This package contains code for minimizing the coherence of complex matrices.
The file trstmi.py contains the main script used for generating low coherence matrices.
| Argument | Description |
|---|---|
dim1 |
lower bound on dimension |
dim2 |
upper bound on dimension |
num1 |
lower bound on number of points |
num2 |
upper bound on number of points |
trials |
number of different starting random initializations |
tol |
stopping gradient tolerance |
proc |
takes the value cpu or gpu |
verbose |
takes the value 0 or 1 |
This code runs on Python 3. You will need PyTorch installed. If you would like to use GPU parallelization you will need a version of PyTorch installed supporting your hardware. The trust region optimizer used in the program is not a built-in optimizer in PyTorch, but can be found here. It may be installed with:
python -m pip install git+https://github.com/vchoutas/torch-trust-ncg
If you have any questions about the code, please contact Carlos Saltijeral (carlossaltrev@gmail.com) or
Josiah Park
Texas A&M University
Department of Mathematics
j.park@tamu.edu