Is it possible to use different number of gpus on different nodes?

[TheShadow29]

I have two machines, one with 4 gpus, and another with 8. Currently, I am able to do training on 4 gpus on both machines, but I can't figure out how to use all 4 on the first and all 8 on the second. This page suggests using the exact same script on both nodes, but gpus argument would be different.

I wanted to know if there is any easy solution here.