-
I want to use two GPUs by starting a process, one for model A and the other for model B. Is there a viable way? |
Beta Was this translation helpful? Give feedback.
Answered by
loadams
Jan 9, 2025
Replies: 1 comment
-
@Wallejd - I recommend opening an issue with your question, we tend to respond there sooner. But to do what you want, you would need to invoke two different Deepspeed commands, one that is only using GPU0 running modelA and one that uses GPU1 with modelB. There is no way to do this with a single command to handle 2 models, each on their own GPU. |
Beta Was this translation helpful? Give feedback.
0 replies
Answer selected by
loadams
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
@Wallejd - I recommend opening an issue with your question, we tend to respond there sooner. But to do what you want, you would need to invoke two different Deepspeed commands, one that is only using GPU0 running modelA and one that uses GPU1 with modelB. There is no way to do this with a single command to handle 2 models, each on their own GPU.