/SNN-NLN

An SNN implementation of the NLN architecture for RFI detection.

Primary LanguageLuaMIT LicenseMIT

SNN-NLN

An SNN implementation of the NLN architecture for RFI detection. Published in PASA

DOI

PWC

PWC

Contains:

  • A PyTorch re-implementation of this work with updated auto-encoder architecture
  • Code to convert trained ANN models to SNNs using SpikingJelly
  • Implementation of the Spiking NLN (SNLN)

All code files are in src/. Replicating all results can be achieved by running replicate.py

Installation

conda create -n snn-nln python=3.10
conda activate snn-nln
pip install -r src/requirements.txt

You may need extra instructions for installing PyTorch with Cuda / Rocm support.

Data Dependencies

The data used in this project is not included in this repository. You will need to download the datasets from zenodo and unzip them into /data.

Dependencies

AOFlagger is also required. Installation instructions can be found here.

Licensing

This code is licensed under the MIT License. See LICENSE for more details.