Elastor: Elastic and Efficient Model Partitioning and Checkpointing for Fault-Tolerant Distributed Training.

Jiawen Niu Xuanyu Wang Fangcheng Fu Haoyang Li 0017 Hao Ge Sheng Lin Jiawen Niu Bin Cui 0001 Elastor: Elastic and Efficient Model Partitioning and Checkpointing for Fault-Tolerant Distributed Training. 398-412 2026 PPoPP https://doi.org/10.1145/3774934.3786445 conf/ppopp/2026 db/conf/ppopp/ppopp2026.html#WangF0GLN026

Haoyang Li 0017 Fangcheng Fu Hao Ge Sheng Lin Xuanyu Wang Jiawen Niu Yujie Wang Hailin Zhang 0004 Xiaonan Nie Bin Cui 0001 Malleus: Straggler-Resilient Hybrid Parallel Training of Large-scale Models via Malleable Data and Model Parallelization. 185:1-185:28 2025 June 3 Proc. ACM Manag. Data 3 https://doi.org/10.1145/3725322 db/journals/pacmmod/pacmmod3.html#LiFGLWNWZNC25 streams/journals/pacmmod

Haoyang Li 0017 Fangcheng Fu Sheng Lin Hao Ge Xuanyu Wang Jiawen Niu Jinbao Xue Yangyu Tao Di Wang 0052 Jie Jiang 0015 Bin Cui 0001 Hydraulis: Balancing Large Transformer Model Training via Co-designing Parallel Strategies and Data Assignment. 1-30 2025 3 Proc. ACM Manag. Data 6 https://doi.org/10.1145/3769802 db/journals/pacmmod/pacmmod3.html#LiFLGWNXTWJC25 streams/journals/pacmmod

Sheng Lin Fangcheng Fu Haoyang Li 0017 Hao Ge Xuanyu Wang Jiawen Niu Yaofeng Tu Bin Cui 0001 LobRA: Multi-tenant Fine-tuning over Heterogeneous Data. 2616-2625 2025 April 18 Proc. VLDB Endow. 8 https://www.vldb.org/pvldb/vol18/p2616-fu.pdf https://doi.org/10.14778/3742728.3742752 db/journals/pvldb/pvldb18.html#LinFLGWNTC25 streams/journals/pvldb

Haoyang Li 0017 Fangcheng Fu Hao Ge Sheng Lin Xuanyu Wang Jiawen Niu Xupeng Miao Bin Cui 0001 Hetu v2: A General and Scalable Deep Learning System with Hierarchical and Heterogeneous Single Program Multiple Data Annotations. 2025 April abs/2504.20490 CoRR https://doi.org/10.48550/arXiv.2504.20490 db/journals/corr/corr2504.html#abs-2504-20490 streams/journals/corr

Sheng Lin Fangcheng Fu Haoyang Li 0017 Hao Ge Xuanyu Wang Jiawen Niu Yaofeng Tu Bin Cui 0001 LobRA: Multi-tenant Fine-tuning over Heterogeneous Data. 2025 September abs/2509.01193 CoRR https://doi.org/10.48550/arXiv.2509.01193 db/journals/corr/corr2509.html#abs-2509-01193 streams/journals/corr

Haoyang Li 0017 Fangcheng Fu Hao Ge Sheng Lin Xuanyu Wang Jiawen Niu Yujie Wang Hailin Zhang 0004 Xiaonan Nie Bin Cui 0001 Malleus: Straggler-Resilient Hybrid Parallel Training of Large-scale Models via Malleable Data and Model Parallelization. 2024 abs/2410.13333 CoRR https://doi.org/10.48550/arXiv.2410.13333 db/journals/corr/corr2410.html#abs-2410-13333 streams/journals/corr

Haoyang Li 0017 Fangcheng Fu Sheng Lin Hao Ge Xuanyu Wang Jiawen Niu Jie Jiang 0015 Bin Cui 0001 Demystifying Workload Imbalances in Large Transformer Model Training over Variable-length Sequences. 2024 abs/2412.07894 CoRR https://doi.org/10.48550/arXiv.2412.07894 db/journals/corr/corr2412.html#abs-2412-07894 streams/journals/corr

Bin Cui 0001 Fangcheng Fu Hao Ge Jie Jiang 0015 Haoyang Li 0017 Sheng Lin Xupeng Miao Xiaonan Nie Yangyu Tao Yaofeng Tu Di Wang 0052 Xuanyu Wang Yujie Wang Jinbao Xue Hailin Zhang 0004