper
  • SISSO++ (πŸ₯‰3 Β· ⭐ 1 Β· πŸ’€) - C++ Implementation of SISSO with python bindings. Apache-2 C++
  • ML-for-CurieTemp-Predictions (πŸ₯‰3 Β· 🐣) - Machine Learning Predictions of High-Curie-Temperature Materials. MIT single-paper magnetism
  • AMP (πŸ₯‰2) - Amp is an open-source package designed to easily bring machine-learning to atomistic calculations. Unlicensed

Representation Learning

Back to top

General models that learn a representations aka embeddings of atomistic systems, such as message-passing neural networks (MPNN).

Deep Graph Library (DGL) (πŸ₯‡38 Β· ⭐ 12K) - Python package built to ease deep learning on graph, on top of existing DL frameworks. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 280 Β· πŸ”€ 2.8K Β· πŸ“¦ 150 Β· πŸ“‹ 2.4K - 13% open Β· ⏱️ 28.09.2023):

     git clone https://github.com/dmlc/dgl
    
  • PyPi (πŸ“₯ 83K / month):

     pip install dgl
    
  • Conda (πŸ“₯ 290K Β· ⏱️ 14.09.2023):

     conda install -c dglteam dgl
    
PyG Models (πŸ₯‡29 Β· ⭐ 19K) - Representation learning models implemented in PyTorch Geometric. MIT general-ml
  • GitHub (πŸ‘¨β€πŸ’» 450 Β· πŸ”€ 3.3K Β· πŸ“‹ 3.2K - 23% open Β· ⏱️ 28.09.2023):

     git clone https://github.com/pyg-team/pytorch_geometric
    
SchNetPack (πŸ₯‡27 Β· ⭐ 640) - SchNetPack - Deep Neural Networks for Atomistic Systems. MIT
  • GitHub (πŸ‘¨β€πŸ’» 31 Β· πŸ”€ 180 Β· πŸ“¦ 59 Β· πŸ“‹ 210 - 1% open Β· ⏱️ 13.09.2023):

     git clone https://github.com/atomistic-machine-learning/schnetpack
    
  • PyPi (πŸ“₯ 420 / month):

     pip install schnetpack
    
e3nn (πŸ₯‡24 Β· ⭐ 750) - A modular framework for neural networks with Euclidean symmetry. MIT
  • GitHub (πŸ‘¨β€πŸ’» 29 Β· πŸ”€ 110 Β· πŸ“¦ 110 Β· πŸ“‹ 150 - 10% open Β· ⏱️ 02.09.2023):

     git clone https://github.com/e3nn/e3nn
    
  • PyPi (πŸ“₯ 49K / month):

     pip install e3nn
    
  • Conda (πŸ“₯ 9.2K Β· ⏱️ 18.06.2023):

     conda install -c conda-forge e3nn
    
kgcnn (πŸ₯‡23 Β· ⭐ 85 Β· πŸ“ˆ) - Graph convolutions in Keras with TensorFlow, PyTorch or Jax. MIT
  • GitHub (πŸ‘¨β€πŸ’» 7 Β· πŸ”€ 25 Β· πŸ“¦ 15 Β· πŸ“‹ 77 - 6% open Β· ⏱️ 28.09.2023):

     git clone https://github.com/aimat-lab/gcnn_keras
    
  • PyPi (πŸ“₯ 170 / month):

     pip install kgcnn
    
NVIDIA Deep Learning Examples for Tensor Cores (πŸ₯‡21 Β· ⭐ 12K) - State-of-the-Art Deep Learning scripts organized by models - easy to train and deploy with reproducible accuracy and.. Custom educational drug-discovery
  • GitHub (πŸ‘¨β€πŸ’» 120 Β· πŸ”€ 2.8K Β· πŸ“‹ 780 - 28% open Β· ⏱️ 05.09.2023):

     git clone https://github.com/NVIDIA/DeepLearningExamples
    
dgl-lifesci (πŸ₯‡21 Β· ⭐ 620) - Python package for graph neural networks in chemistry and biology. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 22 Β· πŸ”€ 130 Β· πŸ“¦ 120 Β· πŸ“‹ 80 - 30% open Β· ⏱️ 16.04.2023):

     git clone https://github.com/awslabs/dgl-lifesci
    
  • PyPi (πŸ“₯ 6.7K / month):

     pip install dgllife
    
MatGL (Materials Graph Library) (πŸ₯‡21 Β· ⭐ 120) - Graph deep learning library for materials. BSD-3
  • GitHub (πŸ‘¨β€πŸ’» 12 Β· πŸ”€ 26 Β· πŸ“¦ 6 Β· πŸ“‹ 39 - 7% open Β· ⏱️ 26.09.2023):

