Skip to content

Latest commit

 

History

History
26 lines (15 loc) · 674 Bytes

README.md

File metadata and controls

26 lines (15 loc) · 674 Bytes

DuNets-RMA

Code for the research paper "Deep Unrolling Networks with Recurrent Momentum Acceleration for Nonlinear Inverse Problems".

A pre-print is available at arXiv, for now please cite as:

Requirements

In order to run the code, you will need the following:

  • PyTorch (>= 1.10.0)
  • Python (>=3.7.xx)
  • PyEIT (link)

Structure

A nonlinear deconvolution problem

  1. simulate data: toy_dataset_simulation.py
  2. train: LPD toy_train_lpd.py; LPGD toy_train_lpgd.py

Electrical impedance tomography

  1. simulate data: eit_circle_simulation.py
  2. train: LPD eit_train_lpd.py; LPGD eit_train_lpgd.py