Can I predict elastic constant through this?

[tngksldia]

How can I import the trained model to predict elastic constant. I want the model trained from m3gnet, chgnet, mace. MEGNET provides the trained model itself. Though, I cannot find the trained model for others.

[janosh]

all models you list will auto-download their respective checkpoints when you install the corresponding PyPI package and run the model. see e.g. CHGNet.load() or mace_mp().

[tngksldia]

Thank you for your consideration! Though, I wonder how can I get 'machine learned Elastic Constant', which can not be directly got from the machine learned potential.

[janosh]

have a look at the ElasticityCalc class in matcalc

[JaGeo]

Or the elastic workflow in atomate2:
https://jageo.github.io/TutorialAtomate2Forcefields/intro.html
😃

[tngksldia]

Thanks everyone! I'll try it!