HOOMD-TF

HOOMD-TF: GPU-Accelerated, Online Machine Learning in the HOOMD-blue Molecular Dynamics Engine - Published in JOSS (2020)

https://github.com/ur-whitelab/hoomd-tf

Science Score: 95.0%

This score indicates how likely this project is to be science-related based on various indicators:

○
CITATION.cff file
✓
codemeta.json file
Found codemeta.json file
✓
.zenodo.json file
Found .zenodo.json file
✓
DOI references
Found 3 DOI reference(s) in README and JOSS metadata
✓
Academic publication links
Links to: joss.theoj.org
✓
Committers with academic emails
3 of 13 committers (23.1%) from academic institutions
○
Institutional organization owner
✓
JOSS paper metadata
Published in Journal of Open Source Software

Keywords

machine-learning molecular-dynamics tensorflow

Scientific Fields

Sociology Social Sciences - 87% confidence

Last synced: 4 months ago · JSON representation

Repository

A plugin that allows the use of Tensorflow in Hoomd-Blue for GPU-accelerated ML+MD

Basic Info

Host: GitHub
Owner: ur-whitelab
License: mit
Language: Python
Default Branch: master
Homepage: https://hoomd-tf.readthedocs.io
Size: 58.3 MB

Statistics

Stars: 32
Watchers: 5
Forks: 8
Open Issues: 9
Releases: 14

Topics

machine-learning molecular-dynamics tensorflow

Created about 7 years ago · Last pushed over 1 year ago

Metadata Files

Readme License

HOOMD-TF

This plugin enables the use of TensorFlow in a HOOMD-blue simulation to compute quantities like forces and collective variables and do learning while running a simulation. You may also use it without hoomd-blue to process trajectories via MDAnalysis. Please see here for documentation for install and usage instructions.

HOOMD-TF can be used for a variety of tasks such as online force-matching, online machine learning in HOOMD-blue simulations, and arbitrary collective variable calculations using TensorFlow tensor operations. Because both HOOMD-blue and TensorFlow are GPU-accelerated, HOOMD-TF was designed with speed in mind, and minimizes latency with a GPU-GPU communication scheme. Of particular interest, HOOMD-TF allows for online machine learning with early termination, rather than the more tradditional batch learning approach for MD+ML.

HOOMD-TF includes several utility functions as convenient built-ins, such as: * RDF calculation * EDS Biasing (See this paper) * Coarse-Grained simulation force matching

In addition to all these, the TensorFlow interface of HOOMD-TF makes implementing arbitrary ML models as easy as it is in TensorFlow, by exposing the HOOMD-blue neighbor list and particle positions to TensorFlow. This enables GPU-accelerated tensor calculations, meaning arbitrary collective variables can be treated in the TensorFlow model framework, as long as they can be expressed as tensor operations on particle positions or neighbor lists.

Tutorials

See example notebooks here to learn about what HOOMD-TF can do.

Prerequisites

The following packages are required to compile:

tensorflow >= 2.3
hoomd >= 2.6
tbb-devel (only for hoomd if installed with conda)

tbb-devel is required when using the HOOMD-blue conda release. It is not automatically installed when installing HOOMD-blue, so use conda install -c conda-forge tbb-devel to install. The TensorFlow version should be TensorFlow 2.3 release. It is recommended you install via pip.

Citation

Please use the following citation:

HOOMD-TF: GPU-Accelerated, Online Machine Learning in the HOOMD-blue Molecular Dynamics Engine. R Barrett, M Chakraborty, DB Amirkulova, HA Gandhi, G Wellawatte, and AD White (2020) Journal of Open Source Software doi: 10.21105/joss.02367

Owner

Name: White Laboratory
Login: ur-whitelab
Kind: organization

Website: http://thewhitelab.org
Repositories: 41
Profile: https://github.com/ur-whitelab

JOSS Publication

HOOMD-TF: GPU-Accelerated, Online Machine Learning in the HOOMD-blue Molecular Dynamics Engine

Published

July 28, 2020

DOI

10.21105/joss.02367

Volume 5, Issue 51, Page 2367