Torch autodiff DFT-D3#

Implementation of the DFT-D3 dispersion model in PyTorch. This module allows to process a single structure or a batch of structures for the calculation of atom-resolved dispersion energies.

Note

This project is still in early development and the API is subject to change. Contributions are welcome, please checkout our contributing guidelines.

Example

>>> import torch
>>> import tad_dftd3 as d3
>>> numbers = d3.util.pack((  # S22 system 4: formamide dimer
...     d3.util.to_number("C C N N H H H H H H O O".split()),
...     d3.util.to_number("C O N H H H".split()),
... ))
>>> positions = d3.util.pack((
...     torch.tensor([  # coordinates in Bohr
...         [-3.81469488143921, +0.09993441402912, 0.00000000000000],
...         [+3.81469488143921, -0.09993441402912, 0.00000000000000],
...         [-2.66030049324036, -2.15898251533508, 0.00000000000000],
...         [+2.66030049324036, +2.15898251533508, 0.00000000000000],
...         [-0.73178529739380, -2.28237795829773, 0.00000000000000],
...         [-5.89039325714111, -0.02589114569128, 0.00000000000000],
...         [-3.71254944801331, -3.73605775833130, 0.00000000000000],
...         [+3.71254944801331, +3.73605775833130, 0.00000000000000],
...         [+0.73178529739380, +2.28237795829773, 0.00000000000000],
...         [+5.89039325714111, +0.02589114569128, 0.00000000000000],
...         [-2.74426102638245, +2.16115570068359, 0.00000000000000],
...         [+2.74426102638245, -2.16115570068359, 0.00000000000000],
...     ]),
...     torch.tensor([
...         [-0.55569743203406, +1.09030425468557, 0.00000000000000],
...         [+0.51473634678469, +3.15152550263611, 0.00000000000000],
...         [+0.59869690244446, -1.16861263789477, 0.00000000000000],
...         [-0.45355203669134, -2.74568780438064, 0.00000000000000],
...         [+2.52721209544999, -1.29200800956867, 0.00000000000000],
...         [-2.63139587595376, +0.96447869452240, 0.00000000000000],
...     ]),
... ))
>>> param = dict(a1=0.5660, s8=0.3908, a2=3.1280)  # ωB97M-D3(BJ) parameters
>>> energy = torch.sum(d3.dftd3(numbers, positions, param), -1)
>>> torch.set_printoptions(precision=7)
>>> print(energy)  # Energies in Hartree
tensor([-0.0124292, -0.0045002])
>>> print(energy[0] - 2*energy[1])
tensor(-0.0034288)