January 12, 2021

155 words 1 min read



Deep learning quantum Monte Carlo for electrons in real space

repo name deepqmc/deepqmc
repo link https://github.com/deepqmc/deepqmc
language Python
size (curr.) 60481 kB
stars (curr.) 164
created 2019-12-06
license MIT License


build coverage python pypi commits since last commit license code style chat doi

DeepQMC implements variational quantum Monte Carlo for electrons in molecules, using deep neural networks written in PyTorch as trial wave functions. Besides the core functionality, it contains implementations of the following ansatzes:


Install and update using Pip.

pip install -U deepqmc[wf,train,cli]

A simple example

from deepqmc import Molecule, evaluate, train
from deepqmc.wf import PauliNet

mol = Molecule.from_name('LiH')
net = PauliNet.from_hf(mol).cuda()

Or on the command line:

$ cat lih/param.toml
system = 'LiH'
ansatz = 'paulinet'
n_steps = 40
$ deepqmc train lih --save-every 20
converged SCF energy = -7.9846409186467
equilibrating: 49it [00:07,  6.62it/s]
training: 100%|███████| 40/40 [01:30<00:00,  2.27s/it, E=-8.0302(29)]
$ ln -s chkpts/state-00040.pt lih/state.pt
$ deepqmc evaluate lih
evaluating:  24%|▋  | 136/565 [01:12<03:40,  1.65it/s, E=-8.0396(17)]
comments powered by Disqus