Joint Automated Repository for Various Integrated Simulations (JARVIS) is an integrated framework for computational science using density functional theory, classical force-field/molecular dynamics and machine-learning.
Joint Automated Repository for Various Integrated Simulations (JARVIS) is an integrated framework for computational science using density functional theory,
classical force-field/molecular dynamics and machine-learning. The jarvis-tools package can be used for high-throughput computation, data-analysis, and training machine-learning models. Some of the packages used in the jarvis-tools package are shown below. JARVIS-official website: https://jarvis.nist.gov
Alternative nix install::
Nix allows a robust and reproducible package for Linux. To generate a Nix environment for using JARVIS, follow the Nix instructions.
Jupyter notebooks
Python for beginners:
JARVIS-DFT data analysis:
JARVIS-ML training:
Comparing ML algorithms:
JARVIS-FF data-analysis:
See more in the plot-gallery below
References
JARVIS-FF:
Evaluation and comparison of classical interatomic potentials through a user-friendly interactive web-interface, Nature: Sci Data. 4, 160125 (2017).https://www.nature.com/articles/sdata2016125
Computational Screening of High-performance Optoelectronic Materials using OptB88vdW and TBmBJ Formalisms, Scientific Data 5, 180082 (2018).https://www.nature.com/articles/sdata201882
Materials science in the artificial intelligence age: high-throughput library generation, machine learning, and a pathway from correlations to the underpinning physics, MRS Comm., 1-18 https://doi.org/10.1557/mrc.2019.95
Similarly, an example calculation for Silicon is given in vasp folder (https://github.com/usnistgov/jarvis/blob/master/jarvis/vasp/examples/runstruct_pyvasp.py). The input is a POSCAR file, which is already provided. executable paths, pseudopotential directory path and Special_POTCAR.yaml path needs to be adjusted in joptb88vdw.py top section. The master.py can be submitted to the queuing system with qsub sub.sh.
ML example:
We trained machine learning models using JARVIS-DFT data on bandgaps, formation energies and elastic modulus and other properties. We used both chemical and structural descriptors during GradientBoostingRegression training. Example of getting 1557 descriptors for a system is given at: https://github.com/usnistgov/jarvis/blob/master/jarvis/sklearn/examples/desc_example.py