     git clone https://github.com/materialsvirtuallab/matgl
    
  • PyPi (πŸ“₯ 390 / month):

     pip install m3gnet
    
ocp (πŸ₯ˆ19 Β· ⭐ 470) - ocp is the Open Catalyst Projects library of state-of-the-art machine learning algorithms for catalysis. MIT
  • GitHub (πŸ‘¨β€πŸ’» 32 Β· πŸ”€ 170 Β· πŸ“‹ 130 - 11% open Β· ⏱️ 14.09.2023):

     git clone https://github.com/Open-Catalyst-Project/ocp
    
ALIGNN (πŸ₯ˆ19 Β· ⭐ 150 Β· πŸ“ˆ) - Atomistic Line Graph Neural Network. Custom
  • GitHub (πŸ‘¨β€πŸ’» 7 Β· πŸ”€ 63 Β· πŸ“¦ 7 Β· πŸ“‹ 46 - 50% open Β· ⏱️ 11.08.2023):

     git clone https://github.com/usnistgov/alignn
    
  • PyPi (πŸ“₯ 490 / month):

     pip install alignn
    
e3nn-jax (πŸ₯ˆ18 Β· ⭐ 120) - jax library for E3 Equivariant Neural Networks. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 4 Β· πŸ”€ 13 Β· πŸ“‹ 9 - 11% open Β· ⏱️ 28.09.2023):

     git clone https://github.com/e3nn/e3nn-jax
    
  • PyPi (πŸ“₯ 2.5K / month):

     pip install e3nn-jax
    
Uni-Mol (πŸ₯ˆ17 Β· ⭐ 420) - Official Repository for the Uni-Mol Series Methods. MIT pre-trained
  • GitHub (πŸ‘¨β€πŸ’» 10 Β· πŸ”€ 76 Β· πŸ“₯ 6K Β· πŸ“‹ 100 - 28% open Β· ⏱️ 26.09.2023):

     git clone https://github.com/dptech-corp/Uni-Mol
    
matsciml (πŸ₯ˆ17 Β· ⭐ 65) - Open MatSci ML Toolkit is a single framework for prototyping and scaling out deep learning models for materials.. MIT workflows
  • GitHub (πŸ‘¨β€πŸ’» 8 Β· πŸ”€ 8 Β· πŸ“‹ 10 - 50% open Β· ⏱️ 27.09.2023):

     git clone https://github.com/IntelLabs/matsciml
    
escnn (πŸ₯ˆ15 Β· ⭐ 220) - Equivariant Steerable CNNs Library for Pytorch https://quva-lab.github.io/escnn/. Custom
  • GitHub (πŸ‘¨β€πŸ’» 9 Β· πŸ”€ 33 Β· πŸ“‹ 55 - 34% open Β· ⏱️ 11.09.2023):

     git clone https://github.com/QUVA-Lab/escnn
    
  • PyPi (πŸ“₯ 1.1K / month):

     pip install escnn
    
Compositionally-Restricted Attention-Based Network (CrabNet) (πŸ₯ˆ12 Β· ⭐ 10) - Predict materials properties using only the composition information!. MIT
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 3 Β· πŸ“¦ 10 Β· πŸ“‹ 16 - 87% open Β· ⏱️ 19.06.2023):

     git clone https://github.com/sparks-baird/CrabNet
    
  • PyPi (πŸ“₯ 89 / month):

     pip install crabnet
    
GDC (πŸ₯ˆ10 Β· ⭐ 220) - Graph Diffusion Convolution, as proposed in Diffusion Improves Graph Learning (NeurIPS 2019). MIT generative
  • GitHub (πŸ‘¨β€πŸ’» 3 Β· πŸ”€ 38 Β· ⏱️ 26.04.2023):

     git clone https://github.com/gasteigerjo/gdc
    
graphite (πŸ₯ˆ10 Β· ⭐ 25) - A repository for implementing graph network models based on atomic structures. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 6 Β· πŸ“¦ 6 Β· ⏱️ 16.08.2023):

     git clone https://github.com/llnl/graphite
    
GATGNN: Global Attention Graph Neural Network (πŸ₯‰9 Β· ⭐ 60 Β· πŸ’€) - Pytorch Repository for our work: Graph convolutional neural networks with global attention for improved materials.. MIT
  • GitHub (πŸ‘¨β€πŸ’» 3 Β· πŸ”€ 18 Β· πŸ“‹ 6 - 50% open Β· ⏱️ 03.10.2022):

