OhmDSP/MLDSD

This dataset includes simulated time-frequency representations of three signal classes (hop, cw, chirp) in various levels of noise. This was a dataset used for the first NVidia tutorial on RFML called Deep Learning for Signal Detection taught online and at GTC conferences starting in 2015. We created the tutorial at a tech startup called KickView, which joined forces with Zeta Associates (Lockheed Martin Subsidiary) in 2022.

A notebook based on the 2015 base notebook is available at https://github.com/ohmdsp/mldsd.