Links to download data sets:
- Processed images and spectroscopic redshifts from Pasquet et al. 2019
- Additional catalogs and redshift predictions from the paper
A breif description of the data files provided:
sdss.npz: NumPy.npzfile containing both the image datacube and the labels used for training and testing the photo-z classifier of Pasquet et al. 2019. The datacube (“cube”) contains images with size 64×64 pixels for 659857 SDSS galaxies in the u,g,r,i and z filters as a NumPy float32 ndarray with shape (659857, 64, 64, 5). The labels array (“labels”) is a NumPy recarray with 659857 lines (one for each galaxy) of 64 columns. More details can be found at the data website or Pasquet et al. 2019.cat_train.npy/.csv: Galaxy property catalog for our training set.cat_dev.npy/.csv: Galaxy property catalog for our validation set.cat_test.npy/.csv: Galaxy property catalog for our test set.performance.txt: Performance metrics comparing our work with Pasquet et al 2019 and Hayat et al 2021.
If you have any questions, comments or want to collaborate, feel free to drop an email to: biprateep@pitt.edu