MuonRec: Shifting the Optimizer Paradigm Beyond Adam in Scalable Generative Recommendation

TL;DR

This paper introduces MuonRec, a novel optimizer framework for recommender systems that outperforms Adam/AdamW by reducing training steps and enhancing ranking quality, especially in generative models.

Contribution

MuonRec is the first to apply the Muon optimizer to RecSys, improving training efficiency and recommendation quality over traditional optimizers.

Findings

Reduces training steps by 32.4% on average.

Achieves 12.6% relative gain in NDCG@10.

Outperforms Adam/AdamW baselines in various settings.

Abstract

Recommender systems (RecSys) are increasingly emphasizing scaling, leveraging larger architectures and more interaction data to improve personalization. Yet, despite the optimizer's pivotal role in training, modern RecSys pipelines almost universally default to Adam/AdamW, with limited scrutiny of whether these choices are truly optimal for recommendation. In this work, we revisit optimizer design for scalable recommendation and introduce MuonRec, the first framework that brings the recently proposed Muon optimizer to RecSys training. Muon performs orthogonalized momentum updates for 2D weight matrices via Newton-Schulz iteration, promoting diverse update directions and improving optimization efficiency. We develop an open-source training recipe for recommendation models and evaluate it across both traditional sequential recommenders and modern generative recommenders. Extensive experiments demonstrate that MuonRec reduces converged training steps by an average of 32.4\% while simultaneously improving final ranking quality. Specifically, MuonRec yields consistent relative gains in NDCG@10, averaging 12.6\% across all settings, with particularly pronounced improvements in generative recommendation models. These results consistently outperform strong Adam/AdamW baselines, positioning Muon as a promising new optimizer standard for RecSys training. Our code is available.