A machine learning software library for computational materials physics
Project description
mechanoChemML library
Developed by the Computational Physics Group at the University of Michigan.
List of contributors (alphabetical order):
- Arjun Sundararajan
- Elizabeth Livingston
- Greg Teichert
- Matt Duschenes
- Sid Srivastava
- Xiaoxuan Zhang
- Zhenlin Wang
- Krishna Garikipati
Overview
mechanoChemML is a machine learning software library for computational materials physics. It is designed to function as an interface between platforms that are widely used for scientific machine learning on one hand, and others for solution of partial differential equations-based models of physics. Of special interest here, and the focus of mechanoChemML, are applications to computational materials physics. These typically feature the coupled solution of material transport, reaction, phase transformation, mechanics, heat transport and electrochemistry.
Version information
This is version 0.1.0.
License
GNU Lesser General Public License (LGPL) v3.0. Please see the file LICENSE for details.
Installation
$ conda create --name mechanochemml python=3.7
$ conda activate mechanochemml
$ (mechanochemml) pip install mechanochemml
Download examples
One can either download the whole mechanoChemML library
$ (mechanochemml) git clone https://github.com/mechanoChem/mechanoChemML.git mechanoChemML-master
Or just download the examples provided by the mechanoChemML library
$ (mechanochemml) svn export https://github.com/mechanoChem/mechanoChemML/trunk/examples ./examples
One needs to run the following command to install the proper TensorFlow version, which is compatible with their CUDA version
$ (mechanochemml) python3 examples/install_tensorflow.py
Documentation and usage
The documentation of this library is available at https://mechanochemml.readthedocs.io/en/latest/index.html, where one can find instructions of using the provided classes, functions, and workflows provided by the mechanoChemML library.
To create a local copy of the documentation, one can use
$ (mechanochemml) cd mechanoChemML-master/doc
$ (mechanochemml) make html
Acknowledgements
This code has been developed under the support of the following:
- Toyota Research Institute, Award #849910 "Computational framework for data-driven, predictive, multi-scale and multi-physics modeling of battery materials"
Referencing this code
If you write a paper using results obtained with the help of this code, please consider citing
- X. Zhang, G.H. Teichert, Z. Wang, M. Duschenes, S. Srivastava, A. Sunderarajan, E. Livingston, K. Garikipati (2021), mechanoChemML: A software library for machine learning in computational materials physics, arXiv preprint arXiv:2112.04960.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for mechanoChemML-0.0.6-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 4522784e624b97d7cbeccb106eb63d2a883335ec3e106509658f9deb57db24e0 |
|
MD5 | 6d37c54d952f1559a5fbb6240bd03852 |
|
BLAKE2b-256 | 930d82cc3dbf8f95e80fdabc067fe6c8d4ba7715fa0220bf26eb7c4e5c36464d |