GASpy is able to create various catalyst-adsorbate systems and then use DFT to simulate the adsorption energies of these systems. GASpy_regressions analyzes GASpy's results to create surrogate models that can make predictions on DFT calculations that we have not yet performed. We then store these predictions in the Mongo collections that we set up in GASpy. Refer to our Jupyter notebooks for examples/specifics.
You will need to first install GASpy. Then to use GASpy_regressions, you will need
to make sure that this repository is cloned into your local repository of GASpy
as a submodule. Then run
via Docker, e.g. docker run -v "/local/path/to/GASpy:/home/GASpy" ulissigroup/gaspy_regressions:latest foo.
Active learning across intermetallics to guide discovery of electrocatalysts for CO2 reduction and H2 evolution. Note that the repository which we reference in this paper is version 0.1 of GASpy_feedback, which can stil be found here.
Current GASpy_regressions version: 0.20
For an up-to-date list of our software dependencies, you can simply check out how we build our docker image here.