In train.ini , what means about 'orbital basis number' in basic <orbital>?

[Lintzuching]

Hi,
When I trying to reproduce the result of TBG in your paper ' Deep-Learning Density Functional Theory Hamiltonian for Efficient ab initio
Electronic-Structure Calculation ' , I'm confused about the ' orbital basis number ' means . Does it means ' Valence electrons ' ? Or something else ?
Regards ,
TzuChing

[mzjb]

DeepH predict DFT Hamiltonian using the numerical localized basis. orbital basis number means the index of the localized basis on an atom. It does not necessarily correspond to the valence electrons of materials or isolated atoms.

[Lintzuching]

Could you give me some instance? In your demo , the orbital basis number is in the region [0,18] . How about TBG ?
What region you suggest to set ?
TzuChing

[mzjb]

For TBG, we use 13 localized basis functions for C to generate the dataset, which is the Standard guideline choice of OpenMX for C. The region in train.ini needs to be consistent with the number of basis functions used when generating the dataset.

[mzjb]

Could you give me some instance? In your demo , the orbital basis number is in the region [0,18] . How about TBG ? What region you suggest to set ? TzuChing

See this link for an example ini file and this link for the corresponding dataset.

[Lintzuching]

I'm idiot .... Thanks your nice and friendly respond !

[ismsgithub]

The link truly gives us advice on PAO selection, but I still feel confused about how to calculate orbital basis number? What's the formula?

[mzjb]

The link truly gives us advice on PAO selection, but I still feel confused about how to calculate orbital basis number? What's the formula?

orbital basis number in the ini file just means the index of the localized basis for a atom. One does not need to calculate it. The first basis is 0, the second basis is 1 and so on.

[ismsgithub]

"For TBG, we use 13 localized basis functions for C to generate the dataset". Why is 13?

[mzjb]

"For TBG, we use 13 localized basis functions for C to generate the dataset". Why is 13?

The Standard guideline choice of OpenMX for C is C6.0-s2p2d1. 2x1+2x3+1x5=13.

[ismsgithub]

Oh, I got it. Thanks for your nice and friendly response!