Capricorn: Efficient In-Memory Checkpointing for MoE Model Training with Dynamicity Awareness.
Wenqian Xie, Zhiquan Lai, Shengwei Li, Weijie Liu, Wei Wang, Yanqi Hao, Dongsheng Li
Browse the full CLUSTER paper archive.
Wenqian Xie, Zhiquan Lai, Shengwei Li, Weijie Liu, Wei Wang, Yanqi Hao, Dongsheng Li
Browse the full CLUSTER paper archive.