     git clone https://github.com/superlouis/GATGNN
    
CGAT (πŸ₯‰9 Β· ⭐ 15 Β· πŸ’€) - Crystal graph attention neural networks for materials prediction. MIT
  • GitHub (πŸ‘¨β€πŸ’» 4 Β· πŸ”€ 7 Β· ⏱️ 10.01.2023):

     git clone https://github.com/hyllios/CGAT
    
hippynn (πŸ₯‰8 Β· ⭐ 42) - python library for atomistic machine learning. Custom workflows
  • GitHub (πŸ‘¨β€πŸ’» 11 Β· πŸ”€ 18 Β· πŸ“‹ 4 - 75% open Β· ⏱️ 05.08.2023):

     git clone https://github.com/lanl/hippynn
    
DeeperGATGNN (πŸ₯‰8 Β· ⭐ 32) - Scalable graph neural networks for materials property prediction. MIT
  • GitHub (πŸ‘¨β€πŸ’» 3 Β· πŸ”€ 7 Β· ⏱️ 19.04.2023):

     git clone https://github.com/usccolumbia/deeperGATGNN
    
UVVisML (πŸ₯‰8 Β· ⭐ 14) - Predict optical properties of molecules with machine learning. MIT optical-properties single-paper probabilistic
  • GitHub (πŸ”€ 4 Β· ⏱️ 26.05.2023):

     git clone https://github.com/learningmatter-mit/uvvisml
    
Equiformer (πŸ₯‰7 Β· ⭐ 120) - [ICLR23 Spotlight] Equiformer: Equivariant Graph Attention Transformer for 3D Atomistic Graphs. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 25 Β· πŸ“‹ 12 - 41% open Β· ⏱️ 21.06.2023):

     git clone https://github.com/atomicarchitects/equiformer
    
AdsorbML (πŸ₯‰7 Β· ⭐ 20) - MIT surface-science single-paper
  • GitHub (πŸ‘¨β€πŸ’» 5 Β· πŸ”€ 5 Β· ⏱️ 31.07.2023):

     git clone https://github.com/Open-Catalyst-Project/AdsorbML
    
ML4pXRDs (πŸ₯‰7) - Contains code to train neural networks based on simulated powder XRDs from synthetic crystals. MIT XRD single-paper
  • GitHub (πŸ”€ 1 Β· πŸ“₯ 2 Β· ⏱️ 14.07.2023):

     git clone https://github.com/aimat-lab/ML4pXRDs
    
MACE-Layer (πŸ₯‰6 Β· ⭐ 25) - Higher order equivariant graph neural networks for 3D point clouds. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 3 Β· ⏱️ 06.06.2023):

     git clone https://github.com/ACEsuit/mace-layer
    
escnn_jax (πŸ₯‰6 Β· ⭐ 22 Β· 🐣) - Equivariant Steerable CNNs Library for Pytorch https://quva-lab.github.io/escnn/. Custom
  • GitHub (πŸ‘¨β€πŸ’» 8 Β· πŸ”€ 2 Β· ⏱️ 28.06.2023):

     git clone https://github.com/emilemathieu/escnn_jax
    
  • PyPi:

     pip install escnn_jax
    
GLAMOUR (πŸ₯‰6 Β· ⭐ 18 Β· πŸ’€) - Graph Learning over Macromolecule Representations. MIT single-paper
  • GitHub (πŸ”€ 6 Β· ⏱️ 31.12.2022):

     git clone https://github.com/learningmatter-mit/GLAMOUR
    
EquiformerV2 (πŸ₯‰5 Β· ⭐ 82 Β· 🐣) - [arXiv23] EquiformerV2: Improved Equivariant Transformer for Scaling to Higher-Degree Representations. MIT
  • GitHub (πŸ‘¨β€πŸ’» 2 Β· πŸ”€ 9 Β· ⏱️ 28.07.2023):

     git clone https://github.com/atomicarchitects/equiformer_v2
    
CraTENet (πŸ₯‰5 Β· ⭐ 6) - An attention-based deep neural network for thermoelectric transport properties. MIT transport-phenomena
  • GitHub (πŸ”€ 1 Β· ⏱️ 05.04.2023):

     git clone https://github.com/lantunes/CraTENet
    
Per-site PAiNN (πŸ₯‰5 Β· 🐣) - Fork of PaiNN for PerovskiteOrderingGCNNs. MIT probabilistic pre-trained single-paper
  • GitHub (πŸ‘¨β€πŸ’» 10 Β· ⏱️ 05.06.2023):

