Session chair: Jacob Gorm Hansen

MEPipe: Democratizing LLM Training with Memory-Efficient Slice-Level Pipeline Scheduling on Cost-Effective Accelerators
Zhenbo Sun (Tsinghua University), Shengqi Chen (Tsinghua University), Yuanwei Wang (Tsinghua University), Jian Sha (Tsinghua University), Guanyu Feng (Zhipu AI), Wenguang Chen (Tsinghua University)
Paper

HybridFlow: A Flexible and Efficient RLHF Framework
Guangming Sheng (The University of Hong Kong), Chi Zhang (ByteDance), Zilingfeng Ye (ByteDance), Xibin Wu (ByteDance), Wang Zhang (ByteDance), Ru Zhang (ByteDance), Yanghua Peng (ByteDance), Haibin Lin (ByteDance), Chuan Wu (The University of Hong Kong)
Paper

Mist: Efficient Distributed Training of Large Language Models via Memory-Parallelism Co-Optimization
Zhanda Zhu (University of Toronto, CentML, Vector Institute), Christina Giannoula (University of Toronto), Muralidhar Andoorveedu (CentML), Qidong Su (University of Toronto, CentML, Vector Institute), Karttikeya Mangalam (UC Berkeley), Bojian Zheng (Independent Researcher), Gennady Pekhimenko (CentML, University of Toronto, Vector Institute)
Paper

Hourglass: Enabling Efficient Split Federated Learning with Data Parallelism
Qiang He (Huazhong University of Science and Technology), Kaibin Wang (Swinburne University of Technology), Zeqian Dong (Swinburne University of Technology), Liang Yuan (University of Adelaide), Feifei Chen (Deakin University), Hai Jin (Huazhong University of Science and Technology), Yun Yang (Swinburne University of Technology)
Paper

FlowCheck: Decoupling Checkpointing and Training of Large-Scale Models
Zimeng Huang (Shanghai Jiao Tong University & Alibaba Cloud), Hao Nie (Alibaba Cloud & Peking University), Haonan Jia (Alibaba Cloud), Bo Jiang (Shanghai Jiao Tong University), Junchen Guo (Alibaba Cloud), Jianyuan Lu (Alibaba Cloud), Rong Wen (Alibaba Cloud), Biao Lyu (Zhejiang University & Alibaba Cloud), Shunmin Zhu (Hangzhou Feitian Cloud & Alibaba Cloud), Xinbing Wang (Shanghai Jiao Tong University)
Paper