← back to paper

Review history

arxiv: 2605.06374 · 2 revisions

ResiHP: Taming LLM Training Failures with Dynamic Hybrid Parallelism

1. 2026-05-12 UNVERDICTED LOW v0.9.0 novelty 4.0
   146511 ms 5525 in 1627 out 2026-05-12T03:57:31.752339+00:00
2. 2026-05-08 UNVERDICTED LOW v0.9.0 novelty 5.0
   56130 ms 5523 in 1109 out 2026-05-08T05:10:41.025858+00:00