MPipeMoE: Memory Efficient MoE for Pre-trained Models with Adaptive Pipeline Parallelism

Zheng Zhang, Donglin Yang, Yaqi Xia, Liang Ding 0006, Dacheng Tao, Xiaobo Zhou, Dazhao Cheng. MPipeMoE: Memory Efficient MoE for Pre-trained Models with Adaptive Pipeline Parallelism. In IEEE International Parallel and Distributed Processing Symposium, IPDPS 2023, St. Petersburg, FL, USA, May 15-19, 2023. pages 167-177, IEEE, 2023. [doi]

@inproceedings{ZhangYXDTZC23,
  title = {MPipeMoE: Memory Efficient MoE for Pre-trained Models with Adaptive Pipeline Parallelism},
  author = {Zheng Zhang and Donglin Yang and Yaqi Xia and Liang Ding 0006 and Dacheng Tao and Xiaobo Zhou and Dazhao Cheng},
  year = {2023},
  doi = {10.1109/IPDPS54959.2023.00026},
  url = {https://doi.org/10.1109/IPDPS54959.2023.00026},
  researchr = {https://researchr.org/publication/ZhangYXDTZC23},
  cites = {0},
  citedby = {0},
  pages = {167-177},
  booktitle = {IEEE International Parallel and Distributed Processing Symposium, IPDPS 2023, St. Petersburg, FL, USA, May 15-19, 2023},
  publisher = {IEEE},
  isbn = {979-8-3503-3766-2},
}