GPU support for custom inference runtimes in MLServer

[koolgax99]

I am trying to use GPU in my custom inference endpoint built using MLserver. I am unable to load the model on gpu. Can you please let me know if this is possible or not?

Thank you