     git clone https://github.com/learningmatter-mit/per-site_painn
    
Show 21 hidden projects...
  • benchmarking-gnns (πŸ₯ˆ14 Β· ⭐ 2.3K Β· πŸ’€) - Repository for benchmarking graph neural networks. MIT single-paper benchmarking
  • Crystal Graph Convolutional Neural Networks (CGCNN) (πŸ₯ˆ12 Β· ⭐ 510 Β· πŸ’€) - Crystal graph convolutional neural networks for predicting material properties. MIT
  • Neural fingerprint (nfp) (πŸ₯ˆ12 Β· ⭐ 54 Β· πŸ’€) - Keras layers for end-to-end learning with rdkit and pymatgen. Custom
  • SE(3)-Transformers (πŸ₯ˆ10 Β· ⭐ 410 Β· πŸ’€) - code for the SE3 Transformers paper: https://arxiv.org/abs/2006.10503. MIT single-paper
  • molecularGNN_smiles (πŸ₯‰9 Β· ⭐ 250 Β· πŸ’€) - The code of a graph neural network (GNN) for molecules, which is based on learning representations of r-radius.. Apache-2
  • ai4material_design (πŸ₯‰9 Β· ⭐ 1) - Code for Kazeev, N., Al-Maeeni, A.R., Romanov, I. et al. Sparse representation for machine learning the properties of.. Apache-2 pre-trained material-defect
  • DTNN (πŸ₯‰7 Β· ⭐ 77 Β· πŸ’€) - Deep Tensor Neural Network. MIT
  • Cormorant (πŸ₯‰7 Β· ⭐ 54 Β· πŸ’€) - Codebase for Cormorant Neural Networks. Custom
  • charge_transfer_nnp (πŸ₯‰6 Β· ⭐ 21 Β· πŸ’€) - Graph neural network potential with charge transfer. MIT electrostatics
  • tensorfieldnetworks (πŸ₯‰5 Β· ⭐ 140 Β· πŸ’€) - MIT
  • Autobahn (πŸ₯‰5 Β· ⭐ 27 Β· πŸ’€) - Repository for Autobahn: Automorphism Based Graph Neural Networks. MIT
  • SCFNN (πŸ₯‰5 Β· ⭐ 14 Β· πŸ’€) - Self-consistent determination of long-range electrostatics in neural network potentials. MIT C++ electrostatics single-paper
  • FieldSchNet (πŸ₯‰5 Β· ⭐ 9 Β· πŸ’€) - MIT
  • Graph Transport Network (πŸ₯‰4 Β· ⭐ 14) - Graph transport network (GTN), as proposed in Scalable Optimal Transport in High Dimensions for Graph Distances,.. Custom transport-phenomena
  • Per-Site CGCNN (πŸ₯‰4 Β· ⭐ 1 Β· 🐣) - Crystal graph convolutional neural networks for predicting material properties. MIT pre-trained single-paper
  • Atom2Vec (πŸ₯‰3 Β· ⭐ 25 Β· πŸ’€) - Atom2Vec: a simple way to describe atoms for machine learning. Unlicensed
  • Element encoder (πŸ₯‰3 Β· ⭐ 5 Β· πŸ’€) - Autoencoder neural network to compress properties of atomic species into a vector representation. GPL-3.0 single-paper
  • Point Edge Transformer (πŸ₯‰2) - Smooth, exact rotational symmetrization for deep learning on point clouds. CC-BY-4.0
  • SphericalNet (πŸ₯‰1 Β· ⭐ 3 Β· πŸ’€) - Implementation of Clebsch-Gordan Networks (CGnet: https://arxiv.org/pdf/1806.09231.pdf) by GElib & cnine libraries in.. Unlicensed
  • gkx: Green-Kubo Method in JAX (πŸ₯‰1 Β· ⭐ 2 Β· 🐣) - Green-Kubo + JAX + MLPs = Anharmonic Thermal Conductivities Done Fast. MIT transport-phenomena
  • atom_by_atom ( ⭐ 2 Β· 🐣) - Atom-by-atom design of metal oxide catalysts for the oxygen evolution reaction with Machine Learning. Unlicensed surface-science single-paper

Unsupervised Learning

Back to top

Projects that focus on unsupervised learning (USL) for atomistic ML, such as dimensionality reduction, clustering and visualization.

