ASCL.net

Astrophysics Source Code Library

Making codes discoverable since 1999

ASCL Code Record

[ascl:2404.014] astroNN: Deep learning for astronomers with Tensorflow

astroNN creates neural networks for deep learning using Keras for model and training prototyping while taking advantage of Tensorflow's flexibility. It contains tools for use with APOGEE, Gaia and LAMOST data, though is primarily designed to apply neural nets on APOGEE spectra analysis and predict luminosity from spectra using data from Gaia parallax with reasonable uncertainty from Bayesian Neural Net. astroNN can handle 2D and 2D colored images, and the package contains custom loss functions and layers compatible with Tensorflow or Keras with Tensorflow backend to deal with incomplete labels. The code contains demo for implementing Bayesian Neural Net with Dropout Variational Inference for reasonable uncertainty estimation and other neural nets.

Code site:
https://github.com/henrysky/astroNN https://astronn.readthedocs.io/en/latest/
Used in:
https://ui.adsabs.harvard.edu/abs/2024AJ....167...73Skc
Described in:
https://ui.adsabs.harvard.edu/abs/2019MNRAS.483.3255L
Bibcode:
2024ascl.soft04014L

Views: 254

ascl:2404.014
Add this shield to your page
Copy the above HTML to add this shield to your code's website.