You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm training an ANI model with a new element using the tutorial and have a few questions:
The tutorial notebook doesn’t use EnergyShifter, yet it seems to be applied by default in the pretrained model. What is the purpose of EnergyShifter, and how can I generate sae_linfit.dat file?
It seems that pbc or cell-batched training and prediction are not supported yet. However, I heard that the ANI-1xnr model is trained with periodic boundary conditions. In this case, is the batch size set to 1?
Sincerely,
kangmg
The text was updated successfully, but these errors were encountered:
Dear TorchANI developer,
I'm training an ANI model with a new element using the tutorial and have a few questions:
The tutorial notebook doesn’t use EnergyShifter, yet it seems to be applied by default in the pretrained model. What is the purpose of EnergyShifter, and how can I generate sae_linfit.dat file?
How can I save the trained model in the neurochem format ?
It seems that pbc or cell-batched training and prediction are not supported yet. However, I heard that the ANI-1xnr model is trained with periodic boundary conditions. In this case, is the batch size set to 1?
Sincerely,
kangmg
The text was updated successfully, but these errors were encountered: