CNN-Encoder-Decoder

This repository contains the code and data used for training CNN-based encoder-decoder model, described in the paper: "Noise reduction in X‑ray photon correlation spectroscopy with convolutional neural networks encoder–decoder models" by T.Konstantinova, L.Wiegart, M.Rakitin, A.M.DeGennaro and A.M.Barbour. https://doi.org/10.1038/s41598-021-93747-y

Files descriptions:

• run_training.py -- the main script to assign the parameters of the model and to run the model. This script calls the initiaion of the data sets and the model (from the file utils.py) and training of the moodel (from the file nets.py);
• nets.py -- the class for the autoencoder model;
• train_and_test.py -- functions for model training, validation and testing;
• utils.py -- auxiliary functions for model assembly, fixing the random seed, data loader, etc.
• requirements.txt files with required libraries for the scripts to run.

To run the script, type in the terminal:

>> conda create --name cnn-training

>> conda activate cnn-training

>> conda install pip

>> pip install -r requirements.txt

>> python run_training.py

Link for the data:

The folder data/ contains the data for training and testing the model.

https://doi.org/10.11578/dc.20210704.1