-
CUDA_VISIBLE_DEVICES=3,4,5 llamafactory-cli webui,开始微调时会把数据放到3、4、5这三张卡上,相当于在这三张卡上同时加载3个模型,每个gpu使用三分之一的数据。但是单卡的显存肯定不够的。如果是llamafactory-cli webui则直接把服务器上的卡全用了。 |
Beta Was this translation helpful? Give feedback.
Answered by
hiyouga
Nov 2, 2024
Replies: 1 comment
-
DeepSpeed zero3 |
Beta Was this translation helpful? Give feedback.
0 replies
Answer selected by
lxlx2084
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
DeepSpeed zero3