[CP2K-user] [CP2K:18997] Re: drastic change in pressure on changing smoothing

Vyacheslav Bryantsev vyacheslav.bryantsev at gmail.com
Tue Jun 27 16:29:25 UTC 2023

The important addition to my previous email is that the DeepMD training was 
done for different volumes, and this is where the data with "smoothing" may 
not be fully compatible if combined. 

On Tuesday, June 27, 2023 at 11:57:21 AM UTC-4 Vyacheslav Bryantsev wrote:

> Dear CP2K Developers,
> If the stress tensor with "smoothing" is not currently correct, I am 
> writing to ask if the forces on individual atoms are still correct and 
> those can be used to fit the machine learning potential for running NPT 
> simulations.
> I am asking because we did the DeepMD fitting with and without "smoothing" 
> and it seems that only without "smoothing" we were able to reproduce the 
> trend in the experimental densities as a function of the temperature. Is 
> not getting the correct stress tensor has something to do with not getting 
> the correct densities with DeepMD? Please advise. 
> Thank you,
> Vyacheslav Bryantsev
> On Monday, February 24, 2020 at 3:40:42 AM UTC-5 Sergi Ruiz-Barragan wrote:
>> Thanks Travis.

You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/cc6c53e2-61d5-45b0-b15a-187ccb286b96n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20230627/840ae361/attachment.htm>

More information about the CP2K-user mailing list