Closed kryczko closed 3 years ago
Currently, the pre-trained models return the averaged energies for the given ensemble. I'm wondering what would be the best way to also include the standard deviation within the forward function.
Hi,
You could use the following code to get the predicted energy from each model, then calculate the std
model.members_energies((species, coordinates)).energies
Perfect! Thanks so much!
Currently, the pre-trained models return the averaged energies for the given ensemble. I'm wondering what would be the best way to also include the standard deviation within the forward function.