Pith. sign in

Rdma over ethernet for distributed AI training at meta scale

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

citation-role summary

background 1

citation-polarity summary

fields

cs.DC 1

years

2025 1

verdicts

REJECT 1

roles

background 1

polarities

support 1

representative citing papers

Distributed Training under Packet Loss

cs.DC · 2025-07-02 · reject · novelty 6.0

The paper proposes unbiased gradient renormalization and stale-parameter broadcasts to keep distributed training convergent under random packet loss, with bounded inter-worker drift.

citing papers explorer

Showing 1 of 1 citing paper.

  • Distributed Training under Packet Loss cs.DC · 2025-07-02 · reject · none · ref 4

    The paper proposes unbiased gradient renormalization and stale-parameter broadcasts to keep distributed training convergent under random packet loss, with bounded inter-worker drift.