FastSpecFit: Spectrophotometric Modeling of 38 Million Extragalactic Targets from the Dark Energy Spectroscopic Instrument Second Data Release
Optical spectrophotometry provides a direct window into the physical conditions, kinematics, and formation histories of galaxies and quasars. FastSpecFit is an open-source Python software package that forward-models each Dark Energy Spectroscopic Instrument (DESI) spectrum jointly with its broadband photometry, convolving stellar-population synthesis and emission-line templates with each fiber's native spectral resolution. This forward-modeling approach preserves DESI's uncorrelated pixel statistics, fits a typical spectrum in ~0.1 s of single-core time, and scales to tens of millions of objects, with uncertainties estimated via Monte Carlo resampling. FastSpecFit measures stellar velocity dispersions, non-parametric star-formation histories, stellar masses, star-formation rates, dust attenuation, K-corrections, rest-frame magnitudes, and fluxes and equivalent widths for 46 emission lines, including separate narrow and broad components for the Balmer lines. We describe these algorithms in detail, validate the measurements against independent codes, and present the resulting value-added catalog for DESI Data Release 2, which provides spectrophotometric and physical properties for 38 million extragalactic targets at 0<z<4. We also release a companion photometry-only catalog and document the catalogs previously released for the DESI Early Data Release and Data Release 1.