The GibbsML package implements a Gaussian process machine learning model for the prediction of temperature-dependent oxide formation free energies. This information can be used, for example, for the construction of Ellingham diagrams. See http://ellingham.energy-materials.org for web app powered by GibbsML.
If you make use of GibbsML or part of the package, please cite the following reference:
J. A. Garrido Torres, V. Gharakhanyan, N. Artrith, T. Hoffmann Eegholm, and A. Urban, "From zero Kelvin quantum mechanics to high-temperature metallurgy with machine learning", (2021) ASAP
J. A. Garrido Torres (jagt@princeton.edu)
A. Urban (a.urban@columbia.edu)
Installation with pip:
pip install --user .
Or in editable (developer) mode:
pip install --user -e .
See the tutorials subdirectory for Jupyter notebooks that demonstrate the usage of the package.