This repository is the source code and data for the paper
B. Peng, F. Siegismund-Poschmann, E. Jorswieck, "RISnet: a dedicated scalable neural network architecture for optimization of reconfigurable intelligent surfaces", International ITG 26th Workshop on Smart Antennas and 13th Conference on Systems, Communications, and Coding, Braunschweig, 2023.
To train the neural network, run train.py
with the following arguments:
1e11
.8e-4
.32, 32
.0.25, 0.25, 0.25, 0.25
.True
if you want to save the tensorboard log and trained models in a folder named after date and time of the beginning of training, False
otherwise.cpu
or cuda
.To test the saved neural network, run test.py
with the path to the saved model in line 46.