DADApy (πŸ₯‡15 Β· ⭐ 72) - Distance-based Analysis of DAta-manifolds in python. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 15 Β· πŸ”€ 8 Β· πŸ“¦ 2 Β· πŸ“‹ 25 - 40% open Β· ⏱️ 14.09.2023):

     git clone https://github.com/sissa-data-science/DADApy
    
  • PyPi (πŸ“₯ 55 / month):

     pip install dadapy
    
ASAP (πŸ₯ˆ13 Β· ⭐ 110) - ASAP is a package that can quickly analyze and visualize datasets of crystal or molecular structures. MIT
  • GitHub (πŸ‘¨β€πŸ’» 6 Β· πŸ”€ 25 Β· πŸ“¦ 4 Β· πŸ“‹ 24 - 25% open Β· ⏱️ 30.08.2023):

     git clone https://github.com/BingqingCheng/ASAP
    
Sketchmap (πŸ₯‰8 Β· ⭐ 39) - Suite of programs to perform non-linear dimensionality reduction -- sketch-map in particular. GPL-3.0 C++
  • GitHub (πŸ‘¨β€πŸ’» 8 Β· πŸ”€ 10 Β· πŸ“‹ 8 - 37% open Β· ⏱️ 24.05.2023):

     git clone https://github.com/lab-cosmo/sketchmap
    
Show 3 hidden projects...

Visualization

Back to top

Projects that focus on visualization (viz.) for atomistic ML.

Chemiscope (πŸ₯‡15 Β· ⭐ 87) - An interactive structure/property explorer for materials and molecules. BSD-3 JavaScript
  • GitHub (πŸ‘¨β€πŸ’» 17 Β· πŸ”€ 27 Β· πŸ“₯ 140 Β· πŸ“¦ 5 Β· πŸ“‹ 110 - 38% open Β· ⏱️ 14.09.2023):

     git clone https://github.com/lab-cosmo/chemiscope
    
  • npm (πŸ“₯ 14 / month):

     npm install chemiscope
    

Wavefunction methods (ML-WFT)

Back to top

Projects and models that focus on quantities of wavefunction theory methods, such as Monte Carlo techniques like deep learning variational Monte Carlo (DL-VMC), quantum chemistry methods, etc.

DeepQMC (πŸ₯‡21 Β· ⭐ 300 Β· πŸ“ˆ) - Deep learning quantum Monte Carlo for electrons in real space. MIT
  • GitHub (πŸ‘¨β€πŸ’» 13 Β· πŸ”€ 56 Β· πŸ“¦ 1 Β· πŸ“‹ 36 - 5% open Β· ⏱️ 27.09.2023):

     git clone https://github.com/deepqmc/deepqmc
    
  • PyPi (πŸ“₯ 140 / month):

     pip install deepqmc
    
FermiNet (πŸ₯ˆ14 Β· ⭐ 580) - An implementation of the Fermionic Neural Network for ab-initio electronic structure calculations. Apache-2
  • GitHub (πŸ‘¨β€πŸ’» 18 Β· πŸ”€ 100 Β· ⏱️ 14.08.2023):

     git clone https://github.com/deepmind/ferminet
    
DeepErwin (πŸ₯‰9 Β· ⭐ 35) - DeepErwin is a python 3.8+ package that implements and optimizes JAX 2.x wave function models for numerical solutions.. Custom
  • GitHub (πŸ‘¨β€πŸ’» 6 Β· πŸ”€ 6 Β· πŸ“‹ 10 - 10% open Β· ⏱️ 21.09.2023):

     git clone https://github.com/mdsunivie/deeperwin
    
  • PyPi (πŸ“₯ 20 / month):

     pip install deeperwin
    
Show 1 hidden projects...
  • SchNOrb (πŸ₯‰5 Β· ⭐ 54 Β· πŸ’€) - Unifying machine learning and quantum chemistry with a deep neural network for molecular wavefunctions. MIT

Others

Back to top

Show 1 hidden projects...

Contribution

Contributions are encouraged and always welcome! If you like to add or update projects, choose one of the following ways:

  • Open an issue by selecting one of the provided categories from the issue page and fill in the requested information.
  • Modify the projects.yaml with your additions or changes, and submit a pull request. This can also be done directly via the Github UI.

If you like to contribute to or share suggestions regarding the project metadata collection or markdown generation, please refer to the best-of-generator repository. If you like to create your own best-of list, we recommend to follow this guide.

For more information on how to add or update projects, please read the contribution guidelines. By participating in this project, you agree to abide by its Code of Conduct.

License

CC0

About

πŸ† A ranked list of awesome atomistic machine learning projects βš›οΈπŸ§¬πŸ’Ž.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published