search

PhasesResearchLab / pySIPFENN

Python python toolset for Structure-Informed Property and Feature Engineering with Neural Networks. It offers unique advantages through (1) effortless extensibility, (2) optimizations for ordered, dilute, and random atomic configurations, and (3) automated model tuning.
https://pysipfenn.org
Other
20 stars 3 forks source link

Output in ESPEI format #7

Open latmarat opened 1 year ago

latmarat commented 1 year ago

Hello,

I just came across this package. Can pySIPFENN output predictions in the format readable by ESPEI? If not, do you know of any converters?

Thanks, Marat

amkrajewski commented 1 year ago

Hi Marat,

Not automatically, as the required output depends on your ESPEI phase description (e.g. sublattice model choice).

That said, we do a lot of SIPFENN -> ESPEI exports in our research projects and we can certainly can provide active support in getting that working for you!

I'm currently at the CALPHAD conference in Boston, but next week I will add a small tutorial to this repository (and pysipfenn.org documentation page) that should get you started with some examples.

We are also planning to cover this in the next pySIPFENN workshop with Material Genome Foundation, around this Fall, alongside some exciting techniques allowing you to locally tune SIPFENN models during your assessment.