Skip to content

DeepErwin is a python 3.8+ package that implements and optimizes JAX 2.x wave function models for numerical solutions to the multi-electron Schrödinger equation. DeepErwin supports weight-sharing when optimizing wave functions for multiple nuclear geometries and the usage of pre-trained neural network weights to accelerate optimization.

License

Notifications You must be signed in to change notification settings

connection-on-fiber-bundles/deeperwin

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

44 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DeepErwin

DeepErwin is python package that implements and optimizes wave function models for numerical solutions to the multi-electron Schrödinger equation.

DeepErwin is based on JAX and supports:

  • Optimizing a wavefunction for a single nuclear geometry
  • Optimizing wavefunctions for multiple nuclear geometries in parallel, while sharing neural network weights across these wavefunctions to speed-up optimization
  • Using pre-trained weights of a network to speed-up optimization for entirely new wavefunctions
  • Using second-order optimizers such as KFAC or L-BFGS

A detailed description of our method and the corresponding results can be found in our recent arxiv publication. When you use DeepErwin in your work, please cite:

M. Scherbela, R. Reisenhofer, L. Gerard, P. Marquetand, and P. Grohs.
Solving the electronic Schrödinger equation for multiple nuclear geometries with weight-sharing deep neural networks.
arXiv preprint arXiv:2105.08351 (2021).

Getting Started

The quickest way to get started with DeepErwin is to have a look at our documentation. It contains a detailed description of our python codebase and a tutorial which should help you to quickly get up-and-running using DeepErwin.

About

DeepErwin is a collaborative effort of Michael Scherbela, Rafael Reisenhofer, Leon Gerard, Philipp Marquetand, and Philipp Grohs.
The code was written by Michael Scherbela, Leon Gerard, and Rafael Reisenhofer.
If you have any questions, freel free to reach out via e-mail.

About

DeepErwin is a python 3.8+ package that implements and optimizes JAX 2.x wave function models for numerical solutions to the multi-electron Schrödinger equation. DeepErwin supports weight-sharing when optimizing wave functions for multiple nuclear geometries and the usage of pre-trained neural network weights to accelerate optimization.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%