分享

MegaTrain: Full Precision Training of 100B+ Parameter Large Language Models on a Single GPU

515
查看论文
热度