There were two main tasks that encompassed the majority of the work for this Epic. The first is still unfinished, which was working on initializing scarlet with partially deconvolved models to help speed up convergence (and runtime). It worked to a certain degree but with some important caveats that still need to be investigated, hence DM-20127 is left open.
The other work involved testing the current version of scarlet on HSC images with fake sources, and was done by Sophie Reed and myself. That work was summarized in a presentation for the DESC Blending Working Group (
DM-22535), which can be used for a more detailed description of the work performed during this epic.