[FEATURE] llama-server --rpc

[arturbacilla]

Llama-server has an optional feature to share memory with multiple GPU and multiple hosts (using the flag --rpc ). But since this is a experimental feature, it is not provided within the docker image but only by building it. I was wondering if there is a way to llama-swap to have a experimental image with this --rpc flag available, or if one could replace the binary inside the container somehow