A value-added catalogue of neural network-based europium abundances for GALAH DR4

Catalogues of europium abundance predictions associated with the paper "A value-added catalogue of neural network-based europium abundances for GALAH DR4".The contents of the three catalogues (cnn_predictions.fits, korg_secondary_catalogue.fits and golden_sample.fits) are described in Appendix A in the paper.Additionally, to maximize reproducibility of our results, we provide the extrapolated normalized spectra used as input to create network predictions (normalized_spectra_prediction_targets.fits), as well as corresponding tables of stellar parameters (training_and_validation_dataset.fits) and extrapolated normalized spectra (normalized_spectra_train_val.fits) for the training/validation dataset.cnn_predictions.fits - the main catalogue, containing CNN-based predictions of [Eu/H] for 118 946 DR4-exclusive stars.korg_secondary_catalogue.fits - a secondary catalogue of [Eu/H] values synthetised with the Korg software for 2500 metal-poor DR4-exclusive stars.golden_sample.fits - a subset of high-precision [Eu/H] predictions for 54 068 DR4-exclusive stars passing strict quality and stellar parameter cuts, using Korg results for metal-poor stars and CNN predictions for all other targets.training_and_validation_dataset.fits - training and validation dataset of 52 147 stars from the DR3/DR4 crossmatch. Columns correspond to those found in the catalogues above, with the addition of 'eu_h_DR3' as the label.normalized_spectra_prediction_targets.fits - stores the 159-pixel input spectra for the DR4-exclusive stars, which can be matched by sobject_id.normalized_spectra_train_val.fits - stores the 159-pixel input spectra for the training/validation stars, which can be matched by sobject_id.