Hacker news

  • Top
  • New
  • Past
  • Ask
  • Show
  • Jobs

Flattening Every Memory Peak in Long-Context Mixture-of-Experts Training (https://arxiv.org)

13 points by gmays 5 days ago | 1 comments | View on ycombinator

ricardobeat 2 days ago |

The abstract reads like slop. Authors working on LLMs should know better than that.