Replies: 2 comments
-
Sorry for the late reply. That is a good point on the per-atom stress. I might change it, thank you. |
Beta Was this translation helpful? Give feedback.
0 replies
-
Many thanks for the clarification! |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I'm seeking some clarification regarding the loss functions implemented in MACE, specifically the differences between
loss=virials
andloss=stress
. My understanding is that these two are not interchangeable. Forloss=virials
, does the framework utilize per-atom virials within the loss function?Regarding
loss=stress
, considering that stress is an intensive property, the concept of per-atom stress as indicated in the loss.py file seems somewhat unconventional. Could you please provide some insight into this?Thank you for your assistance in understanding these aspects of MACE.
Beta Was this translation helpful? Give feedback.
All reactions