---

RecBole (伯乐) 3.0

“世有伯乐，然后有千里马。千里马常有，而伯乐不常有。”——韩愈《马说》

RecBole 1.0 | HomePage | Datasets | Paper

Docs: Run | Architecture | Develop

To support the emerging recommendation paradigms, we develop RecBole3.0, an LLM-inspired recommendation library covering representative up-to-date approaches.

1) Highlights

• Flexible Architecture for Systematic Co-design: We introduce a flexible architecture that decouples data, model, and optimization into self-contained units, facilitating both development and usage while adapting to emerging recommendation paradigms.
• Comprehensive Model Support for Emerging Paradigms: RecBole3.0 currently supports representative models across 5 emerging paradigms, establishing a unified and reproducible benchmarking environment for fair comparisons.
• Configurable Pipelines for Reproducible Experimentation: RecBole3.0 enables flexible composition of datasets, models, trainers, and evaluators through configuration for consistent and reproducible experimentation across recommendation paradigms.

2) Implemented Models

Our library includes algorithms covering five major categories:

• Generative Recommendation: TIGER, LETTER, LC-Rec, ETEGRec, RPG, MiniOneRec, and CARE.
• Scaling-based Recommendation: RankMixer, HSTU, and LSRM.
• Latent Reasoning Recommendation: ReaRec, LARES, and CARE.
• LLM-based Recommendation: LLMRank, LLM4RS, LlamaRec, LC-Rec, MiniOneRec, E4SRec, and BIGRec.
• Agent-based Recommendation: AgentCF, AgentCF++, and STARec.

3) The Team

RecBole3.0 is developed and maintained by members from RUCAIBox, the main developers are Enze Liu (@BishopLiu), Zhuoxuan Li (@ZhuoxuanLi-CS), Dongze Wu (@Joyful-bh), Jiale Xu (@JialeXu627), Xiaolei Wang (@wxl1999), Bowen Zheng (@zhengbw0324), Bingqian Li (@Fotiligner), Kesha Ou (@TayTroye), and Chenghao Wu (@wuchenghao0215).

4) Experimental Results

Evaluation Settings:

model:
  history_max_length: 20
trainer:
  eval:
    protocol: full
    exclude_history: false

Musical_Instruments:

Models	Recall@5	Recall@10	NDCG@5	NDCG@10
LSRM	0.0326	0.0529	0.0190	0.0255
HSTU	0.0377	0.0614	0.0224	0.0300
HSTU-Large	0.0410	0.0659	0.0253	0.0333
TIGER	0.0360	0.0562	0.0236	0.0301
LETTER	0.0354	0.0552	0.0230	0.0294
RPG	0.0369	0.0547	0.0244	0.0301
LC-Rec	0.0329	0.0518	0.0216	0.0276
E4SRec	0.0333	0.0529	0.0210	0.0273
LARES	0.0388	0.0610	0.0246	0.0318
ReaRec	0.0345	0.0546	0.0219	0.0284
LlamaRec	0.0407	0.0611	0.0269	0.0335

Industrial_and_Scientific:

Models	Recall@5	Recall@10	NDCG@5	NDCG@10
LSRM	0.0240	0.0391	0.0142	0.0191
HSTU	0.0288	0.0466	0.0165	0.0222
HSTU-Large	0.0325	0.0509	0.0196	0.0256
TIGER	0.0271	0.0435	0.0177	0.0229
LETTER	0.0261	0.0405	0.0170	0.0216
RPG	0.0257	0.0384	0.0174	0.0215
LC-Rec	0.0259	0.0401	0.0175	0.0220
E4SRec	0.0242	0.0372	0.0156	0.0197
LARES	0.0296	0.0466	0.0182	0.0236
ReaRec	0.0237	0.0390	0.0153	0.0202
LlamaRec	0.0324	0.0454	0.0225	0.0267

RecBole Family Projects

The following table summarizes the open-source contributions of RecBole family projects on GitHub.

Projects	Stars	Forks	Issues	Pull requests
RecBole
RecBole2.0
RecBole3.0
RecSysDatasets

Cite

If you find RecBole useful for your research or development, please cite the following papers: RecBole, RecBole2.0 and RecBole3.0.

@inproceedings{recbole,
  author    = {Wayne Xin Zhao and Shanlei Mu and Yupeng Hou and Zihan Lin and Yushuo Chen and Xingyu Pan and Kaiyuan Li and Yujie Lu and Hui Wang and Changxin Tian and Yingqian Min and Zhichao Feng and Xinyan Fan and Xu Chen and Pengfei Wang and Wendi Ji and Yaliang Li and Xiaoling Wang and Ji{-}Rong Wen},
  title     = {RecBole: Towards a Unified, Comprehensive and Efficient Framework for Recommendation Algorithms},
  booktitle = {{CIKM}},
  pages     = {4653--4664},
  publisher = {{ACM}},
  year      = {2021}
}

@article{recbole2.0,
  author    = {Wayne Xin Zhao and Yupeng Hou and Xingyu Pan and Chen Yang and Zeyu Zhang and Zihan Lin and Jingsen Zhang and Shuqing Bian and Jiakai Tang and Wenqi Sun and Yushuo Chen and Lanling Xu and Gaowei Zhang and Zhen Tian and Changxin Tian and Shanlei Mu and Xinyan Fan and Xu Chen and Ji{-}Rong Wen},
  title     = {RecBole 2.0: Towards a More Up-to-Date Recommendation Library},
  journal   = {arXiv preprint arXiv:2206.07351},
  year      = {2022}
}