Reproduce “RecoGEM: A Serving System for Production-Scale Generative Recommendation” 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
Xiangrui Yang, Yunfeng Zhao, Jiawei Yang, Fangyu Zhang, Enrui Zhao, Liang Zeng, Cheng Peng, Minhui Xie and 3 more
RecoGEM serves generative recommendation by coordinating candidate transitions, state, and live updates in a recommendation-aware runtime, while a specialized engine executes step-specific decoding profiles for beam search over semantic identifier sequences. For OneModel at batch size 32, it reports 1,053 QPS and 30.8 ms P99 versus GeneralGen’s 366 QPS and 89.0 ms; this measures serving performance, not recommendation quality.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.