Heng Pan, Penglai Cui, Zhenyu Li 0001, Ru Jia, Penghao Zhang, Leilei Zhang, Ye Yang, Jiahao Wu, Mathy Lauren, Gaogang Xie. Zebra: Accelerating Distributed Sparse Deep Training With in-Network Gradient Aggregation for Hot Parameters. In 32nd IEEE International Conference on Network Protocols, ICNP 2024, Charleroi, Belgium, October 28-31, 2024. pages 1-11, IEEE, 2024. [doi]
Abstract is missing.