1 paper
Yuxiao Wang, Yuedong Xu, Qingyang Duan +4
The rapid growth of large language models (LLMs) and the continuous release of new GPU products have significantly increased the demand for distributed training across heterogeneou…