flarenet is a convolutional neural network used to predict flares in 20-second cadence data from NASA's Transiting Exoplanet Survey Satellite (TESS).

With a single line of code, you can generate flare predictions for any TESS 20-s target.

Installation

With pip:

pip install flarenet20

For developers:

git clone https://github.com/veraberger/flarenet.git
cd flarenet
pip install .

Usage

Tutorial notebook: docs/example.ipynb

Use our model trained on real TESS light curves with injected false-positives, or train your own with the Flarenet class.

The model outputs likelihoods only - choose your own confidence threshold for flares.

If you use this code in publications, please cite Berger, Schanche, et al. (2025).

MIT License