1 paper · 1 filter
Qingyuan Li, Bo Zhang, Hui Kang +4
Nowadays, communication bottlenecks have emerged as a critical challenge in the distributed training and deployment of large language models (LLMs). This paper introduces FlashComm…