Introduction

BM4Ckit

Batched Machine-learning for Catalysts kit

A Python-based package containing various tools for machine learning applied in catalytic chemistry, including reading, writing, filtering, modifying and converting batched structure files, training ML models, model predictions, ML-potential-based batched structure optimisation as well as molecular dynamics, etc.

Manual webpages(Tutorials) are now under construction using Chirpy theme by Cotes Chung to whom our sincere gratitude is given. Current manuals can be found in Manual.

Please note that the project is still in its early version and may change significantly in the future.

Third-party Libraries

Used:

• NumPy (BSD-3-Clause License), Copyright © 2005-2025 NumPy Developers.
• Joblib (BSD-3-Clause License), Copyright © 2008-2021, The joblib developers.
• PyTorch (BSD-3-Clause License), Copyright © 2016-present Facebook Inc.

Optional:

• ASE (LGPL-v2.1 License)
• DGL (Apache-2.3 License)
• PyTorch Geometric (MIT License)

Latest Version

0.9a

Licenses

See LICENSES for full license texts.