MiniOneRec: An Open-Source Framework for Scaling Generative Recommendation

TL;DR

MiniOneRec introduces an open-source generative recommendation framework that demonstrates scalable, efficient performance on public benchmarks by replacing traditional embeddings with compact SID sequences and employing a minimal post-training recipe.

Contribution

It is the first fully open-source framework for generative recommendation, providing an end-to-end workflow from SID construction to reinforcement learning, validating scaling laws, and improving performance with a lightweight post-training pipeline.

Findings

Scaling laws hold on public benchmarks with increasing model size.

Generative approach is parameter-efficient and improves ranking accuracy.

Post-training techniques enhance recommendation diversity and performance.

Abstract

The recent success of large language models (LLMs) has renewed interest in whether recommender systems can achieve similar scaling benefits. Conventional recommenders, dominated by massive embedding tables, tend to plateau as embedding dimensions grow. In contrast, the emerging generative paradigm replaces embeddings with compact Semantic ID (SID) sequences produced by autoregressive Transformers. Yet most industrial deployments remain proprietary, leaving two fundamental questions open: (1) Do the expected scaling laws hold on public benchmarks? (2) What is the minimal post-training recipe that enables competitive performance? We present MiniOneRec, to the best of our knowledge, the first fully open-source generative recommendation framework, which provides an end-to-end workflow spanning SID construction, supervised fine-tuning, and recommendation-oriented reinforcement learning. We generate SIDs via a Residual Quantized VAE and post-train Qwen backbones ranging from 0.5B to 7B parameters on the Amazon Review dataset. Our experiments reveal a consistent downward trend in both training and evaluation losses with increasing model size, validating the parameter efficiency of the generative approach. To further enhance performance, we propose a lightweight yet effective post-training pipeline that (1) enforces full-process SID alignment and (2) applies reinforcement learning with constrained decoding and hybrid rewards. Together, these techniques yield significant improvements in both ranking accuracy and candidate diversity.