Xinyi Liu

PhD student, Computer Science and Technology, Peking University

Joined

August 2024

Names

Xinyi Liu (Preferred)

Emails

****@stu.pku.edu.cn (Confirmed)

Personal Links

Homepage

ORCID

Career & Education History

PhD student

Computer Science and Technology, Peking University (stu.pku.edu.cn)

2023 – 2028

Undergrad student

Computer Science and Technology, Harbin Institute of Technology (stu.hit.edu.cn)

2019 – 2023

Advisors, Relations & Conflicts

PhD Advisor

Bin CUI

Present

Expertise

MLSys

Present

Publications

LAER-MoE: Load-Adaptive Expert Re-layout for Efficient Mixture-of-Experts Training
Xinyi Liu, Yujie Wang, Fangcheng Fu, Xuefeng Xiao, Huixia Li, Jiashi Li, Bin Cui
- Crossref
- Readers: Everyone
FlexSP: Accelerating Large Language Model Training via Flexible Sequence Parallelism.
Yujie Wang, Shiju Wang, Shenhan Zhu, Fangcheng Fu, Xinyi Liu, Xuefeng Xiao 0001, Huixia Li, Jiashi Li, Faming Wu, Bin Cui 0001
- ASPLOS (2) 2025
- Readers: Everyone
NetMoE: Accelerating MoE Training through Dynamic Sample Placement.
Xinyi Liu, Yujie Wang, Fangcheng Fu, Xupeng Miao, Shenhan Zhu, Xiaonan Nie, Bin Cui 0001
- ICLR 2025
- Readers: Everyone
Galvatron: An Automatic Distributed System for Efficient Foundation Model Training.
Xinyi Liu, Yujie Wang, Shenhan Zhu, Fangcheng Fu, Qingshuo Liu, Guangming Lin, Bin Cui 0001
- CoRR 2025
- Readers: Everyone
Data-Centric and Heterogeneity-Adaptive Sequence Parallelism for Efficient LLM Training.
Yujie Wang, Shiju Wang, Shenhan Zhu, Fangcheng Fu, Xinyi Liu, Xuefeng Xiao 0001, Huixia Li, Jiashi Li, Faming Wu, Bin Cui 0001
- CoRR 2024
- Readers: Everyone
FlexSP: Accelerating Large Language Model Training via Flexible Sequence Parallelism
Yujie Wang, Shiju Wang, Shenhan Zhu, Fangcheng Fu, Xinyi Liu, Xuefeng Xiao, Huixia Li, Jiashi Li, Faming Wu, Bin Cui
- Crossref
- Readers: Everyone
NetMoE: Accelerating MoE Training through Dynamic Sample Placement
Xinyi Liu, Yujie Wang, Fangcheng Fu, Xupeng Miao, Shenhan Zhu, Xiaonan Nie, Bin CUI
- ICLR 2025 Spotlight
- Readers: Everyone

Xinyi Liu

Names

Emails

Personal Links

Career & Education History

Advisors, Relations & Conflicts

Expertise

Publications

LAER-MoE: Load-Adaptive Expert Re-layout for Efficient Mixture-of-Experts Training

FlexSP: Accelerating Large Language Model Training via Flexible Sequence Parallelism.

NetMoE: Accelerating MoE Training through Dynamic Sample Placement.

Galvatron: An Automatic Distributed System for Efficient Foundation Model Training.

Data-Centric and Heterogeneity-Adaptive Sequence Parallelism for Efficient LLM Training.

FlexSP: Accelerating Large Language Model Training via Flexible Sequence Parallelism

NetMoE: Accelerating MoE Training through Dynamic Sample Placement

Co-Authors