search

mir-group / nequip

NequIP is a code for building E(3)-equivariant interatomic potentials
https://www.nature.com/articles/s41467-022-29939-5
MIT License
611 stars 135 forks source link

🌟 [FEATURE]How to Train and Validate on Separate Datasets #378

Closed Guackkk closed 11 months ago

Guackkk commented 11 months ago

Hello!

I'm currently working on training Nequip. Is it possible to train and validate the model on different datasets? Any insights and guidance on this matter would be greatly appreciated. Thank you!

Linux-cpp-lisp commented 11 months ago

Hi @Guackkk ,

Thanks for your interest in our code!

During training, see https://github.com/mir-group/nequip/blob/main/configs/full.yaml#L154-L157.

Otherwise, you can use nequip-evaluate to evaluate error metrics and predictions of a trained model on any test or evaluation dataset specified using our YAML syntax in a separate YAML file passed with the --dataset-config option.

Guackkk commented 11 months ago

Thanks for your quick and kind reply!