Data from "flarenet: A Convolutional Neural Network for Stellar Flare Detection with TESS"

flarenet allows users to generate stellar flare predictions for each timestep for a TESS target of choice with only a single line of code. This DOI contains the TESS 20-second cadence observations used as the training set for the flarenet model in Berger, Schanche et al. (2026). The sample comprises ~1300 light curves selected to have fewer than 40% of flux measurements 3 sigma above the baseline, and no visible flares upon visual inspection, to train the model on real TESS systematics and non-flare stellar variability.