SparDL: Distributed Deep Learning Training with Efficient Sparse Communication

Documenti analoghi

• Communication-Efficient Distributed Learning via Sparse and Adaptive Stochastic Gradient
  di: Deng, Xiaoge, et al.
  Pubblicazione: (2021)
• LoCoDL: Communication-Efficient Distributed Learning with Local Training and Compression
  di: Condat, Laurent, et al.
  Pubblicazione: (2024)
• PacTrain: Pruning and Adaptive Sparse Gradient Compression for Efficient Collective Communication in Distributed Deep Learning
  di: Wang, Yisu, et al.
  Pubblicazione: (2025)
• SpecInF: Exploiting Idle GPU Resources in Distributed DL Training via Speculative Inference Filling
  di: Lv, Cunchi, et al.
  Pubblicazione: (2025)
• Hecate: Unlocking Efficient Sparse Model Training via Fully Sharded Sparse Data Parallelism
  di: Qing, Yuhao, et al.
  Pubblicazione: (2025)