Is it a common case to use systems with different number of atoms in training a DP model? #1720
Answered
by
njzjz
ipe-zhangyz
asked this question in
Q&A
-
|
Different number of atoms in different systems may cause frenquent change of the shape of many tensors in the training network, which may hurt the performance of training. So I wonder if it is a common case in practice? Thanks! |
Beta Was this translation helpful? Give feedback.
Answered by
njzjz
May 19, 2022
Replies: 1 comment
-
|
No performance degradation has been observed yet. |
Beta Was this translation helpful? Give feedback.
0 replies
Answer selected by
ipe-zhangyz
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
No performance degradation has been observed yet.