A Catalog of High Confidence Very Metal-poor Stars from DESI Based on Deep Learning Method

This repository provides the catalog used in the paper “A Catalog of High Confidence Very Metal-poor Stars from DESI Based on Deep Learning Method”. It is associated with manuscript number "AAS73235". It contains 2,569 high-confidence very metal-poor star candidates identified from DESI spectra using a deep-learning framework, including 1,377 newly identified candidates. The table includes DESI target IDs, coordinates, and predicted stellar parameters ($T_{\rm eff}$, $\log g$, and $\mathrm{[Fe/H]}$). Column description targetid: unique DESI target identifier ra: right ascension (J2000, in degrees) dec: declination (J2000, in degrees) pred_logg: predicted surface gravity, $\log g$, in dex pred_teff: predicted effective temperature, $T_{\rm eff}$, in K pred_fe_h: predicted metallicity, $\mathrm{[Fe/H]}$, in dex