DeepMD simulate monolayer adsorption

[nsysu7433111]

Dear developers and team,
Thanks to the developers and development team of DeePMD for creating a very accurate potential model, which has allowed us to take a huge step forward in our progress on the atomic scale.

I have a question about how to set up a DeepMD simulation for the adsorption of a 2D material monolayer.
Our simulation workflow is as follows: after the DLP training process is complete, we use DLP to perform a minimization (Figure 1). We then use the final minimized structure to perform an NPT simulation (Figures 2 and 3).
However, when I use DP-Lammps to run the NPT simulation for the monolayer adsorption, the structure becomes unstable (it doesn't hold the whole surface), as shown in Figure 3.

My training set :
1.Substrate 50,300K (BC2N,BC2As,BC2P)
2.Metal Lithium 300K (Adsorption Atom)
3.Dimer (Contain Substrate and Adsorption Atom)
4.Full of Li Adsorption on Substrate(One&Both side )
5.Two of Li adsorption on Substrate(Only one side)

Descriptor Settings
Descriptor: hybrid
type: se_e2_a
rcut: 6.26130000000001
rcut_smth: 2.4455000001
type: se_e3
rcut: 4.00000001
rcut_smth: 2.00000001

Loss Function Weights
start_pref_f : 1000
limit_pref_f : 1
start_pref_v : 100
limit_pref_v : 0.1
start_pref_e : 0.001
limit_pref_e : 1

Fig.1 Minimize final data

Fig.2 Npt simulation(Before Crash) Fig.3 Npt simulation(Crash)

[wanghan-iapcm]

please refer to #4802 (comment)

[nsysu7433111]

Thank you, Dr. Wang Han, for your reply. I will proceed with the work based on your suggestions.