Reproduce “reComposer: Editable Planning for Controllable Image Generation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Susie S.Y. Li, Angie Boggust, Fred Hohman, Donghao Ren, Dominik Moritz, Yannick Assogba
reComposer lets creators develop images through an editable plan before generation. A planning agent surfaces ambiguities, users can revise semantic elements with decision-specific editors, and a synthesizer combines the plan into a generation prompt. In an eight-person within-subject study, participants made 11.4 versus 4.8 edits on average; this shows more iteration, not faster work or better image quality.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.