Table 1 Hyperparameters for the NEP training

From: A comprehensive exploration of thermal transport at Cu/diamond interfaces via machine learning potentials

Parameter

Value

Parameter

Value

version

NEP4

type

6 Cu C Cr Mo Ti W

\({r}_{C}^{R}\)

6 Å

\({r}_{C}^{A}\)

4 Å

\({n}_{\max }^{R}\)

8

\({n}_{\max }^{A}\)

8

\({N}_{{basis}}^{R}\)

10

\({N}_{{basis}}^{A}\)

10

\({l}_{\max }^{3b}\)

4

\({l}_{\max }^{4b}\)

2

\({l}_{\max }^{5b}\)

0

\({N}_{{neuron}}\)

50

\({N}_{{batch}}\)

1200

\({N}_{{generation}}\)

300,000